Comprehensive Handbook of Psychological Assessment

Volume 2: Personality Assessment

Part 6: Special Populations and Settings

Comprehensive Handbook of Psychological Assessment, Volume 2: Personality Assessment Mark J. Hilsenroth (Editor), Daniel L. Segal (Editor), Michel Hersen (Editor-in-Chief) ISBN: 978-0-471-41612-8

September 2003

Part Six: Special Populations and Settings

40 The Use of Projective Tests in Assessing Neurologically Impaired Populations 539 Arpi Minassian and William Perry

41 Projective Assessment of Malingering 533 Jon D. Elhai, Bill N. Kinder, and B. Christopher Frueh

42 Projective Assessment of Personality in Forensic Settings 562 Joseph T. McCann

43 Cultural Applications of the Rorschach, Apperception Tests, and Figure Drawings 573 Barry Ritzler

44 Collaborative Exploration with Projective Techniques: A Life-World Approach 586 Constance T. Fischer, Emilija Georgievska, and Michael Melczak

Part Seven: Applications for Children and Adolescents

45 Sentence Completion Measurement of Psychosocial Maturity 595 P. Michiel Westenberg, Stuart T. Hauser, and Lawrence D. Cohn

46 Assessment of Object Representation in Children and Adolescents: Current Trends and Future Directions 617 Francis D. Kelly

47 Projective Assessment of Affect in Children’s Play 628 Sandra W. Russ

Author Index 643

Subject Index 661

PART SIX

SPECIAL POPULATIONS AND SETTINGS

CHAPTER 40

The Use of Projective Tests in Assessing Neurologically Impaired Populations

ARPI MINASSIAN AND WILLIAM PERRY

USING PROJECTIVE TESTS TO ASSESS THE PERSONALITY CHARACTERISTICS OF NEUROLOGICALLY IMPAIRED INDIVIDUALS 540 Head-Injured Patients 540 Cerebrovascular Dysfunction 541 Normal Aging and Dementia 541 Impaired Children and Adolescents 542

Psychologists who specialize in personality assessment are often asked to differentiate between psychological and cognitive impairment and to assess “organically impaired” individuals. These clinicians, however, have relied on use of their traditional assessment battery, which includes the Wechsler scales, “objective” personality tests such as the Minnesota Multiphasic Personality Inventory (MMPI), and “projective” tests, in order to determine whether the patient is cognitively impaired. Despite the widespread use of traditional personality assessment instruments in assessing neurologically impaired individuals, there has been relatively little direction provided to those who specialize in personality assessment on how to interpret the results obtained from their traditional testing battery, either to diagnose organicity or to assess cognitive functions. Furthermore, advances in the field of neuropsychology have led to a divergence of the two assessment traditions. In reality, however, clinical neuropsychologists endorse that cognitive status is inextricably linked with per-

NEUROPSYCHOLOGICAL FUNCTIONING 544 Historical Perspectives on the Rorschach as a Neuropsychological Instrument 544 The Contributions of the Comprehensive System 546 The Process Approach 548 REFERENCES 549

THE USE OF PROJECTIVE ASSESSMENTS TO ASSESS

sonality style, coping mechanisms, and emotional states and that a thorough assessment of neuropsychological functioning should include the assessment of personality. This may be best exemplified in a study by Zillmer (1994), who found that, when evaluating patients referred for neuropsychological assessment, 48% of clinical neuropsychologists responding to a test-usage survey reported assessing personality features using measures such as the MMPI, as well as projective tests such as the Rorschach test (Rorschach, 1942), the Thematic Apperception Test (TAT; Murray, 1963), projective drawings, and the Sentence Completion Test (SCT; Loevinger & Wessler, 1970). Similarly, personality assessment instruments (e.g. the NEO Personality Inventory) are being used with increased frequency to assess brain-behavior relationships. The field of personality assessment, in turn, has a growing understanding that behavior and personality are products of brain functioning (Zillmer & Perry, 1996). In fact, given our recognition that psychological factors are mediated by brain functioning, it is difficult to understand how a full assessment of one’s personality can be made without taking into consideration the cognitive functioning of the individual. In this regard, Zillmer and Perry (1996) have asserted that personality is mediated by brain functions and, consequently, the awareness of an individual’s cognitive capabilities is essential to the understanding of personality. Notwithstanding the fact that personality assessment instruments are commonly used by both psychodiagnosticians and neuropsychologists to assess the personality features and neuropsychological functioning of cognitively impaired patients, there has been little attempt to include “projective” instruments in this area and to integrate the results of these two assessment traditions.

The present chapter is a review of the existing literature on the use of projective assessment instruments in neurologically impaired patients. The literature reviewed was published primarily in the Journal of Personality Assessment and Assessment, as well as clinical neuropsychology journals. There is a broader and rich literature that has not been translated into English and therefore was not included in this review. Two approaches to using projective tests with neurologically impaired patients are discussed. First we review the use of projective instruments in neurologically impaired populations where the major effort is to understand the personality functioning of these individuals. Next we discuss the use of what are traditionally regarded as projective instruments in the assessment of neuropsychological functioning. Finally, we provide recommendations for the integration of personality and neuropsychological assessment.

USING PROJECTIVE TESTS TO ASSESS THE PERSONALITY CHARACTERISTICS OF NEUROLOGICALLY IMPAIRED INDIVIDUALS

Neurological dysfunction, apart from causing intellectual and cognitive impairment, is frequently associated with emotional and personality disturbance (Reitan & Wolfson, 1997). There is a long tradition of using projective measures such as the Rorschach to understand personality functioning in organically impaired individuals (Caputo, 1999; Cattelani et al., 1998; De Mol, 1973; Epstein, 1998; Hall, Hall, & Lavoie, 1968; Hybler, 1990; Leon-Carrion et al., 2001; Malmgren, Bilting, Frobarj, & Lindqvist, 1997; Sinacori, 2000). This work has involved a variety of specific neuropsychological populations such as head-injured patients, patients with cerebral dysfunctions, older individuals and dementia patients, and neurologically impaired children. The following sections will turn to a discussion of the use of projective testing in these populations.

Head-Injured Patients

Head injuries are one of the most common causes of brain damage, and their incidence rate is estimated to be from half a million to almost two million occurrences a year (Lezak, 1995). Multiple psychological and personality factors are important to consider in cases of head-injured patients, including the patient’s withdrawal from social settings, changes in the patient’s sense of identity, affective dysmodulation, reduced motivation, and many others. These factors play a crucial role in the patient’s participation in short-term treatment, longterm rehabilitation, resumption of role functioning, and potential development of serious psychiatric symptoms such as depression and suicidal ideation (Leon-Carrion et al., 2001; Parker, 1996). Trauma to the frontal lobes especially can be associated with distinct personality changes that may include increases in impulsivity, poor judgment and future planning, and difficulty in monitoring one’s own behavior.

Projective assessment of head-injured patients has primarily centered around the Rorschach. Two approaches have been employed; the first is a qualitative examination of Rorschach responses, that is, interpreting Rorschach imagery on individual responses (De Mol, 1973; Parker, 1996), while the second has used quantitative measures such as Rorschach ratios and variables from the Comprehensive System (CS; Exner, 1993) to generate group profiles (Epstein, 1998; Exner, Colligan, Boll, Stischer, & Hillman, 1996; Leon-Carrion et al., 2001; Sinacori, 2000).

In a qualitative analysis of Rorschach responses, Parker (1996) examined the records of 33 subjects who had suffered a minor head injury after a motor vehicle accident and had sustained whiplash and often a brief (less than 5 minutes) loss of consciousness. Their Rorschach responses reflected several themes of emotional and personality change. Mood changes such as increased anxiety and post-traumatic stress reactions were reflected in imagery that was labeled by the patients as frightening, threatening, and sinister. Increases in depression were indicated by responses containing morbid features such as dead humans and animals or experiences of unhappiness and pain. Another prominent personality theme that was noted was that of “identity change,” expressed via Rorschach images of being vulnerable, disfigured, victimized, defenseless, and incompetent. A high degree of bodily concern and feelings of defectiveness were also observed in another study of traumatic brain injury patients (Epstein, 1998). Additionally patients evidenced a change in their world outlook, regarding their environment as threatening and offering responses that were replete with aggressive and destructive themes. Parker suggested that this population used defense mechanisms such as retreat into fantasy to cope with their changed environment and altered sense of identity.

Although a drawback of Parker’s study was the lack of a comparison group, his findings were partially consistent with that of Exner et al.’s (1996) quantitative approach. Exner and colleagues assessed 60 subjects with mild to moderate closed head injury, approximately two thirds of whom had experienced some loss of consciousness following their head trauma. According to the authors, these subjects also evidenced a preoccupation with changes in their self-image and identity, offering a greater percentage of reflection responses than non-head-injured subjects. Unlike Parker’s subjects, however, the subjects in Exner et al.’s study did not evidence significantly increased indices of affective disturbance; for example, the percentage of morbid responses was not significantly higher in head-injured patients as compared to the noninjured controls. In fact, these subjects produced impoverished Rorschach protocols that were interpreted as a tendency to restrict and avoid emotional content. For example, the head-injured patients generated a low frequency of achromatic and color responses, resulting in a lower than average affective ratio (Afr), which prompted the authors to suggest that head-injured patients were experiencing “a sort of emotional shutdown” (p. 325).

Exner and colleagues also focused on coping and decisionmaking abilities and reported that head-injured subjects appeared to have fewer available coping resources than normal comparison subjects (low EA) and were inconsistent in their decision-making style (higher than average frequency of ambitent style). Head-injured patients also had significantly higher Lambda scores, which the authors interpreted as a more simplistic response style and a tendency to avoid complexity and ambiguity. The authors suggested that this simplistic style may either be a consequence of cognitive impairment, a practical defense against being overwhelmed, or a trait that predated the head injury and possibly contributed to the patient finding him- or herself in the high-risk situation that caused the injury.

Exner et al.’s (1996) findings echoed those of an earlier study on severe closed-head-injured patients (Ellis & Zahn, 1985). The 35 head-injured subjects in this study demonstrated significantly higher Lambda scores and lower Afr ratios than normal comparison subjects, prompting the authors to conclude that “the controlled display and experience of emotion were difficult for the CHI [Closed Head Injury] survivor” (p. 127). Ellis and Zahn suggested that high scores on the Suicide Constellation Index provided evidence that headinjured subjects were experiencing painful emotions such as depression, anxiety, and irritability, but that this “affect was turned inward towards the self” (p. 127).

The high incidence of Rorschach CS indicators of depression and suicidal ideation among head-injured patients has

been noted in several other studies (Leon-Carrion et al., 2001; Sinacori, 2000). Sinacori, for example, found that 25% of the head-injured subjects in her study were classified as depressed based upon the Depression (DEPI) index. Leon-Carrion and colleagues’ recent study notes still a higher percentage (49%) of head-injured patients that were classified as depressed according to DEPI index criteria. Thirtythree percent of the sample also were positive for the Suicide Constellation (S-CON) index. The authors discussed several contributors to depression and suicidal ideation in headinjured subjects, noting that the change in mood could be a direct “psychological” consequence of head injury and the resulting physical and cognitive limitations. They also point out, however, that neurobiological factors are important to consider. For example, depression appears to be more common in patients with head injuries that result in lesions to serotonin-modulating circuitry, which has been implicated in the regulation of mood.

Cerebrovascular Dysfunction

An informative approach to assessing personality deficits in a specific group of organic patients has been illustrated by Vitale, Pulos, Wollitzer, and Steinhelber (1974), who studied 138 patients with cerebrovascular insufficiency disease, defined as the pathology of cerebrovascular circulation resulting in changes in brain tissue and usually involving stroke. Patients were administered an extensive psychological battery including projective tests such as the Rorschach, TAT, and Sentence Completion, and a cluster analysis was performed. The three projective tests loaded on two clusters, one that the authors labeled “social conformity” and that included variables from the TAT and Sentence Completion. The other factor was labeled “defensiveness” and included Rorschach and TAT variables. The authors did not identify the specific projective test measures that comprised these clusters but discussed the results in terms of the social rigidity and conformity that reflected poststroke patients’ concerns about fitting in with social norms. Their conclusion is similar to that in Exner et al.’s (1996) study of head-injured patients where the impoverished quality of the protocols is interpreted as a defensive coping mechanism, perhaps involving denial of the serious complications of the illness. An alternative explanation for Vitale et al.’s finding, however, is that strokes often cause significant language deficits, which may have contributed to the impoverished responses the authors observed.

Normal Aging and Dementia

The changes in personality features across the life span and into old age have historically been a focus of attention in Rorschach research (Ames, 1960a, 1966; Light & Amick, 1956), and variables from the Rorschach continue to be useful in characterizing older subjects. Early studies such as Light and Amick’s (1956) offered an overly gloomy and pathological picture of the personality of the 65- to 85-yearold: “The aged are . . . suspicious, anxious, and evasive . . . show a somewhat immature and introversive inner life which is colored by fantasy and unreality . . . little awareness of affectional needs . . . signs of inflexibility, stereotypy, and intellectual impotence” (pp. 194–195). Longitudinal studies indicated that there were noticeable trends in Rorschach protocols as subjects advanced into old age (Ames, 1960a, 1960b, 1966), such as decreased number of responses, increased Whole versus Detail responses in late old age, increased pure Form responses, decreased good Form responses, decreased Human Movement responses, decreased Human and increased Animal responses, and increased card rejections.

Other studies have refuted the hypothesis that aging is consistent with pathology, finding, for example, that elderly patients do not exhibit more defensiveness than younger ones, contrary to some expectations (Tamkin & Hyer, 1983). In contrast to Light and Amick’s results, Shimonaka and Nakazato (1991) found that among elderly Japanese subjects, social interest and sensitivity were preserved up until death. Several studies on elderly Finnish subjects have also challenged the earlier results of Ames, finding that perceptual accuracy (F#%) was maintained in 71-year-old subjects and no significant differences in content of responses were found between older subjects and their younger counterparts (Mattlar, Knuts, & Virtanen, 1985). The authors found similar results when testing 80-year-old males, concluding that “for the majority of subjects, personality was characterized by intact ambitions, reasonable independence, and sufficient resources” (Mattlar, Carlsson, Forsander, & Karppi, 1992, p. 41). Mattlar and colleagues offer several reasons for the discrepancy between these findings and the rather pessimistic results of earlier researchers, noting that the activity and functioning level of today’s older individuals are markedly improved compared to the aged population in the 1950s and 60s, and that early studies tended not to differentiate institutionalized and possibly demented older subjects from noninstitutionalized ones.

Indeed, when elderly subjects with dementia are compared to age-matched but nondemented subjects, results from projective tests such as the Rorschach, TAT, and the Draw-A-Person Test can significantly differentiate the groups (Belloni Sonzogni, Carabelli, Curioni, & Fumagalli, 1999; Benson, 2000; Insua & Loza, 1986; Knapp, 1994; Muzio & Luperto, 1999; Perry, Potterat, Auslander, Kaplan, & Jeste, 1996). Dementia patients have fewer human Movement (M) responses, and the human movements that do appear have been interpreted as representative of “weak energy” or more passive movement, reflecting the demented person’s sense of mental and physical deterioration and lack of strength (Insua & Loza, 1986). Although the authors suggest that the Rorschach can be used to accurately characterize the psychological status of dementia patients, they appropriately include the caveat that movement scores are not independent of other confounding factors such as verbal, cognitive, and intellectual abilities that are diminished in demented states. Perry and colleagues (1996), for example, also found that Dementia of the Alzheimer’s Type (DAT) patients had fewer Human Movement responses but additionally noted that the Rorschach protocols were indicative of marked cognitive and intellectual impoverishment, characterized by high Lambda, few blends, a low number of responses, few Populars, increased card rejections, and so on. Other studies of dementia patients have suggested that these individuals have fewer psychological resources (low EA) and poorer reality testing (higher X–%) than controls (Muzio & Luperto, 1999), as well as more primitive levels of object relations that could not be accounted for simply by decreased ability to provide verbal responses to projective tests (Benson, 2000).

Auslander, Perry, & Jeste (2002) studied older, stable schizophrenia patients and compared them to a cohort of nonpsychiatric subjects. Individuals in this study ranged from 45 to 100 years old. In addition to comparing the two groups on the Ego Impairment Index (EII; Perry & Viglione, 1991), a Rorschach measure of cognitive and perceptual disturbance, they compared subjects’ Rorschach responses to performance on the Dementia Rating Scale (DRS; Mattis, 1973), a neuropsychological screening battery. They found a slight elevation in EII scores in both groups when compared to younger normal subjects but also noted that the older normal group differed from the older schizophrenia group in that they produced richer and more complex responses. Consequently, the index scores between the two groups were similar but their qualitative responses were distinct. Furthermore, the EII scores of the older schizophrenia patients were significantly correlated with DRS scores such that increased cognitive and perceptual disturbance was associated with poorer neuropsychological functioning. Their finding supports the caveat that psychological interpretation of projective measures in aging patients must take into account their neuropsychological status, since impoverished protocols may reflect cognitive impairment rather than psychological defensiveness.

Impaired Children and Adolescents

The research on projective testing of neurologically impaired children and adolescents has included data from the Rorschach (Acklin, 1990; Barison & Del Monaco Carucci, 1973; De Negri & Saccomani, 1975; Horowitz, 1981; Hybler, 1990; Upadhyaya & Sinha, 1974), the TAT (De Negri & Saccomani, 1975; Horowitz, 1981; Koykis, 1985; Upadhyaya & Sinha, 1974), and other projective instruments such as the Hand Test (Wagner, Rasch, & Marsico, 1990). A significant portion of this work has been conducted in Europe. Impaired children’s responses to the TAT stimuli suggest a preoccupation with their self-image and their image of others. For example, children and teenagers with cerebral dysfunctions that may manifest in motor disabilities experience anxiety-laden and deprecatory self-images (De Negri & Saccomani, 1975). They also appear to be preoccupied with how to cope in social situations that involve nondisabled peers (Koykis, 1985). Developmentally delayed (DD) adolescents show similar profiles to the children with cerebral dysfunction; in addition, DD subjects seem to be focused on their basic needs and regard their external environment as threatening (Upadhyaya & Sinha, 1974).

Learning disabilities (LDs) are fairly prevalent among children (2 to 10% incidence rate; American Psychiatric Association, 1994), and emotional disturbances in LD children are thought to be quite common (Sattler, 1982). Although the disorder is likely to be a frequent referral for child assessment, there has been little published using projective tests with this population. Acklin (1990) studied the records of 41 children classified as either spatial disordered or language disordered. Although he did not find significant differences between these two classifications of LD, there were notable Rorschach differences between the LD children as a group and non-LD children. LD children had higher perceptual inaccuracy scores (higher X–%); higher Afr scores, indicating greater constriction or avoidance of emotional stimuli; a trend toward higher Lambda scores, reflecting increased rigidity and simplicity in responding; and a trend toward more White Space responses, which he suggested reflected their greater oppositionality in interpersonal situations. Not surprisingly these subjects also demonstrated increased levels of psychological stress as manifested in a greater frequency of the Comprehensive System variables of D and Adjusted D scores less than zero.

In summary, the findings from the use of projective instruments in understanding personality functions among neurologically impaired groups can be interpreted in several different ways: (1) The neurological insult has influenced the brain systems responsible for the regulation of emotion, (2) the psychological stress associated with neurological impairment and the accompanying cognitive problems gives rise to emotional disturbances, social withdrawal, and changes in self-identity, and (3) the individual who eventually becomes a neurological patient, for example, by sustaining a head trauma, has suboptimal personality and emotional resources to begin with, making high-risk behavior and resulting neurological trauma more likely (Reitan & Wolfson, 1997). The works discussed previously are representative of how neurologically impaired individuals perform on projective tests such as the Rorschach; however, the psychological inferences that have been drawn must be accepted with caution. All of the above studies are cross-sectional analyses, therefore causal statements about preexisting personality features are impossible to make with great certainty. Additionally, the results have been compared to the performance of neurologically intact individuals, without recognition of the likelihood that neurologically impaired patients may have “abnormal” results not necessarily due to personality features secondarily associated with head injuries or other insults, but rather due directly to specific cognitive dysfunctions. For example, in Exner et al.’s (1996) study, the authors eliminated 12 out of 72 subjects who produced protocols with fewer than 14 responses, citing that these protocols were probably “not interpretively valid.” (p. 319). Additionally, the remaining subjects, as well as the subjects in Ellis and Zahn’s (1985) study, produced significantly high Lambda scores. The authors suggested that as a group these patients may have been withdrawn, guarded, and avoidant of complexity and ambiguity. Alternatively, head-injured patients, as a direct result of trauma to the brain regions responsible for executive functioning, decision making, and cognitive flexibility, may actually be unable to cognitively process complex and ambiguous stimuli and accurately articulate their responses. Similarly, Vitale et al. (1974) propose that the protocols of patients with cerebrovascular dysfunction are suggestive of the personality characteristic of defensiveness. Here again, their interpretation based upon the impoverished protocols does not take into consideration the patients’ neuropsychological impairment, which may prevent these subjects from generating rich, complex, and well-integrated verbal responses.

Although these examples highlight the difficulties of making psychological interpretations with projective tests in neurologically impaired groups, it remains possible that the descriptions of the patient’s psychological functioning are nevertheless accurate of the patients’ self-experience. Thus, one could argue that a distinction between psychological and neuropsychological disturbance is artificial and that, as authors have pointed out, personality factors are inextricably linked with neuropsychological capabilities so that the two cannot be parsed out separately. Since the use of “objective” and “projective” tests to assess neurologically impaired individuals is so widespread, clearly clinicians are finding the assessment of personality important when attempting to understand how neurologically impaired individuals cope with everyday life.

THE USE OF PROJECTIVE ASSESSMENTS TO ASSESS NEUROPSYCHOLOGICAL FUNCTIONING

When using a “projective” test to measure neuropsychological functioning, it is necessary to reframe the role of the instrument (Perry et al., 1996). Each of the projective tests provides a different test environment that taps into specific cognitive functions (Johnson, 1994). Thus, each test becomes a problem to be solved. For example, the Rorschach problem is one of integrating complex, abstract visual stimuli in a meaningful manner. This “problem-solving” approach lends itself to assessing cognitive functions under more complex situations and the resulting responses can offer rich examples of the cognitive deficits of neurologically impaired individuals. In this regard the tests change from personality assessment instruments to neuropsychological assessment measures. This approach, however, requires that the diagnostician sufficiently understand neuropsychological functioning so that the interpretations developed can be “brain-behavior” based rather than personality based. For example, if a patient with DAT offers only 13 simple responses to the Rorschach, all with pure form, one might interpret the resulting protocol as reflecting a tendency to avoid or withdraw from complexity and defend against the overwhelming stimuli. If these responses were characterized by poor form quality, the interpretation may suggest that this hypothetical individual sees things differently than others and misperceives the environment. While these interpretations may be true, this same protocol can be illustrative of neurological deficits when viewed as a response to a complex problem-solving test. For example, the low number of responses may reveal the inability of the demented individual to integrate complex visual stimuli while the tendency for poor form quality may be an indication of visual dysgnosia (impairment in the ability to recognize the symbolic significance of objects presented visually) associated with parietal lobe impairment. Thus, using a neuropsychological approach, the responses to projective tests provide illustrations of underlying cognitive deficits and the “personality-based” volitional interpretations are given secondary value.

The significant majority of the work conducted with projective tests to assess neuropsychological functioning has been with the Rorschach Inkblot test. This area of study can further be divided into two domains. The first and most popular is the use of the Rorschach in the identification and diagnosis of neurologically impaired individuals, and the second area is the use of the test to assess specific cognitive functions. Caputo (1999) reviewed the trends in the use of the Rorschach as a neuropsychological instrument and identified the neurological populations that have historically been the focus of study. According to her count, of the 75 total Rorschach studies from 1920 to the 1990s that have examined seizure disorder, electroconvulsive therapy (ECT), or Tourette’s syndrome, 42 of them took place in the 1940s and 1950s, while another neuropsychological population of great interest was the head-injured patient, who was becoming increasingly common as a result of war-related physical trauma. The popularity of psychosurgery and the resulting increasing numbers of patients with lobotomies, lobectomies, and hemispherectomies created a third group for study. The tradition of using the Rorschach to assist in diagnosing neurological disease begins, however, with Hermann Rorschach’s original work.

Historical Perspectives on the Rorschach as a Neuropsychological Instrument

Hermann Rorschach regarded the inkblot test as a complex task. His conceptualizations exceeded the popular notion that generating responses to the inkblots was a fanciful game and an exercise in creativity: “The interpretation of the chance forms falls in the field of perception and apperception rather than imagination” (Rorschach, 1942, p. 16). According to Rorschach, three different processes of perception were involved in the interpretation of the inkblots: sensation, memory, and association. He emphasized that the content of subjects’ responses was a secondary consideration in interpretation, while the process by which they arrived at their responses (reaction time, use of form vs. movement or color, use of the whole vs. part of the blots) was of primary importance. Indeed, Rorschach’s set of recommendations for how to interpret inkblot responses was among the earliest works to emphasize the process of problem solving over the outcome answer. This method, later referred to as the “process approach,” was further developed by Heinz Werner, applied to the field of neuropsychology by Edith Kaplan (1988), and has become one of the leading approaches in clinical neuropsychological assessment.

Rorschach’s initial studies using the inkblot test included significant numbers of subjects with neurological deficits, including patients with epilepsy, dementia, and Korsakoff’s amnesia. He observed consistent and distinct patterns of responding in these subjects; for example, noting that Korsakoff’s patients gave large numbers of responses that he attributed to confabulation, and that dementia patients produced a high number of perseverative responses. Rorschach, although initially tentative about the test’s ability to make definitive psychiatric diagnoses, saw potential in this realm: “After a further period of development it should be possible . . . to come to a definite conclusion as to whether the subject is normal, neurotic, schizophrenic, or has organic brain dysfunction” (p. 120). He, however, due to his early death, was unable to further pursue this line of research.

Zygmunt Piotrowksi advanced Rorschach’s notion that the inkblot test could diagnose organicity. To that effect, he wrote:

It is known of course that the so called higher mental processes change in a rather uniform manner following an organic disturbance of the central nervous system . . . By means of the Rorschach method we can measure the higher processes. Therefore, studying a group of organic cases, it should be possible to recognize the change in their higher mental functions using the Rorschach approach. (Piotrowski, 1936, p. 28)

Piotrowski, with the help of Rorschach’s work and that of Rorschach’s collaborator Emil Oberholzer, introduced 10 “signs” derived from Rorschach responses that he suggested could be helpful in differentiating patients with cerebral organic brain dysfunction from those patients with nonorganic psychiatric disturbances. These signs were: (1) R (total number of responses is less than 15), (2) T (average time per response exceeds 1 minute), (3) M (no more than one movement response), (4) Cn (Color Denomination or the Naming of colors without integration into the response), (5) F% (percentage of good form responses is below 70), (6) P% (percentage of Popular responses is below 25), (7) Rpt (the repetition of the same response to several inkblots), (8) Imp (Impotence or giving a response in spite of the subject’s recognition of its inadequacy), (9) Plx (perplexity or the subject’s requests for reassurance), and (10) AP (automatic phrases or the mechanistic, indiscriminate use of a pet phrase).

After examining the responses of 33 subjects, Piotrowski suggested that if at least 5 of his 10 signs were present, the existence of a cerebral disturbance was supported. He provided a striking case example of the sensitivity of the sign he termed Color Denomination, or the simple naming of colors on the blots with no attempt to integrate them into a response. A patient with epilepsy appropriately incorporated color in her response to Card IX of the Rorschach during a preseizure testing session; however, 1 day following a seizure episode her reaction to the same blot was only to name the colors it illustrated. Piotrowksi was careful not to exaggerate the specificity of these signs to organic disturbance exclusively, noting that some of them, poor form quality, for example, were also observed frequently in schizophrenia patients who at the

time were not popularly thought of as neurologically impaired patients.

Although Piotrowski approached the Rorschach as a neuropsychological instrument, that is, as a means of assessing cognitive functioning, he worded his inferences in crude “psychological” terms, such as the “weakening of inner life” and the “general law of mental disintegration” (p. 37) when describing the neurological patient. Thus the language that he used reflects the simplistic understanding of the neuropsychological functioning of the cognitively impaired patient at the time of his work.

The sign approach has been used successfully to study a variety of neurologically impaired patients, including headinjured or brain-damaged subjects (Chaudhury, John, Bhatoe, & Rohatgi, 1999; Nakamura, 1983; Nedelcu & Zellingher, 1971), patients with epilepsy (Chaudhury, John, & Rohatgi, 1998; Sampaio, 1977), and demented patients (Dorken & Kraal, 1951). For example, Dorken and Kraal (1951) studied 35 inpatient subjects with senile dementia and found that over half of them had five or more of the Piotrowski signs, as well as other hallmark signs of organic impairment such as confabulation. Notably, the subjects also had a high percentage of Pure Form responses, which the authors argued was probably not reflective of the popular interpretation of constriction and guardedness, but more likely due to the apathy that is commonly observed in demented patients.

Other early studies of Rorschach indicators of organic disturbance had limited success. In a study by Reitan (1955), brain-damaged subjects were differentiated from nonorganic neurotic subjects with only two of Piotrowki’s signs: (1) the total number of responses and (2) the Impotence sign, or the subject’s recognition of the inadequacy of the response but failure to improve it. Patients with Tourette’s syndrome had Rorschach records that were characterized as simplistic, stereotypic, and unproductive; however, the quantitatively based sign approach derived from the Rorschach failed to identify patients as “organic” while other tests such as the WAIS and the Bender Gestalt were found to be more sensitive to brain dysfunction related to Tourette’s syndrome (Shapiro, Shapiro, & Clarkin, 1974).

The sign approach to making organic diagnoses has been criticized by Frank (1991) on several grounds: (1) heterogeneous patients were grouped together based on a gross impairment (e.g., encephalopathy) but without respect to specific etiology of the dysfunction (e.g., trauma, disease, tumor, deterioration), thus making the interpretation of the results difficult, and (2) the usefulness of these signs has been highly inconsistent across studies so that it is very difficult to identify a specific, quantitative psychometric pattern that is characteristic of all patients with organic dysfunction. Furthermore, Piotrowski’s signs have been noted to yield a high rate of false negatives, that is, the signs sometimes fail to detect organicity in subjects who actually do have a neurological condition (DeCato, 1993). Caputo (1999) also noted some methodological flaws in the use of Piotrowski’s sign approach, the most significant of which is the lack of comparison groups in most of the studies. When comparison subjects were included, they had typically been taken from medical or psychiatric populations versus from a heterogeneous normal population. Additionally subject numbers for these studies had typically been very low, some as low as 5 subjects.

The marginal results obtained with the Rorschach in making diagnoses of organicity as well as the development of sophisticated neurodiagnostic technologies in the 1960s (i.e., EEG) resulted in a waning of the popularity of using the test with neurologically impaired patients (Caputo, 1999). The advances in the field of neurology were the impetus for a new school of neuropsychologists, primarily in North America, who favored highly experimental and objective assessments and rejected the Rorschach as being “unscientific.” In fact, a presumably comprehensive reference published in 1983, the Clinical Psychology Handbook, did not include any mention of the Rorschach in a discussion about evaluation and treatment of patients (Elias, 1989). More recently, however, there has been a resurgence of the use of the Rorschach in neuropsychological studies as modern neuropsychologists “rediscover” that the test taps specific, measurable cognitive functions (Acklin & Wu-Holt, 1995; Exner et al., 1996).

The Contributions of the Comprehensive System

John Exner’s reframing of the Rorschach and development of the Comprehensive System helped to extend the use of the test for neuropsychological purposes (Elias, 1989). Exner responded to the criticisms of the psychologists who regarded the Rorschach as “unscientific” in two ways. First he gathered empirical data on large samples of patient and normal populations; second, he reconceptualized the test as more than a test of projection. Exner asserted that most subjects were fully aware the stimuli were inkblots and in order to fulfill the task of identifying the stimulus, they essentially had to misperceive it and “convert the blot into something that it is not” (Exner, 1993, p.29). As such the Rorschach creates a problem-solving situation that requires a chain of complex decision-making operations. Exner outlined the response process, suggesting that solving the Rorschach problem “What might this be?” involved the processing of visual information, then the encoding and classifying of the stimulus. Projection, he argued, probably played only a minor role in this initial stage of perception and encoding. This “reintroduction” of the Rorschach as a task of perception and visual information processing has again helped bridge the gap between personality and neuropsychological assessment. In formulating a cognitive framework for understanding the test, Acklin and Wu-Holt (1995) assert that the “status of the Rorschach as a cognitive and/or neuropsychological assessment measure is currently underdeveloped, though we contend that it may have great potential” (p. 170).

Full appreciation of the Rorschach as a neuropsychological instrument requires an understanding of brain function. Each phase in the response process involves complex cognitive functions that are mediated by specific neural circuitry. Failures in specific aspects of the response process, therefore, lend themselves to hypotheses regarding which brain regions are impaired. First, the subject’s ability to understand the task and mentally retain the instructions throughout the duration of the task can reveal information about attention, orientation, and working memory (Perry & Potterat, 1997). Next, the subject must visually scan the blot and organize and process the percepts. Inferences regarding hemispheric integrity can be made since the left hemisphere of the brain is thought to be responsible for processing of localized, fine-detail stimuli while the right hemisphere processes global or whole-percept visual information (Delis, Kiefner, & Fridlund, 1988). The two neural tracts, dorsal and ventral, that are involved in the processes of visual scanning and processing can also be assessed (Acklin & Wu-Holt, 1995). The visual-scanning “strategy” of the subject can reveal important information about potential dysfunction along these neural pathways as well as indicate hemisphere-specific deficits.

Once the visual stimuli have been scanned, encoded, and organized, the subject must reorient to the instructed question, “What might this be?” This demand to categorize the percept requires the complex function of memory and its subcomponents, such as retrieval. Although memory is not restricted to a single brain region (Perry & Potterat, 1997), areas such as the hippocampus, the thalamus (Acklin & Wu-Holt, 1995), the posterior cortical areas, and the frontal lobes (Perry & Potterat, 1997) are involved in storing and retrieving information and associating the percepts with encoded iconic stores. Finally, the integration of these operations and the generation of a comprehensible and perceptually accurate or “reasonable” response is a function governed by the frontal lobes. In summary, the “path” of the Rorschach response from start to finish can be described in neuropsychological terms and can even provide some insight into putative impairment in specific anatomical regions.

As an example of this approach, Perry and Potterat (1997) present the case of a subject with a right-hemisphere stroke. Of the subject’s 13 Rorschach responses, 11 of them had Dd locations, meaning that he focused primarily on unusual and infrequently used areas of the blots. This finding may be suggestive of his deficit in perceiving the whole or gestalt of the stimuli. In addition, almost all of these responses involved only the right side of the inkblot, not unexpected given that his brain damage involved a lesion in the right hemisphere and subsequently his left visual field. Similar results were found by Daniels, Shenton, Holzman, and Benowitz (1988); their sample of 23 subjects with right-hemisphere cortical damage tended to produce responses comprised mostly of details of the Rorschach blots instead of integrating the whole stimulus in their responses. The impact of hemispherespecific deficits was further illustrated by Belyi (1983), who noted that patients with right-hemisphere tumors focused on unusual details of the Rorschach blots when giving responses, again suggestive of the importance of the right hemisphere’s role in the “formulation of visual gestalts” (Belyi, 1983, p. 410). When right-hemisphere-lesioned patients do offer whole responses, it is frequently in the context of combining parts of the stimulus inappropriately and offering percepts that are perceptually inaccurate and bizarre (Hall et al., 1968).

Several authors have applied a neuropsychological approach to the Rorschach in dementia patients (Dorken & Kraal, 1951; Insua & Loza, 1986). Perry and colleagues (1996) identified variables from the CS that theoretically corresponded to the cardinal neuropsychological deficits of patients with DAT and found that DAT patients evidenced significant differences from normal elderly subjects on the majority of these variables. The hallmark feature of aphasia, for example, was operationalized by high Lambda, low Blends, high Deviant verbalizations, and a low number of responses. DAT patients significantly differed from their nondemented counterparts in three of these four variables. Similar results were found for variables purportedly tapping the other cardinal signs of DAT, such as amnesia, agnosia, apraxia, and abstraction difficulties.

Other specific groupings of variables from the CS have evidenced sensitivity to neuropsychological dysfunction. One example is the Cognitive Triad, comprised of processing, mediation, and ideation (Colligan, 1997; Exner et al., 1996). Colligan (1997) used this triad to describe the deficits of a patient with suspected organic impairment. The patient approached the Rorschach stimuli in an overly conservative and irregular way and demonstrated marked disturbances in thinking, poor judgment, and faulty decision-making capabilities. Exner and colleagues (1996) studied a much larger sample, 60 subjects with mild to moderate closed head injuries, and found that head-injured patients could be characterized as “underincorporators” (signified by low Zd scores), engaging in ineffective scanning of stimuli and not attending carefully to details. The cluster of CS variables used to measure information and response processing (i.e., Zd) was found to load with other neuropsychological tests in a factor analytic study, whereas variables from the MMPI loaded separately from any of the neuropsychological tests (Zillmer & Perry, 1996).

In children, Acklin (1990) used the Rorschach to characterize the cognitive deficits of learning disability. He reported that perceptual inaccuracy and distortion does appear to characterize the Rorschach profiles of LD children, who were found to have significantly elevated Form Quality minus responses. In addition these subjects evidenced a coping style closer to ambitent than non-learning-disabled children, suggestive of the inconsistent and inefficient approach to problem solving that has been behaviorally observed in LD. As in the case of adult head-injured subjects, scanning of the stimuli was characterized as inefficient and underincorporative as evidenced by significantly lower Zd scores, again reflecting haphazard, ineffective scanning and inattention to detail areas.

Cognition and problem-solving activity on the Rorschach have also been assessed using a slightly different conceptualization, that of response complexity. Response complexity has been defined as the degree of productivity or “richness” in a Rorschach protocol. It has been operationalized in varying ways by Rorschach researchers but is usually comprised of elements such as determinant blends, the use of multiple contents, and degree of organizational activity (Viglione, 1999). Conversely, response simplicity is typically characterized by pure form, animal content only, perseverations, and card rejections. Viglione concluded that, based on an overview of the literature, response simplicity has been associated with “negative characteristics and problem groups” (p. 259) and also is frequently observed among neurologically impaired groups such as LD patients, head-injured patients, dementia, and in the rapid decline in functioning shortly preceding death.

The construct validity of response complexity has been supported by the finding that increased complexity was significantly related to pupil dilation, a putative psychophysiological measure of cognitive effort and resource allocation (Minassian, 2001). Minassian studied schizophrenia patients and found that with an increase in pupil dilation, their Rorschach responses became more complex and productive, involving multiple areas of the blot and multiple determinants in formulating a response. Another study (Felger, 1996) has suggested that reduced resource allocation in schizophrenia as measured by the skin conductance orienting response was associated with impoverished and simplistic Rorschach responses. In addressing the utility of response complexity, Viglione asserts that the complexity of Rorschach responses is highly associated with neuropsychological functions and can therefore be informative as to a subject’s cognitive limitations.

The use of the Rorschach to characterize the nature and degree of neurological disturbance is not limited to the CS. Malmgren et al. (1997) note that some Rorschach signs that characterized patients with Korsakoff ’s amnesia in their longitudinal study are not included in the CS. The fact that neurologically impaired populations often produce short, impoverished records also makes it difficult to use the CS variables in a traditional manner. Malmgren and colleagues conclude that “this casts some doubts on the sufficiency of Exner’s system for an adequate analysis of ‘organic’ cases” (Malmgren et al., 1997, p. 19).

Perry et al.’s (1996) DAT study suggested that the CS could be a useful indicator of dementia-related neuropsychological disturbance; however, “there were several important exceptions” (p. 359). The CS measure of perseveration (PSV) failed to differentiate DAT patients from normal comparison subjects, a noteworthy finding since DAT patients, due to their deficits in memory and executive functioning, demonstrate perseverative behavior in a number of other contexts (Lezak, 1995). Additionally, one of the CS variables thought to be a marker for aphasia, the deviant verbalization score, was not significantly higher in DAT patients as expected.

The Process Approach

These limitations of the CS in assessing impoverished protocols have prompted several researchers to approach the Rorschach test in a novel way, yet consistent with the process approach that Hermann Rorschach himself described. The process approach is based on the notion that problem solving involves numerous different cognitive subroutines, and that individuals use these subroutines in varying ways depending on their skill, level of intellect, or neurological deficits. Consequently, it is based on the “careful systematic observations of the problem-solving strategies used by patients” (Milberg, Hebben, & Kaplan, 1996, p. 60).

Perry et al. (1996) applied this approach to linguistic errors and executive perseverations observed on Rorschach protocols. With the revision of taxonomies utilized by others (Barr, Bilder, Goldberg, & Kaplan, 1989; Johnston & Holzman, 1979; Sandson & Albert, 1984; Solovay, Shenton, Gasperetti, & Coleman, 1986), the authors found significant differences between DAT patients and normal elderly individuals where the CS failed to do so. Linguistic errors such as semantic paraphasias, word-finding circumlocutions, and use of superordinate categories, all cardinal features of DAT, significantly differentiated DAT patients from normal comparison subjects. Additionally, all three executive perseveration errors from Perry et al.’s scale occurred significantly more in DAT patients than in normals. In a later study, Perry, Felger, & Braff (1998) found that one of the three perseveration errors, stuck-in-set perseveration, was elevated in a subgroup of schizophrenia patients who were skin conductance nonresponders, thought to be an indication of frontal lobe impairment. Furthermore, stuck-in-set perseverations were highly correlated (r ” .47) with the perseverations measure from the Wisconsin Card Sorting Test (WCST), a test thought to be sensitive to frontal lobe dysfunction (Perry & Braff, 1998). The application of neuropsychological principles as well as associations with psychophysiological phenomena has served to broaden the utility of the Rorschach as a neuropsychological instrument and has “helped to bridge the gap between the fields of neuropsychology and personality assessment” (Perry et al., 1996, p. 360).

The process approach has been used with personality measures besides the Rorschach to characterize neurocognitive deficits. Patients with organic brain disorder were distinguished from nonorganic patients by their sequence of responding to a personality inventory, evidencing long sequences of the same response and an inability to shift to a new response (Frankle, 1995). Furthermore, this high frequency of perseveration appeared to be unrelated to verbal IQ, circumventing the concern that some personality measures may not be able to differentiate brain-damaged profiles from low IQ profiles (Wagner & Maloney, 1980).

The adoption of a qualitative approach to analyzing TAT responses has been fruitful in assessing neuropsychological deficits in DAT patients (Johnson, 1994). When compared to other psychiatric patients, DAT patients used significantly fewer words to describe the cards (an example of aphasia) and demonstrated a loss of instructional set between cards, simply describing the picture instead of generating a story as they were instructed (impairment in working memory). These features may help inform interpretation of the content responses from the TAT. Similar results were found when using the Hand Test with developmentally delayed adults; those subjects who had mental retardation in the absence of psychiatric illness tended to simply describe the test stimulus instead of responding to it in a meaningful way (Panek & Wagner, 1993). The Hand Test was also sensitive to unproductive and perseverative responses in severely behaviorally handicapped children (Wagner et al., 1990).

In summary, in this chapter we have reviewed the use of projective tests in assessing personality characteristics and cognitive functions of neurologically impaired individuals. The majority of the literature that we have reviewed has focused on the Rorschach test. As we have presented, the Rorschach offers rich information that may be useful in developing specialized rehabilitation treatment plans for neurologically impaired patients. Additionally, when viewing the Rorschach as a cognitive problem-solving test and interpreting responses from a neuropsychological perspective, the test may provide unique advantages over other cognitive assessment instruments. Perry and Potterat (1997) assert that “The Rorschach, used as an abstract problem-solving test, can help clinicians characterize the distinct nature of different cognitive processes as well as the unique psychological features of the individuals who are inflicted with these diseases” (p. 568). Clearly there is a valuable role for the Rorschach to be used as both an instrument for assessing personality characteristics as well as a means of describing cognitive functioning.

What, then, is the future role of “projective testing” in the field of neuropsychology? No doubt, using the process approach to interpreting traditional personality assessments will continue to be informative in understanding brain-behavior relationships. In addition, there is a long tradition of pairing Rorschach test performance with psychophysiological measures of neurocognitive functioning (Assael, Kohen-Raz, & Alpern, 1967; Kohen-Raz & Assael, 1966), and this line of work will continue to characterize the brain functions involved in processing and responding to projective stimuli. Perhaps one of the most ambitious endeavors in this regard has been the recent attempt by Peled and Geva (2000) to generate a neural network model for how Rorschach inkblots are perceived, supporting the notion that biological changes in the brain such as alterations in neuronal firing can influence the “psychological event” of integrating and interpreting a Rorschach stimulus. There have been several recent studies using the Rorschach in combination with neuro-imaging techniques. For example, DeFelipe-Oroquieta (personal communication, September 2001) has conducted several studies combining the Rorschach with positron emission tomography (PET) in epileptic patients. In a recent published study, thought-disturbed responses to Rorschach stimuli were correlated with increased levels of activity in brain areas associated with speech production as illustrated by the functional magnetic resonance imaging (fMRI) technique (Kircher et al., 2001). These examples demonstrate that the field of neuroscience can extend the use of what have been traditionally referred to as “projective tests” to reach beyond personality assessment.

Personality diagnosticians and clinical neuropsychologists will continue to find projective instruments useful in understanding stress, trauma, depression, and the “adaptive trajectory” (Rutter, 1985) of neurologically impaired patients. However, in the future we are likely to see that clinicians will need to have an increased understanding of neuropsychology that will help them to modulate their clinical interpretations. Only then can the field of assessment move away from a dualistic (mind vs. body) understanding of personality to an integrated appreciation of brain-behavior relationships.

REFERENCES

• Acklin, M.W. (1990). Personality dimensions in two types of learning-disabled children: A Rorschach study. Journal of Personality Assessment, 54, 67–77.
• Acklin, M.W., & Wu-Holt, P. (1995). Contributions of cognitive science to the Rorschach technique: Cognitive and neuropsychological correlates of the response process. Journal of Personality Assessment, 67, 169–178.
• American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author.
• Ames, L.B. (1960a). Age changes in the Rorschach responses of a group of elderly individuals. Journal of Genetic Psychology, 97, 257–285.
• Ames, L.B. (1960b). Age changes in the Rorschach responses of individual elderly subjects. Journal of Genetic Psychology, 97, 287–315.
• Ames, L.B. (1966). Changes in Rorschach response throughout the human life span. Genetic Psychology Monographs, 74, 89–125.
• Assael, M., Kohen-Raz, R., & Alpern, S. (1967). Developmental analysis of EEG abnormalities in juvenile delinquents. Diseases of the Nervous System, 28, 49–54.
• Auslander, L.A., Perry, W., & Jeste, D.V. (2002). Assessing disturbed thinking and cognition using the Ego Impairment Index in older schizophrenia patients: Paranoid vs. nonparanoid distinction. Schizophrenia Research, 53, 199–207.
• Barison, F., & Del Monaco Carucci, S. (1973). [Phenomenologic aspects of the Rorschach test in minimal brain dysfunction]. Revue de Neuropsychiatrie Infantile et d’Hygie`ne Mentale de l’Enfance, 21, 55–62.
• Barr, W.B., Bilder, R.M., Goldberg, E., & Kaplan, E. (1989). The neuropsychology of schizophrenic speech. Journal of Communication Disorders, 22, 327–349.
• Belloni Sonzogni, A., Carabelli, S., Curioni, M., & Fumagalli, A. (1999). L’immagine di se’ nella demenza. Ricerche di Psicologia, 23(3), 29–53.
• Belyi, B.I. (1983). Reflection of functional hemispheric asymmetry in some phenomena of visual perception. Human Physiology, 8, 410–417.
• Benson, S.L. (2000). An empirical analysis of object relations in persons with Alzheimer’s disease. Dissertation Abstracts International, 61, 1625B.
• Caputo, J. (1999, July). The Rorschach as a neuropsychological instrument: Historical precedents and future use. Paper presented at the 16th International Rorschach Association Congress, Amsterdam.

550 The Use of Projective Tests in Assessing Neurologically Impaired Populations

• Cattelani, R., Patruno, M., Catellani, A., Cesana, L., Lombardi, F., & Mazzucchi, A. (1998). Il metodo Rorschach in soggetti traumatizzati cranici. Archivio di Psicologia, Neurologia e Psichiatria, 59, 314–326.
• Chaudhury, S., John, T.R., Bhatoe, H.S., & Rohatgi, S. (1999). Evaluation of Piotrowski’s organic signs of head injury. Journal of Projective Psychology and Mental Health, 6, 53–57.
• Chaudhury, S., John, T.R., & Rohatgi, S. (1998). Evaluation of Piotrowski’s organic signs in epilepsy. Journal of Projective Psychology and Mental Health, 5, 127–130.
• Colligan, S.C. (1997). The neuropsychology of the Rorschach: An M.D. with M.B.D. In J.R. Meloy, M.W. Acklin, C.G. Gacono, J.F. Murray, & C.A. Peterson (Eds.), Contemporary Rorschach interpretation (pp. 535–547). Mahwah, NJ: Erlbaum.
• Daniels, E.K., Shenton, M.E., Holzman, P.S., & Benowitz, L.I. (1988). Patterns of thought disorder associated with right cortical damage, schizophrenia, and mania. American Journal of Psychiatry, 145, 944–949.
• De Mol, J. (1973). Psychological dynamics in cranial trauma through symbolism on Card 7 of the Rorschach. Annales Medico-Psychologiques, 1, 609–618.
• De Negri, M., & Saccomani, L. (1975). Psychodynamics of a minimal cerebral dysfunction. Revue de Neuropsychiatrie Infantile et d’Hygie`ne Mentale de l’Enfance, 23, 701–713.
• DeCato, C.M. (1993). Piotrowski’s enduring contributions to the Rorschach: A review of Perceptanalysis. Journal of Personality Assessment, 61, 584–595.
• Delis, D.C., Kiefner, M.G., & Fridlund, A.J. (1988). Visuospatial dysfunction following unilateral brain damage: Dissociations in hierarchical hemispatial analysis. Journal of Clinical and Experimental Neuropsychology, 10, 421–431.
• Dorken, H., & Kraal, V.A. (1951). Psychological investigation of senile dementia. Geriatrics, 6, 151–163.
• Elias, J.Z. (1989). The changing American scene in the use of projective techniques: An overview. British Journal of Projective Psychology, 34(2), 31–39.
• Ellis, D.W., & Zahn, B.S. (1985). Psychological functioning after severe closed head injury. Journal of Personality Assessment, 49, 125–128.
• Epstein, M. (1998). Traumatic brain injury and self-perception as measured by the Rorschach using Exner’s Comprehensive System. Dissertation Abstracts International, 59, 870B.
• Exner, J.E. (1993). The Rorschach: A comprehensive system: Volume 1. Basic foundations (3rd ed.). New York: Wiley.
• Exner, J.E., Colligan, S.C., Boll, T.J., Stischer, B., & Hillman, L. (1996). Rorschach findings concerning closed head injury patients. Assessment, 3, 317–326.
• Felger, T.E. (1996). Allocation of attentional resources and thought disorder in schizophrenia patients. Dissertation Abstracts International, 56, 6386B.
• Frank, G. (1991). Research on the clinical usefulness of the Rorschach: 2. The assessment of cerebral dysfunction. Perceptual and Motor Skills, 72, 103–111.
• Frankle, A.H. (1995). A new method for detecting brain disorder by measuring perseveration in personality inventory responses. Journal of Personality Assessment, 64, 63–85.
• Hall, M.M., Hall, G.C., & Lavoie, P. (1968). Ideation in patients with unilateral or bilateral midline brain lesions. Journal of Abnormal Psychology, 73, 526–531.
• Horowitz, H.A. (1981). Psychiatric casualties of minimal brain dysfunction in adolescents. Adolescent Psychiatry, 9, 275–294.
• Hybler, I. (1990). Le Rorschach et l’image du corps dans les traumatismes craniens severes. Bulletin de Psychologie, 43(396), 720–725.
• Insua, A.M., & Loza, S.M. (1986). Psychometric patterns on the Rorschach of healthy elderly persons and patients with suspected dementia. Perceptual and Motor Skills, 63, 931–936.
• Johnson, J.L. (1994). The Thematic Apperception Test and Alzheimer’s Disease. Journal of Personality Assessment, 62, 314–319.
• Johnston, H., & Holzman, P.S. (1979). Assessing schizophrenic thinking. San Francisco: Jossey-Bass.
• Kaplan, E. (1988). A process approach to neuropsychological assessment. In Clinical neuropsychology and brain function: Research, measurement, and practice (pp. 127–167). Washington, DC: American Psychological Association.
• Kircher, T.T.J., Liddle, P.F., Brammer, M.J., Williams, S.C.R., Murray, R.M., & McGuire, P.K. (2001). Neural correlates of formal thought disorder in schizophrenia: Preliminary findings from a functional magnetic resonance imaging study. Archives of General Psychiatry, 58, 769–774.
• Knapp, N.M. (1994). Research with diagnostic drawings for normal and Alzheimer’s subjects. Art Therapy, 11, 131–138.
• Kohen-Raz, R., & Assael, M. (1966). EEG and Rorschach findings in a group of juvenile delinquents suspect of organic brain disorder. Acta Paedopsychiatrica, 33(8), 251–258.
• Koykis, F. (1985). Ve´cu corporel et inte´gration institutionnelle et sociale chez des adolescents handicape´s. Revue Belge de Psychologie et de Pe´dagogie, 47(189), 9–15.
• Leon-Carrion, J., De Serdio-Arias, M.L., Cabezas, F.M., Roldan, J.M., Dominguez-Morales, R., Martin, J.M., et al. (2001). Neurobehavioural and cognitive profile of traumatic brain injury patients at risk for depression and suicide. Brain Injury, 15, 175– 181.
• Lezak, M.D. (1995). Neuropsychological assessment (3rd ed.). New York: Oxford University Press.
• Light, B.H., & Amick, J.H. (1956). Rorschach responses of normal aged. Journal of Projective Techniques, 20, 185–195.
• Loevinger, J., & Wessler, R. (1970). Measuring ego development. San Francisco: Jossey-Bass.
• Malmgren, H., Bilting, M., Frobarj, G., & Lindqvist, G. (1997). A longitudinal pilot study of the Rorschach as a neuropsycholog-

ical instrument. In A.M. Carlsson, et al. (Eds.), Research into Rorschach and projective methods (pp. 117–139). Stockholm: Swedish Rorschach Society.

• Mattis, S. (1973). Dementia Rating Scale. Odessa, FL: Psychological Assessment Resources.
• Mattlar, C.-E., Carlsson, A., Forsander, C., & Karppi, S.-L. (1992). Rorschach and old age: Personality characteristics for a group of physically fit 80-year-old men. British Journal of Projective Psychology, 37(2), 41–51.
• Mattlar, C.-E., Knuts, L., & Virtanen, E. (1985). Personality structure on the Rorschach for a group of healthy 71-year-old females and males. Projective Psychology, 30, 3–8.
• Milberg, W.P., Hebben, N., & Kaplan, E. (1996). The Boston process approach to neuropsychological assessment. In I. Grant & K.M. Adams (Eds.), Neuropsychological assessment of neuropsychiatric disorders (2nd ed., pp. 58–101). New York: Oxford University Press.
• Minassian, A. (2001). Information processing, thought disturbance, and schizophrenia: The relationship between pupil dilation, eye movements, and Rorschach responses. Dissertation Abstracts International, 62, 1590B.
• Murray, H.A. (1963). Thematic Apperception Test manual. Cambridge, MA: Harvard University Press.
• Muzio, E., & Luperto, L. (1999). De´mence et fonctionnement de la personnalite´ a` travers le Rorschach chez un groupe de femmes age´es hospitalise´es. European Review of Applied Psychology/ Revue Europe´enne de Psychologie Applique´e, 49, 227–236.
• Nakamura, H. (1983). Rorschach test on organic brain disease: Comparing with schizophrenics. Kyushu Neuro-psychiatry, 29, 93–103.
• Nedelcu, A., & Zellingher, R. (1971). [Comparative results obtained with the organic integrity test in recent craniocerebral injuries]. Annales Me´dico-Psychologiques, 2, 521–526.
• Panek, P.E., & Wagner, E.E. (1993). Hand Test characteristics of dual diagnosed mentally retarded older adults. Journal of Personality Assessment, 61, 324–328.
• Parker, R.S. (1996). The spectrum of emotional distress and personality changes after minor head injury incurred in a motor vehicle accident. Brain Injury, 10, 287–302.
• Peled, A., & Geva, A.B. (2000). The perception of Rorschach inkblots in schizophrenia: A neural network model. International Journal of Neuroscience, 104, 49–61.
• Perry, W., & Braff, D.L. (1998). A multimethod approach to assessing perseverations in schizophrenia patients. Schizophrenia Research, 33, 69–77.
• Perry, W., Felger, T., & Braff, D. (1998). The relationship between skin conductance hyporesponsivity and perseverations in schizophrenia patients. Biological Psychiatry, 44, 459–465.
• Perry, W., & Potterat, E. (1997). Beyond personality assessment: The use of the Rorschach as a neuropsychological instrument in

patients with amnestic disorders. In J.R. Meloy, M.W. Acklin, C.G. Gacono, J.F. Murray, & C.A. Peterson (Eds.), Contemporary Rorschach interpretation (pp. 557–575). Mahwah, NJ: Erlbaum.

• Perry, W., Potterat, E., Auslander, L., Kaplan, E., & Jeste, D. (1996). A neuropsychological approach to the Rorschach in patients with dementia of the Alzheimer’s type. Assessment, 3, 351–363.
• Perry, W., & Viglione, D.J. (1991). The Ego Impairment Index as a predictor of outcome in melancholic depressed patients treated with tricyclic antidepressants. Journal of Personality Assessment, 56, 487–501.
• Piotrowski, Z. (1936). On the Rorschach method and its application in organic disturbances of the central nervous system. Rorschach Research Exchange, 1, 23–39.
• Reitan, R.M. (1955). Evaluation of the postconcussion syndrome with the Rorschach Test. Journal of Nervous and Mental Disease, 121, 463–467.
• Reitan, R.M., & Wolfson, D. (1997). Emotional disturbances and their interaction with neuropsychological deficits. Neuropsychology Review, 7, 3–19.
• Rorschach, H. (1942). Psychodiagnostics: A diagnostic test based on perception. Bern, Switzerland: Hans Huber.
• Rutter, M. (1985). Resilience in the face of adversity: Protective factors and resistance to psychiatric disorder. British Journal of Psychiatry, 147, 598–611.
• Sampaio, A. (1977). Temporal lobe psychopathology. Neurobiologia, 40, 27–40.
• Sandson, J., & Albert, M.L. (1984). Varieties of perseveration. Neuropsychologia, 22, 715–732.
• Sattler, J. (1982). Assessment of children’s intelligence and special abilities (2nd ed.). Boston: Allyn & Bacon.
• Shapiro, E., Shapiro, A.K., & Clarkin, J. (1974). Clinical psychological testing in Tourette’s syndrome. Journal of Personality Assessment, 38, 464–478.
• Shimonaka, Y., & Nakazato, K. (1991). Aging and terminal changes in Rorschach responses among the Japanese elderly. Journal of Personality Assessment, 57, 10–18.
• Sinacori, D.R. (2000). Depression in a brain-injured sample: An investigation of indicators on the Rorschach and MMPI-2. Dissertation Abstracts International, 60, 4251B.
• Solovay, M.R., Shenton, M.E., Gasperetti, C., & Coleman, M. (1986). Scoring manual for the Thought Disorder Index. Schizophrenia Bulletin, 12, 483–496.
• Tamkin, A.S., & Hyer, L.A. (1983). Defensiveness in psychiatric elderly persons: Fact or fiction. Psychological Reports, 52, 455–458.
• Upadhyaya, S., & Sinha, A.K. (1974). Some findings on psychodiagnostic tests with young retarded adults. Indian Journal of Clinical Psychology, 1, 73–79.

552 The Use of Projective Tests in Assessing Neurologically Impaired Populations

• Viglione, D.J. (1999). A review of recent research addressing the utility of the Rorschach. Psychological Assessment, 11, 251– 265.
• Vitale, J.H., Polus, S.M., Wollitzer, A.O., & Steinhelber, J.C. (1974). Relationships of psychological dimensions to impairment in a population with cerebrovascular insufficiency. Journal of Nervous and Mental Disease, 158, 456–467.
• Wagner, E.E., & Maloney, P. (1980). Efficacy of three projective techniques in differentiating brain damage among subjects with normal IQs. Journal of Clinical Psychology, 36, 968–973.
• Wagner, E.E., Rasch, M.A., & Marsico, D.S. (1990). Hand Test characteristics of severely behavior handicapped children. Journal of Personality Assessment, 54, 802–806.
• Zillmer, E.A. (1994, April). The neuropsychology of personality I: Recent research. Paper presented at the annual meeting of the Society for Personality Assessment, Chicago, IL.
• Zillmer, E.A., & Perry, W. (1996). Cognitive-neuropsychological abilities and related psychological disturbance: A factor model of neuropsychological, Rorschach, and MMPI indices. Assessment, 3, 209–224.

Projective Assessment of Malingering

JON D. ELHAI, BILL N. KINDER, AND B. CHRISTOPHER FRUEH

RORSCHACH INVESTIGATIONS IN THE DETECTION OF MALINGERING 553 Early Rorschach Malingering Studies 554 Rorschach Studies of Malingering Since 1980 554 OTHER PROJECTIVE ASSESSMENTS USED TO DETECT MALINGERING 556 Thematic Apperception Test (TAT) 556 Group Personality Projective Test (GPPT) 557

According to the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV), malingering involves the conscious fabrication or exaggeration of physical and/or psychological symptoms, with the desire to achieve an external goal (American Psychiatric Association, 1994). While the prevalence of malingering is not known with precise accuracy, estimates have ranged from 7% to 8% in outpatient settings and 16% to 17% in forensic settings (Rogers, Salekin, Sewell, Goldstein, & Leonard, 1998; Rogers, Sewell, & Goldstein, 1994).

Since the 1930s, projective personality tests have been used in a number of empirical studies of malingering. The majority of these investigations have used the Rorschach Inkblot test. In addition, a very limited number of studies have used the Thematic Apperception Test (TAT), the Group Personality Projective Test (GPPT), and the 136-item Sentence Completion Test (SCT-136). Most projective studies of malingering have used simulation research designs, comparing individuals instructed to simulate some form of maladjustment with individuals instructed to answer honestly or simulate a good impression on a test. Very few projective malingering studies have used known-groups comparison methodology (comparing individuals judged to be malingerers with those judged to be honest reporters) or differential prevalence designs (comparing individuals who have a large incentive to malinger with those who do not, e.g., disability and nondisability claimants).

Malingering research using projective personality tests is important because these measures do not include validity

Sentence Completion Test-136 (SCT-136) 557 SUMMARY 557 RESEARCH AND METHODOLOGICAL ISSUES 558 MINIMIZATION OR CONCEALMENT OF PSYCHOPATHOLOGY 559 IMPLICATIONS FOR CLINICAL PRACTICE 559 REFERENCES 559

scales that assess the examinee’s response set. Therefore, investigating overreported response sets and providing indices of symptom exaggeration in projective measures would be extremely helpful. This chapter reviews the empirical research on projective personality assessment in the detection of malingering. First, early research utilizing the Rorschach to detect malingering will be presented, followed by a review of more recent Rorschach studies. Next, malingering research on other projective personality measures will be reviewed. Last, methodological issues in the investigation of malingering with projective assessments will be considered. Although few studies have been conducted on this issue since Perry and Kinder (1990) and Schretlen (1997) published their literature reviews, the aim of the present review is to comprehensively discuss both the relevant literature and methodological issues in the projective assessment of malingering. It should be emphasized that because this chapter reviews the projective research on malingering detection, the focus is therefore on studies that have assessed exaggerated or fabricated psychopathology; this chapter does not explore in detail the studies that have assessed the minimization or concealment of psychopathology.

RORSCHACH INVESTIGATIONS IN THE DETECTION OF MALINGERING

First, we discuss the early malingering studies using the Rorschach.

Early Rorschach Malingering Studies

The early Rorschach malingering studies primarily used simulation designs with repeated measures analyses, comparing individuals’ test responses under standard versus malingering instruction sets. Fosberg (1938, 1941, 1943) conducted the earliest malingering studies using the Rorschach. Fosberg (1938) was interested in exploring the Rorschach’s reliability. Using a repeated measures simulation design, he instructed two participants to complete the test under four conditions: (1) standard instructions, (2) instructions to make the best impression, (3) instructions to make the worst possible impression, and (4) instructions to look for determinants. He found no significant differences in chi-square analyses between conditions for variables involving location, determinants, or content, concluding that the Rorschach cannot be feigned.

Next, Fosberg (1941) examined Rorschach data from 129 male and female college professors and students, under a variety of standard, fake good, and fake bad instructional sets. Using a similar repeated measures design, resulting pairwise correlations demonstrated a high degree of association between conditions, and Fosberg again concluded that the test could not be feigned. In a subsequent study, Fosberg (1943) explored the manner in which his participants feigned on the Rorschach and concluded that if his “test-wise” collegeaffiliated participants could not effectively feign the Rorschach, “Rorschach naı¨ve” participants would not be able to malinger more effectively. However, Cronbach (1949) sharply criticized Fosberg’s studies for his use of inappropriate statistical analyses, thus compromising the studies’ conclusions.

During the 1940s, clinical observations of malingering on the Rorschach were noted by Benton (1945) and Hunt (1946). Both authors noted that signs of malingering may include a small number of responses (R), and frequent card rejections. In addition, Benton (1945) noted a slower reaction time, an attitude of perplexity and pained compliance with testing, and a smaller number of Popular (P) responses, while Hunt (1946) noted perseveration and a larger number of P responses. However, no statistical analyses were conducted to test these hypotheses. Although no control group was included, Rosenberg and Feldberg (1944) provided malingering signs of 93 soldiers suspected of malingering, including decreased R, vague form, increased P, perseveration, delayed response time, card rejections, and repeated questioning about test directions.

In the 1950s, two notable Rorschach malingering studies were conducted by Carp and Shavzin (1950) and Feldman and Graley (1954). Carp and Shavzin (1950) implemented a repeated measures simulation design and first instructed 20 male undergraduate students to feign a bad impression in order to avoid military conscription, and then a good impression in order to be released from a psychiatric hospital, on two successive Rorschach administrations. An additional 20 male students were given the same instructions in reverse order. Analyzing a number of Rorschach variables and indices, the only between-group difference found was that of organization (Z). While inappropriately using a chi-square analysis, they did find that some participants scored differently across test conditions.

Next, Feldman and Graley (1954) instructed two groups of undergraduate participants (N ” 72) to complete the Rorschach (group administration) while simulating the worst possible impression, while a subgroup of 30 participants were asked on a previous occasion to take the test under standard instructions. Using Kloper’s scoring method, 35 standard Rorschach variables were examined, and under simulation conditions participants yielded significantly lower P, but significantly more responses involving inanimate movement (m), color-form plus pure color (CF # C), form-color (FC), and sex-anatomy. However, group administrations may not be generalizable to individual administrations because of differences in their factor structures (Shaffer, Duszynski, & Thomas, 1981). Several qualitative signs were found to be prevalent in the experimental conditions but not in the control conditions, including self-references, aggressive and sexual references, symmetry remarks, absurd content, dysphoria, and expressions of personal feelings. Strategies reported for feigning included: (1) avoiding the normal response, (2) using sexual responses, (3) endorsing symptoms of maladjustment, (4) having a specific mental disorder in mind, (5) having a nonspecific mental disorder in mind, and (6) emphasizing aggressive and gory components.

More than a decade later, Easton and Feigenbaum (1967) used a repeated measures simulation design, administering the Rorschach to 11 college students under standard instructions on two different occasions. An additional 11 students were asked to complete the Rorschach under standard instructions and then again with instructions to simulate an unfavorable impression in order to avoid military conscription. The authors found that under simulation instructions, of 12 standard Rorschach variables, significantly lower scores were obtained on common detail responses (D), form (F), Obj, P, and R. However, test repetition alone resulted in decreases in whole animal (A) and P, and increases in D, whole human (H), animal detail (Ad), and R.

Rorschach Studies of Malingering Since 1980

Rorschach malingering studies changed significantly after 1980. Specifically, most researchers began including patient comparison groups, and many instructed participants to simulate specific types of psychopathology, rather than giving general directions to simulate a good or bad impression. Additionally, known-groups comparison designs began appearing in the literature.

Bash and Alpert (1980) conducted the first known-groups comparison of malingering on the Rorschach, examining additional measures as well. The authors compared four groups of prisoners (N ” 120): (1) prisoners judged to be malingering auditory hallucinations, (2) prisoners diagnosed with schizophrenia and suffering from genuine auditory hallucinations, (3) prisoners diagnosed with schizophrenia who were not reporting auditory hallucinations, and (4) nonpsychotic prisoners. The authors found that an index of malingering, consisting of variables including R, number of cards rejected, reaction time, P, number of Easy P responses, perseveration, ratio of human movement to pure color (M:C), aggressive responses, C%, animal and inanimate movement responses, bizarre responses, Percent of Whole Locations (W%), and aspirational ratio (W:M) discriminated groups. However, no information was provided on the accuracy of these variables individually.

Seamons, Howell, Carlisle, and Roe (1981) presented a counterbalanced, repeated measures simulation study of 48 male forensic patients. The authors asked nonschizophrenic, latent schizophrenic, residual schizophrenic, and schizophrenicpsychotic patients (n ” 12 per group) to fake welladjustment, and then malinger mental illness on subsequent administrations of the Rorschach. Results demonstrated that under instructions to feign mental illness, of 48 variables explored, patients only obtained fewer P, significantly more dramatic responses (i.e., depression, sex, blood, gore, mutilation, confusion, hatred, fighting, and decapitation themes), higher experience potential (ep), and more inappropriate combination responses.

Albert, Fox, and Kahn (1980) conducted one of the earliest studies with coached and uncoached simulators. They administered the Rorschach with standard instructions to six inpatients diagnosed with paranoid schizophrenia and six college students, while providing instructions to feign paranoid schizophrenia to six college students coached on paranoid schizophrenia and six students who did not receive such coaching. After soliciting fellows of the Society of Personality Assessment, 46 fellows judged the protocols. Findings revealed that based on Rorschach protocols, the fellows diagnosed uncoached simulators and schizophrenic patients with psychosis equally, while the coached simulators were diagnosed with psychosis more frequently. Additionally, judges reported being equally confident of their diagnoses across groups.

Pettigrew, Tuma, Pickering, and Whelton (1983) used a between-groups simulation design, testing 62 college students instructed to simulate psychosis, while providing honest instructions to 75 students, 29 civilly committed psychotic patients, and 26 forensic psychotic inpatients on a multiplechoice group format of the Rorschach. The multiple-choice version contained all of the blots reproduced on one page, with possible answer choices corresponding to good form with bizarre wording, good form without bizarre wording, poor form without bizarre wording, and poor form without elaboration. The advantage of the multiple-choice format is that it eliminates the confounding effects of between-condition differences in R found in many Rorschach malingering studies (since differences in R affect a number of standard Rorschach indices). Results demonstrated that simulators gave significantly more good form with bizarre wording responses.

Meisner (1988) conducted a between-groups simulation study using the Rorschach, the first study of its kind to use a cash incentive in order to motivate simulators to convincingly malinger. He used 29 nondepressed undergraduate students under standard instructions and provided training on depression to 29 nondepressed undergraduates instructed to feign severe depression, with a $50 incentive for the most convincing feigning performance. The investigator also used the Beck Depression Inventory (BDI) in order to verify that participants were able to convincingly feign depression. He analyzed a number of Rorschach variables related to depression and anxiety, and although lacking a patient comparison group, the author found that simulators scored higher on morbid content (MOR), and lower on blood (Bl) and R.

Several studies in the 1990s investigated malingered psychosis on the Rorschach. Netter and Viglione (1994) compared honest Rorschach protocols of 20 schizophrenic inpatients and 20 nonpatients with those of 20 nonpatients instructed to simulate schizophrenia. Simulators were provided training on schizophrenia and were told that successful simulation would result in the reward of two movie tickets. The authors were interested in numerous standard Rorschach variables, as well as modified responses (i.e., responses that have a circumstantial quality, those that call attention to personal distress or the bizarreness of a response, or spoiling or modifying a response with poor form quality). Compared to schizophrenic patients, simulators had longer reaction times. Patients and simulators did not differ on the schizophrenia index (SCZI), with nearly half of the simulators scoring 4 or higher on SCZI, indicating a moderate likelihood of schizophrenia. When eliminating modified responses from SCZI, simulators did differ significantly from patients. Additionally, simulators scored lower on modified distorted form (MOD X!%) than schizophrenic patients.

Perry and Kinder (1992) also investigated the simulation of schizophrenia on the Rorschach using a between-groups simulation design. The authors provided 20 undergraduate males with a description of schizophrenia and instructed them to simulate the disorder on the Rorschach, without making their simulation appear obvious. Their protocols were compared with a control group of 20 male undergraduates instructed to complete the test under standard directions. Nonparametric analyses were conducted on Rorschach variables that lacked normal distributions. After finding that the simulation group produced a significantly smaller number of responses, the authors attempted to control for these differences by equating groups on R. Although lacking a schizophrenia patient comparison group, Perry and Kinder demonstrated that of a number of variables, the simulators yielded a significantly greater reaction time, more dramatic responses, and scored higher on the weighted sum of special scores (WSum6), SCZI, distorted form (X!%), and human responses with poor form (M!). However, simulators scored lower on P, Easy P, and conventional form (X#%).

Ganellen, Wasyliw, Haywood, and Grossman (1996) conducted a known-groups comparison of malingered psychosis on the Rorschach. The authors tested 48 forensic patients referred for competency and/or sanity evaluations and divided them into groups based on Minnesota Multiphasic Personality Inventory (MMPI) F scale T-scores of greater than 90 (“malingered” group), and 90 or less (“honest” group). Results indicated that from a variety of Rorschach variables, only dramatic content differed significantly between groups, with the “malingered” group scoring higher. However, this study relied solely on F scale scores to classify individuals as malingerers and honest reporters, and a T-score cutoff of 90 may result in classifying too many honest responders as malingerers (Greene, 1997).

Simulated depression on the Rorschach was investigated by Caine, Kinder, and Frueh (1995). Using a between-groups simulation design, the authors instructed 20 depressed (nonpsychotic) female inpatients and 20 nondepressed female undergraduates to complete the Rorschach under standard instructions. An additional 20 nondepressed female undergraduates were provided a description of depression and were instructed to simulate the disorder on the Rorschach, without making their simulation appear too obvious. The authors used the MMPI-2’s Depression (DEP) scale as a manipulation check to ensure that simulators had the ability to feign depression. Nonparametric analyses were used for nonnormally distributed Rorschach variables. R was not revealed to differ between groups. Using a variety of Rorschach depressionrelated variables, the authors did not find any significant differences between simulators and patients, revealing only that simulators and patients scored significantly higher on MOR than controls.

Frueh and Kinder (1994) explored simulated combat-related post-traumatic stress disorder (PTSD) in a between-groups simulation design. They instructed 20 male undergraduates and 20 male Vietnam combat veterans diagnosed with PTSD to complete the Rorschach in an open and honest manner. An additional group of 20 male undergraduates were provided with information about combat PTSD and were asked to feign combat PTSD on the Rorschach, with a cash incentive promised to successful simulators. Groups did not differ on R. Nonparametric analyses were implemented for nonnormally distributed variables. Of the standard scores and the dramatic score, simulators scored significantly higher than patients and controls on dramatic, pure color total (Sum C), CF # C, and X!%. Simulators scored higher than patients, but not higher than controls, on M!. Simulators scored lower than controls on Lambda (the ratio of pure form responses with all other responses) and X#%.

OTHER PROJECTIVE ASSESSMENTS USED TO DETECT MALINGERING

Clearly, a number of studies have explored malingering using the Rorschach. Next, we discuss the handful of malingering studies that have employed other types of projective tests.

Thematic Apperception Test (TAT)

Several studies have used the TAT to detect malingering. Kaplan and Eron (1965) conducted a between-groups simulation study with 36 undergraduate students who were naive to projective testing and were instructed either to complete the TAT with standard directions or with directions to simulate being an aggressive and hostile person. An additional 36 graduate students in clinical psychology (with projective test experience) were instructed to complete the TAT with either standard or aggressive/hostile directions. Although no patient comparison group was implemented, results indicated that of six TAT variables, simulators scored significantly lower on emotional tone (ET) and outcome (O), and scored higher on aggressive themes. Additionally, graduate students with projective training obtained significantly more unusual formal characteristics (UFC) than naive subjects when under simulation instructions.

Hamsher and Farina (1967) implemented a between-groups simulation design with the TAT. The authors instructed 61 undergraduate students to complete the TAT (abbreviated version) with instructions to either simulate openly seeking psychological treatment or simulate being guarded in completing the test. On TAT ratings of openness, subjects in the open condition were rated as significantly more open than those in the guarded condition (especially for the blank card). For females, the open and guarded groups differed more than that seen in males.

Holmes (1974) conducted the most recent malingered TAT study, presenting findings from two simulation experiments. In Experiment #1, 29 undergraduate students were asked to respond honestly to the TAT, and then in a subsequent test administration they were instructed to simulate high achievement and motivation. An additional 31 students were given the same instructions but in reverse order. Using analyses of variance, Holmes found significant between- and withingroup differences, concluding that subjects could alter their responses accordingly on the TAT. In Experiment #2, the author instructed 27 students to complete the TAT honestly on two successive test administrations. Thirty-one students were instructed to first complete the TAT honestly and then with directions to conceal their personalities as much as possible on a subsequent test administration. Based on results indicating a very small degree of association between the honest and faking conditions, the author concluded that the simulation group was able to inhibit their personality projections.

Group Personality Projective Test (GPPT)

Two studies (Brozovich, 1970; Cassel & Brauchle, 1959) have explored malingering on the GPPT. The GPPT is a 90 item multiple-choice test in which the examinee is presented stick figure drawings and must select an answer choice that describes what he or she believes is taking place. It should be noted that neither of these studies included a patient comparison group.

Cassel and Brauchle (1959) implemented a repeated measures simulation design, instructing 50 high school seniors to complete the GPPT on three successive test administrations with (1) standard directions, (2) directions to fake a poor and disturbed personality, and last, (3) directions to feign a good personality with minimal disturbance. Based on t tests, of six of the GPPT’s part scores, when simulating a poor-disturbed personality students scored significantly higher on the tension reduction quotient, withdrawal needs, succorance needs, and total score, while scoring lower on nurturance needs and affiliation-psychosexual needs. Under directions to simulate a good personality, participants were also able to alter their responses.

Brozovich (1970) conducted a similar repeated measures simulation study with the GPPT, including 38 graduate students in introductory group testing courses. The author implemented three successive GPPT administrations with standard directions and counterbalanced directions to either feign welladjustment or emotional disturbance. Using analyses of variance, when faking an emotional disturbance students scored significantly higher on the tension reduction quotient, succorance needs, and total score, while scoring lower on nurturance needs, affiliation-psychosexual needs, and withdrawal needs. When faking well-adjustment, students scored significantly lower on tension reduction quotient, neuroticism needs, and the total score. Thus, it appears that the GPPT is susceptible to malingered response sets.

Sentence Completion Test-136 (SCT-136)

Timmons, Lanyon, Almer, and Curran (1993) developed the SCT-136 and examined the protocols of 51 patients in litigation for personal injury, social security disability, or worker’s compensation claims. From previous literature, the authors developed a scoring system based on signs that distinguish malingerers from nonmalingerers in disability settings. Implementing a factor analysis, three factors were found to represent malingering signs and were named “angry negativity/ no fair deal,” “disability exaggeration/overinvestment,” and “excessive honesty/virtue.” When correlated with MMPI indices of somatization and overreporting, although small correlations were found, the authors claimed that the results “give strong support to the validity of the total malingering score and Factor I, and some support to Factor II” (p. 30).

In the same paper, Timmons et al. (1993) presented crossvalidation results, conducting a simulation experiment with 39 undergraduate students using an abbreviated version of the SCT-136 (called the SCT-39) that represented each factor. The SCT-39 was administered four times. First, students were instructed to simulate successful recovery from an accident. In the second, third, and fourth administrations, students were given a case vignette involving a minor accident in which the victim malingered to collect money. They were read descriptions of the three factors of malingering strategies on the SCT and asked to separately use those individual strategies to simulate on the successive test administrations. Students were successful in altering their factor scores based on the different malingering strategies. However, no patient comparison group was included, thus restricting the conclusions.

SUMMARY

This paper reviewed malingering studies using projective tests, the vast majority of which used the Rorschach. Although the Rorschach studies differed in methodology, scoring systems, and results, some overall conclusions can be made. Of the studies noting clinical signs of malingerers on the Rorschach (Benton, 1945; Hunt, 1946; Rosenberg & Feldberg, 1944), the majority suggests that malingered protocols include a small number of responses (R), frequent card rejections, and significant perseveration. Across the controlled studies, several indices resulted in significant differences between simulator and comparison groups. A number of investigations revealed that simulators obtained higher scores on dramatic responses, reaction time, CF # C, and X!%, while obtaining lower scores on X#%, R, and P. When examining specific types of simulation studies, those involving the simulation of psychosis revealed a trend of higher reaction time and higher dramatic scores, and lower scores on Easy P for simulators. However, the few remaining studies of specific psychopathology do not seem to yield comparable results.

Very few studies found have used projective tests other than the Rorschach in malingering detection. The few studies that have incorporated other projective tests are not easily comparable, as they examined different variables. The exception is the GPPT, for which two studies examined the same variables. Conclusions from these GPPT studies suggest that when simulating maladjustment, simulators tend to score higher on tension reduction, succorance needs, and the total score, while scoring lower on nurturance needs and affiliation-psychosexual needs.

RESEARCH AND METHODOLOGICAL ISSUES

Several methodological issues should be mentioned with regard to projective assessment studies of malingering. First, the great majority of these studies use simulation designs. While simulation designs do provide more experimental rigor in the investigation of malingering, we do not know if simulators’ malingering performances are similar to those of actual malingerers (Iverson, Franzen, & Hammond, 1995). Conducting more known-groups comparison studies should help elucidate this issue. In simulation studies, while malingering experts have encouraged researchers to offer incentives to simulators in order to improve motivation and better approximate the feigning performances of actual malingerers (Rogers, 1997), only the projective studies of Meisner (1988), Netter and Viglione (1994), and Frueh and Kinder (1994) have included incentives. Additionally, many projective studies have included healthy control groups rather than clinical comparison groups, limiting the meaningfulness of results.

Only a few projective malingering studies (Carp & Shavzin, 1950; Easton & Feigenbaum, 1967; Feldman & Graley, 1954; Hamsher & Farina, 1967; Holmes, 1974) have provided simulators with specific malingering scenarios or contexts in which they might malinger, in an attempt to put simulators in a “malingering frame” so they may more effectively simulate and to improve external validity. However, the provision of such scenarios has been criticized as it may likely lead to a decrease in the generalizability of results (Rogers, 1997).

Regarding the type of psychopathology that is malingered, projective test studies have rarely included instructions to feign a specific type of psychiatric disorder. Instead, ambiguous simulation instructions are often given, resulting in less experimental rigor and perhaps reduced generalizability. However, Albert et al. (1980), Caine et al. (1995), Frueh & Kinder (1994), Ganellen et al. (1996), Meisner (1988), Netter and Viglione (1994), Perry & Kinder (1992), and Pettigrew et al. (1983) all provided specific types of psychiatric disorders to be simulated. Of these studies, fewer have provided simulators with training on symptom information specific to the disorder being simulated.

Several projective studies (Caine et al., 1995; Holmes, 1974; Perry & Kinder, 1992) have included instructions to simulators to be believable in their feigning performance, without appearing too obvious as malingerers. However, it is unknown if actual malingerers attempt to attenuate their overreporting style in order to appear believable, as many may simply attempt to feign global and severe psychopathology (Elhai, Gold, Sellers, & Dorfman, 2001).

Last, malingering investigations using projective tests have relied primarily on between-groups statistical tests in order to examine how well groups are separated on a projective test variable. However, none of these studies have used cutting scores to alert clinicians of scores that may indicate malingering. Additionally, discriminant analyses have not been used in projective studies in order to examine the effects of combinations of test variables in predicting malingered response sets (Rogers, 1997).

A note should be made in reference to using projective versus objective personality tests in assessing malingered psychopathology. The objective assessment literature in malingering has grown at a much more rapid rate and has become better established than the projective malingering literature. For example, the MMPI-2 has grown to be considered a standard assessment of malingering (Greene, 1997), with a metaanalysis indicating strong effect sizes for its malingering scales (Rogers, Sewell, & Salekin, 1994). Future projective malingering investigations should therefore explore the incremental validity of projective measures over using objective measures alone.

Several limitations specific to Rorschach malingering studies should be mentioned. First, differences in R across conditions have been found in a number of studies. Since R determines a number of other scores from the Rorschach Comprehensive System, it is important to either match subjects on R or to statistically control for the confounding effects of R, which few studies have done (Perry & Kinder, 1990). Second, while Rorschach variables typically involve nonnormal data distributions (Perry & Kinder, 1990), many Rorschach malingering studies have inappropriately used parametric tests in detecting malingering. However, in many cases, nonparametric tests would be more appropriate.

Future empirical studies using projective tests in detecting malingering should implement several strategies. First, simulation studies should continue using improved experimental rigor, by including patient comparison groups instead of solely using within-groups fake-honest conditions. Second, simulators should be instructed and trained to simulate specific psychiatric disorders. Specifically, disorders other than schizophrenia and psychotic disorders should be implemented, as these disorders are overrepresented by previous projective simulation studies (Perry & Kinder, 1990). Future studies should also implement cutting score analysis to examine the sensitivity and specificity of projective test variables in simulation detection. Discriminant analysis should be used to assess the effects of combinations of test variables in predicting simulated from honest clinical response sets. In terms of Rorschach studies, R should be controlled for, and nonparametric statistical tests should be used in situations where normal data are not present.

MINIMIZATION OR CONCEALMENT OF PSYCHOPATHOLOGY

One important issue should be addressed. As stated earlier, this chapter’s main focus was to review the research assessing malingered response sets on projective personality tests and to discuss methodological considerations. However, related to the topic of malingering, and under the rubric of dissimulation, is the issue of minimizing or concealment of psychopathology. Minimizing psychopathology is similar to malingering, in that both represent a form of misrepresenting one’s emotional condition. Like malingering, the issue of minimized psychopathology is of great interest in clinical and forensic settings, and it may frequently occur in such settings as employment screenings, civil competency evaluations, and child custody evaluations, just to name a few. Thus from a clinical standpoint, while being able to detect genuinely healthy individuals who have overreported their psychiatric symptoms is important, it is also crucial to be able to detect genuinely disturbed patients who have underreported their psychiatric symptoms.

In recent years, several papers have explored whether projective tests can effectively detect the minimization of psychopathology. These studies examined Rorschach variables in conjunction with MMPI/MMPI-2 validity scales in participants with a motivation to conceal their emotional problems. However, results from these investigations are mixed with regard to the Rorschach’s ability to detect the minimization of psychopathology. Two of these studies revealed lower test scores on the MMPI but not the Rorschach for individuals with an incentive to minimize their emotional problems (Ganellen, 1994; Grossman, Wasyliw, Benn, & Gyoerkoe, 2002), suggesting that the MMPI-2 may be more susceptible to the minimizing of psychopathology. However, another study found no advantage in using the Rorschach over the MMPI in detecting such a response trend (Wasyliw, Benn, Grossman, & Haywood, 1998). The interested reader is referred to these articles for more information on this topic.

IMPLICATIONS FOR CLINICAL PRACTICE

At the present time, no cutoff scores are available for projective tests in discriminating malingered from genuine protocols. Therefore, it would be inadvisable to provide clinical decision rules of malingering to clinicians who use projective tests. Schretlen (1997) also concluded that a clinical decision rule for conclusively detecting malingering using projective tests is inappropriate. However, he did provide signs that (although not conclusive) may arouse a suspicion of malingering when using the Rorschach, including a context in which the examinee has an incentive to malinger; Rorschach malingering signs involving few responses and/or frequent card rejections; marked paucity of popular responses; numerous dramatic, morbid, or bizarre responses; and repeated questions about the purpose of testing or pained compliance with testing. Last, it should be emphasized that there is no gold standard for detecting malingered psychopathology. Therefore, rather than using one assessment measure, numerous instruments should be used, including objective testing, clinical interviews, and physiological assessments, and the convergence of data should be used to speculate about possible malingering.

REFERENCES

• Albert, S., Fox, H.M., & Kahn, M.W. (1980). Faking psychosis on the Rorschach: Can expert judges detect malingering? Journal of Personality Assessment, 44, 115–119.
• American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author.
• Bash, I.Y., & Alpert, M. (1980). The determination of malingering. Annals of the New York Academy of Sciences, 347, 87–99.

560 Projective Assessment of Malingering

• Benton, A.L. (1945). Rorschach performances of suspected malingerers. Journal of Abnormal and Social Psychology, 40, 94–96.
• Brozovich, R. (1970). Fakability of scores on the Group Personality Projective Test. Journal of Genetic Psychology, 117, 143–148.
• Caine, S.L., Kinder, B.N., & Frueh, B.C. (1995). Rorschach susceptibility to malingered depressive disorders in adult females. In J.N. Butcher & C.D. Spielberger (Eds.), Advances in personality assessment, Vol. 10 (pp. 165–174). Hillsdale, NJ: Erlbaum.
• Carp, A.L., & Shavzin, A.R. (1950). The susceptibility to falsification of the Rorschach psychodiagnostic technique. Journal of Consulting Psychology, 14, 230–233.
• Cassel, R.N., & Brauchle, R.P. (1959). An assessment of the fakability of scores on the Group Personality Projective Test. Journal of Genetic Psychology, 95, 239–244.
• Cronbach, L.J. (1949). Statistical methods applied to Rorschach scores: A review. Psychological Bulletin, 46, 393–429.
• Easton, K., & Feigenbaum, K. (1967). An examination of an experimental set to fake the Rorschach test. Perceptual and Motor Skills, 24, 871–874.
• Elhai, J.D., Gold, S.N., Sellers, A.H., & Dorfman, W.I. (2001). The detection of malingered posttraumatic stress disorder with MMPI-2 fake bad indices. Assessment, 8, 221–236.
• Feldman, M.J., & Graley, J. (1954). The effects of an experimental set to simulate abnormality on group Rorschach performance. Journal of Projective Techniques, 18, 326–334.
• Fosberg, I.A. (1938). Rorschach reactions under varied instructions. Rorschach Research Exchange, 3, 12–30.
• Fosberg, I.A. (1941). An experimental study of the reliability of the Rorschach psychodiagnostic technique. Rorschach Research Exchange, 5, 72–84.
• Fosberg, I.A. (1943). How do subjects attempt fake results on the Rorschach test? Rorschach Research Exchange, 7, 119–121.
• Frueh, B.C., & Kinder, B.N. (1994). The susceptibility of the Rorschach Inkblot test to malingering of combat-related PTSD. Journal of Personality Assessment, 62, 280–298.
• Ganellen, R.J. (1994). Attempting to conceal psychological disturbance: MMPI defensive response sets and the Rorschach. Journal of Personality Assessment, 63, 423–437.
• Ganellen, R.J., Wasyliw, O.E., Haywood, T.W., & Grossman, L.S. (1996). Can psychosis be malingered on the Rorschach? An empirical study. Journal of Personality Assessment, 66, 65–80.
• Greene, R.L. (1997). Assessment of malingering and defensiveness by multiscale personality inventories. In R. Rogers (Ed.), Clinical assessment of malingering and deception (2nd ed.; pp. 169– 207). New York: Guilford Press.
• Grossman, L.S., Wasyliw, O.E., Benn, A.F., & Gyoerkoe, K.L., (2002). Can sex offenders who minimize on the MMPI conceal psychopathology on the Rorschach? Journal of Personality Assessment, 78, 484–501.
• Hamsher, J.H., & Farina, A. (1967). “Openness” as a dimension of projective test responses. Journal of Consulting Psychology, 31, 525–528.
• Holmes, D.S. (1974). The conscious control of thematic projection. Journal of Consulting and Clinical Psychology, 42, 323–329.
• Hunt, W.A. (1946). The detection of malingering: A further study. Naval Medical Bulletin, 46, 249–254.
• Iverson, G.L., Franzen, M.D., & Hammond, J.A. (1995). Examination of inmates’ ability to malinger on the MMPI-2. Psychological Assessment, 7, 118–121.
• Kaplan, M.F., & Eron, L.D. (1965). Test sophistication and faking in the TAT situation. Journal of Projective Techniques and Personality Assessment, 29, 498–503.
• Meisner, S. (1988). Susceptibility of Rorschach distress correlates to malingering. Journal of Personality Assessment, 52, 564–571.
• Netter, B.E.C., & Viglione, D.J. (1994). An empirical study of malingering schizophrenia on the Rorschach. Journal of Personality Assessment, 62, 45–57.
• Perry, G.G, & Kinder, B.N. (1990). The susceptibility of the Rorschach to malingering: A critical review. Journal of Personality Assessment, 54, 47–57.
• Perry, G.G., & Kinder, B.N. (1992). Susceptibility of the Rorschach to malingering: A schizophrenia analogue. In C.D. Spielberger & J.N. Butcher (Eds.), Advances in personality assessment, Vol. 9 (pp. 127–140). Hillsdale, NJ: Erlbaum.
• Pettigrew, C.G., Tuma, J.M., Pickering, J.W., & Whelton, J. (1983). Simulation of psychosis on a multiple-choice projective test. Perceptual and Motor Skills, 57, 463–469.
• Rogers, R. (1997). Researching dissimulation. In R. Rogers (Ed.), Clinical assessment of malingering and deception (2nd ed.; pp. 398–426). New York: Guilford Press.
• Rogers, R., Salekin, R.T., Sewell, K.W., Goldstein, A., & Leonard, K. (1998). A comparison of forensic and nonforensic malingerers: A prototypical analysis of explanatory models. Law and Human Behavior, 22, 353–367.
• Rogers, R., Sewell, K.W., & Goldstein, A. (1994). Explanatory models of malingering: A prototypical analysis. Law and Human Behavior, 18, 543–552.
• Rogers, R., Sewell, K.W., & Salekin, R.T. (1994). A meta-analysis of malingering on the MMPI-2. Assessment, 1, 227–237.
• Rosenberg, S.J., & Feldberg, T.M. (1944). Rorschach characteristics of a group of malingerers. Rorschach Research Exchange, 8, 141–158.
• Schretlen, D.J. (1997). Dissimulation on the Rorschach and other projective measures. In R. Rogers (Ed.), Clinical assessment of malingering and deception (2nd ed.; pp. 208–222). New York: Guilford Press.
• Seamons, D.T., Howell, R.J., Carlisle, A.L., & Roe, A.V. (1981). Rorschach simulation of mental illness and normality by psychotic and nonpsychotic legal offenders. Journal of Personality Assessment, 45, 130–135.
• Shaffer, J.W., Duszynski, K.R., & Thomas, C.B. (1981). Orthogonal dimensions of individual and group forms of the Rorschach. Journal of Personality Assessment, 45, 230–239.
• Timmons, L.A., Lanyon, R.I., Almer, E.R., & Curran, P.J. (1993). Development and validation of sentence completion test indices of malingering during examination for disability. American Journal of Forensic Psychology, 11, 23–38.
• Wasyliw, O.E., Benn, A.F., Grossman, L.S., & Haywood, T.W., (1998). Detection of minimization of psychopathology on the Rorschach in cleric and noncleric alleged sex offenders. Assessment, 5, 389–397.

References 561

Projective Assessment of Personality in Forensic Settings

JOSEPH T. MCCANN

ADMISSIBILITY OF EXPERT TESTIMONY 563 Legal Standards 563 Professional Standards 564 EVALUATING THE ADMISSIBILITY OF PROJECTIVE ASSESSMENT METHODS 565 Case Law 565 Admissibility Criteria 567

With increasing frequency, psychologists are providing forensic consultation in cases with complex legal issues. In criminal cases, a defendant’s mental state at the time of offense (Shapiro, 1999), competency to stand trial (Grisso, 1988), or risk for future violence (Meloy, 2000) is often the focus. In civil matters, psychologists are often asked to consult on issues pertaining to the presence of psychological injury and trauma (Simon, 1995) or neuropsychological impairment (Doerr & Carlin, 1991). In matters pertaining to families and juveniles, psychologists are often asked to provide opinions about child custody (Gould, 1998), termination of parental rights (Dyer, 1999), child sexual abuse (Kuehnle, 1996), and juvenile offending (Grisso, 1998).

Forensic psychological evaluations require the consultant to rely on multiple sources of data when formulating opinions and drawing conclusions. While traditional psychological assessment methods such as self-report inventories, structured interviews, and projective methods are often used, these methods often lack direct relevance to the legal issue and instead provide information on a defendant or litigant’s psychological functioning that is secondary or only peripherally related to the ultimate issue in a case (Heilbrun, 1992). Often the appropriateness of psychological assessment methods in forensic cases is determined by how well the expert connects testing data to the legal issue (Marlowe, 1995).

The use of projective personality assessment methods in forensic settings has been marked with controversy. In the same way that it is difficult to discuss the reliability and validity of projective techniques without specifying a technique or a particular purpose for which a technique is used, it is also difficult to make general assertions about the approGUIDELINES FOR PROJECTIVE ASSESSMENT IN FORENSIC SETTINGS 569 CONCLUSION 570 REFERENCES 570

priate application of projective methods in forensic settings. Further complicating the issue is the fact that the use of psychological assessment methods in forensic settings must be evaluated in light of both professional and legal standards, yet the law and behavioral sciences often approach issues in divergent ways.

With respect to projective methods, there have been favorable views on use of the Rorschach in forensic settings (McCann, 1998; Meloy, 1991; Meloy, Hansen, & Weiner, 1997). Meloy (1991) noted that in a forensic context the Rorschach “is a scientifically valid and psychologically meaningful task that may contribute to the trier of fact’s ability to answer certain psycholegal questions,” although Meloy noted that the Rorschach is not “dispositive of any criminal or civil legal issue” (p. 232). Likewise, McCann (1998) concluded that the Rorschach meets both legal and professional standards of admissibility for expert testimony. Meloy et al. (1997) conducted a search of appellate court opinions and found that the Rorschach was viewed favorably by most courts and when Rorschach testimony was deemed inadmissible, it was due to invalid inferences being made by the expert witness. Favorable views of forensic applications of projective techniques have dealt primarily with the Rorschach.

More recently, there have been a number of strong objections to the use of projective methods in both clinical and forensic settings. Several concerns have been raised that psychologists will continue to use projective methods despite the purported lack of reliability and validity of these methods (Dawes, 1994). The forensic applications of specific projective methods have also been recently addressed. For instance, Lally (2001) analyzed admissibility of human figure drawings using Heilbrun’s (1992) guidelines for forensic professionals and the legal criteria outlined by the U.S. Supreme Court in Daubert v. Merrell Dow Pharmaceuticals, Inc. (1993). Lally concluded that methods for interpreting human figure drawings meet neither professional nor current legal standards for admissibility because of the lack of standardized scoring methods, weak reliability, questionable relevance, poor normative data, and weak error rates. In addition, Grove and Barden (1999) voiced strong opposition to use of the Rorschach in forensic settings by arguing that it failed a Daubert analysis, in opposition to McCann’s (1998) more favorable analysis.

An extensive analysis of the scientific status of projective methods by Lilienfeld, Wood, and Garb (2000) concluded that there is “ample justification for skepticism concerning most widely used projective techniques” (p. 53). Based on their review of the three most commonly used methods, namely the Rorschach, Thematic Apperception Test, and projective drawings, Lilienfeld and his colleagues drew seven basic conclusions about use of projective methods in the courtroom. More specifically, they concluded that (1) judges and juries should be informed that projective techniques are considered highly controversial in the field of psychology; (2) acknowledgment should be given to the fact that projective methods are susceptible to faking and situational influences; (3) projective methods are often used for purposes that are not supported by research; (4) scoring of projective techniques can be unreliable; (5) norms for projective techniques are often poor, misleading, or lacking; (6) projective methods may be biased against minority groups; and (7) projective techniques may not be admissible under Daubert criteria.

Although these conclusions were formed based on an extensive review of literature and raise important considerations for use of projective methods in forensic settings, they may not necessarily be true of all projective methods. As will be argued later in this chapter, it is not particularly useful to make broad statements about the admissibility or appropriateness of projective methods in forensic settings. Rather, each individual method needs to be evaluated in terms of both the empirical literature and the particular forensic issue that is being addressed. Before outlining general guidelines for using projective methods in forensic settings, it is necessary to first examine the legal standards for appraising admissibility of expert testimony.

ADMISSIBILITY OF EXPERT TESTIMONY

There are two major sources of information that can be used to inform professionals on the admissibility of expert testimony: legal standards and the professional literature.

Legal Standards

For most of the twentieth century, admissibility of expert testimony in the United States was guided by the principles outlined in United States v. Frye (1923). Commonly known as the Frye test, the standard viewed expert testimony as admissible if the theory or procedure forming the basis of the expert’s opinion “was sufficiently established to have gained general acceptance in the field in which it belongs” (p. 1014). The Frye test also came to be known as the “general acceptance” test and was adopted in other federal courts, as well as many state courts. In the 1970s, however, an attempt was made to provide legislation governing all forms of evidence presented in federal courts, resulting in adoption of the Federal Rules of Evidence (FRE; McCann, 1998; McCann, Shindler, & Hammond, 2003).

The original version of FRE 702 governed “Testimony of Experts” and stated: “If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education, may testify thereto in the form of opinion or otherwise.” The basic meaning of FRE 702 is that the admissibility of expert testimony is based on whether such testimony is helpful to the trier of fact. According to FRE 703: “The facts or data in the particular case on which an expert bases an opinion or inference may be those perceived by or made known to the expert at or before the hearing. If of a type reasonably relied upon by experts in the particular field in forming opinions or inferences upon the participant, the facts or data need not be admissible in evidence.” This rule implies that any method might be admissible if it is of a type reasonably relied upon by others in one’s professional community (McCann, 1998).

The Frye “general acceptance” test and FRE “helpfulness” standard were competing legal tests until the U.S. Supreme Court decision in Daubert v. Merrell Dow Pharmaceuticals (1993). The Daubert case involved a dispute over admissibility of expert testimony on the cause of birth defects in mothers who had ingested an antinausea drug during pregnancy. The main question was whether admissibility of expert testimony should be evaluated according to the FRE or Frye test. As such, the Daubert opinion had important ramifications for all forms of expert testimony. The court held that “Nothing in the [Federal] Rules [of Evidence] as a whole or in the text and drafting history of Rule 702, which specifically governs expert testimony, gives any indication that ‘general acceptance’ is a necessary precondition to the admissibility of scientific evidence” (Daubert, p. 2790). In short, the Court held that the FRE superseded Frye and Rule

564 Projective Assessment of Personality in Forensic Settings

702 governs the admissibility of expert testimony in all federal courts. The Daubert opinion also outlined a number of criteria to assist trial judges in determining if a theory or technique is “scientific knowledge”: (1) Has the theory or technique been tested or is it capable of being tested; (2) Has the theory or technique been peer reviewed and published; (3) What is the known or potential rate of error in applying the particular scientific theory or technique; and (4) To what extent has the theory or technique been generally accepted in the relevant scientific community? It is also worth noting that endorsement of the FRE over Frye can be interpreted as favoring a more liberal “helpfulness” standard over what may be perceived as a more conservative “general acceptance” standard. It remains unclear whether the Daubert opinion has made it easier or more difficult for certain forms of scientific testimony to be admitted in federal courts.

Two subsequent U.S. Supreme Court cases have clarified and affirmed Daubert. In General Electric Co. v. Joiner (1997), the Court affirmed the role of the trial judge as “gatekeeper” in making decisions on the admissibility of expert testimony. In Kumho Tire Co., Ltd. v. Carmichael (1999), the Court held that the Daubert standard and FRE apply to all forms of expert testimony, not just scientific expert testimony. Therefore, in federal courts the admissibility of expert testimony is determined by the trial court judge using the standards outlined in the FRE and Daubert opinion. The Joiner and Kumho decisions give trial judges considerable discretion to admit or reject expert testimony; the four Daubert criteria guide determinations of admissibility or judges can use additional or other criteria to evaluate expert testimony (O’Conner & Krauss, 2001).

In response to these recent legal developments, FRE 702 has been revised and now reads: “If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education, may testify thereto in the form of an opinion or otherwise, if (1) the testimony is based upon sufficient facts or data, (2) the testimony is the product of reliable principles and methods, and (3) the witness has applied the principles and methods reliably to the facts of the case” (italicized part added in revision; O’Conner & Krauss, 2001, p. 2). As such, expert testimony based on the use of projective methods will be subjected to an analysis of the reliability of these methods as well as the manner in which the expert has applied his or her data to the relevant legal issue.

While the FRE are the prevailing standard in all federal courts, the Frye test continues to be applicable in many state court jurisdictions (Hamilton, 1998; McCann et al., 2003). Some state appellate courts have explicitly rejected Daubert in favor of Frye, including New York, Florida, Nebraska, and Washington (Hamilton, 1998). In a review of individual state evidentiary rules, Hamilton (1998) noted that 27 states held Daubert to be either helpful or controlling, 11 states rejected Daubert in favor of the Frye test, 5 states rejected Daubert in favor of their own unique standard, and 7 states had not decided the issue. The implication of these findings is that expert witnesses should know the prevailing legal standards in the court where the expert will testify. The FRE, informed by Daubert, will apply in all federal courts; the state rule of evidence (i.e., Frye, FRE/Daubert, or some other test) will apply in state and local courts.

Professional Standards

Aside from case law, professional guidelines have been offered to assist mental health professional in their selection of psychological assessment instruments in forensic settings. Although these standards do not represent formal standards adopted by a professional organization and they do not carry legal weight, they offer assistance in evaluating the applicability of projective methods in legal settings.

Heilbrun (1992) noted that the choice of individual psychological assessment methods should be guided by relevance of a particular method to the legal issue in a particular case. He outlined two types of relevance that should be considered. The first type of relevance pertains to instruments that are a direct measure of some legal construct (e.g., competence to stand trial). The second form of relevance pertains to instruments that measure a psychological construct (e.g., reality testing, impulsivity) that is critical to understanding a psychological construct that is related to a legal standard (e.g., dangerousness, mental illness). This latter form of relevance applies to the application of traditional projective methods in forensic settings. That is, instruments such as inkblot methods, apperception tests, projective drawings, and sentence completion tests should measure a relevant construct or facet of personality that is related to a psychological construct that is present in a legal standard.

According to Heilbrun (1992), the selection of psychological assessment instruments in forensic settings should be guided by seven principles: (1) the test should be commercially available, adequately documented in a manual, and peer reviewed; (2) reliability should be established, with a coefficient of 0.80 advised or explicit justification for lower coefficients; (3) the test should be relevant to the legal issue or some psychological construct underlying the legal issue; (4) the test should have a standard method of administration; (5) the test should be applicable to the population and purpose for which it is used; (6) objective tests and actuarial data applications are preferred; and (7) there should be some method for interpreting test results within the context of response style. These guidelines do not favor use of projective methods in forensic settings. However, it should be noted that actuarial data might be used to establish the appropriateness of a projective method in forensic cases.

Another set of guidelines for selecting psychological tests in forensic settings has been outlined by Marlowe (1995). In a hybrid model using both psychometric principles and legal evidentiary standards, Marlowe provided a flowchart to evaluate seven basic questions when determining if psychometric evidence is admissible. These seven questions are (1) Does the expert possess scientific, technical, or special knowledge that will assist the trier of fact; (2) Is the theory or data collection procedure used by the expert recognized in the field, time tested, falsifiable, and committed to refinement; (3) Does the instrument have adequate reliability, standardized norms, and relevance to the case: (4) Does the instrument have justified norms and standardized administration and scoring procedures; (5) Is data from the instrument relevant, prejudicial, or duplicative; (6) Does data from the instrument violate prevailing social policies; and (7) Are the expert’s conclusions derived from a valid and empirically grounded line of reasoning?

Many of these standards overlap with legal standards, as well as those outlined by Heilbrun (1992). It should be recognized that several of the guidelines outlined by Marlowe reflect psychometric considerations, such as test validity and rate of error when drawing conclusions from test data. Furthermore, some of the questions in his model must be answered by the trial judge, such as whether psychometric data are relevant or prejudicial in a given case.

EVALUATING THE ADMISSIBILITY OF PROJECTIVE ASSESSMENT METHODS

In light of legal and professional standards for evaluating use of psychological assessment methods in forensic cases, an important issue is whether projective methods meet these standards. Answering this critical question is complicated by the fact that one cannot make global statements about projective methods; instead, projective methods must be examined individually. Some projective methods have been subjected to critical analysis (Lally, 2001; McCann, 1998; Meloy, 1991), while other projective methods have not.

A useful place to begin an analysis of projective methods in forensic cases is the issue of their general acceptance in the field of psychology, particularly since general acceptance in the field lies at the heart of the Frye test and remains an important consideration in a Daubert/FRE analysis. Using the criterion of use in clinical and forensic settings as a gauge of general acceptance, some projective methods appear to be widely accepted. Surveys have shown that the Rorschach is utilized by over 80% of mental health agencies (Lubin, Larsen, & Matarazzo, 1984; Lubin, Wallis, & Paine, 1971; Piotrowski & Keller, 1989) and practicing clinical psychologists who are engaged in psychological assessment services (Watkins, Campbell, Nieberding, & Hallmark, 1995). In the survey by Watkins and colleagues (1995), 5 of the 10 instruments most frequently administered by clinical psychologists were projective methods. In forensic settings, various surveys have shown that some projective methods are used frequently. Borum and Grisso (1995) found that the Rorschach was used in 32% of criminal responsibility evaluations and 30% of competency to stand trial evaluations. Ackerman and Ackerman (1997) found that in child custody evaluations, the Rorschach was the second most widely used instrument with adults and the sixth most commonly used instrument with children. Human figure drawings have also been cited as among the more widely used methods in forensic practice (Lally, 2001). Oberlander (1995) sampled mental health professionals in Massachusetts who specialized in child sexual abuse evaluations and found that 54.8% of the sample believed that standard projective tests were useful, 90.3% of the sample believed that children’s drawings were useful, and 87% of the sample believed that play sessions were useful. Overall, it appears that certain projective methods, such as the Rorschach, Thematic Apperception Test, sentence completion tests, and projective drawings, are used frequently in forensic settings. Another method for appraising the appropriateness of projective methods in forensic evaluations is to examine case law to determine how courts view the adequacy of these methods.

Case Law

Despite the proliferation of legal and professional commentary about the admissibility of various psychological assessment methods, it is important to recognize that judges remain the ultimate gatekeepers of whether expert testimony based on projective methods is admitted in court (General Electric Co. v. Joiner, 1997; Kumho Tire Co. v. Carmichael, 1999). As such, the case law must be used to clarify how Daubert and other legal standards will be applied (Grove & Barden, 1999).

Several judicial opinions cite expert testimony that includes reference to projective methods, however no analysis is given on the appropriateness or admissibility of individual methods. For example, some court opinions have made passing reference to expert witnesses using projective drawings, the Thematic Apperception Test, Rorschach Inkblot Method, and incomplete sentence tests without any legal analysis of the general acceptance, scientific adequacy, or admissibility of these methods (e.g., Lucas v. Shalala, 1993; State v. Scott, 1992). In Maddox v. Lord (1991), a federal appeals court held that a criminal defendant was denied effective assistance of legal counsel because expert testimony that tended to support the mitigating defense of extreme emotional disturbance was not presented at trial and thus a new trial was ordered. The expert testimony at issue in Maddox involved the opinion of a psychologist who had employed the Rorschach, Thematic Apperception Test, and several other projective methods; however, the psychologist was unable to testify at trial. Two experts reviewed his diagnostic methods and testified that they were acceptable. Once again, the court did not rule on the admissibility of specific projective methods and focused instead on the failure of the defendant’s attorney to present this expert testimony at trial. In Fielitz v. Fielitz (1996), a family court judge viewed expert testimony as “extremely enlightening” and helpful in deciding a child custody case. The expert’s evaluation included use of the Rorschach, but the judge did not evaluate the admissibility of this particular method and instead focused on the psychologist’s opinion and testimony as a whole.

The case of Finley v. Apfel (1999) is interesting and involves the appeal of an administrative law judge’s adverse ruling on a claimant’s disability insurance benefits. A psychological evaluation, based in part on projective methods, was made available after the judge’s ruling denying benefits. The appellate court noted that the evaluation in question was based on the MMPI-2 and “lesser known” tests such as the House-Tree-Person and incomplete sentence techniques. Despite the court viewing these projective methods as less common, it held that the case should be sent back to the administrative law judge to consider this new evidence that had not been presented at the original hearing.

In a federal case from the Southern District of Georgia, Breda v. Wolf Camera (1998), the trial judge appeared to have a relatively favorable view of the specific tests an expert wanted to administer in a civil case. The major issue in Breda was a discovery matter in which a determination was rendered as to whether the plaintiff was required to submit to a mental examination by a defense expert. Although controversy surrounding the use of projective techniques has been cited in the behavioral science literature as a basis for the inadmissibility of projective methods (Lilienfeld et al., 2000), the judge in Breda had a different view. Judge Smith cited Ziskin and Faust’s (1988) criticism of projective methods, but still noted: “The Court is hesitant to dictate diagnostic procedures to [the] Dr. . . . Even a cursory scan of the caselaw reveals that both the MMPI-2 and Rorschach are acceptable diagnostic indicators in a mental examination . . . Even though [the plaintiff ’s] clinicians opted to forego this testing, the Court expects reasonable professionals to differ in their treatment and diagnostic practices. [The] Dr.’s request [to administer the Rorschach] falls within normal parameters of psychological examinations, and a mandate by the court to deny him his usual methods of diagnosing would be a presumptuous intrusion into a highly specialized field” (p. 433). Thus, the trial court judge endorsed a view that continuing controversy of a particular method is not sufficient to support a motion to deny an expert witness from using a particular assessment method.

In another post-Daubert criminal case, Rorschach-based expert testimony was offered by the defense to negate the mens rea element of robbery (United States v. Towns, 1998). The defendant had been diagnosed with schizoaffective disorder and claimed to have no intention of robbing a bank; instead, he asserted that his reason for robbing the bank was to obtain mental health treatment if he was caught by police. Although the issue of admissibility of Rorschach testimony was never addressed directly, the trial judge in Towns noted a difference between expert testimony offered for one purpose (i.e., whether the defendant intended to rob) and expert testimony offered for another purpose (i.e., demonstrating the defendant suffered from a mental condition and that his psychological difficulties supported his contention of motive). The judge ruled the testimony admissible for the purpose of supporting the contention of motive but not for establishing the issue of the defendant’s intent. This case highlights an important issue when selecting psychological assessment methods, including projective techniques, in forensic settings. Expert testimony must be relevant and must fit the facts of the case. Aside from the issue of admissibility of projective methods in general, their use must have some relevance to a legal issue being litigated. In some instances, projective methods may have relevance, such as helping to determine if a defendant or litigant has a thought disorder in a criminal responsibility evaluation. In other cases, projective methods may have limited or no relevance, such as the direct effect a litigant’s thought disorder might have on the parent-child relationship in custody evaluations. The Towns decision highlights the importance of making sure projective methods are applied relevantly in a given case.

The idiosyncracies of judicial decisions on the admissibility of expert testimony based on projective methods are apparent in Usher v. Lakewood Engineering (1994). A post-Daubert and pre-Kumho case, Usher involved the trial judge issuing a protective order that barred a defense expert from administering a battery of psychological tests that included several standard instruments, including projective methods such as the Rorschach and Thematic Apperception Test. The judge employed a balancing test that essentially resulted in experts for the plaintiff and defendant using only a clinical interview, and no psychological tests, to form a basis for their opinions. In the view of the judge, precluding the use of psychological tests would “have the effect of providing a level playing field for the parties. Usher’s psychologist expert will be testifying based upon his clinical evaluation conducted without the disputed testing, and Lakewood’s psychiatrist expert will have the identical opportunity. It will then be the province of the trier of fact to resolve the ‘battle of the experts’” (Usher v. Lakewood Engineering, 1994, p. 414). Interestingly, the trial judge never addressed the issue of whether an unstructured clinical interview met any of the Daubert criteria for admissibility of expert testimony. Although Usher is idiosyncratic in certain respects, it illustrates that there is likely to be variability across trial judges as to the admissibility of expert testimony based on projective methods.

Another case highlights the idiosyncracies of legal tests of admissibility. It was noted earlier that some state courts have explicitly rejected the Daubert standard and have chosen to retain the Frye test’s general acceptance standard. One state that has chosen to retain the Frye test is Florida (Hamilton, 1998). In Wordsworth Irving v. State of Florida (1998), a defendant appealed his conviction for sexual battery and lewd assault on a child based on the argument that the trial judge erred in allowing a clinical psychologist to testify that the victim exhibited symptoms consistent with a child that had been sexually abused. The appellate court held that two projective tests used by the expert constitute “diagnostic standards” that must pass the Frye test in order to be admitted at trial. Therefore, the court held that the case should be reversed, the defendant given a new trial, and the projective methods subjected to a Frye analysis to determine if they are generally accepted in the field of psychology.

Another court opinion conveyed a more critical view of projective methods, particularly in their inability to reliably and validly identify individuals who have been sexually abused. The court in J.H.C. v. State of Florida (1994) stated that projective tests provide information that might support the assertions of an alleged victim, but that such data go to the truthfulness of a victim, about which experts are generally not allowed to testify. The court held further that an expert’s statement about whether the victim fit the profile of a sexually abused child was overly reliant on the issue of victim credibility and was improperly admitted.

In general, very few cases address the admissibility, reliability, or validity of projective methods and there is diversity in the way individual judges view projective methods. Most court opinions make passing or superficial mention of projective methods as among the techniques a particular expert employed during the course of an evaluation. In addition, there is variability in the legal standards used to evaluate admissibility of projective methods, including Daubert and Frye, but the trial judge is ultimately responsible for deciding the admissibility of expert testimony. Finally, courts appear to focus more on the scope, nature, and relevance of an expert’s opinion as a whole, rather than on individual methods employed by that expert. The relevance of an expert’s testimony to the issue being decided appears to be a major factor in deciding admissibility, including whether the expert’s opinion inappropriately intrudes into areas that are reserved for the trier of fact, such as credibility of a witness or victim or a criminal defendant’s intent. Whether an expert’s opinion will be of help to the trier of fact also appears to be a consideration in determining admissibility of expert testimony.

Admissibility Criteria

One set of factors that is important to consider in evaluating the appropriateness of projective methods in forensic settings includes the issue of standardization of norms, scoring, and methods of administration. In this regard, there is more variability in how projective methods fare. Lally (2001) has noted that there is considerable diversity in the standardization of administration and scoring of human figure drawings. Although scoring systems for human figure drawings that focus on overall ratings of pathology fare better than systems focusing on global impressions or specific diagnostic signs, Lally concluded that human figure drawings do not appear to offer any additional information beyond that which could be derived from other assessment methods. Lilienfeld et al. (2000) also concluded that human figure drawings have highly questionable incremental validity, although these authors noted that some human figure drawing methods appear to have normative data available.

Another critical issue in evaluating the utility of projective methods in forensic settings is their inability to identify response sets, including malingering. The issue of response sets has been cited by Heilbrun (1992) as important when evaluating the suitability of psychological assessment methods in forensic settings. In general, projective methods have fared poorly as measures of malingering and other forms of deception. Fauteck (1995) noted in a review of the detection of malingered psychosis among offenders that there was a lack of research on the accuracy of human figure drawings and the Thematic Apperception Test in identifying malingering. Rogers and Shuman (2000) noted that projective methods have not been shown to be accurate in identifying malingering and should not be used for that specific purpose in insanity evaluations. Although traditional lore has held that the Rorschach is immune from faking because content but not structure could be faked (Fosberg, 1938), there is limited research to support this contention. Perry and Kinder (1990) noted that there have been problems in the statistical analyses of early Rorschach studies on malingering. While Seamons, Howell, Carlisle, and Roe (1981) found certain configurations in the Rorschach data to be associated with malingering, such as normal form quality and Lambda combined with a high number of special scores and responses with bizarre content, more recent research has yielded unconvincing results (Albert, Fox, & Kahn, 1980; Bash & Alpert, 1980; Kahn, Fox, & Rhode, 1988; Perry & Kinder, 1990). Overall, “the studies conducted on dissimulation and the Rorschach have established trends in structural summary data that are suggestive of biased responding. However, no clear diagnostic cutoffs or operating characteristics have been established” (McCann, 1998, p. 140). It appears that no projective method has been shown to validly and consistently identify malingering, meaning that other assessment methods must be used for this specific purpose.

The psychometric properties of projective methods, including reliability, validity, and the rate of diagnostic error, constitute another broad set of factors that must be considered in evaluating the suitability of these assessment methods in forensic evaluations. Considerable controversy has reemerged about the scientific adequacy of projective methods. The most comprehensive and critical review has been provided by Lilienfeld et al. (2000). These authors concluded that projective techniques are highly controversial, susceptible to faking, poorly validated, unreliably scored, have inadequate norms, and that it is doubtful projective methods meet the criteria for admissibility as outlined under Daubert. It is worth pointing out, however, that Lilienfeld and his colleagues believe it is possible to construct a reliable and valid projective assessment method and cite Loevinger’s (1976, 1998) Washington University Sentence Completion Test as an example of a “successful” projective technique with very good reliability and validity. Furthermore, Lilienfeld and his colleagues noted that some indices from the Rorschach and Thematic Apperception Test and human figure drawings have empirical support. However, only the Rorschach has received favorable endorsement for use in forensic settings by other writers (McCann, 1998; Meloy, 1991; Meloy et al., 1997).

There has been an intensification of debate over the scientific merits of the Rorschach. Evidence of this debate is seen in the exchange between Wood, Nezworski, and Stejskal (1996a, 1996b) and Exner (1996), as well as other critical exchanges in the literature (cf. Garfield, 2000a, 2000b; Lerner, 2000; Weiner, 2000; Wood, Lilienfeld, Garb, & Nezworski, 2000a, 2000b). Debate over clinical and forensic applications of the Rorschach has generated controversy, with Garb (1999) calling for a moratorium on use of the Rorschach in clinical and forensic settings and Acklin (1999) and Weiner (1999) pointing out the utility of the Rorschach as a psychological assessment method. A complete review of the issues raised in these debates is beyond the scope of this chapter. However, it is worth noting that the presence of controversy alone does not preclude the admissibility of specific methods in court. As noted earlier in the Breda opinion, courts often expect that professionals in a particular field will differ from one another in their approaches to treatment and diagnosis. Thus, while debate over projective methods may be fruitful and helpful if it leads to further development and refinement of these methods, professional controversy itself does not constitute a formal legal test of admissibility.

An additional issue that must be recognized when projective methods are considered for use in forensic settings involves the issue of relevance. Because traditional projective methods are used to evaluate personality constructs and have not been developed or validated to measure specific psycholegal issues, the clinician must question whether a method under consideration will provide information and data that are relevant to an issue being evaluated. The clinician must also balance the importance of data provided by projective methods against the potential for misuse by legal professionals and other mental health professionals involved in a case. For instance, the use of projective techniques for detecting child sexual abuse is an area where there is potential for considerable misuse. Conflicting findings have been rendered on the validity of projective drawings to detect child sexual abuse, with West (1998) providing a supportive view and Garb, Wood, and Nezworski (2000) providing evidence of poor support. Lilienfeld and his colleagues (2000) cited evidence of publication bias in the literature on use of projective methods in detecting child sexual abuse. More importantly, however, is the fact that no psychological assessment method has been established to ascertain whether a particular individual has been exposed to sexual abuse. Furthermore, courts frown on the use of expert testimony to prove or disprove the credibility of a witness or victim in sexual assault cases. Therefore, projective methods (like other assessment methods) should not be offered as evidence of a specific ultimate issue in forensic cases.

In light of the specific controversies and problems that have surrounded use of projective methods in forensic cases, the question arises as to whether these methods should be used in legal settings. A review of the case law, as well as professional literature, reveals that while this issue is surrounded by controversy, no legal and professional consensus exists as to their proper or improper role in forensic matters. The following section, therefore, outlines issues that experts may consider when deciding whether to employ projective assessment methods in forensic settings.

GUIDELINES FOR PROJECTIVE ASSESSMENT IN FORENSIC SETTINGS

The expert witness will find that his or her testimony is likely to be analyzed both quantitatively and qualitatively in legal proceedings (O’Conner & Krauss, 2001). A quantitative analysis examines the underlying facts or data (including projective testing) to determine if they are sufficient to support an expert opinion. A qualitative analysis looks at the principles and methods upon which the opinion is based, as well as the helpfulness of the opinion in light of the particular facts of the case. One problem that is apt to arise is the variability in how receptive individual courts and judges are to psychological expert testimony and projective methods.

Recently, the National Conference of Commissioners on Uniform State Laws drafted a set of standards for applying the FRE in federal courts (O’Conner & Krauss, 2001). The result of this conference was a set of guidelines for interpreting FRE 702, although it should be remembered that these guidelines are not legally binding and serve only as a guide to courts. According to guidelines offered by the National Conference, courts may allow an expert to provide an opinion if (1) the testimony will help in understanding evidence or determining a fact; (2) the witness is qualified by knowledge, skill, experience, training, or education; (3) the testimony is based on methods or principles that are reasonably reliable; (4) the testimony is based on sufficient and reliable facts or data; and (5) the witness has applied the principles or methods reliably to the facts of the case. Furthermore, courts can use a number of principles for evaluating the reliability of a method or principle, including controlling legislation or judicial decision establishing reliability of the method or principle; a presumption of reliability; substantial acceptance of the method in the relevant scientific, technical, or specialized field; the extent to which the principle has been tested; adequacy of the research methods employed; peer reviews; error rate; and experience of the witness in application of the principle or method.

Guidelines for Projective Assessment in Forensic Settings 569

In light of these legal guidelines, as well as a review of the professional status of projective methods, some general conclusions and guidelines are offered for the use of projective methods in forensic settings. First, expert witnesses are likely to encounter considerable variability across judges, courts, and jurisdictions as to how projective methods will be received. This variability is likely to be a function of the relevance of the expert’s proffered testimony to the facts of the case and the effectiveness with which the expert can convey an opinion to the judge. In addition, the expert should be cognizant of the specific standard of admissibility that is operating in the specific legal jurisdiction in which the testimony is being offered.

Second, continuing controversy over projective methods does not constitute a specific criterion for establishing admissibility or inadmissibility in court. While the issue of controversy is relevant to the issue of acceptance of projective methods in the field of clinical psychology, documentation of the wide use of these methods in the field constitutes strong evidence of their acceptance, particularly in the specialized field of psychological assessment. There is considerable diversity in the field and assessment methods vary within individual subspecialities in the field. For instance, assessment practices are likely to differ between behavioral and psychodynamic therapists, yet such diversity or even outright disagreement among professionals does not necessarily convey a lack of acceptance of particular methods. Judges (e.g., Breda v. Wolf Camera, 1998) and the National Conference of Commissioners of Uniform State Laws appear willing to accept such disagreement when they note that a principle or method may be presumed reasonably reliable if it has substantial acceptance in a specialized community of professionals (O’Conner & Krauss, 2001).

Third, there is considerable variability among projective methods as to whether they provide a sound basis for expert testimony. Much of this variability is a function of the extent to which an individual method has standardized administration, scoring, and norms. Some methods such as the Rorschach have been standardized and endorsed for use in forensic settings (McCann, 1998; Meloy, 1991). Other projective methods have standardized methods of administration and scoring (e.g., Roberts Apperception Test for Children, McArthur & Roberts, 1982; Washington University Sentence Completion Test, Loevinger, 1976, 1998). However, some projective methods may have highly specialized or limited applications depending on the constructs being measured. A few projective methods, such as projective drawings and the Thematic Apperception Test, have been faulted for their lack of standardized methods for administrating, scoring, and interpreting the

test. As such, the forensic application of these methods is less clear.

Fourth, forensic use of projective methods requires that psychologists be able to support the reliability and validity of the technique. Once again there is considerable variability among individual methods, as well as among individual indices and scales on different projective tests. It is worth noting, however, that many conclusions or interpretations derived from projective methods are capable of being tested and falsified, while others are not. For example, if one wants to test the hypothesis that reflection responses on the Rorschach are associated with narcissism, one can establish a reliable criterion (e.g., ratings of narcissism) and examine the number of reflection responses in those who are both high and low on the criterion. Some impressions derived from projective methods are much more difficult, if not impossible, to falsify (e.g., the presence of an octopus percept on an inkblot test is associated with an internalized cathexis of rage and anger toward a controlling mother object representation).

Finally, and most importantly, the expert must recognize that careful scrutiny will be given to the use of projective methods in forensic settings. The nature of this scrutiny will consist of both a quantitative analysis, in which adequacy of the data and methods will be examined, and qualitative analysis, in which reliability of the opinion will be examined. In other words, the data and opinion must fit the facts of the case. Although some projective methods may be reliable and valid for one purpose (e.g., identifying disordered thinking; measuring ego development; evaluating object relationships), they have not been demonstrated to be reliable and valid for other purposes (e.g., predicting dangerousness; evaluating competencies). Nevertheless, some projective methods might provide useful information on relevant issues in an overall forensic assessment, such as measuring reality testing in a criminal responsibility evaluation. The expert must make a scientifically and theoretically sound connection between the data and the opinion or legal issue being evaluated without stretching or overinterpreting the data. Two guiding questions in this regard might be (1) How much additional or useful information will the projective method provide in the evaluation; and (2) Is there empirical support for using the projective method for the intended purpose?

CONCLUSION

The use of projective methods in forensic evaluations is surrounded by controversy. However, the presence of disagreement or controversy does necessarily preclude forensic application. Similarly, the wide use of some methods does not necessarily support their use in forensic settings if those methods have questionable empirical support. The selection of assessment methods in forensic settings requires careful analysis of the empirical literature as well as recognition of prevailing legal standards of admissibility. Perhaps it is appropriate to conclude by reiterating Justice Blackmun’s optimistic view outlined in the Daubert opinion of how integrity of the legal system can be preserved. The amici curiae briefs filed in the case had raised concerns that abandonment of the “general acceptance” test in favor of the FRE “helpfulness” standard might create a “free for all” of questionable scientific evidence. Justice Blackmun countered by stating that “Vigorous cross-examination, presentation of contrary evidence, and careful instruction on the burden of proof are the traditional and appropriate means of attacking shaky but admissible evidence” (p. 278). Both this approach to evaluating expert testimony and continued efforts to refine and improve projective methods will assist in differentiating those methods and forensic applications that are appropriate from those methods and applications that should be abandoned.

REFERENCES

• Ackerman, M.J., & Ackerman, M.C. (1997). Custody evaluations practices: A survey of experienced professionals (revisited).Professional Psychology: Research and Practice, 28, 137–145.
• Acklin, M.W. (1999). Behavioral science foundations of the Rorschach test: Research and clinical applications. Assessment, 6, 319–324.
• Albert, S., Fox, H.M., & Kahn, M.W. (1980). Faking psychosis on the Rorschach: Can expert judges detect malingering? Journal of Personality Assessment, 44, 115–119.
• Bash, I.Y., & Alpert, M. (1980). The determination of malingering. Annals of the New York Academy of Sciences, 347, 86–99.
• Borum, R., & Grisso, T. (1995). Psychological test use in criminal forensic evaluations. Professional Psychology: Research and Practice, 26, 465–473.
• Breda v. Wolf Camera, 78 Fair Empl. Prac. Cas. 433 (S.D. Ga. 1998).
• Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786 (1993).
• Dawes, R.M. (1994). House of cards: Psychology and psychotherapy built on myth. New York: Free Press.
• Doerr, H.O., & Carlin, A.S. (Eds.). (1991). Forensic neuropsychology: Legal and scientific bases. New York: Guilford Press.
• Dyer, F.J. (1999). Psychological consultation in parental rights cases. New York: Guilford Press.
• Exner, J.E. (1996). A comment on “The Comprehensive System for the Rorschach: A critical examination.” Psychological Science, 7, 11–13.
• Fauteck, P.K. (1995). Detecting the malingering of psychosis in offenders. Criminal Justice and Behavior, 22, 3–18.
• Fielitz v. Fielitz, 1996 LEXIS 12 (Del. Fam. Ct. 1996).

Finley v. Apfel, 1999 LEXIS 14173 (S.D. Ala. 1999).

• Fosberg, I. (1938). Rorschach reactions under varied conditions. Rorschach Research Exchange, 3, 12–30.
• Garb, H.N. (1999). Call for a moratorium on the use of the Rorschach Inkblot test in clinical and forensic settings. Assessment, 6, 313–317.
• Garb, H.N., Wood, J.M., & Nezworski, M.T. (2000). Projective techniques and the detection of child sexual abuse. Journal of the American Professional Society on the Abuse of Children, 5, 161–168.
• Garfield, S.L. (2000a). The Rorschach test in clinical diagnosis. Journal of Clinical Psychology, 56, 387–393.
• Garfield, S.L. (2000b). The Rorschach test in clinical diagnosis: A brief commentary. Journal of Clinical Psychology, 56, 431–434.
• General Electric Co. v. Joiner, 118 S.Ct. 512 (1997).
• Gould, J.W. (1998). Conducting scientifically crafted child custody evaluations. Thousand Oaks, CA: Sage.
• Grisso, T. (1988). Competency to stand trial evaluations: A manual for practice. Sarasota, FL: Professional Resource Exchange.
• Grisso, T. (1998). Forensic evaluation of juveniles. Sarasota, FL: Professional Resource Press.
• Grove, W.M., & Barden, R.C. (1999). Protecting the integrity of the legal system: The admissibility of testimony from mental health experts under Daubert/Kumho analyses. Psychology, Public Policy, and Law, 5, 224–242.
• Hamilton, H.G. (1998). The movement from Frye to Daubert: Where do the states stand? Jurimetrics, 38, 201–213.
• Heilbrun, K. (1992). The role of psychological testing in forensic assessment. Law and Human Behavior, 16, 257–272.
• J.H.C. v. State of Florida, 642 So.2d 601 (Fla. Dist. Ct. App. 1994).
• Kahn, M., Fox, H., & Rhode, R. (1988). Detecting faking on the Rorschach: Computer versus expert clinical judgment. Journal of Personality Assessment, 52, 516–523.
• Kuehnle, K. (1996). Assessing allegations of child sexual abuse. Sarasota, FL: Professional Resource Press.
• Kumho Tire Co., Ltd. v. Carmichael, 119 S.Ct. 1167 (1999).
• Lally, S.J. (2001). Should human figure drawings be admitted into court? Journal of Personality Assessment, 76, 135–149.
• Lerner, P.M. (2000). A nonreviewer’s comment: On the Rorschach and baseball. Journal of Clinical Psychology, 56, 439.
• Lilienfeld, S.O., Wood, J.M., & Garb, H.N. (2000). The scientific status of projective techniques. Psychological Science in the Public Interest, 1, 27–66.
• Loevinger, J. (1976). Ego development: Conceptions and theories. San Francisco: Jossey-Bass.
• Loevinger, J. (1998). Technical foundations for measuring ego development: The Washington University Sentence Completion Test. Mahwah, NJ: Erlbaum.
• Lubin, B., Larsen, R.M., & Matarazzo, J. (1984). Patterns of psychological test usage in the United States: 1935–1982. American Psychologist, 39, 451–454.
• Lubin, B., Wallis, R.R., & Paine, C. (1971). Patterns of psychological test usage in the United States. Professional Psychology, 2, 70–74.
• Lucas v. Shalala, 1993 LEXIS 20308 (N.D. Ill. 1993).
• Maddox v. Lord, 1991 LEXIS 17837 (S.D. NY 1991).
• Marlowe, D.R. (1995). A hybrid decision framework for evaluating psychometric evidence. Behavioral Sciences and the Law, 13, 207–228.
• McArthur, D.S., & Roberts, G.E. (1982). Roberts Appreciation Test for Children manual. Los Angeles: Western Psychological Services.
• McCann, J.T. (1998). Defending the Rorschach in court: An analysis of admissibility using legal and professional standards. Journal of Personality Assessment, 70, 125–144.
• McCann, J.T., Shindler, K.L., & Hammond, T.R. (2003). The science and pseudoscience of expert testimony. In S.O. Lilienfeld, J. Lohr, & S.J. Lynn (Eds.), Science and pseudoscience in contemporary clinical psychology. (pp. 77–108) New York: Guilford Press.
• Meloy, J.R. (1991). Rorschach testimony. Journal of Psychiatry and Law, 8, 221–235.
• Meloy, J.R. (2000). Violence risk and threat assessment: A practical guide for mental health and criminal justice professionals. San Diego, CA: Specialized Training Services.
• Meloy, J.R., Hansen, T.L., & Weiner, I.B. (1997). The authority of the Rorschach: Legal citations during the past 50 years. Journal of Personality Assessment, 69, 53–62.
• Oberlander, L.B. (1995). Psycholegal issues in child sexual abuse evaluations: A survey of forensic mental health professionals. Child Abuse and Neglect, 19, 475–490.
• O’Conner, M. & Krauss, D. (2001). Legal update: New developments in Rule 702. APLS News, 21, 1–4, 18.
• Perry, G., & Kinder, B. (1990). The susceptibility of the Rorschach to malingering: A critical review. Journal of Personality Assessment, 54, 47–57.
• Piotrowski, C., & Keller, J.W. (1989). Psychological testing in outpatient mental health facilities: A national study. Professional Psychology: Research and Practice, 20, 423–425.
• Rogers, R., & Shuman, D.W. (2000). Conducting insanity evaluations (2nd ed.). New York: Guilford Press.
• Seamons, D., Howell, R., Carlisle, A., & Roe, A. (1981). Rorschach simulation of mental illness by psychotic and nonpsychotic legal offenders. Journal of Personality Assessment, 45, 130–135.
• Shapiro, D.L. (1999). Criminal responsibility evaluations: A manual for practice. Sarasota, FL: Professional Resource Press.
• Simon, R.I. (Ed.). (1995). Posttraumatic stress disorder in litigation: Guidelines for forensic assessment. Washington, DC: American Psychiatric Press.

572 Projective Assessment of Personality in Forensic Settings

State v. Scott, 148 Ill.2d 479, 594 N.E.2d 217 (Ill. 1992).

United States v. Frye, 293 F. 1013 (D.C. Cir., 1923).

United States v. Towns, 19 F.Supp.2d 67 (W.D.N.Y. 1998).

Usher v. Lakewood Engineering, 158 F.R.D. 411 (N.D. Ill. 1994).

• Watkins, C.E., Campbell, V.L., Nieberding, R., & Hallmark, R. (1995). Contemporary practice of psychological assessment by clinical psychologists. Professional Psychology: Research and Practice, 26, 54–60.
• Weiner, I.B. (1999). What the Rorschach can do for you: Incremental validity in clinical applications. Assessment, 6, 327–338.
• Weiner, I.B. (2000). Using the Rorschach properly in practice and research. Journal of Clinical Psychology, 56, 435–438.
• West, M.M. (1998). Meta-analysis of studies assessing the efficacy of projective techniques in discriminating child sexual abuse. Child Abuse and Neglect, 22, 1151–1166.
• Wood, J.M., Lilienfeld, S.O., Garb, H.N., & Nezworski, M.T. (2000a). Limitations of the Rorschach as a diagnostic tool: A

reply to Garfield (2000), Lerner (2000), and Weiner (2000). Journal of Clinical Psychology, 56, 441–448.

• Wood, J.M., Lilienfeld, S.O., Garb, H.N., & Nezworski, M.T. (2000b). The Rorschach test in clinical diagnosis: A critical review, with a backward look at Garfield (1947). Journal of Clinical Psychology, 56, 395–430.
• Wood, J.M., Nezworski, T., & Stejskal, W.J. (1996a). The Comprehensive System for the Rorschach: A critical examination. Psychological Science, 7, 3–10.
• Wood, J.M., Nezworski, T., & Stejskal, W.J. (1996b). Thinking critically about the Comprehensive System for the Rorschach: A reply to Exner. Psychological Science, 7, 14–17.
• Wordsworth Irving v. State of Florida, 705 So.2d 1021 (Fla. Dist. Ct. App. 1998).
• Ziskin, J., & Faust, D.M. (1988). Coping with psychiatric and psychological testimony (4th ed.). Los Angeles: Law and Psychology Press.

CHAPTER 43

Cultural Applications of the Rorschach, Apperception Tests, and Figure Drawings

BARRY RITZLER

INTRODUCTION 573 EARLY STUDIES 574 RACIAL DIFFERENCES 574 CULTURE FREE VERSUS CULTURE SENSITIVE 575 THE RORSCHACH AS A CULTURE-FREE METHOD 576 Empirical Evidence 576 A Caveat 576 Other Studies 577 The Modernism Effect 577 THE RORSCHACH AS A CULTURE-SENSITIVE METHOD 577 Acculturation 578

The preliminary title of this chapter was “Cultural Applications of Projective Techniques,” but two considerations resulted in revision. First, projective techniques is an outdated, inaccurate label that fails to capture the nature of frequently used personality assessment methods such as the Rorschach (Ritzler, 1999); a better term may be self-expression methods for differentiating previously labeled projective techniques from self-report methods previously referred to as objective methods. Exner (1993), E. Schachtel (1966), and others (e.g., Weiner, 1998) have emphasized the nonprojective nature of much of the Rorschach response process. Even though the apperception and figure drawing methods have a greater “projective” component, they also easily fit under the rubric “self-expression method.”

The second reason for revising the title emerged after a search of the personality assessment literature over the last 50 years revealed almost no information about research on cultural applications of self-expression methods of personality assessment other than the Rorschach, apperception tests, and figure drawings. Since the primary intention of this presentation was to use empirical findings to support the discussion, other self-expression methods such as sentence completion and early memories procedures were excluded.

CULTURE-FREE AND CULTURE-SENSITIVE ASPECTS OF APPERCEPTION TESTS 578 Culture-Relevant Pictures 578 TEMAS 578 Culture-Free Findings 579 Cultural Sensitivity 579 CULTURE-FREE AND CULTURE-SENSITIVE ASPECTS OF FIGURE DRAWINGS 579 Culture-Free Aspects 579 Culture-Sensitive Aspects 580 Modernism 580 CONCLUSIONS 580 NOTE 581 REFERENCES 581

INTRODUCTION

The Rorschach is a standardized personality assessment method second only to the Minnesota Multiphasic Personality Inventory-2 (MMPI-2) in frequency of use by professional psychologists in the United States (Watkins, Campbell, Nieberding, & Hallmark, 1995). It also has a substantial body of empirical studies supporting its validity. This empirical foundation was strengthened by the development of the Comprehensive System (Exner, 1993) that involves a widely disseminated set of specific procedures and guidelines for standardized administration, reliable coding, and systematic interpretation. Although not without flaws and limitations, the Rorschach method now consists of sufficiently operationalized procedures that enable psychologists to study it more effectively, communicate its results more clearly, and apply it more broadly than ever before.

Thematic apperceptions methods (e.g., Thematic Apperception Test [TAT], Morgan & Murray, 1935; and Tell-Me-A-Story [TEMAS], Costantino, Malgady, & Rogler, 1988) are second to the Rorschach in popularity among self-expression methods (Watkins et al., 1995). Although empirical support continues to grow, the lack of a practical comprehensive scor-

574 Cultural Applications of the Rorschach, Apperception Tests, and Figure Drawings

ing system for the TAT and the recency of scorable methods such as the TEMAS probably are the main reasons why apperception methods lag behind the Rorschach in popularity. Nevertheless, many psychologists see considerable promise in these methods (cf. Cramer, 1996).

Figure drawings have a history in personality assessment that compares with the Rorschach and apperception methods (Riethmiller & Handler, 1997). They have been included because of the amount of empirical validation that has been published over the last 50 years.

In spite of the popularity and potential effectiveness of the Rorschach, apperception tests, and figure drawings, they share one inescapable limiting characteristic for cultural applications; that is, they were developed and standardized on a specific culture. Furthermore, the nonpatient normative samples designed to provide information on an average level of psychological effectiveness primarily are comprised of individuals who represent the majority populations from that culture—namely, Caucasian, middle-class city or suburban dwellers with at least a high school education. These limitations make it difficult to draw conclusions on the appropriateness of these methods for cultural applications. Nevertheless, previously existing empirical evidence and recent research findings show that each method is at least potentially effective for cultural applications.

The discussion of the cultural effectiveness of the Rorschach, apperception tests, and figure drawings will focus entirely on their stimulus properties and interpretation techniques, including administration and coding. This emphasis is not meant to diminish the importance of nonmethod factors necessary for effective multicultural and cross-cultural personality assessment—for instance, appropriate training, knowledge of the client’s culture, and meaningful applications of assessment information in a specific cultural context. These issues have been treated effectively by Dana (1993, 1998, 2000) and others (Allen, 1998; Allen & Walsh, 2000; Cuellar, 1998; Lindsey, 1998; Malgady, 2000; Morris, 2000; Okazaki, 1998; and van de Vijer, 2000) in a broader context of multicultural assessment and intervention. These important discussions, however, have made greater mention of self-report methods. Consequently, this chapter will primarily focus on unique characteristics of the Rorschach, apperception tests, and figure drawings (i.e., self-expression methods) to evaluate their potential for cultural application.

EARLY STUDIES

Almost from their introduction to professional psychologists, self-expression methods were used with enthusiasm by anthropologically oriented researchers. In particular, the Rorschach was favored as a method for assessing the influence of culture on personality (see Hallowell, 1941, for an early discussion of the rationale). Schachtel, Henry, and Henry (1942) presented an early study of six (!) Pilaga Indian children with little apparent concern for the low N. In the same year, Hallowell (1942) published a much more extensive study of the effects of acculturation on an American Indian sample (N ” 102). It was the first study to show that acculturation has a significant impact on personality functioning (see later). Cook (1942) was the first to publish data from a primitive (Samoan) culture (see modernism sections). Hallowell (1945a) began to appreciate the “culture-free” aspects of the Rorschach (see later) when he reviewed 3,684 responses from 151 Saultreaux Indian children and adults and found no major differences from responses of Caucasians. His complete work with Native Americans was summarized in a report later that year (Hallowell, 1945b) based on a collection of over 1,000 Rorschachs. Other, equally ambitious studies appeared in the early literature; for example, Billig, Gillin, and Davidson (1947a, 1947b) compared the Rorschachs of 7,500 Mayan Indians and Ladinos and found the Ladinos to be “more constricted.” In another example, DuBois (1944) published an extensive study of the people of the East Indian island of Alor using the Rorschach and other methods.

Even though some methodology problems were recognized (e.g., Lantz, 1948), the Rorschach emerged from the first decade of anthropological studies with an optimistic rating (Abel, 1948). At the same time, the TAT (Henry, 1947) and, a little later, figure drawings (Haward, 1958) were viewed with similar enthusiasm.

Even later, the Rorschach and TAT received a positive review for use in psychological anthropology (Bock, 1988), but the emphasis remained on primarily psychoanalytic interpretation. The early studies, with several extensive samples, could have yielded much more information if contemporary methods of coding and interpretation had been available. One wonders if some of those samples still exist in available archives.

RACIAL DIFFERENCES

Before the cultural relevance of the Rorschach, apperception tests, and drawings is evaluated, the issue of racial differences per se will be addressed. Race and culture are related, but they are different. To date, the few studies comparing races on self-expression methods have not tried to isolate cultural factors. Consequently, the information obtained from those studies more clearly focuses on race per se. However, very few studies comparing races have been conducted. Frank (1992) found only seven over 60 years comparing African Americans and Caucasians in the United States. Frank observed that six of the seven studies (all but the earliest; Hunter, 1937) drew participants from school populations (universities, high schools, and elementary schools) with care taken to control for intelligence. Frank suggests that such selection may have limited cultural differences. The results across the studies were very similar: There were few, if any, differences in structural or content variables. The primary consistent difference was that African Americans gave fewer responses than Caucasians and, as a consequence, produced somewhat less complex protocols. Frank argued that this difference could not be attributed to intelligence, because those differences were controlled in all studies, or depression, because the groups did not differ on other indicators of depression on the Rorschach. Frank’s conclusion was that the difference in productivity was a result of reluctance toward self-disclosure by the African Americans, especially in research conducted by Caucasian examiners.

Elsewhere, in a recent study (Presley, Smith, Hilsenroth, & Exner, 2001) of the nonpatient adult sample assessed by Exner (2002) for the Comprehensive System, African Americans differed from others on only one of dozens of variables. African Americans showed significantly fewer Cooperative Movement responses, a finding interpreted by the authors as a consequence of the African American sample’s reluctance to relate to Caucasian examiners—the same conclusion drawn by Frank.

Also, in a recent analysis of the Rorschach Comprehensive System protocols of 432 patients in one clinical setting, Meyer (2002) found no differences between European Americans and several ethnic minorities including African Americans. Actually, he found that on four variables predicting psychotic disorder, regression lines favored minorities over European Americans. Meyer concluded that “the available data support using the Comprehensive System across ethnic groups.”

Caudill (1952) reported no Rorschach differences between acculturated Japanese American participants and middle-class Caucasian Americans with 70 participants in each sample.

The Presley et al. (2001), Meyer (2002), and Caudill (1952) results support Frank’s (1992) earlier conclusion that there are no differences attributable to race in personality functioning as measured by the Rorschach. Frank went on to argue that no study has been conducted to assess Rorschach differences between middle-class Caucasian culture and the urban, African American culture celebrated by “black pride” advocates. For such a comparison, Frank notes, participants would need to be selected who typify the personality styles of the two cultures—not participants from university samples or high schools in the same community.

Studies of racial comparisons using the TAT and figure drawings, although scarce, also show no meaningful differences. Although Johnson and Sikes (1965) found Rorschach and TAT differences between Mexican Americans and others, they found no differences between blacks and whites. Even when the TAT was modified so that people in the drawings had clearly African American features (Riess, Schwartz, & Cottingham, 1950), there were no differences between the stories of white and black college students, although Northern participants (both black and white) told more sophisticated stories than Southern participants.

In another study, while black children rated the African American version of the TAT as more desirable and told more positive stories to the altered pictures (Dlepu & Kimbrough, 1982), there were no differences in the personality data yielded by the stories when compared with stories of white children.

In human figure drawings, Wise (1969) found no differences between African American and Caucasian adolescents.

In an apperception test study, when the cultural correlates of racial identity were included in sample selection, differences occurred. In their development of the TEMAS method, Costantino et al. (1988) and their colleagues have clearly shown that urban-dwelling minority children tell different stories than suburban Caucasian children. However, when the picture stimuli are drawn to be more consistent with the urban, minority cultures, the differences are eliminated.

Even though more studies need to be conducted, it seems reasonable at this point to conclude that race per se is not a factor in personality functioning. Cultural correlates of race, however, may account for some differences.

CULTURE FREE VERSUS CULTURE SENSITIVE

There are two ways that a personality assessment method developed in one culture can be effective for application with another culture. The first way is if results in the first culture are the same as results in the second culture; for example, do adult nonpatients yield the same results in both cultures? Do children at the same age levels appear the same in both cultures? Do patients with the same types of problems appear the same? And so on.

The second way for a method to have cross-cultural relevance is for the differences in the results between two cultures to be consistent with otherwise observed psychological differences between the cultures. For example, if one culture is characterized by intense, overt expression of emotions and another is characterized by a more reserved, indirect expression of feelings, does the method yield information that correctly differentiates the two cultures on emotional expressivity? The extent to which a method yields consistent data across cultures is an indication of the extent to which it is culture sensitive. Both qualities may be useful and the same method may possess degrees of each. The following discussions will evaluate the culture-free and culture-sensitive qualities of the Rorschach, apperception tests, and figure drawings. The related issues of modernism and acculturation also will be discussed.

THE RORSCHACH AS A CULTURE-FREE METHOD

A personality assessment method is more likely to be culture free if it presents stimuli and requires tasks that are not frequently encountered in specific cultures. The Rorschach may qualify on both criteria. It is unlikely that there are many cultures that have the specific task of inkblot interpretation as a familiar activity. Furthermore, the Rorschach cards are sufficiently ambiguous that their shapes, colors, and shadings may be encountered anywhere in the world, by anyone, on an everyday basis. In contrast, the casual activity of seeing meaning in cloud formations or other ambiguous stimuli (Tea leaves? Land formations? Finger painting?) probably is familiar to most cultures. Looked at in another way, the Rorschach may have enough structure to assure similarity across cultures. For example, most people from any culture would agree that the bottom D area of Card IX is pink, or that Card VI has shading nuances, or that Card I resembles an animal with wings. Also, the task of responding to the instruction “What might this be?” has similarities across all cultures. To respond to the question, any subject must (1) look at the card and process its information, (2) generate possible responses to the question, (3) select from these alternatives those that seem most appropriate, and (4) communicate these responses to the examiner. Although cultural and environmental factors will influence these component phases of the Rorschach response process, they probably occur in most responses given by most people regardless of culture.

Empirical Evidence

As stated previously, the moderately ambiguous nature of the Rorschach, coupled with its basically simple structure, make it a candidate for culture-free personality assessment. Indeed, recent evidence exists to indicate that the stimulus properties of the Rorschach cards and the standardized administration and coding procedures result in relatively culture-free data (Ritzler, 2001). At the International Rorschach Society meeting in Amsterdam in 1999, Comprehensive System data for nonpatient adults were presented from eight countries (Argentina, Belgium, Finland, Japan, Peru, Portugal, Spain, and the United States). There was remarkable consistency in the results. For instance, when selected key variables from the Comprehensive System are considered (M, SumC, Lambda, X!%, Egocentricity Index, Isolation Index, EA, and es), the only significant differences consist of a higher X!% in Japan compared to Finland and Portugal.1

A Caveat

Although the Amsterdam studies provide compelling evidence of the Rorschach as a culture-free method, caution should be exercised in considering the Amsterdam data as representative of average effective psychological functioning across cultures. The Amsterdam datasets, with a combined total of 2,053 participants, yield results that are more simplistic (e.g., mean Lambda ” 1.00), less emotionally rich (mean SumC ” 2.95; 10% extratensive style), and more functionally deficient (mean EA ” 6.79; 50% ambitent; mean X!% ” 18.8) than the original Comprehensive System norms (N ” 600) reported by Exner (1993; Lambda ” 0.58; SumC ” 4.52; 38% extratensive, EA ” 8.87; 19% ambitent; and X!% ” 7.0). A replication of Exner norms is under way (current N ” 175) and a preliminary report (Exner, 2002) revealed notable similarities with the original data (Lambda ” 0.61; SumC ” 4.69; 31% extratensive; EA ” 9.43; 19% ambitent; and X!% ” 10.0). The differences between the Amsterdam and Exner norms may be attributable to two factors: participant selection procedures and examiner experience. Exner (1993) drew his nonpatient participants from organizations such as church congregations, PTAs, service clubs, and so forth. In return for their participation, Exner contributed money to their organizations. Consequently, his nonpatients are “joiners”; that is, individuals who become actively involved in organizations outside the home. The Amsterdam studies, on the other hand, sampled from a broader base of nonpatients and obtained fewer “joiners.” Consequently, the Amsterdam samples were more likely to include individuals from lower levels of effectiveness—perhaps enough to explain the overall differences in effectiveness and complexity between the Amsterdam and Exner norms.

In addition to the differences in selection procedures, the Amsterdam and Exner norms differ in the amount of experience of the examiners. The Amsterdam studies typically employed well-trained students with modest experience. Exner, however, used only highly experienced examiners and no students. This difference may also have accounted for the more limited Amsterdam results since the less experienced examiners may not have inquired sufficiently or may not have as adequately prepared their participants for the study. Consequently, while the Amsterdam studies provide evidence of the culture-free nature of the Rorschach Comprehensive System, they do not necessarily provide accurate estimates of average effective functioning.

Other Studies

Beyond Amsterdam, other studies have shown that the Rorschach yields similar results across cultures. For instance, it is quite clear that Oriental nonpatients show results that correspond closely to U.S. norms (e.g., Korea—Moon & Cundick, 1983: Taiwan—Yang, Chen, & Hsu, 1965; and Hong Kong—Sachs & Lee, 1992).

In an earlier study, Kaplan, Rickers-Ovsiankina, and Joseph (1956) used 24 Rorschachs equally distributed across Spanish American, Navaho, Zuni, and Mormon cultures and found that Rorschach experts without knowledge of the specific cultures could not significantly categorize the protocols according to culture. Even with detailed information about the cultures, experts only correctly sorted an average of 13 of 24 cases.

Even earlier, C. Adcock (1951) found no Rorschach differences between Cook Island native children (n ” 88) and New Zealand mainland children (n ” 30) except for a slight constriction in the protocols of the island children responding to a mainland examiner. Later, A. Adcock and Ritchie (1958) found no differences between Maori and white New Zealand adults. Curiously, they interpreted this to mean that the Rorschach was “invalid” and failed to mention the possibility of culture- and/or race-free qualities.

Using the Rorschach to study conforming characteristics of Congolese students, Claeys (1967) found that these educated natives gave protocols very similar to those of U.S. students.

Mattlar, Carlson, and Forsander (1993) found that the only difference between Finnish nonpatients and U.S. nonpatients was a few culture-specific popular responses (e.g., on Card II, Finnish participants frequently saw “Christmas elves,” a common icon in that country). In contrast, Bourguignon and Nett (1955) found no differences in Populars (or other variables) between Haitian and U.S. participants.

In a particularly revealing study, Pires (2000) noted that Portuguese nonpatient participants with low levels of education gave protocols that frequently differed from U.S. norms, whereas more highly educated Portuguese showed no differences when compared with the Comprehensive System norms. This raises the issue of the “modernism effect” discussed next.

The Modernism Effect

Modernism is a likely confounding factor in personality assessment comparing cultures. With the now long-standing availability of information technology, easier travel, higher education, and employment opportunities, many traditional cultural behavior patterns have been blunted. Less advantaged, more primitive environments are more likely to have citizens who show behavior patterns that preserve the “old ways.” Consequently, when Rorschachs of individuals in relatively primitive living situations are compared with those from more modern situations, any differences observed may result from the modernism effect rather than other cultural influences. Since many Rorschach studies that claim to have identified cultural differences appear to have modernism as a primary confound, they have been listed in Table 43.1 for convenient display. They do not include the studies covered in the Early Studies section. In each case, the study shows the protocols from the more primitive “cultures” to have the more primitive results.

THE RORSCHACH AS A CULTURE-SENSITIVE METHOD

While there is considerable evidence to indicate that the Rorschach is not affected by cultural differences, some evidence attests to its cultural sensitivity. Focusing on content analysis, DeVos and Boyer (1989) documented many symbolic differences across cultures. Also, in spite of the previously cited studies to the contrary, Chen, Gong, Li, and Jie (1997) recently found significant differences on the Comprehensive System when comparing 666 Chinese nonpatient adults with the Comprehensive System norms.

In some specialized studies, Rorschach differences not apparently associated with modernism were identified. Guelfi (1981) found different Rorschach patterns of depression in inpatients of French and Maugrabin descent. Hsu, Watrous, and Lord (1961) found more adaptive Rorschach scores in adolescents from Hawaii compared to adolescents from Chicago. Lichter and Rothman (1981) found Jewish adults identified with the New Left movement to have different Rorschach results from traditionally oriented Jewish adults. And Thapa (1983) found Rorschach differences between married and unmarried Nepalese; these differences seem peculiar to the specific culture of Nepal, because they have not appeared in other studies comparing married and unmarried participants.

In a rare study finding significant differences between different modern cultures, Georgas and Vassiliou (1967) found

Author(s)	Journal	Sample(s)
Bagh (1958)	Indian Journal of Psychology	Rural Muslims with low, medium, and high levels of education
Billig, Gillin, & Davidson (1947a)	Journal of Personality	Rural Guatemalans
DeAcosta (1966)	Revista de Psicologia	“Depressed village” inhabitants
DeVos and Miner (1958)	Sociometry	Oasis and urban Arabs
Doob (1960)	(book)	Three nonliterate African tribes and urban Jamaicans
Honigman (1949)	(book)	Canadian Indians
Joseph, Spicer, & Chesky (1949)	(book)	Popago Indians
Matsui, Horike, & Ohashi (1980)	Tohoku Psychologia Folia	Okinawan shamans
Meernhout & Mukendi (1980)	Bulletin de Psychologie et d’Orientation	Rural Zairois
Metraux & Abel (1957)	American Journal of Orthopsychiatry	Peasant community in Monserrat, B.W.I.
Preston (1964)	Genetic Psychology Monographs	Northwest coast Alaskan Eskimos
Ray (1955)	Indian Journal of Psychology	Abor and Gallong (India)
Rey (1955)	Ofakim	Deprived children in Morocco
Thompson (1951)	Psychiatry: Journal for the Study of Interpersonal Processes	Popago, Navaho, and Hopi Indians

TABLE 43.1 Cross-Cultural Studies Illustrating the “Modernism Effect” Confound

significant differences between Athenians and norms from the United States.

A few studies have claimed to have obtained Rorschach results consistent with the hypothesized “national character” of the countries in question. For instance, Asthana (1956) used the Rorschach and TAT to identify signs of the Mexican “machismo” national character in the Rorschachs of 900 villagers. And Rausch de Trauenberg (1988) traced the effects of trauma in the Rorschachs of Vietnamese adolescents.

Acculturation

A major confound also exists for the study of cultural sensitivity of self-expression assessment methods—acculturation. There may be no more surefire hypothesis than the following: As acculturation increases, self-expression personality assessment results increasingly resemble norms from the culture that the participants have entered. Rorschach data seem to confirm this hypothesis. Table 43.2 summarizes several studies that have shown the equalizing effect of acculturation in the contexts of different cultures.

CULTURE-FREE AND CULTURE-SENSITIVE ASPECTS OF APPERCEPTION TESTS

The cultural relevence of apperception tests has been accentuated by the designing of culture-relevant picture stimuli. Perhaps the best example is the TEMAS (Tell-Me-A-Story) method. Most culturally relevant research, however, has been done with the standard TAT. While some evidence exists for the TAT as a culture-free method, most studies have shown results that indicate it is primarily culture sensitive.

Culture-Relevant Pictures

Investigators have attempted to enhance the cultural sensitivity of apperception test pictures by presenting the characters with culture-specific dress and physical characteristics and by depicting scenes relevant to central issues in the culture (Dana, 1982). This methodology has met with some success. Table 43.3 summarizes studies in which specifically designed apperception pictures have elicited stories that have specific cultural relevance. Most of these studies also have provided norms for common themes occurring in response to the specialized pictures.

TEMAS

Perhaps the best example of an apperception test specifically designed for multicultural application is the Tell-Me-A-Story method (TEMAS) developed by Costantino and his colleagues (Costantino, Malgady, & Rogler, 1988). Through a series of validation studies (e.g., Costantino, Malgady, Colon-Malgady, & Bailey, 1992; Costantino, Malgady, Rogler, & Tsui, 1988; Costantino, Malgady, & Vazquez, 1981) they have demonstrated the method’s usefulness in assessing culturally relevant personality issues for African American and Hispanic

Author(s)	Journal	Culture	Participants
Abel & Hsu (1949)	Rorschach Research Exchange	American	Chinese immigrants
Armstrong & Eng (1978)	Journal of Social Psychology	Malaysian	Aborigines
Boyer (1965)	International Journal of Social Psychology	American	Apaches
Ponzo, E. (1966)	Psicologia Generale e del Lavoro	Brazilian	Tukanos
Regini (1986)	Psychologia: An International Journal of Psychology in the Orient	Nepal	Gurungs
Singh, V. (1976)	Asian Journal of Psychology and Education	Indian	Males with contact with other cultures
Spindler & Spindler (1958)	American Anthropologist	American	Menomini Indians

TABLE 43.2 Studies Demonstrating That Acculturation Affects Rorschach Results in the Direction of the Assimilated Culture

TABLE 43.3 Studies Demonstrating the Use of Specially Designed Apperception Pictures

Author(s)	Date	Journal	Culture	Norms?
Dana	(1982)	(book)	Lakota Sioux	Yes
Dhapola	(1974)	Journal of Psychogical Resources	East Indian (Kumaoni)	Yes
Lessa & Spiegelman	(1954)	U. of CA Publications in Culture and Society	Ulithi Atol (Micronesia)	Yes
Ma & Kong	(1998)	Psychologia Science China	Chinese and Japanese	Yes
Sabeau	(1976)	Confrontations Psychiatriques	Merina (Madagascar)	Yes
Sherwood	(1957)	Journal of Social Psychology	Swazi	Yes
Wang	(1969)	Acta Psychologia Taiwanica	Taiwan	Yes
Zeidner, Klingman, & Itskowitz	(1993)	Journal of Personality Assessment	Israeli Persian Gulf (children)	No

children and adolescents. While no studies have been published using the TEMAS with adults, the multicolored pictures, many of which depict social interaction between adults and children, might also be effective with older participants. The TEMAS has an extensive coding system with corresponding systematic interpretation methodology—a rare asset for apperception tests. The major limitation is that little, if any, research validation has come from outside the Costantino team (Ritzler, 1993).

Culture-Free Findings

In contrast to the Rorschach studies, most TAT cultural research indicates that the method has primary cultural sensitivity, particularly when the pictures are selected for cultural relevance. Furthermore, modernism does not appear to lurk as a confound.

Only one TAT study in the PsycINFO database had a clear, “culture-free” result. Botla (1971) assessed achievement motivation in 86 South Africans, 298 Arabs, and 80 U.S. undergraduates, all women, and found no differences. Of course, modernism may play a role here since all subjects were educated and from middle-class, urban backgrounds.

Cultural Sensitivity

The preponderance of TAT cultural research indicates that it is primarily a culturally sensitive method. The most convincing evidence for this evaluation comes from studies that have demonstrated TAT differences between cultures within the same country. These studies are summarized in Table 43.4.

CULTURE-FREE AND CULTURE-SENSITIVE ASPECTS OF FIGURE DRAWINGS

The empirical studies of the cultural relevance of figure drawings have been less frequent than Rorschach and apperception test studies and have yielded very mixed results. Modernism, however, emerges as a major modifying variable in many studies.

Culture-Free Aspects

Culture-relevant research on figure drawings with controls for modernism provides evidence for considering drawings as having some freedom from culture bias. For instance, van der Vijfeijken and Vedder (2000) found no differences between the human figure drawings of 7- and 8-year-old Dutch

Author(s)	Date	Journal	Country	Cultures Differentiated
Anderson et al.	(1961)	American Journal of Orthopsychiatry	Israel	Kibbutz and nonkibbutz children
Cohen-Emerique	(1979)	Psychologie Francaise	France	Moroccan Jews
Eisenman & Foulks	(1970)	Psychological Reports	Guatemala	Ladinos, Mengalas, and Indians
Fisher & Fisher	(1960)	Journal of Projective Techniques	U.S. (Texas)	Jews and non-Jews
Gladwin	(1953)	Transactions of the NY Academy of Sciences	Truk	Men and women
Hibbard et al.	(2000)	Journal of Personality Assessment	U.S.	Asian Americans
Johnson & Sikes	(1965)	Journal of Projective Techniques	U.S.	African Americans and Mexican Americans
Leblanc	(1958)	Journal of Social Psychology		Katangese
Melikian	(1964)	Journal of Social Psychology		Arab university students
Rigney, Smith, & Douglas	(1961)	(book)	U.S.	Beatniks
Sundberg et al.	(1992)	Journal of Personality Assessment	U.S. (Oregon)	Rajneeshees
Tedeschi & Kian	(1962)	Journal of Social Psychology	Persia	U.S. and Persians
Yu	(1980)	International Journal of Social Psychiatry	China	Chinese collective

TABLE 43.4 Studies Showing TAT Differences Between Cultures Within the Same Country

(n ” 53) and immigrant (n ” 71) children. Halliman (1988) found no differences on body image in the drawings of Muslim (n ” 94) and Judaic-Christian (n ” 168) women in the United States. Other groups not differing on drawings from their U.S. counterparts were Vietnamese children (Delatte, 1978), Orthodox Jewish adolescent girls (Spero, 1985), bilingual Spanish-speaking children (Hamilton, 1984), Thai children (Gardiner, 1974), and Nigerian children (Zaid, 1979). Also, Cox, Koyasu, Huanima, and Perara (2001) found that the only difference between the drawings of Japanese and United Kingdom children was that the Japanese had better artistic skills.

Culture-Sensitive Aspects

In somewhat fewer studies (but with no less convincing results), figure drawings seem to have some cultural relevance. In two studies of specific drawing characteristics, Gardiner (1969, 1972) found a wide range of differences in the frequency of hostile elements (Gardiner, 1969) and smiling (Gardiner, 1972). In studying the drawings of 2,382 boys 11 to 13 years of age from 26 cultures, he found the following frequencies for hostile elements (incomplete list): Thailand, 35%; Germany, 26%; Taiwan, 25%; Yugoslavia, 17%; Algeria, 15%; Mecca, 6%; Syria, 3%; Brooklyn non-Orthodox Jews, 1%; and Japan, 1%.

In the study of drawn smiles, Gardiner (1972) assessed 1,043 drawings of 9- to 17-year-old children and found the following evidence supporting the reputation of Thailand as “the land of smiles”: Thailand, 64%; Japan, 31%; Mexico, 20%, Chiapas Indians, 5%.

In other studies, Laosa, Swartz, and Diaz-Guerrero (1974) found differences in sexuality and overall differentiation on the drawings of Mexican and Anglo American children. Munroe and Munroe (1983) found major differences in the drawings of adolescents from three African tribes—the Kipaigis, the Logoli, and the Gusii.

Modernism

Most of the studies supporting the cultural sensitivity of figure drawings are confounded by modernism. Table 43.5 presents a summary of those studies. In each of the studies, the less advantaged samples drew the less developed drawings (i.e., the modernism effect).

CONCLUSIONS

This review of the empirical evidence regarding cultural application of self-expression methods brings several issues to light. First, there is a need for more research, particularly in regard to studies that identify specific cultural samples without the confounds of race, modernism, or acculturation. Figure drawings, in particular, need more study.

The second issue is the culture-free or culture-sensitive status of the self-expression methods reviewed in this chapter. It appears that the Rorschach is a major, culture-free assess-

Author(s)	Date	Journal	Samples
Andersson	(1996)	Child Study Journal	Tanzanian, Swedish, and exiled Swedish children
Dennis	(1960)	Journal of Social Psychology	Bedouins
Honigman & Carrera	(1959)	American Anthropologist	Eskimo and Cree Indian children
Hunsberger	(1978)	Canadian Journal of Behavioral Sciences	Canadian White and Indian children
Koppitz	(1969)	Journal of Clinical Psychology	Middle-class and lower-class U.S. children
Koppitz & DeMoreau	(1968)	Revista Interamericana de Psicologia	Middle-class and lower-class U.S. and Mexican children
Martlew & Connolly	(1996)	Child Development	Schooled and unschooled New Guinea children
Pfeffer	(1985)	Social Behavior and Personality	Low-income girls in Nigeria
Pfeffer & Olowu	(1986)	Journal of Social Psychology	Middle- and low-income children in Nigeria
Weller & Sharan	(1971)	Child Development	Middle- and lower-class Israeli children

TABLE 43.5 Figure Drawing Studies Showing the Effect of Modernism

ment method that is likely to yield similar results across a wide range of cultures. In only a few circumstances does the Rorschach seem sensitive to cultural differences. The apperception tests, on the other hand, seem to be quite sensitive to cultural influences and only come close to being culture free when the stimulus pictures are altered to include culturerelevant characteristics. However, since the altered pictures tend to focus on culture-specific situations, they often yield different results across different cultures.

Figure drawings seem to occupy a middle ground in regard to cultural sensitivity with both culture-free and culturesensitive results appearing in the literature in about equal proportions. Consequently, these three methods seem well suited for a battery of assessment methods aimed at differentiating cultural from universal personality characteristics.

Next, the issue of confounding factors has been considered. When cultures differ greatly in the extent to which they are modern, advantaged cultures, differences on culture-sensitive methods are dramatic. Even the relatively culture-free Rorschach yields significant differences between individuals from primitive and modern cultures. Consequently, the confound of modernism has been identified.

Another confound is that of acculturation. Almost without exception, when acculturation occurs, the assessment results become more similar to those from the new culture than the old. Consequently, acculturation levels must be assessed and controlled in most studies of cultural assessment.

Finally, the many studies reviewed in this chapter illustrate the usefulness of these self-expression methods for understanding cultural issues in personality development and functioning. As more standardized and reliable procedures for using the Rorschach, apperception tests, and figure drawings continue to develop, they will become even more useful in multicultural and cross-cultural psychological study and intervention.

NOTE

1. These differences may be a function of different content frequencies across these three countries. See discussion of crosscultural content differences.

REFERENCES

• Abel, T. (1948). The Rorschach test in the study of culture. Rorschach Research Exchange, 12, 79–93.
• Abel, T., & Hsu, S. (1949). Chinese personality revealed by the Rorschach. Rorschach Research Exchange, 13, 285–301.
• Adcock, A., & Ritchie, J. (1958). Intercultural use of the Rorschach. American Anthropologist, 60, 881–892.
• Adcock, C. (1951). A factorial approach to Rorschach interpretation. Journal of Genetic Psychology, 44, 261–272.
• Allen, J. (1998). Personality assessment with American Indians and Alaska natives: Instrument considerations and service delivery style. Journal of Personality Assessment, 70, 17–42.
• Allen, J., & Walsh, J. (2000). A construct-based approach to equivalence: Methodologies for cross-cultural/multicultural personality assessment research. In R. Dana (Ed.), Handbook of cross-cultural and multicultural personality assessment. Mahwah, NJ: Erlbaum.
• Anderson, H., Anderson, G., Rabin, A., Elonen, A., Abel, T., & Diaz-Guerrero, R. (1961). Culture components as a significant factor in child development. American Journal of Orthopsychiatry, 31, 481–520.
• Andersson, S. (1996). Social scaling in children’s family drawings: A comparative study in three cultures. Child Study Journal, 25, 97–121.
• Armstrong, H., & Eng, K. (1978). Body-image barrier perception as a function of assimilation within the Malaysian aborigines. Journal of Social Psychology, 105, 165–173.

582 Cultural Applications of the Rorschach, Apperception Tests, and Figure Drawings

• Asthana, H. (1956). Some aspects of personality structuring in Indian (Hindu) social organization. Journal of Social Psychology, 44, 155–163.
• Bagh, D. (1958). An experimental study of Rorschach characteristics of different cultural groups of rural Bengal. Indian Journal of Psychology, 33, 55–66.
• Billig, O., Gillin, J., & Davidson, W. (1947a). Aspects of personality and culture in a Guatemalan community: Ethnological and Rorschach approaches. Part I. Journal of Personality, 16, 153– 187.
• Billig, O., Gillin, J., & Davidson, W. (1947b). Aspects of personality and culture in a Guatemalan community: Ethnological and Rorschach approaches. Part II. Journal of Personality, 16, 326– 368.
• Bock, P. (1988). Rethinking psychological anthropology: Continuity and change in the study of human action. New York: W.H. Freeman and Co.
• Botla, E. (1971). The achievement motive in three cultures. Journal of Social Psychology, 85, 163–170.
• Bourguignon, E., & Nett, E. (1955). Rorschach populars in a sample of Haitian protocols. Journal of Projective Techniques, 19, 117–124.
• Boyer, L. (1965). Effects of acculturation on the personality traits of the old people of the Mescalero and Chiricahua Apaches. International Journal of Social Psychology, 11, 264–271.
• Caudill, W. (1952). Japanese-American personality and acculturation. Genetic Psychology Monographs, 45, 3–102.
• Chen, M., Gong, Y., Li, S., & Jie, Y. (1997). The cross-cultural comparison of the results of Rorschach test between American and Chinese. Chinese Mental Health Journal, 11, 209–212.
• Claeys, W. (1967). Conforming behavior and personality variables in Congolese students. International Journal of Psychology, 2, 13–23.
• Cohen-Emerique, M. (1979). Study of identity changes as shown by the TAT. Psychologie Francaise, 24, 249–257.
• Cook, P. (1942). The application of the Rorschach test to a Samoan group. Rorschach Research Exchange, 6, 51–60.
• Costantino, G., Malgady, R., Colon-Malgady, G., & Bailey, J. (1992). Clinical utility of the TEMAS with nonminority children. Journal of Personality Assessment, 59, 433–438.
• Costantino, G., Malgady, R., & Rogler, L. (1988). Tell-Me-A-Story (TEMAS) manual. Los Angeles: Western Psychological Services.
• Costantino, G., Malgady, R., Rogler, L. & Tsui, E. (1988). Discriminant analysis of clinical outpatients and public school children by TEMAS: A thematic apperception test for Hispanics and Blacks. Journal of Personality Assessment, 52, 670–678.
• Costantino, G., Malgady, R., & Vazquez, C. (1981). A comparison of the Murray-TAT and a new Thematic Apperception Test for urban Hispanic children. Hispanic Journal of Behavioral Sciences, 3, 291–300.
• Cox, M., Koyasu, M., Huanima, H., & Perara, J. (2001). Children’s figure drawings in the United Kingdom and Japan: The effects

of age, sex, and culture. British Journal of Developmental Psychology, 19, 275–292.

• Cramer, P. (1996). Storytelling, narrative, and the Thematic Apperception Test. New York: Guilford Press.
• Cuellar, I. (1998). Psychological assessment of Hispanic Americans. Journal of Personality Assessment, 70, 71–86.
• Dana, R. (1982). Picture-story cards for Sioux/Plains Indians. Fayetteville: University of Arkansas.
• Dana, R. (1993). Culturally competent assessment practice in the United States. Journal of Personality Assessment, 66, 472–487.
• Dana, R. (1998). Cultural identity assessment of culturally diverse groups. Journal of Personality Assessment, 70, 1–16.
• Dana, R. (2000). Culture and methodology in personality assessment. In I. Cuellar & R. Paniagua (Eds.), Handbook of multicultural mental health (pp. 97–120). San Diego: Academic Press.
• DeAcosta, T. (1966). Psychological aspects of academic problems in “depressed villages.” Revista de Psicologia, 3, 11–26.
• Delatte, J. (1978). Human figure drawings of Vietnamese children. Child Study Journal, 8, 227–234.
• Dennis, W. (1960). The human figure drawings of Bedouins. Journal of Social Psychology, 52, 209–219.
• DeVos, G., & Boyer, L. (1989). Symbolic analysis cross-culturally: The Rorschach test. Berkeley: University of California Press.
• DeVos, G., & Miner, H. (1958). Algerian culture and personality in change. Sociometry, 21, 255–268.
• Dhapola, T. (1974). The use of a shortened verbal form of the TAT for the study of Kumaoni character. Journal of Psychological Resources, 18, 9–13.
• Dlepu, O., & Kimbrough, C. (1982). Feeling tone and card preferences of Black elementary children for the TCB and TAT. Journal of Non-White Concerns in Personnel and Guidance, 10, 50–56.
• Doob, L. (1960). Becoming more civilized: A psychological exploration. New Haven: Yale University Press.
• DuBois, C. (1944). The people of Alor: A socio-psychological study of an East Indian island. Minneapolis: University of Minnesota Press.
• Eisenman, R., & Foulks, E. (1970). Usefulness of Museen’s TAT scoring system: I. Differences among Guatemalan Indians, Ladinos, and Mengalas on a modified TAT. Psychological Reports, 27, 179–185.
• Exner, J. (1993). The Rorschach: A comprehensive system: Volume I. Basic foundations (3rd ed.). New York: Wiley.
• Exner, J. (2002). A new nonpatient sample for the Rorschach Comprehensive System: A progress report. Journal of Personality Assessment, 78, 391–404.
• Fisher, S., & Fisher, R. (1960). A projective test analysis of ethnic subculture themes in families. Journal of Projective Techniques, 24, 366–369.
• Frank, G. (1992). The response of African Americans to the Rorschach: A review of research. Journal of Personality Assessment, 59, 317–325.
• Gardiner, H. (1969). A cross-cultural comparison of hostility in children’s drawings. Journal of Social Psychology, 79, 261–263.
• Gardiner, H. (1972). The use of human figure drawings to assess a cultural value: Smiling in Thailand. Journal of Psychology, 80, 203–204.
• Gardiner, H. (1974). Human figure drawings as indicators of value development among Thai children. Journal of Cross-Cultural Psychology, 5, 124–133.
• Georgas, J., & Vassiliou, V. (1967). A normative Rorschach study of Athenians. Journal of Projective Techniques and Personality Assessment, 31, 31–38.
• Gladwin, T. (1953). The role of man and woman on Truk: A problem in personality and culture. Transactions of the New York Academy of Sciences, 15, 303–309.
• Guelfi, J. (1981). A comparative study of depressive symptomatology on French and Maugrabin inpatients. Psychopathologie Africaine, 17, 171–189.
• Halliman, C. (1988). Muslim and Judaic-Christian perception of desirable body shape. Perceptual and Motor Skills, 67, 80–82.
• Hallowell, A. (1941). The Rorschach test as a tool for investigating cultural variables and individual differences in the study of personality in primitive societies. Rorschach Research Exchange, 5, 31–34.
• Hallowell, A. (1942). Acculturation processes and personality changes as indicated by the Rorschach technique. Rorschach Research Exchange, 6, 42–50.
• Hallowell, A. (1945a). “Popular” responses and cultural differences: An analysis based on frequencies in a group of American Indian subjects. Rorschach Research Exchange, 9, 153–168.
• Hallowell, A. (1945b). The Rorschach technique in the study of personality and culture. American Anthropologist, 47, 195–210.
• Hamilton, L. (1984). Human figure drawings as measures of selfconcept development in bilingual children. Journal of Instructional Psychology, 11, 28–36.
• Haward, L. (1958). Extra-cultural differences in drawings of the human figure by African children. Ethnos, 3–4, 220–230.
• Henry, W. (1947). The Thematic Apperception Technique in the study of culture-personality relations. Genetic Psychology Monographs, 35, 3–135.
• Hibbard, S., Tang, P., Latko, R., Park, J., Munn, S., Bolz, S., et al. (2000). Mechanism manual for the TAT between Asian Americans and Whites. Journal of Personality Assessment, 75, 351– 372.
• Honigman, J. (1949). Culture and ethos of Kaska society. New Haven: Yale University Press.
• Honigman, J., & Carrera, R. (1959). Cross-cultural use of Machover’s Figure Drawing Test. American Anthropologist, 59, 650–654.
• Hsu, F., Watrous, B., & Lord, E. (1961). Culture pattern and adolescent behavior. International Journal of Social Psychology, 7, 33–53.
• Hunsberger, B. (1978). Racial awareness and preference in White and Indian Canadian children. Canadian Journal of Behavioral Sciences, 10, 176–180.
• Hunter, M. (1937). Responses of comparable white and negro adults to the Rorschach Test. Journal of Psychology, 3, 173–182.
• Johnson, D., & Sikes, M. (1965). Rorschach and TAT responses of Negro, Mexican-American, and Anglo psychiatric patients. Journal of Projective Techniques and Personality Assessment, 29, 183–188.
• Joseph, A., Spicer, R., & Chesky, J. (1949). The desert people: A study of the Popago Indians. Chicago: University of Chicago Press.
• Kaplan, B., Rickers-Ovsiankina, M., & Joseph, A. (1956). An attempt to sort Rorschach records from four cultures. Journal of Projective Techniques, 20, 172–180.
• Koppitz, E. (1969). Emotional indicators on Human Figure Drawings of boys and girls from lower and middle-class backgrounds. Journal of Clinical Psychology, 25, 432–434.
• Koppitz, E., & DeMoreau, M. (1968). A comparison of emotional indicators on Human Figure Drawings of children from Mexico and from the United States. Revista Interamericana de Psicologia, 2, 41–48.
• Lantz, H. (1948). Rorschach testing in pre-literate cultures. American Journal of Orthopsychiatry, 18, 287–291.
• Laosa, L., Swartz, J., & Diaz-Guerrero, R. (1974). Perceptualcognitive and personality development of Mexican and Anglo-American children as measured by human figure drawings. Developmental Psychology, 10, 131–139.
• Leblanc, M. (1958). Acculturation of attitude and personality among Katangese women. Journal of Social Psychology, 47, 257–264.
• Lessa, W., & Spiegelman, M. (1954). Ulithian personality as seen through ethnological materials and thematic test analysis. University of California Publications in Culture and Society, 2, 243–301.
• Lichter, R., & Rothman, S. (1981). Jewish ethnicity and radical culture: A social psychological study of political activists. Political Psychology, 3, 116–157.
• Lindsey, M. (1998). Culturally competent assessment of African American clients. Journal of Personality Assessment, 70, 43–53.
• Ma, Q., & Kong, K. (1998). A study of the group TAT. Psychologia Science China, 21, 126–130.
• Malgady, R. (2000). Myths about the null hypothesis and the path to reform. In R. Dana (Ed.), Handbook of cross-cultural and multicultural personality assessment (pp. 49–62). Mahwah, NJ: Erlbaum.
• Martlew, M., & Connolly, K. (1996). Human figure drawings by schooled and unschooled children in Papua New Guinea. Child Development, 67, 2743–2762.

584 Cultural Applications of the Rorschach, Apperception Tests, and Figure Drawings

• Matsui, H., Horike, K., & Ohashi, H. (1980). The Rorschach responses of Okinawan shamans “yuta.” Tohoku Psychologia Folia, 39, 61–78.
• Mattlar, C., Carlson, A., & Forsander, C. (1993). The issue of the popular response: Definition and universal vs. culture-specific popular responses. British Journal of Projective Techniques, 38, 53–62.
• Meernhout, M., & Mukendi, N. (1980). Rorschach responses from another culture: A sample of rural Zairois. Bulletin de Psychologie et d’Orientation, 29, 61–70.
• Melikian, L. (1964). The use of selected TAT cards among Arab university students: A cross-cultural study. Journal of Social Psychology, 62, 3–19.
• Metraux, R., & Abel, T. (1957). Normal and deviant behavior in a peasant community: Montserrat, B.W.I. American Journal of Orthopsychiatry, 27, 167–184.
• Meyer, G. (2002). Exploring possible ethnic differences and bias in the Rorschach Comprehensive System. Journal of Personality Assessment, 78, 104–129.
• Moon, T., & Cundick, B. (1983). Shifts and constancies in Rorschach responses as a function of culture and language. Journal of Personality Assessment, 97, 345–349.
• Morgan, C., & Murray, H. (1935). A method for investigating fantasies: The Thematic Apperception Test. Archives of Neurological Psychiatry, 34, 289–306.
• Morris, E. (2000). An Afrocentric perspective for clinical research and practice. In R. Dana (Ed.), Handbook of cross-cultural and multicultural personality assessment (pp. 17–41). Mahwah, NJ: Erlbaum.
• Munroe, R., & Munroe, R. (1983). Drawings and values in three East African societies. Journal of Social Psychology, 119, 135– 136.
• Okazaki, S. (1998). Assessment of Asian Americans: Research agenda for cultural competency. Journal of Personality Assessment, 70, 54–70.
• Pfeffer, K. (1985). Sex identification and sex-typing in some Nigerian children’s drawings. Social Behavior and Personality, 13, 69–72.
• Pfeffer, K., & Olowu, A. (1986). Perceptions of dress style in some Nigerian children’s drawings. Journal of Social Psychology, 126, 287–290.
• Pires, A. (2000). National norms for the Rorschach normative study in Portugal. In R. Dana (Ed.), Handbook of cross-cultural and multicultural personality assessment (pp. 367–392). Mahwah, NJ: Erlbaum.
• Ponzo, E. (1966). Psychological studies on the acculturation of primitive populations: II. Acculturation and detribalization: The Tukanos. Comparative psychological observations with other populations. Revista di Psicologie Sociale e Archivo Italiano di Psicologie Generale e del Lavoro, 33, 41–107.
• Presley, G., Smith, C., Hilsenroth, M., & Exner, J. (2001). Clinical utility of the Rorschach with African Americans. Journal of Personality Assessment, 77, 491–507.
• Preston, C. (1964). Psychological testing with northwest coast Alaskan Eskimos. Genetic Psychology Monographs, 69, 323–419.
• Rausch de Trauenberg, N. (1988). Psychological study of Vietnamese adolescents: Value of the Rorschach: Cultural specificity, traumatic experience, and modes of adaptation. Enfance, 41, 95–104.
• Ray, A. (1955). The tensional feelings among the Abors and Gallongs as indicated by the Rorschach technique. Indian Journal of Psychology, 30, 95–103.
• Regini, M. (1986). The culture-contact and the personality structure of the Gurungs: A Rorschach study. Psychologia: An International Journal of Psychology in the Orient, 29, 50–55.
• Rey, A. (1955). Resume of studies concerning children from North Africa. Ofakim, 9, 376–382.
• Riess, B., Schwartz, E., & Cottingham, A. (1950). Further critical evaluations of the Negro version of the TAT. Journal of Abnormal and Social Psychology, 45, 700–709.
• Riethmiller, R., & Handler, L. (1997). Problematic methods and unwarranted conclusions in DAP research: Suggestions for improved research procedures. Journal of Personality Assessment, 69, 459–475.
• Rigney, R., Smith, J., & Douglas, L. (1961). The real bohemia: A sociological and psychological study of the “Beats.” New York: Basic Books.
• Ritzler, B. (1993). TEMAS: A new variation on an old theme. Journal of Psychoeducational Research, 22, 322–325.
• Ritzler, B. (1999). Objective and projective: Personality assessment’s frightful misnomers. Paper presented at the Society for Personality Assessment Midwinter Meeting, New Orleans, LA.
• Ritzler, B. (2001). Multicultural usage of the Rorschach. In L. Suzuki, J. Ponterotto, & P. Meller (Eds.), Handbook of multicultural assessment (pp. 237–252). San Francisco: Jossey-Bass.
• Sabeau, J. (1976). Essay of transcultural psychology applied to the ethnic group, Merina. Controntations Psychiatriques, 14, 5–54.
• Sachs, J., & Lee, H. (1992). Dual scaling analysis of Rorschach responses in a Hong Kong Chinese sample. Psychologia: An International Journal of Psychology in the Orient, 35, 84–95.
• Schachtel, A., Henry, J., & Henry, Z. (1942). Rorschach analysis of Pilaga Indian children. American Journal of Orthopsychiatry, 12, 679–713.
• Schachtel, E. (1966). Experiential foundations of Rorschach’s test. Hillsdale, NJ: Analytic Press.
• Sherwood, E. (1957). On the designing of TAT pictures with special reference to a set for an African people assimilating Western culture. Journal of Social Psychology, 45, 161–190.
• Singh, V. (1976). Culture-contact and personality structure. Asian Journal of Psychology and Education, 1, 13–19.
• Spero, M. (1985). The clinical significance of Orthodox Jewish cultural content in idiographic responses to diagnostic psychological tests. Journal of Psychology and Judaism, 9, 86–113.
• Spindler, L., & Spindler, G. (1958). Male and female adaptation in culture change. American Anthropologist, 60, 217–233.
• Sundberg, N., Goldman, M., Rotter, N., & Smyth, D. (1992). Comparative TATs of high-achieving Rajneeshees. Journal of Personality Assessment, 59, 326–339.
• Tedeschi, J., & Kian, M. (1962). Cross-cultural study of the TAT assessment for achievement motivation: Americans and Persians. Journal of Social Psychology, 58, 227–234.
• Thapa, S. (1983). The development of personality as related to the marital status in Nepalese culture as revealed by the Rorschach responses. Psychologia: An International Journal of Psychology in the Orient, 26, 252–257.
• Thompson, L. (1951). Perception patterns in three Indian tribes. Psychiatry: Journal for the Study of Interpersonal Processes, 14, 255–263.
• Van de Vijer, F. (2000). The nature of bias. In R. Dana (Ed.), Handbook of cross-cultural and multicultural personality assessment (pp. 87–106). Mahwah, NJ: Erlbaum.
• Van der Vijfeijken, K., & Vedder, P. (2000). The cultural sensitivity of the Human Figure Drawing test. Pedagogische-Studies, 77, 166–172.
• Wang, M. (1969). Report on the revision of the TAT. Acta Psychologica Taiwanica, 11, 24–41.
• Watkins, C., Campbell, V., Nieberding, R., & Hallmark, R. (1995). Contemporary practice of psychological assessment by clinical psychologists. Professional Psychology, 26, 54–60.
• Weiner, I. (1998). Principles of Rorschach interpretation. Mahwah, NJ: Erlbaum.
• Weller, L., & Sharan, S. (1971). Articulation of the body concept among first-grade Israeli children. Child Development, 42, 1553–1559.
• Wise, J. (1969). Self-reports by Negro and white adolescents to the Draw-A-Person. Perceptual and Motor Skills, 28, 193–194.
• Yang, K., Chen, W., & Hsu, C. (1965). Rorschach responses of normal Chinese adults: IV. The speed of production. Acta Psychologica Taiwanica, 7, 34–51.
• Yu, E. (1980). Chinese collective orientation and need for achievement. International Journal of Social Psychiatry, 26, 184–189.
• Zaid, S. (1979). Values expressed in Nigerian children’s drawings. International Journal of Psychology, 14, 163–169.
• Zeidner, M., Klingman, A., & Itskowitz, R. (1993). Coping under threat of missile attack: A semiprojective assessment procedure. Journal of Personality Assessment, 60, 435–456.

CHAPTER 44

Collaborative Exploration with Projective Techniques: A Life-World Approach

CONSTANCE T. FISCHER, EMILIJA GEORGIEVSKA, AND MICHAEL MELCZAK

THE LIFE WORLD 587 PROJECTIVE TECHNIQUES WITHIN THE LIFE WORLD 588 COLLABORATIVE PRACTICES 589 Contextualize 589 Intervene 590 Describe via Structural Re-Presentation 591 CONFLUENCE OF COLLABORATIVE ASSESSMENT WITH DIVERSE THEORIES 592 Cognitive Behavioral Psychology 592

Narrative Psychology 592 Depth Psychology 592 Gestalt Psychotherapy 593 Developmental Theories 593 Classic and Contemporary Assessment Approaches 593 REFERENCES 594

In collaborative assessment, the client and the assessor work together to revise and refine the assessor’s initial impressions that were formed in light of the interview, test data, and referral issues. Their goal is to understand particular life events in everyday terms, using assessment data as points of access to the client’s world. Although nonprojective test data also are used in this process, this chapter will focus on the use of projective techniques. Here is an introductory example, which includes use of the Minnesota Multiphasic Personality Inventory:

• Psychologist: I think I’m beginning to make sense of something. You told me that you’re not depressed, but that you do seem to be dragging, that you’re not as enthusiastic as usual [client nods]. Well, you remember the long true-false test? [client nods]. On it you seem to have said “false” to many items about crying, being sad, blue, and so on [scale 2 ” 56]. I would guess [from Rorschach patterns] that you push away any such potential feelings, in order to remain strong?
• Client: Hmm. My father would call it being a man. I guess so.
• Psy: May I ask what your wife calls it?
• Cl: [laughing] Being stoic! Otherwise she calls it—calls me, distant.

Psy: [smiling] I can see that. Now, on the inkblots, it struck me that although you weren’t letting yourself feel sorry for yourself, and you worked hard at coming up with plenty of “answers,” you were nevertheless also attuned to the sad, worn out, hurt aspects of the cards—“a sad sheep,” “drying out leaf,” “diseased vertebra.”

Cl: Is that bad?

• Psy: I wouldn’t say “bad,” but I would say that even though you try not to let people see it, some aspects of your current situation do indeed have you dragging. I would guess from some of your responses [m ” 4] that something’s hanging overhead, keeping you in a holding pattern?
• Cl: I’ll say! [psychologist nods and gestures for examples; client describes having to wait for his employers to decide whether they were going to go out of business, worrying that his wife’s chemotherapy would not be successful, and not knowing whether to confront their son about suspected drug abuse for fear that he would quit school again. Only the job situation had been mentioned in the interview.]
• Psy: Uh! . . . I can see that you know what your goals are but that you find yourself stuck, having to wait, being

powerless to take action, when normally you’re a takecharge person [a:p ” 7:2; zd ” #5.0. Client nods, with surprised expression]. And you’re feeling the strain, and are attuned to possible bad outcomes. I imagine that Dr. Callah sensed some of this, which she referred to as “depression.”

• Cl: Well, I don’t want to be sick, to have to take medication.
• Psy: Surely your body has been expressing your being so concerned over time. Medication probably would help. However, I think that just talking, like we are now, could help by itself. Or both. Fortunately, we’ve seen that with a bit of encouragement, you know how to talk about your situations.

The above collaboration not only confirms the psychologist’s general impressions but also brings them down to earth, where psychologist, client, and other relevant parties (referring physician, later counselor, wife) can converse within a shared frame of reference. Also, the client has been engaged as an active participant, which affirms his capacity to engage his difficulties.

Of course the collaborative process also allows for revision of the psychologist’s initial interview and test-based impressions. Here is an example, with the same client:

• Psy: Well, despite your being despondent, given your circumstances, you seem to be maintaining a high level of intellectual functioning [R ” 29; F#% ” 72; zf ” 20; zd ” #5; Thematic Apperception Test stories fully followed directions to present what is happening now, what led up to this, what will happen next, and so on].
• Cl: No, that’s not so. It takes me twice as long to do the daily crossword puzzle. My son even commented on that. And working my way through TurboTax, which is pretty automatic, took me much longer this year. I made a lot of dumb mistakes. It’s easier to appear alert when I’m interacting with people, like here doing the testing.
• Psy: OK. Then I’d say that although compared with people in general, you’re cognitively astute [Cl: What?] intellectually sharp, you’re not as sharp or focused as you used to be, especially when you’re on your own. [Cl: Right.]

As illustrated above, collaboration goes beyond unilateral feedback to the client. The results of collaborative assessment are highly individualized and, as later examples will illustrate, the process is constructive, sometimes therapeutic. The

assessment usually explores not just the client’s present state but also his or her viable options.

This chapter next will address what is meant by “lifeworld approach,” and then will address how projectives “work” within a life-world approach. The principles and further examples of collaborative use of projectives will be provided. These sections will be followed by a discussion of the confluence of collaborative assessment with a sample of psychological theories. Finally, issues such as limitations, relation to cultural context, and ethics will be addressed.

THE LIFE WORLD

The life world is the everyday place where we go about living our lives—at home, work, play, and in community. For better and worse, psychologists for much of the twentieth century attempted to build our discipline as a science in the tradition of the natural sciences. We eschewed humans’ particular circumstances, experiences, and perspectives in our laboratory search for universal laws of behavior. Only gradually are we, in effect, replacing the goals of “predict and control” with goals more like “understand and influence.” In assessment work the authors of this chapter give priority to the life world over that of the laboratory. Our point of departure into test scores, norms, research data, diagnostic categories, and personality theories is life events—events that call to be understood in their contexts in order to be relevant to decisions and action. Psychologists’ technical realm is not more real than these life circumstances. For example, a diagnosis of psychosis organizes our thinking, but does not explain the labeled person’s ongoing life. Hence as collaborative assessors we regard the life world as being our point of return, too; our results are not scores, categories, and abstractions, but rather they are revised understandings of the person who is going about his or her life. Beyond that, results are individualized suggestions to client and to helpers. These suggestions are grounded in explorations in which the client has participated; they are already tailored.

En route to these outcomes, observed behavior and test patterns serve not only as evidence for conclusions, but also as instances of client comportment. They provide points of access for exploring life-world instances and their contexts. The contexts in which such instances do not occur also are explored (e.g., when is John not resentful of authority?).

The assessor is the expert on tests and patterns, and the client and his or her involved others are the life-world experts. Collaboratively, they can develop life examples, contexts, options. The psychologist can now write maximally useful reports—individualized, concrete, understandable. The

psychologist’s impressions are not just statistically probabilistic, but are accompanied by actual examples provided by the client. Readers bring additional examples into play from their associations to the presented examples. Speculation is greatly reduced. Challenges to findings are curtailed in that probability statements are accompanied by confirmable past behaviors and their contexts. Moreover, with findings already in life-world terms, the psychologist does not have to “translate” the report for attorneys, judges, family, caretakers, or other readers.

What of clients who dissemble, or who do not wish to cooperate? Through our efforts to collaborate, we evaluate the person’s modes of self-presentation. We wind up somewhat ahead of having only our direct impressions and our test data. As with all assessments, the client’s circumstances invite or limit our contribution. Actually, it is the overly agreeable client who requires extra effort to develop discriminating examples that counter as well as match the assessor’s profferings.

PROJECTIVE TECHNIQUES WITHIN THE LIFE WORLD

The client encounters the projective task, whether to report what the inkblots might be, to tell stories for the Thematic Apperception Test pictures, to draw humans, to copy Bender Gestalt designs, or to complete sentence stems, similarly to how he or she takes up other ambiguous tasks when being observed. As always, the person brings his or her life into the present, on the way to anticipated, hoped for, and feared futures. The assessor catches glimpses of habitual (taken for granted), unowned, conflicted, and conscious styles of coping. Likewise, the client finds that living through each projective technique evokes structurally similar events, which thus become available for sharing and exploration. The following is an example [Bender Gestalt]:

• Psy: I’m surprised by how you changed sizes, and your strategy as you went along—a tiny copy here, a whole page for this last one, counting dots here but just guessing here. I guess I’m surprised because of the contrast with the letter that you sent to me, which is carefully organized and consistent in style . . . ??
• Cl: Well, that’s because I worked on my letter by myself; no one was watching. And here, there was only one chance. You’d laugh if you saw my notes for the letter. They’re like these [copied designs]—scattered, all over the place. [long pause] These [Bender pages] remind me of being self-conscious, like when Mr. Burton [supervisor] came in to observe my orientation session for

new staff. Actually, my final outline was super, but as soon as he came in, I lost my place, and felt like I did back in high school in phys ed—the whole class hooted when the teacher criticized my form. She said my body didn’t seem to listen to my brain.

• Psy: So your having become “discombobulated” [client’s term] has been not so much because the task was open ended, but more because you lost your place in the face of being scrutinized.
• Cl: Yeah, that’s it: losing my place. You know, just like when I was drawing those, I feel sweaty when that’s happening, just like in phys ed. Anxious as all get out.

Interviewing alone, or even use of “objective” personality tests such as the Minnesota Multiphasic Personality Inventory and the Personality Assessment Inventory, do not evoke past events as powerfully as do projective techniques and other personally administered tests. We find that talking with the client about the experience of dealing with a task while that example is still fresh maximizes the emergence of experience-near understandings. Such discussion bypasses abstract conceptualizations and instead focuses on the jointly observed client contending with the task. Discussing impressions from each test or projective technique as it is completed offers the advantage that the data that follow are addressed with some life-world clarifications already having been accomplished. It is true, though, that if the Bender, for example, has been discussed prior to administration of the Rorschach, then the client may undertake that task in a more cooperative manner than might otherwise have been the case. The assessor takes that context into account. In addition, the client, anticipating discussion, is likely to offer remarks about the Rorschach experience right after the inquiry is completed and to recall the experience more vividly during postscoring discussion. There are many variations of the point at which collaborative discussion may be fruitful.

In regard to the nature of projective techniques, we think not so much in terms of historical notions of the client projecting unconscious material into the ambiguous task, but more in terms of the client propelling him- or herself into the task in habitual ways. Those ways may be just taken for granted, they may be dynamic means of not noticing potential meanings or conflict, or they may be efforts to find positive possibilities. Moving fairly quickly into cooperative discovery, as in the previous examples, is not at all unusual within collaborative assessment, even when the material was not focally available to the client earlier. Precursors to productive joint exploration of “deep” psychological material include (1) the client’s recognition of the assessor’s genuine respect, (2) shared understandings about the purposes of the assessment, (3) the building of consensual comprehensions starting with material readily available to the client and moving from there into typically unavailable material, and (4) the assessor having available a life-world repertoire of behavior and experience that may be suggested by scores and patterns obtained via the projective techniques.

In this life-world approach to assessment, “clinical inference” includes deductive moments, but it is mostly a process of coming to understand what it may be like to be the client, and what the implications of that understanding are for decision making. One’s own life, experience of other clients, and exposure to novels and other arts are resources. Developmental and personality theory heighten attunement to nuances, but do not serve as interpretive grids. Interpretive discipline is hermeneutic: One cycles through data asking what impressions cohere across readings, revising in light of new data and understandings, cycling again, checking for coherence and revising again and again. The assessor is not looking for “who is this person really,” but rather for always provisional understandings of who this person has been, under various circumstances. The assessor is keenly aware that her or his understanding is necessarily perspectival.

The assessor takes textbook, manual, and computer printouts seriously but not absolutely. Our reports do not claim that “this instrument says” such and such; rather, we report that the instrument’s patterns led us to explore such and such, and that we formed the following impressions corroborated by x, y, and z examples provided by the client and/or involved others and any case records. En route, as we study the projective data, we find it helpful to regard printouts, manual interpretations, and clinical lore as respected colleagues with whom we converse while taking into account their biases. It is the psychologist who ultimately “says” what his or her assessment impressions are.

Along the way, we ask ourselves how “chance” occurrences might have affected the picture we evolve. For example: On a Rorschach, what if borderline CFs had been scored as FCs? What if the custodial parent had not prepped the kids before they were interviewed? Rather than regarding chance occurrences as invalidators, we regard them as openings to greater precision. One of us (Fischer) remembers years ago as a trainee administering a Rorschach in a child guidance clinic to a 10-year-old boy who had bicycled to the clinic. The assessment seemed to be providing examples of a healthy, happy youngster, but when presented with Card VI he shuddered and pushed the card away. His percept: “a squished frog, its insides splattering all over.” Fischer was surprised and later asked, “Do you know where that squished frog came from?” The boy shuddered again and spoke of having accidentally run over a frog while joyfully riding his

bike through every puddle on his way to the clinic. He worried that the frog had suffered and felt guilty that he had not realized from seeing a dead frog earlier that puddles might contain other frogs. With that context available, Fischer convinced her supervisor that in this instance anatomy and blood content, F!, and death (today, MOR) did not indicate psychopathology, but instead were an instance of sensitivity and of a sense of responsibility.

Yes, this contextualizing and refining do take time, which is a price of locating our findings in their life-world contexts. And, yes, we have to choose which of our impressions to explore collaboratively. It seems to us that psychologists are practicing in this way more and more often, finding that ultimately the extra effort expended during collaborative assessment is efficient in regard to outcomes for client and readers of the report. These practices seem to be followed more often by persons who feel free to do so by virtue of working in independent practice and by psychologists who also are psychotherapists.

COLLABORATIVE PRACTICES

Collaborative practices can be named and carried out in many ways. Contextualizing, trying out alternative action (intervening), and describing via re-presenting behavior are particularily congruent with a life-world orientation.

Contextualize

The following are further examples of contextualizing, of exploring the circumstances, as taken up by the client and as viewed by others, in which particular experience and action have occurred. The assumption here is that humans are relational; we shape and are shaped by the ways we perceive and take up our circumstances. This assumption is consistent with contemporary psycho-dynamic notions of interpersonal psychotherapy.

In this example, the setting is a management consulting office, where an upper-middle-level manager is undergoing evaluation as part of his corporation’s long-range planning of his career. Mr. Groarity flew into the city that morning and spent 5 hours taking paper-and-pencil tests. Secretaries and clerks, keen observers of these management assessees, have agreed that this fellow has been unusually unpretentious and genuinely considerate. Upon beginning the direct assessment, the psychologist has the same impression. She recalls a Thematic Apperception Test story (Card 10) written by Mr. Groarity that morning: “This is a long-married couple, dancing, feeling wonderfully close. When the musicians take a

break, they return to their table and toast each other. They’ve had their difficulties but they’ve learned humility and they’ve grown in appreciation of the other.”

But the psychologist also recalls Mr. Groarity’s reported earliest memory: “I’m about 4 years old, and my Dad cheers and looks real proud when I knocked down Tony, a neighborhood boy a year older, who was reaching for my bike.” For the psychologist, that earliest memory presented an apparent contradiction to impressions of an empathic, considerate person. The Thematic Apperception Test story for Card 4 also weighed in: “He’s refusing to put in a good word for her brother—he just pulls away from her. He’s thinking, ‘You’re both losers. I’m not going to jeopardize my upward progression by looking soft-hearted.’” Later, while administering the Rorschach, the psychologist is struck by many positive features, such as numerous Human Movement responses, several COP responses, good Form Quality, high P, and high zd. But she also is troubled by three references to diseased or decaying spines.

• Psy: You know, I’ve been impressed by what I’m seeing as your maturity, your nondefensive interaction. But I’m puzzled by your attunement on these inkblot cards to diseased backbones. Would you mind telling me what you know about those?
• Cl: That’s easy. They’re about my cancer of the spine. I was diagnosed 6 years ago and told that I had 6 months to live. And here I am now with no signs of cancer. Of course I check in regularly. I can tell you that 6 years ago I wouldn’t have looked so good on your tests. Now I live each day as though it were my last—a miniature of what I want my life to have been.
• Psy: Six years ago I would have seen more of the fellow refusing to help his brother?
• Cl: That’s right; exactly. I’m not proud of that.
• Psy: I imagine that your precancer Dan Groarity is still around?
• Cl: Sure, as an antagonist. The cancer is still around, too, like on those [the Rorschach cards]. Urges to be king of the mountain arise, but now I smile or cringe, and I try to live my “last day.” Sometimes I miss the competition, but then I remember that now I’m competing with cancer and with precancer Dan. Hmm, that is a good name for the antagonist. [Also see example of Mr. Turner, later.]

Of course contextualizing also clarifies less adaptive shaped/shaping. Example:

• Psy: You’ve certainly been helpful in getting me to understand your side of this custody problem, Ms. Bellini. Now can you help me to understand how Mr. Bellini [estranged husband] has come to think that your temper could be a problem if you had primary custody?
• Cl: He’s wrong. I don’t have a temper problem. . . . Sometimes I just have to stand up for myself, so I get heard.
• Psy: Is that like on this item [pointing to last sentence completion stem], “Forms like this:” [completed with “SUCK!!”]?

Cl: Well, they do.

• Psy: But didn’t you think that I might be put off by that? [Client shrugs and nods.] So was this an instance of standing up for yourself, no matter?
• Cl: I guess so. I didn’t know you’d be nice.
• Psy: So you’ve stood up for yourself in a temper sort of way when things seemed unfair?
• Cl: That’s so!
• Psy: [several minutes later] Tell me an example of an unfair situation where you stood up for yourself in a nontemper way. [Client reports how, with earlier coaching from her attorney, she had calmly explained to a hearing officer some discrepancies between her and her husband’s accounts of an incident.]

The above excerpt also illustrates the feasibility of identifying the when-nots of problematic action. This identification humanizes the person; that is, it shows him or her in greater variability and as capable of choice, while also pointing to a place from which he or she can less problematically move toward goals.

Intervene

To assess a client’s personally viable options, the psychologist intervenes into his or her problematic style. An example (continuing with Ms. Bellini):

• Psy: So it’s been easier to stand up for yourself in a nontemper way when you have a helper or a coach, someone who also offers moral support?
• Cl: Yeah, but I don’t get a lot of moral support from nobody.
• Psy: Maybe in the future you could remember past support, and times you’ve stood up for yourself without attacking anyone. Maybe I can be a coach right now, like your lawyer. Okay? So now when you feel inclined to attack, or protest in a temper way, when your fists clench and you want to shout [client smiles and

nods], then you use those signs as a traffic signal to slow down and to imagine a coach suggesting safer ways to get to your destination [client: “yeah, a copilot”]. Here [assessor hands client a blank sentence completion form, points to the last item, and offers the client a pencil]; remember how you felt when you first read “forms like this”? [Client grins, and hits the table surface with her fist; “Right!!”] Now slow down, remember that you’re on your way to a custody evaluation, and take your time; stand up without temper. Go ahead.

Cl: [reading what she now wrote, with some cross-outs] “seem unfair because you don’t know what somebody’s going to make out of what you say.”

This exercise, in effect, was a learning experience for Ms. Bellini. But its primary goals were to assess whether she could accept coaching and how readily she could identify landmarks of her beginning to behave in problematic ways and then find alternative ways of moving toward her personal goals. Life-world assessment is concerned with a person’s viable options, rather than with just naming the person’s status (indeed, as though it were static) on a battery of tests and techniques.

Here is another example of intervening to assess and to encourage options, returning to the woman whose Bender was chaotic:

• Psy: Let’s try an experiment. I’ll bet you can keep your place on this task [Bender] even with a judge (me) watching. [Client exaggerates a groan.] Remember what you told me about how you’ve kept your place when you’re “in a groove,” like playing the piano? So get in the groove now by recalling that you’ve already done this before, that each design turned out fine. So what do you do now to “keep your place”? [Psychologist presents the first card.]
• Cl: [chuckles, takes a blank sheet of paper and accepts a pencil] Well, you’ll give me the cards in order, so I just have to remember to draw the pictures pretty much their size, and I think I’ll go ahead and count the dots on those first ones. I’ll take deep breaths and let myself slow down.
• Psy: [after client finishes copying the designs] Hey, you kept your place! If this were your orientation presentation, Mr. Burton would have been impressed.
• Cl: Yeah. You know, I got a little sweaty just now, but I just kept going, reminding myself that I was familiar with this and to take my time. My “landmarks” are going to be getting sweaty and wanting to escape.

Intervening does take time, but it is productive time both for thoroughness and usefulness of the assessment and for enhancing the client’s confidence that he or she can take initiative to change course. Moreover, too often the client’s experience of being evaluated is one of being seen as an object whose immutable characteristics are being measured. Collaborative, interventional assessment counters that danger while providing useful life-world instances with which a report can be written.

Describe via Structural Re-Presentation

When the assessment process has been collaborative, contextual, and interventional, then the assessment report can represent the client via re-presented concrete life instances as witnessed during the assessment and as reported by the client. This representation is structural, or holistic, in the sense that referral issues are addressed by providing an understanding of how events have occurred and the circumstances under which the person has found alternative means of coping. Explanation is not a reduction to abstract traits or other personality characteristics. Adequate descriptions leave no gaps or incongruities in addressing the referral issues. When uncertainties and disagreements between client and assessor remain, these are spelled out.

For example, instead of saying, as at first recommended by an assessment supervisor, “The applicant’s resistance to authority was evident in his initial refusal to draw human figures,” an individualized report said,

Mr. Turner at first said he would not draw a person, because he couldn’t draw well and he didn’t know how his drawing would be interpreted. When I explained that we would discuss his drawing together, just as we had done with the Bender, he tentatively rendered a diminutive marine in the upper left corner of the paper; the marine held a rifle. Later we agreed that he sometimes has been perceived as uncooperative and even as obstructionistic, when to the contrary, he has felt self-conscious and that he had to defend himself. However, Mr. Turner was decidedly opposed to exploring ways of helping other people to understand his perspective. He insisted, “it’s on them to change.”

This life-world description also presents Mr. Turner dynamically and holistically rather than explaining parts of the picture in terms of other parts.

Yes, collaborative reports typically are longer than other reports, but they also typically are more concrete and useful. First-person form encourages readers to imagine their similarities and differences in regard to the assessor and to imagine what findings might have emerged with themselves or another assessor. Description via action verbs with context provided, rather than via psychological adjectives and constructs, helps readers to picture the client and to imagine how he or she might create different contexts more conducive to satisfying outcomes. Inclusion of when-nots discourages readers from totalizing the client as one type of person. All of these features lead to suggestions to the client and to helpers that are doable and expandable by client and helpers.

An example from the “Suggestions” section of a report:

2. Doris [Bellini], when you find yourself moving toward a temper stance (tight fists, squinted eyes, wanting to shout), remember to slow down as you did when you rewrote that sentence. Ask yourself where you want to get, imagine your coach or copilot being there for you, and take a nontemper course, like when you put those ideas in the suggestion box instead of being mad.

This form of suggestion re-presents and evokes past action that now can become available for the client in difficult situations.

Additional examples and guiding principles can be found in Fischer’s publications, such as Individualizing Psychological Assessment (1994b)—a textbook that includes sample reports and tables on writing to individualize. Fischer (1994a, 1994c, 1998b) provides a life-world approach to the Rorschach, a philosophical base for collaborative, individualized assessment (1979, 1998a), and overviews of principles and practices (2000, 2001).

CONFLUENCE OF COLLABORATIVE ASSESSMENT WITH DIVERSE THEORIES

The authors of this chapter found their way to collaborative, life-focused assessment through postmodern philosophy, primarily European phenomenology, hermeneutic epistemology, and existentialism. Ours is a specifically human-science psychology whose empirical data are shareable life-world events. However, clinicians from many theoretical orientations develop their own collaborative practices, especially when the focus is on the life world. Indeed, the theories we touch on in the following sections have influenced our practices.

Cognitive Behavioral Psychology

Cognitive behavioral psychologists such as Albert Ellis (1973) and Aron Beck (1976) have respected the conjoint involvement of cognition, affect, behavior, and physiology, as has human-science psychology. “Cognition” is increasingly coming to include personal experience and meaning, and “behavior” is increasingly regarded as action as well as consequence (cf. Mahoney, 1991), similar to what we referred to as shaped/ shaping comportment.

In both cognitive behavioral therapy and collaborative assessment, the psychologist helps the client to discover that a problematic comportment is not as pervasive or absolute as he or she has believed. Identifying the when-nots of problematic behavior opens the way for the client to reframe negative anticipations and to assert him- or herself in productive whennot manners. In both cognitive behavioral therapy and collaborative assessment, the psychologist and client explore what the client means by such presentations as “anxiety pervades my life,” and “I lack self-esteem.” In both enterprises, life-world examples, contexts, and when-nots are explored. Assumptions are lightened, and opportunities to try revised behavior are created.

Narrative Psychology

McLeod (1997) writes that narrative underpins all psychotherapy; that whatever theoretical approach is taken, presentations are in terms of telling and retelling stories. Although “narrative psychology” is relatively new, narratives—stories are one of the principal ways in which humans make sense of themselves, others, and the world. This chapter’s excerpts of collaborative dialogue can be seen as joint efforts to story the client’s situation and his or her past behavior and experience, and to develop options. We also have written assessment reports for children in the form of fables, stories whose metaphors are understood by both child and parents as they read and reread the fable together.

Both narrative psychology and collaborative assessment regard the client as co-author (often with unchosen contingencies) of his or her life. As in narrative therapy, the assessment client discovers while traveling through projective assessment material with the psychologist that he or she is the protagonist, an agent, who participates in even what happens to him or her. The Thematic Apperception Test is a wonderful device through which the client may become aware of habitual co-authoring, especially when the assessor requests that an alternative story be told.

Depth Psychology

Carl Jung (1957) and James Hillman (1983) are the major authors of depth psychology, but they both draw on the philosophers of antiquity, such as Finino, Vico, and Heraclitus, as well as on Renaissance figures and themes. Depth psychology gives primacy to the soul of both humans and the worlds in which they live. The mode of access to the soul is images, both psychic and those encountered in the world. Projective materials provide opportunities for understanding the client’s life-world situations as assessment situations evoke images, fantasies, and recollections of earlier events. In this process, understandings and personal change occur through the client’s and the professional’s compassionate receptivity to these images.

Gestalt Psychotherapy

Fritz Perls (1969) drew on classical analysis, gestalt perceptual theory, Zen philosophy, and the work of Jung, Reich, and Moreno. He thought of his work as a new gestalt, an integration that went beyond the sum from earlier traditions. Gestalt therapy emphasizes the relationship between client and therapist as a here-and-now experience within a caring I-Thou (Buber, 1970) frame. In-session exercises bring attention to both verbal and nonverbal aspects of experience. The client becomes aware of his or her posture and voice and of the responses elicited from others, including the therapist. Dialogue about differences between the therapist’s and the client’s perspectives increases the client’s awareness of how he or she has experienced various situations and has contributed to having become stuck in them.

The following two interventions highlight some similarities between collaborative assessment and gestalt therapy. A client has drawn himself, during the Draw-A-Person technique, as smaller and less facially expressive than the drawings of “a person” and of “a person of the other gender.” After discussion, the assessor asks the fellow to draw himself on the same sheet with “a person,” but this time with himself larger and expressive. During this exercise, the client announces that he has discovered that holding himself back makes him feel scared even when there’s no reason to feel that way. A gestalt therapist says to a client, “I noticed that your voice changed when you talked to your mother [empty chair technique]. Your voice seemed low and fearful. Would you like to try something new, here, where you feel safe? Could you say loudly and distinctly what you just said to your mother?”

Developmental Theories

Heinz Kohut (cf. Kahn, 1991) from within a self-psychology tradition addresses preoedipal development and a continuing lifelong development of the self. Fullest development occurs when three needs are met: the needs to be mirrored, to idealize, and to be like others and to belong (twining). In collaborative assessment, mirroring occurs as the assessor checks impressions with the client’s experience, for example, “I would guess that you’re encountering a lot of danger these days, not

too different psychologically from the monsters, clawed bats, and fanged animal head.” In regard to idealization, during collaborative exploration, the client experiences the assessor’s competence as the latter bridges theory, techniques, and life world in a trustworthy manner. Nondefensively acknowledged misguesses (corrected by the client) allow the client both to identify with the power and knowledge of the assessor and to discover his or her own power and knowledge. In regard to the need to belong, the client feels recognized and understood as he or she is treated as a person with this other person, jointly engaged in the assessment enterprise.

L.S. Vygotsky’s (1987) notion of the “zone of proximal development” is implicit in collaborative assessment. This zone is one in which the child is developmentally ready but has not yet mastered a skill. Collaborative learning exercises assist the person’s appropriation of principles. An effective assessment intervention evaluates how ready the client is to appropriate an understanding or course of action. Collaborative trying-out, like drawing oneself larger, suggests how the client might practice a skill in life, moving from approximation to actualization.

Classic and Contemporary Assessment Approaches

As is often the case in the evolution of the social sciences, the early figures in psychological assessment were intrigued by the whole individual coping with a complex world. Henry Murray’s (1938) personality theory and use of his Thematic Apperception Test portrayed the person as being pushed forward in unchosen ways from the past, as well as purposefully projecting self into an environment that both accommodated and pressed back. Gordon Allport (1937, 1961) encouraged a case study method of engaged involvement with the subject. He distinguished such “ideographic” efforts to describe and understand a particular event or individual from “nomothetic” efforts to gather data from many people in order to develop general laws or patterns. George Kelly (1955) proposed that the purpose of testing should be to “survey the pathways along which the subject is free to move.” That subject could explore, reflect on, and revise his or her ways of construing the world. Bruno Klopfer (e.g., et al. 1954) characterized his approach to the Rorschach as phenomenological, and his students celebrated this projective technique for being a means of going beyond sterile aptitude testing, toward assessing the whole, dynamic individual—the perceiving/thinking/feeling/ defending/acting person. Molly Harrower (1956) promoted projective counseling, wherein the client and counselor discussed the former’s inkblot responses, allowing the client to gain insight while helping the assessor to refine his or her psychodynamic exploration. Harrower’s writings are filled with the excitement of clients’ liberating self-discoveries via the Rorschach testing process. (Fischer, 1994a, p. 202)

594 Collaborative Exploration with Projective Techniques: A Life-World Approach

On the contemporary scene, Stephen Finn (1996, 1997; Finn & Tonsager, 1992) has been a major innovator and teacher of therapeutic psychological assessment. His presentations are in demand by international assessment groups. See also the Handler and Hilsenroth (1998) volume Teaching and Learning Personality Assessment, in which the work of many authors intersects with this chapter; see especially chapters by P. Erdberg, S.E. Finn, M. Hilsenroth, H. Lerner, P. Lerner, and H. Potash. The Humanistic Psychology division of the American Psychological Association has published a special double issue of its journal The Humanistic Psychologist (2002), on “Humanistic Approaches to Psychological Assessment.”

These brief characterizations of theoretical confluence indicate touch points across perspectives, through which psychologists can ground and expand their own ways of pursuing collaborative assessment.

REFERENCES

• Allport, G.W. (1937). Personality: A psychological interpretation. New York: Holt.
• Allport, G.W. (1961). Pattern and growth in personality. New York: Holt, Rinehart & Winston.
• Beck, A. (1976). Cognitive therapy and the emotional disorders. New York: International Universities Press.
• Buber, M. (1970). I-Thou. New York: Scribner’s.
• Ellis, A. (1973). Humanistic psychotherapy: The rational-emotive approach. New York: McGraw-Hill.
• Finn. S.E. (1996). A manual for using the MMPI-2 as therapeutic intervention. Minneapolis: University of Minnesota Press.
• Finn, S.E., & Martin, H. (1997). Therapeutic assessment with the MMPI-2 in managed health care. In J.N. Butcher (Ed.), Objective psychological assessment in managed health care: A practitioner’s guide (pp. 131–152). New York: Oxford University Press.
• Finn, S.E., & Tonsager, M.E. (1992). Therapeutic effects of providing MMPI-2 test feedback to college students awaiting therapy. Psychological Assessement, 4, 278–287.
• Fischer, C.T. (1979). Individualized assessment and phenomenological psychology. Journal of Personality Assessment, 43, 115– 122.
• Fischer, C.T. (1994a). Humanizing psychological assessment. In F. Wertz (Ed.), The humanistic movement: Recovering the person in psychology (pp. 202–214). Lake Worth, FL: Gardner Press.
• Fischer, C.T. (1994b). Individualizing psychological assessment. Mahwah, NJ: Erlbaum. (Original work published in 1985)
• Fischer, C.T. (1994c). Rorschach scoring questions as access to dynamics. Journal of Personality Assessment, 62, 515–525.
• Fischer, C.T. (1998a). Phenomenological, existential, and humanistic foundations for psychology as a human science. In M. Hersen & A. Bellack (Eds.), Comprehensive clinical psychology: Volume 1. Foundations (pp. 449–472). London: Elsevier Science.
• Fischer, C.T. (1998b). The Rorschach and the life-world: Exploratory exercises. In L. Handler & M. Hilsenroth (Eds.), Teaching and learning personality assessment (pp. 347–358). Hillsdale, NJ: Erlbaum.
• Fischer, C.T. (2000). Collaborative, individualized assessment. Journal of Personality Assessment, 74, 2–14.
• Fischer, C.T. (2001). Collaborative exploration as an approach to personality assessment. In K.J. Schneider, J.F.T. Bugenthal, & J.F. Pierson (Eds.), The handbook of humanistic psychology: Leading edges in theory, research, and practice (pp. 449–472). Thousand Oaks, CA: Sage.
• Fischer, C.T. (Guest ed.). (2002). Humanistic approaches to psychological assessment [special issue]. Humanistic Psychologist, 30.
• Handler, L., & Hilsenroth, M. (Eds.) (1998). Teaching and learning personality assessment (pp. 347–358). Hillsdale, NJ: Erlbaum.
• Harrower, M. (1956). Projective counseling: A psychotherapeutic technique. American Journal of Psychotherapy, 20, 74–86.
• Hillman, J. (1983). Healing fiction. Barrytown, NY: Station Hill.
• Jung, C. (1957). The undiscovered self (R.F.C. Hull, Trans.). New York: New American Library.
• Kahn, M. (1991). Between the therapist and client: The new relationship, New York: W.H. Freeman.
• Kelly, G. (1955). The psychology of personal constructs (Vols. 1 & 2). New York: W.W. Norton.
• Klopfer, B., Ainsworth, M. D., Klopfer, W.G., & Holt, R.R. (1954). Developments in the Rorschach technique: I. Theory and technique. Yonkers-on-Hudson, NY: World Book.
• Mahoney, M.J. (1991). Human change processes. New York: Basic Books.
• McLeod, J. (1997). Narrative and psychotherapy. London: Sage.
• Murray, H.A. (1938). Explorations in personality. New York: Oxford University Press.
• Perls, F. (1969). Gestalt therapy verbatim. Moab, UT: Real People Press.
• Vygotsky, L.S. (1987). Problems of general psychology. New York: Plenum.

PART SEVEN

APPLICATIONS FOR CHILDREN AND ADOLESCENTS

CHAPTER 45

Sentence Completion Measurement of Psychosocial Maturity

P. MICHIEL WESTENBERG, STUART T. HAUSER, AND LAWRENCE D. COHN

TEST DESCRIPTION 596 Test Protocols and Administration Procedure 596 Scoring Manuals and Rating Procedure 596 THEORETICAL BASIS 598 Jane Loevinger’s Conceptualization of Ego Development 598 Ego Development Stages 598 Age and Ego Development 600 Ego Development in Children and Adolescents 600 TEST DEVELOPMENT 601 Washington University Sentence Completion Test 601 Sentence Completion Test for Children and Youth 602 PSYCHOMETRIC CHARACTERISTICS 603

The use of a projective rather than an objective test is justified by the fact that, in this context, [psychosocial] development is taken to mean (or perhaps to reflect) the person’s frame of reference; thus a format requiring the subject to project his or her own frame of reference is preferable to providing a clearly stated set of questions, reflecting the test constructor’s frame of reference.

—Loevinger (1998, p. 30.)

The assessment of psychological and social maturity has attracted increasing attention during the past two decades due to the relevance of these constructs for clinical practice, organizational settings, and research protocols. Yet few psychoRANGE OF APPLICABILITY AND LIMITATIONS 604 CROSS-CULTURAL ISSUES 605 ACCOMMODATION FOR POPULATIONS WITH DISABILITIES 605 LEGAL AND ETHICAL CONSIDERATIONS 605 COMPUTERIZATION 606 CURRENT RESEARCH STATUS 606 USE IN CLINICAL AND ORGANIZATIONAL PRACTICE 608 FUTURE DEVELOPMENTS 611 NOTES 612 REFERENCES 612

metrically sound instruments have been developed to assess peychosocial maturity. A sentence completion test for measuring maturity in adults was developed by Loevinger and her colleagues: the Washington University Sentence Completion Test (WUSCT; Loevinger, 1985). A version for use with children and youth (ages 8 and older) was recently developed by Westenberg and his colleagues: the Sentence Completion Test for Children and Youth (SCT-Y; Westenberg, Treffers, & Drewes, 1998). Both instruments are based on Jane Loevinger’s theory of ego development, which portrays personality growth as a series of developmental advances in impulse control, interpersonal relations, and conscious preoccupations. Advances in these domains are depicted in terms

596 Sentence Completion Measurement of Psychosocial Maturity

of “stages,” a term that implies an underlying coherence and structure to personality growth.

The WUSCT and SCT-Y are both scored using empirically based scoring manuals based on research with thousands of respondents. Research has indicated excellent reliability, construct validity, and clinical utility. A critical examination of the scientific status of projective techniques noted that the WUSCT “is arguably the most extensively validated projective technique” (Lilienfeld, Wood, & Garb, 2000, p. 56). The two instruments illustrate how projective measures can (and should) meet rigorous psychometric standards.

This chapter provides an overview of the theoretical and empirical basis of the WUSCT and SCT-Y as measures of psychosocial maturity. The chapter also reviews the practical uses of these measures in clinical and organizational settings.

TEST DESCRIPTION

The WUSCT was originally constructed for use with adults (see Loevinger, 1998); the SCT-Y was recently constructed for use with children and adolescents (Westenberg, Treffers, et al., 1998). Both instruments consist of a sentence completion protocol as well as empirically based scoring manuals with detailed administration and scoring procedures.

Test Protocols and Administration Procedure

The WUSCT contains 36 sentence stems (Table 45.1). Several versions of the instrument have been published since 1970. The most recent version of the test, Form 81, is the form that is currently recommended for use because it served as the basis for the revised scoring manual that was published in 1996 (Hy & Loevinger, 1996; Loevinger, 1985, 1998). The SCT-Y (Westenberg, Treffers, et al., 1998) contains 32 items; these items are also presented in Table 45.1. Twentyone of the items on the SCT-Y are identical to the items on Loevinger’s Form 81 of the WUSCT; five additional items were revised slightly and six items were newly constructed in order to create a form that was appropriate for use with older children and adolescents.

The items of the WUSCT and SCT-Y address a variety of issues, including how respondents perceive and respond to personal relationships (e.g., “My mother and I—”), authority (e.g., “Rules are—”), frustration (e.g., “If I can’t get what I want—”), and everyday issues (e.g., “Raising a family—”). The instrument is considered semiprojective because the sentence stems (items) provide respondents with some initial structure while also providing respondents with an opportunity to “project” their viewpoint or frame of reference when completing the sentence stems. These instructions are printed on the top of page 1 of the WUSCT: “Complete the following sentences.” The SCT-Y instructs respondents to “Complete the following sentences in any way that you wish.” The phrase “in any way that you wish” was added because young respondents frequently ask for additional instructions. If subjects request further information, the administrator should make a nonsuggestive answer, such as “There are no right or wrong answers.” Instructions aimed to motivate subjects to show their best selves make it a different test and should be avoided (see the Psychometric Characteristics section).

The WUSCT is printed on two pages: the first 18 items are presented on page 1, and the remaining 18 items are presented on page 2. The SCT-Y is also printed on two pages (16 items on page 1 and 16 items on page 2). Each set of items can be used as an independent short form of the test, although this strategy should be avoided when possible because of reduced reliability (e.g., Drewes & Westenberg, 2001; Novy & Francis, 1992). The use of one test half is not recommended in the assessment of individual clients. Testing individuals requires optimal reliability and stability of test scores.

The WUSCT and SCT-Y have separate forms for males and females (see Table 45.1). For the WUSCT, 30 items are identical on male and female forms (e.g., “When I am criticized—”) and 6 items are closely comparable (e.g., Female form: “The worst thing about being a woman—”; Male form: “The worst thing about being a man—”). For the SCT-Y 28 items are identical on both forms (e.g., “If I were in charge—”) and 4 items are closely comparable (e.g., Female form: “When I am with a boy—”; Male form: “When I am with a girl—”).

The WUSCT and the SCT-Y can be administered individually or in group settings. The instruments can be administered in written format or they can be administered orally. The written format represents the standard administration procedure and it is least likely to be subject to response bias (e.g., subjects seeking to respond in a socially desirable manner). That is, the written procedure fits best with the purpose of the test: to reveal the respondent’s frame of reference without distortion that might arise from the presence of the administrator. However, an oral administration, conducted with care, does not appear to distort test scores (e.g., McCammon, 1981; Westenberg, van Strien, & Drewes, 2001). If administered orally, the investigator should abstain from making comments or posing follow-up questions. In other words, the oral procedure should closely mimic the written procedure.

Scoring Manuals and Rating Procedure

Detailed scoring manuals are available for the WUSCT (Hy & Loevinger, 1996) and for the SCT-Y (Westenberg et al.,

Washington University Sentence Completion Test (WUSCT; Loevinger, 1985, 1998)	Sentence Completion Test for Children and Youth (SCT-Y; Westenberg, Treffers, et al., 1998)
First Page 1. When a child will not join in group activities 2. Raising a family 3. When I am criticized 4. A man’s job 5. Being with other people 6. The thing I like about myself is 7. My mother and I 8. What gets me into trouble is 9. Education 10. When people are helpless 11. Women are lucky because 12. A good father 13. A girl has a right to 14. When they talked about sex, I 15. A wife should 16. I feel sorry 17. A man feels good when 18. Rules are	First Page 1. When a child will not join in group activities 2. Raising children 3. When I am criticized 4. If I were in charge 5. Being with other people 6. The thing I like about myself is 7. My mother and I 8. What gets me into trouble is 9. Education 10. When people are helpless 11. When I am afraid 12. A good father 13. My biggest fear 14. I feel sorry 15. When they avoided me 16. Rules are
Second Page	Second Page
19. Crime and delinquency could be halted if 20. Men are lucky because 21. I just can’t stand people who 22. At times he (she) worried about 23. I am 24. A woman feels good when 25. My main problem is 26. A husband has a right to 27. The worst thing about being a man (woman) 28. A good mother 29. When I am with a woman (man) 30. Sometimes he (she) wished that 31. My father 32. If I can’t get what I want 33. Usually he (she) felt that sex 34. For a woman a career is 35. My conscience bothers me if	17. Crime and delinquency could be halted if 18. Women (Men) are lucky because 19. I just can’t stand people who 20. At times I worry about 21. I am 22. A boy (girl) feels good when 23. My main problem is 24. Good friends 25. The worst thing about being a man (woman) 26. A good mother 27. When I am with a girl (boy) 28. Sometimes I wished that 29. My father 30. If I can’t get what I want 31. My conscience bothers me if 32. I felt proud that I

TABLE 45.1 Sentence Completion Measures of Ego Development

Note. Some items have male and female forms. The female form is placed in parentheses.

2000).1 Both instruments are scored using a two-step procedure: First, each response (i.e., sentence completion) on a protocol is assigned to a developmental level independently of every other response on the protocol. Thus the WUSCT would yield 36 independent ratings and the SCT-Y would yield 32 independent ratings. Second, the distribution of item response ratings is converted into a single total protocol rating (TPR). Both Loevinger and Westenberg provide an algorithm (rule) for converting the distribution of item response ratings into a single TPR (Hy & Loevinger, 1996, pp. 38– 39; Westenberg et al., 2000, pp. 77–80). This single TPR represents the respondent’s core level of psychosocial maturity. Alternatively, an item sum score could be used if a continuous rating scale is required for data analytic purposes in the context of a research study.

Assigning each response (sentence completion) to a developmental level is accomplished by using empirically derived scoring manuals. Separate manuals have been constructed for each item (sentence stem). Each scoring manual is composed of verbatim responses that were collected from heterogeneous samples of individuals; each of these responses has been empirically assigned to one of eight developmental levels (see the Test Development section). The “sentence completion” responses that are presented in the scoring manual are not organized haphazardly; rather, the responses are grouped according to the content of the response (e.g., interpersonal, behavioral), which facilitates the rating process.

Raters first seek to determine if a client’s response (sentence completion) is identical (or nearly identical) to the examples that are presented in the scoring manual. Due to the

598 Sentence Completion Measurement of Psychosocial Maturity

detail presented in the scoring manual, a surprising number of such matches are obtained. When a client’s response cannot be matched to an example in the scoring manual then the rater searches for other aspects of the response that are characteristic of psychosocial development, such as the presence of time perspective, qualification, and contingency. The scoring manual details many additional signs of psychosocial development that are used to score responses not readily matched with the examples in the manual.

It is important to emphasize that rating the WUSCT and SCT-Y is an empirical, not intuitive, process. Self-training exercises have been carefully developed to help novice raters master the scoring procedure (Hy & Loevinger, 1996, pp. 41– 87). These exercises can probably be completed in 2 to 3 weeks if raters practice 2 hours each day. Training is essential: “persons who tried to bypass the training . . . were almost never correct in their ratings” (Hy & Loevinger, 1996, p. 32).

THEORETICAL BASIS

The theoretical basis of the WUSCT and SCT-Y is provided by Loevinger’s conceptualization of ego development.2 Her conceptualization of ego development was mostly based on sentence completion responses from (young) adult samples (see Loevinger, 1998). Sentence completion data obtained from a large sample of children and adolescents necessitated a revised description to adequately conceptualize ego development in children and youth (Westenberg, Jonckheer, Treffers, & Drewes, 1998).

Jane Loevinger’s Conceptualization of Ego Development

Loevinger (1976, 1997) portrays psychosocial maturation as a series of changes in impulse control, interpersonal relations, and conscious preoccupations. Developmental advances in these domains are depicted in terms of stages, a term that implies an underlying coherence and structure to personality. Eight developmental stages have been identified by Loevinger and her colleagues, and each stage (described in the next section) is defined by a characteristic set of capacities (e.g., impulse control) and milestone developments (e.g., a concern with self-evaluated standards). More generally, each developmental stage is defined by a characteristic way of perceiving and responding to the social world. The term ego refers to a “striving to master, to integrate, [and] to make sense of experience” (Loevinger, 1976, p. 59). For Loevinger the “ego” is an abstraction, not an extant structure; thus, she describes the ego informally, referring to it as “a frame of reference” or “lens” through which individuals perceive their world (ego development thus represents a change in one’s frame of reference).

Despite its psychometric origins (see the Test Development section), Loevinger’s developmental model is often linked to stage theories that have Piagetian origins, such as Kohlberg’s (1969) model of moral reasoning and Selman’s (1980) model of the growth of social cognition (see Kegan, 1982; Snarey, 1998). Several of these models do indeed show some similarities with Loevinger’s stages of ego development. Yet in many respects Loevinger’s model does not fit nicely in the cognitive-developmental tradition because it is not a model of cognitive growth or reasoning per se; instead, ego development is primarily concerned with “impulses and methods for controlling impulses, personal preoccupations and ambitions, interpersonal attitudes and social values what psychologists normally call personality” (Blasi, 1998, p. 15).

Ego Development Stages

Loevinger has identified the milestone achievements that seem to characterize each developmental stage (Loevinger, 1976, 1997). A brief description of these stages is provided here; an overview is presented in Table 45.2.

The first stage depicted by Loevinger is labeled the Impulsive stage (E-2).3 Individuals at this stage are characterized by aggressive and sexual impulsivity, egocentrism, and the pursuit of immediate desires. Other people are expected to attend to one’s needs and desires; frustration is not easily accepted and it is reacted to impulsively. Impulsive individuals are also oppositional and defiant; they view rules as arbitrary and punishment as retaliatory; hence they require external constraints for regulating their impulses. Individuals at the Impulsive stage understand their emotional world (inner life) in simple, somewhat impoverished, terms. Likewise, their social world is typically described in simple dichotomies; for example, people are described as either good or bad, nice or mean. Examples of sentence stems and their completions include:4 “If I can’t get what I want—I get pissed off”; “A good father—should give his daughter anything she wants”; “A good mother—is nice.”

The next developmental stage (labeled the Self-Protective stage, E-3) is characterized by a preoccupation with issues of control, trouble, opportunism, and the manipulation of other people. Relationships are often exploitative and manipulative, an interpersonal style that encourages the Self-Protective person to be punitive and wary of the intentions of other people. Finally, hedonism is paramount during the Self-Protective period, and the easy life is perceived as the good life. Examples of sentence stems and their completions include: “If

Ego Level	Characteristics
	Impulse Control	Interpersonal Mode	Conscious Preoccupations
E-2. Impulsive	Impulsive	Egocentric, dependent	Bodily feelings
E-3. Self-Protective	Opportunistic	Manipulative, wary	“Trouble,” control
E-4. Conformist	Respect for rules	Cooperative, loyal	Appearances, behavior
E-5. Self-Aware	Exceptions allowable	Helpful, self-aware	Feelings, problems, adjustment
E-6. Conscientious	Self-evaluated standards, self-critical	Intense, responsible	Motives, traits, achievements
E-7. Individualistic	Tolerant	Mutual	Individuality, development, roles
E-8. Autonomous	Coping with conflict	Interdependent	Self-fulfillment, psychological causation
E-9. Integrated		Cherishing individuality	Identity

TABLE 45.2 Some Characteristics of Ego Development Levels

Note. From Technical foundations for measuring ego development (p. 5), by J. Loevinger, 1998. Mahwah, NJ: Erlbaum.

I can’t get what I want—I beg and crying works with my father “;”Raising a family—I want my family to obey me”; “A man feels good—when he has fun.”

The fourth developmental stage is labeled the Conformist stage (E-4). This stage is characterized by a concern with social norms and social approval. Individuals at this stage are attentive to the expectations and opinions of other people and believe that everyone is, or ought to be, the same, just as rules apply to everyone. Individual differences in beliefs, personality, and behavior are regarded as inappropriate and undesirable. Social approval is highly valued, while social disapproval is feared and avoided. Note, however, that conformity should not be equated with conventionality: Persons at the Conformist ego level might adhere rigidly to nonconventional norms. Examples of sentence stems and their completions include: “If I can’t get what I want—I look sad and pout”; “Being with other people—is good for everyone”; “A woman should always—be friendly and nice.”

Loevinger’s fifth developmental stage is labeled the Self-Aware stage (E-5). It is characterized by an awareness of being different from other people and of having private feelings, views, opinions, and ideas. This self-awareness is accompanied by a growing appreciation of individual differences among people. Both types of advances represent different sides of the same developmental coin: Increasing self-awarenessand examination of inner life is accompanied by an increasing awareness of being different from other people. Distinctions between public and private aspects of oneself are recognized, and there is an increasing belief that it is important to be true to one’s inner self. Individuals at the Self-Aware stage begin to recognize that morality is not absolute and that exceptions to the rules are acceptable. Examples of sentence stems and their completions include: “If I can’t get what I want—I fight for it, then I get disappointed”; “A woman should always be true to what she feels inside”; “A good mother—comes in many different packages.”

The sixth developmental stage described by Loevinger is labeled the Conscientious stage (E-6). The pursuit of selfevaluated standards is one of the hallmarks of this stage, a pursuit that is characterized by a preoccupation with goals, accomplishments, ideals, and issues of conscience. Conscientious individuals are often self-critical and concerned with self-improvement, two traits that again reflect an underlying concern with self-defined goals and self-evaluated standards. Individuals at this stage also display a strong sense of responsibility for their actions and choices in life, as well as the actions and choices of other people. An increasing capacity for psychological awareness and time perspective is also characteristic of this stage. Examples of sentence stems and their completions include: “If I can’t get what I want— I sometimes act immaturely”; “I am—a procrastinator”; “A good mother—loves, cares for, and takes on a great responsibility in raising her children.”

The seventh developmental stage is referred to as the Individualistic stage (E-7). This stage is characterized by a clear sense of personal identity, psychological development, and psychological causation. Individuals at this stage have a complex understanding of personal relationships and the multiple roles that a person may simultaneously have in life (e.g., daughter, mother, spouse, professional). Instead of trying to change oneself and other people to fit an ideal image (a pursuit that is more characteristic of the Conscientious stage) people at this stage also recognize and comment on their own contradictory emotions, motivations, and related inner conflicts. Examples of sentence stems and their completions include: “If I can’t get what I want—I get annoyed or sometimes I resign myself, depends on the situation”; “My mother and I—are probably more alike than I tend to admit”; “I am—emotionally independent and physically dependent.”

The eighth stage is labeled the Autonomous stage (E-8). Individuals at this stage are no longer preoccupied with issues of achievement, goals, or morality; thus, individuals at this

600 Sentence Completion Measurement of Psychosocial Maturity

stage are no longer preoccupied with evaluating their own actions and the actions of other people. Indeed, this stage is characterized by a respect for other people’s need for autonomy, finding their own way, and making their own mistakes. Individuals at this stage also display an appreciation of life’s paradoxes, contradictions, and inconsistencies, an appreciation that gets expressed in existential rather than hostile humor. Examples of sentence stems and their completions include: “My mother and I—love each other enough to respect each other’s private life”; “At times she worried about—the future so much she forgot to enjoy the present”; “The worst thing about being a man—is that you can easily come to expect too much of yourself and others.”

The ninth and last ego development stage is labeled the Integrated stage (E-9). Loevinger suggests that Maslow’s self-actualizing person may best capture the personality characteristics of this stage. Details of this stage are limited because of the rarity of subjects; Loevinger speculates that fewer than 1% of urban residents reach this developmental milestone. She includes this stage in her model (and scoring manual) for theoretical reasons but the details of this stage are not relevant for most clinical practice or research purposes.

Age and Ego Development

Many developmental theories assume a close association between age and developmental advances, as if age is a sufficient condition for development. In contrast, Loevinger’s conception of ego development is only loosely related to age. An individual’s pace, and extent, of ego development depends on many influences beyond the mere passage of time. Recent studies have identified some of these influences, both environmental and hereditary (e.g., Allen, Hauser, Bell, & O’Connor, 1994; Hauser et al., 1984; Newman, Tellegen, & Bouchard, 1998). Loevinger proposes that age may be a necessary but not sufficient condition for development. Thus within any age cohort (e.g., 22-year-olds) there will be individuals at a range of ego levels. Likewise, across any time period (e.g., 5 years, 10 years) one may find different trajectories of ego development; for example, some individuals may advance from the Self-Protective to the Conformist stage, while other individuals may advance from the Self-Protective to the Conscientious stage (Westenberg & Gjerde, 1999).

Loevinger suggests that ego development “is at once a developmental sequence and [italics added] a dimension of individual differences in any age cohort” (Loevinger, 1976, p. 13). From this perspective ego development can be viewed as a personality typology: Any age cohort will include a range of ego levels and these different ego levels, in effect, represent different types of people. However, it is only a quasi-typology because individuals may mature and, thus, change, moving from one position in the typology to another (Loevinger, 1976). The loose relationship between age and ego level is clearly reflected in Loevinger’s scoring manual, which instructs raters to ignore a respondent’s age when evaluating the developmental level of a response. The loose association between age and ego level is also reflected in Loevinger’s portrayal of ego levels, which are never described in agespecific terms.

To what extent, however, are Loevinger’s depiction and measurement of ego development stages truly independent of age? Loevinger and her colleagues relied mostly on adult samples when describing the characteristics of each developmental stage (see Loevinger, 1985, 1993, 1998; Loevinger & Wessler, 1970). The absence of child and adolescent samples raises two critical questions: (1) Can the characteristic descriptions of each ego stage be regarded as truly ageindependent, and (2) Can a scoring manual that was derived mainly from adult responses be used to assess the ego level of children and adolescents? Both questions highlight a more fundamental question: What is the “normal” course of ego development from middle childhood through late adolescence, and how can it be measured? Studies conducted by Westenberg and his colleagues address the latter questions.

Ego Development in Children and Adolescents

A sentence completion test for assessing ego development in children and adolescents was recently developed and crossvalidated by Westenberg and his colleagues, based on data obtained from more than 2,700 children, adolescents, and young adults, ages 8 to 25 (SCT-Y; Westenberg, Jonckheer, et al., 1998; Westenberg, Treffers, et al., 1998; Westenberg et al., 2000). The scoring manual for the SCT-Y was constructed using the same psychometric procedures employed by Loevinger in the development of her manual for assessing adult development (see the Test Development section). Notably the new scoring manual introduced important changes into our understanding of the characteristic signs of ego development displayed by youth. These changes were most notable at the Impulsive, Self-Protective, and Conformist stages, where children and adolescents display more positive signs of development, and less malignancy, than revealed in Loevinger’s studies of adults. The major revisions introduced by Westenberg are described in the following sections.

Impulsive Stage

Unlike adults, children at the Impulsive stage are not characterized by antisocial attitudes, blatant aggression, or the absence of empathic tendencies. Nor are children at this stage characterized by an oppositional-defiant attitude, as is characteristic of Impulsive adults. Instead, the sentence completion responses of children at this stage show some capacity for prosocial interactions, a receptive attitude toward rules, and a dependent coping style. The latter themes do not typically appear in adult protocols until the Conformist stage of development. For example, the response “When people are helpless—I help” is categorized at the Conformist level in Loevinger’s manual, but is one of the most reliable indicators of the Impulsive level in the Westenberg manual. Socially, Impulsive children are also characterized by a preoccupation with impulses; these children also display a tendency to dichotomize their social world (e.g., judging people to be either good or bad, nice or mean). Impulsive children are also characterized by an absence of cognitive complexity, which leads children to reduce abstract ideas to concrete examples. The latter signs of ego development are consistent with Loevinger’s description of this stage. Examples of sentence stems and their completions include: “If I can’t get what I want—I ask my father”; “My conscience bothers me if—I have done something wrong”; “Being with other people—is doing fun things.”

Self-Protective Stage

Unlike adults, Self-Protective youths are not characterized by “an outright manipulative and exploitative attitude toward other people.” Likewise, interpersonally malignant forms of controlling other people are not characteristic of youth at this stage although, like their adult counterparts, Self-Protective youths are preoccupied with issues of control. Self-Protective youth emphasize self-reliance and self-sufficiency. Self-Protective youth are also characterized by a simple laissez-faire orientation toward independence, believing that everyone should be allowed “to do one’s own thing.” For example, in response to the sentence stem “When a child will not join group activities—” Self-Protective youth sometimes write: “I don’t mind” or “it’s up to him.” Notably, in Loevinger’s scoring manual these responses were empirically located at the Conformist stage, whereas in Westenberg’s manual for youth these same responses are empirically located at the Self-Protective stage. Additional examples of sentence stems and their completions include: “If I can’t get what I want—I couldn’t care less”; “I feel sorry—for poor people”; “A good father—is mine.”

Conformist Stage

In Loevinger’s manual a helpful interpersonal style is attributed to the Self-Aware stage, while in Westenberg’s manual for youth this helpful orientation is more characteristic of the Conformist stage. Conformist adults primarily describe interpersonal interactions in terms of actions rather than feelings. In contrast, Conformist youngsters appear to put greater emphasis on feelings rather than concrete actions in specific situations. In addition, Conformist youths are not just geared toward concrete rules in specific situations but are also geared toward more abstract interpersonal norms and values, such as reciprocity and equality. Otherwise, the overlap between Loevinger’s and the revised description of the Conformist ego level is considerable. The overlap for the Self-Aware ego level and beyond is almost perfect. Examples of sentence stems and their completions at the Conformist ego level include: “If I can’t get what I want—I’m unhappy”; “The thing I like about myself—I like to help others”; “My mother and I—get along with each other.”

TEST DEVELOPMENT

In principle, sentence completion tests can be constructed by almost anyone: Identify a set of important themes and then draft a set of sentence stems that address these themes. However, the distinguishing feature of an SCT is not the items per se but the validity and reliability of the scoring procedure. It is the latter issue that distinguishes the WUSCT and the SCT-Y from many other sentence completion tests that are currently used in research or clinical practice. Loevinger was trained as a psychometrician and her continued attention to psychometric rigor guided the development of the WUSCT and accompanying scoring manual.

Washington University Sentence Completion Test

In their “Recommendations for Building a Valid Projective Technique,” Lilienfeld et al. (2000, pp. 55–56) referred to the WUSCT as an exemplar of a psychometrically sound projective test. They argue that a good projective technique should meet three basic criteria: (1) test scores should be based on aggregate scores derived from responses to multiple items (because aggregation “averages out” measurement error), (2) test items should be directly pertinent to the construct being measured, and (3) construction of the instrument should be based on an iterative and self-correcting process. Lilienfeld et al. propose that the WUSCT meets all three criteria: (1) the ego level of a respondent is based on the aggregation of the 36 independent item ratings (32 in the case of the SCT-Y), (2) the items (sentence stems) were selected to reveal a respondent’s general frame of reference or ego level, and (3) the construction of the WUSCT, and the description of each ego

stage, is based on a self-correcting process that Loevinger refers to as micro-validation (Loevinger, 1993).

The micro-validation technique lies at the heart of the selfcorrecting feedback loop linking the measurement of ego development and the evolution of the construct. On the basis of an ingenious feedback loop, Loevinger moved between theory, item selection, scoring manual construction, data, theory revision, instrument revision, new data, and so forth. The WUSCT gradually evolved using this self-correcting process (see Loevinger, 1993). A brief historical account will serve to illustrate the process and the research procedures (for a more elaborate account, see Loevinger, 1993, 1998).

In the 1960s Loevinger and her colleagues investigated an issue that, at that time, received very little research attention: authoritarian and nonauthoritarian child-rearing styles. She concluded that authoritarianism was not the low point on a developmental continuum (as was believed at that time) but, rather, a midpoint on a developmental scale that ranged from a chaotic and impulsive parenting style to a more democratic and flexible approach to child rearing. This conclusion was based on a mixture of clinical insights and research findings. Additional findings led Loevinger to conclude that she was inadvertently measuring a construct that was much broader in scope than her original target (i.e., child-rearing style). She labeled this variable ego development because it appeared to encompass many of the divergent but interrelated aspects of “the self”: moral development, interpersonal relations, and conceptual complexity.

The sentence completion technique was selected because it could elicit responses that were relevant to all aspects of ego development and still yield a profile of scores suitable for psychometric analyses. Initially Loevinger and her colleagues constructed a broad pool of items (sentence stems) to employ in the sentence completion measure; responses were assigned to a developmental level using a quasi-stage model proposed by Sullivan, Grant, and Grant (1957). Sullivan et al. had proposed four levels of “interpersonal maturity and interpersonal integration”: Impulsive, Conformist, Conscientious, and Autonomous. Loevinger initially derived individual item ratings and total protocol ratings based on the conceptual similarity between a sentence completion response (or protocol) and the description of each stage provided by Sullivan et al. Sentence completion responses that were conceptually similar and rated at the same ego level were grouped into a single response category. (The current scoring manual consists of over 2,000 response categories.)

Next, Loevinger employed her micro-validation procedure: Taking base rates into account, she determined if a response category that had been intuitively assigned to a developmental level was empirically more likely to appear at that level compared to any other developmental level. Based on this feedback loop, Loevinger adjusted the stage location of response categories, providing her scoring manual and developmental model with a rare empirical grounding (Westenberg employed this identical procedure when constructing his manual for youth). Finally, an algorithm was devised to convert the profile of 36 item ratings into a total protocol rating (this algorithm is often referred to as the “ogive rules” in published papers).

The resulting scoring manual was cross-validated and refined in successive studies spanning a 25-year period, involving approximately 2,800 respondents. “There was a long period of mini-experiments, for example, trying different sentence stems, different arrangements of items, different lengths of the test” (Loevinger, 1998, p. 7). The initial version of the WUSCT was published in 1970 (Loevinger & Wessler, 1970), a revised version of the test protocol was published in 1985 (Loevinger, 1985), and the revised version of the scoring manual was published in 1996 (Hy & Loevinger, 1996; Loevinger, 1998).

The process of micro-validation and the resulting feedback loop enabled Loevinger to empirically modify the ego development construct itself. Thus Sullivan et al.’s (1957) four-stage model was initially expanded to a five-stage model when Loevinger’s micro-validation procedure revealed another developmental stage: the Self-Protective stage (emerging between the Impulsive and Conformist stages). Eventually Loevinger’s research identified eight developmental stages, currently assessed using the revised scoring manual that was published in 1996.

Sentence Completion Test for Children and Youth

The most recent advance in the measurement of ego development was the construction and publication of the SCT-Y⁵ (Westenberg, Treffers, et al., 1998). As noted earlier, the scoring manual for the WUSCT was constructed primarily on the basis of adult responses; at least 85% of Loevinger’s “manual construction” sample was older than age 16. In contrast, Westenberg’s manual construction sample was mainly comprised of children and adolescents (approximately 65% of the participants were age 16 or younger). The youthful composition of Westenberg’s sample was essential for developing a valid scoring manual for children and adolescents; it was also essential for revealing genuine signs of ego development during childhood and adolescence. The negative tone that is often associated with adult protocols at low ego levels may be due, in part, to the fact that low-level adults are delayed in their development, a delay that has been associated with various forms of adult psychopathology (see the Use in Clinical and Organizational Practice section). Hence it is likely that Loevinger’s portrayal of the earliest ego development stages is not characteristic of children and adolescents at these same stages.

Westenberg and his colleagues sought to develop a scoring manual that was specifically constructed for assessing ego development in youth. Toward this end he initiated a large research project involving more than 2,700 participants, all of whom completed a revised version of the WUSCT. Approximately 80% of the responses generated by Westenberg’s sample could be rated (i.e., assigned to a developmental level) using Loevinger’s manual. However, Westenberg’s own manual construction and micro-validation revealed that approximately 50% of these ratings had to be corrected; that is, the response categories had to be reassigned to a developmental stage that was different than the one indicated in Loevinger’s manual (Westenberg, Jonckheer, et al., 1998). The majority of changes involved reassigning response categories that were empirically located at the Conformist ego stage in the Loevinger manual to the Self-Protective or Impulsive ego stage in the scoring manual for the SCT-Y. For example, some responses indicating unrestrained emotions dropped from the Conformist level to the Impulsive level (e.g., “If I can’t get what I want—I get angry”). This change appears to be consistent with the Loevinger description of the Impulsive person and, therefore, strengthens the connection between the theory and the measure. Similar downshifts at first seemed inconsistent (e.g., “When people are helpless—I help them” dropped from Conformist to Impulsive), but in the end fit well in a more balanced description of the Impulsive ego level in youths.

New response categories constructed on the basis of the 20% unratable responses also emerged mostly at the lowest ego levels. Thus, the downshift of response categories and the addition of new categories resulted in a more detailed scoring system for the lowest ego levels and contributed to a more balanced and “normal” picture of the earliest three ego levels (see the Theoretical Basis section). The downshift of response categories to the Impulsive and Self-Protective ego level is attributed to the greater number of low-level individuals in the normative samples. The revision of the WUSCT scoring manual was based on responses obtained from 67 pre-Conformist individuals; the SCT-Y scoring manual was based on responses obtained from 1,141 pre-Conformist individuals. The importance of the sheer number of low-level subjects was anticipated by Loevinger (1993): “If there are almost no cases at the extremes, one will never be able to verify what responses belong there. Responses typical for extreme cases will occur occasionally at median levels and, if there are few or no extreme cases, be wrongly

assigned there” (p. 9). It may therefore be concluded that the scoring manual for the SCT-Y provides a more sensitive and accurate measure of ego development in adolescents and children over 8 years of age.

PSYCHOMETRIC CHARACTERISTICS

Psychometric studies of the WUSCT and SCT-Y invariably report high levels of interrater reliability. Perfect interrater agreement per item averages about 85%, and interrater agreement within one stage (i.e., disagreement not larger than one stage) is often close to 95% (e.g., Cohn, 1984; Drewes & Westenberg, 2001; Westenberg & Gjerde, 1999). Westenberg, van Strien, et al. (2001) reported an average kappa of .80 (p # .001). The WUSCT and SCT-Y also display high internal consistency: Most studies report a Cronbach’s alpha of .90 or higher (e.g., Loevinger, 1998; Novy & Francis, 1992; Westenberg et al., 2000). In addition, a principal component analysis showed only one major component, another demonstration of the supposed unity of the test.

Several studies suggest that the split-half reliability of the WUSCT and SCT-Y is about .80, and, if disattenuated for the greater unreliability of the two test halves, the correlation between the two halves approached unity (e.g., Novy & Francis, 1992; Westenberg et al., 2000). Test-retest reliability of the WUSCT and SCT-Y is also high, and test-retest correlations are often about .80 (e.g., Jurich & Holt, 1987; Westenberg et al., 2000). Interrater agreement, internal consistency, and test-retest stability also appear adequate when the WUSCT or SCT-Y are administered to clinical populations (e.g., Weiss, Zilberg, & Genevro, 1989; Westenberg, Siebelink, Warmenhoven, & Treffers, 1999).

Several studies have examined the sensitivity of ego level scores to changes in the administration of the instrument or accompanying instructions. Most studies used the same general design: one half of the items were presented in the standard mode, the other half were presented in the modified mode (i.e., split-half within-subjects test-retest design). In general, findings suggest that the oral administration of the WUSCT or SCT-Y does not appear to affect the assessment of a respondent’s ego level, regardless of the respondent’s age, gender, reading and writing skills, IQ, and preference for either presentation mode (see McCammon, 1981; Streich & Swensen, 1985; Westenberg, van Strien, et al., 2001), at least not when administrator and respondents were in direct contact. However, when the WUSCT was administered as part of a telephone interview, then the oral mode did yield significantly lower ego level scores (Hansell, Sparacino, Ronchi, & Strodtbeck, 1985). Additional research needs to confirm that the demand characteristics of a telephone interview negatively affect ego level scores.

Some studies instructed respondents to “be candid” or present a “good face” (Blumentritt, Novy, Gaa, & Liberman, 1996; Drewes & Westenberg, 2001; Jurich & Holt, 1987; Redmore, 1976). In these studies, participants were administered one half of the items after receiving the standard instructions (“Complete the following sentences”); participants then completed the remaining items after being instructed to “be candid” or “make a good impression.” In principle, such instructions should not influence the assessment of ego level because their meaning is partially shaped by one’s developmental level. That is, the instructions were not conceptually relevant to the assessment of ego level. Several studies suggest that such instructions do not appear to influence ego level ratings (e.g., Drewes & Westenberg, 2001; Redmore, 1976).

In contrast, three other types of instructions, each with conceptual relevance to ego level, had a modest but significant impact on ego level scores. In one study, Blumentritt et al. (1996) asked respondents to complete their second administration of the WUSCT in the most complex and thoughtprovoking way. In related research, Redmore (1976) and Blumentritt et al. (1996) obtained ego level scores from respondents under standard assessment conditions; participants were then provided with brief descriptions of each ego stage and instructed to complete the sentence stems as they would be completed by a person at the highest ego levels. Finally, Drewes and Westenberg (2001) instructed respondents to complete sentence stems in as adult and mature a manner as they could. In each study, the modified instructions had a modest positive (and predicted) effect on ego level scores, although the average increase was no more than one half a stage.

In line with Vygotsky’s (1978) notion of a proximal zone of development, Drewes and Westenberg (2001) interpreted the increased ego level scores, obtained under conceptually relevant instructions, as an indication of optimal ego level, whereas the performance under the standard instructions reflected a person’s functioning ego level (for the distinction between optimal and normal functioning, see also Lamborn, Fischer, & Pipp, 1994). The developmental psychologist’s distinction between functional and optimal ego level resonates with the psychometrician’s distinction between “characteristic” versus “maximum” performance (Jackson, 1993).

RANGE OF APPLICABILITY AND LIMITATIONS

The sentence completion method should probably not be used with children younger than age 8, although 6-year-olds have produced meaningful responses to the SCT-Y when the instrument was presented orally. By about age 8 or 9, most children are capable of writing down their own sentence completions. If children are not capable of writing responses (e.g., due to paralysis) then the oral presentation of the stems (and subsequent transcription of responses) does not seem to distort ego level scores (Westenberg, van Strien, et al., 2001). If the testing situation permits, it might be good testing practice to ask children ages 8 through 11 if they would prefer the written or oral format, because a slight advantage was noted for young respondents who indicated a preference for an oral presentation (Westenberg, Van Strien, et al., 2001). No studies have examined if there is an upper age limit beyond which the WUSCT should not be administered. The WUSCT has been administered to individuals older than age 70 (e.g., Labouvie-Vief, Hakim-Larson, & Hobart, 1987).

The WUSCT and SCT-Y have demonstrated incremental validity over measures of intelligence and cognitive development. It is doubtful, however, that SCTs are meaningful tools for assessing personality development in individuals with very low verbal ability and low mental capacities. In our own research we have not recruited individuals with IQ scores lower than 70.

The WUSCT and the SCT-Y can be administered individually or in group settings. Individual administration will be most typical in clinical or assessment settings, while group testing will be typical of research settings. During group administration respondents should not discuss their responses nor look at anyone else’s responses. Group administration should be conducted using the standard written procedure. Individual administration of the WUSCT or SCT-Y could be conducted in either a written or oral format (assuming the test administrator refrains from making comments and asking follow-up questions). Mail or telephone administration procedures are not typically advised. Mail procedures do not allow for the direct supervision needed to ensure an uninterrupted and independent completion of the sentences, and one study indicated that a telephone administration yielded significantly lower ego level scores (Hansell et al., 1985).

The WUSCT and SCT-Y can be administered in the context of a larger assessment battery; however, the sequencing of instruments should be planned carefully, for at least two reasons. First, sentence completion responses might be influenced by the content of other personality questionnaires or interviews. Second, most respondents find completing sentence stems to be a novel task, requiring more effort and motivation than multiple-choice surveys and related questionnaires. Therefore, the SCT should be one of the first and not one of the last instruments to be completed.

CROSS-CULTURAL ISSUES

In principle, the SCT method is suitable for individuals from virtually any culture or country because the sentence stems refer to universal issues (e.g., “Raising a family—”) and social interactions (e.g., “My mother and I—”). Studies in at least 11 non-English-speaking countries support the crosscultural value and applicability of the ego development concept, measure, and scoring procedure (see Carlson & Westenberg, 1998). Cultural adaptations of the measure have been used to examine the relationship between ego development and numerous variables: achievement motivation and economic development (Papiamentu, Curac¸ao), models of achievement in women (Norwegian, Norway), professional training activities (French, Quebec), effects of trauma (Hebrew, Israel), personality correlates of corporate subcultures (Japanese, Japan), the subcultural aspects of academic disciplines (Portuguese, Portugal), religion and caste (Kannada, India), fluency among bilinguals (Vietnamese, United States), psychosocial process correlates of various psychiatric diagnostic groups (German, Germany), and normal fears and anxiety disorders (Dutch, the Netherlands). The English version of the instrument has also been used in the United States with ethnically diverse populations, such as Puerto Rican, Vietnamese, and African American groups. (See Carlson & Westenberg, 1998, for the references to these and other crosscultural studies.)

Findings obtained from studies conducted outside of the United States are similar to the findings obtained from studies conducted within the United States. These studies provide support for the cross-cultural applicability of the ego development construct. However, several linguistic and cultural issues arise when administering the WUSCT to non-U.S. populations, and test administrators should be cognizant of these issues. For example, Dhruvaragan (1981) and Hy (1986) note that the English language offers relatively few ways of encoding social standing, whereas this social attribute is central to several other languages. In Vietnamese several different pronouns can be used to signify “the self” and each of these terms contains different information related to the social characteristics of the speaker, the addressee, and their relationship (Hy, 1986). Notably, English makes no such distinctions. Thus the English word for “I” is not easily translated into Vietnamese because information about the social attributes of the individual is missing. In related research, Kusatsu (1977) observed that the response category “The thing I like about myself is—nothing at all” was most characteristic of the Self-Aware ego level (E-5) in Japan, whereas this response is classified at the Impulsive level (E-2) in the American scoring manual. Kusatsu (1977) argued that “this is

because of the Japanese norm of humiliation of the individual ego” (p. 66). To adequately deal with such cultural differences, Sasaki (1981) developed a Japanese scoring manual. It is unclear, however, whether they have used the microvalidation procedure described earlier.

ACCOMMODATION FOR POPULATIONS WITH DISABILITIES

For most respondents, completing the SCT is not a difficult task. If respondents are unable to write their own responses, then the oral administration of the sentence stems (and subsequent transcription of responses) may be appropriate, provided that the oral procedure mimics the written procedure. Limited experience with sign language indicates that it can replace the written or oral procedures, provided that great care is given to the issues of notation, interpretation, and reliability.

LEGAL AND ETHICAL CONSIDERATIONS

The WUSCT and the SCT-Y have strong psychometric properties. When these instruments are administered and scored properly they yield reliable and valid assessments of ego level. In clinical and organizational settings, however, neither instrument should be used as the sole assessment technique. These measures should always be used in conjunction with other assessment procedures in order to obtain a clear and accurate clinical picture. The more serious the implications of the client’s evaluation, the more one would want corroborating information before speaking strongly from the data. Clinicians might utilize interview procedures and other tests for particular characteristics that are key to different ego levels. One should also check the degree to which the description of the ego level as assessed from the SCT is met with agreement and examples from persons who are close to the client. The WUSCT and the SCT-Y have not yet been evaluated for the purposes of forensic psychology and should therefore be used with great caution in such contexts.

Another ethical consideration concerns “informed consent” and how much information is needed to qualify as “informed,” The standard instructions accompanying both instruments (e.g., “Complete the following sentences”) are intentionally vague, providing respondents with no information about the actual purpose of the test. Jurich and Holt (1987) argued that it would be “ethically desirable . . . to be direct rather than mysterious or devious about what one is trying to do with a test, if it is not necessary to be vague or indirect in order to

obtain valid measurement” (p. 193). In most testing situations it will be sufficient to explain the purpose of the test by saying “that it is a way to understand your ideas (or your child’s) on different topics and the way you see your world.” In our experience most people understand those concepts easily enough, even most children. If the testing situation or client requires more specific information about the purpose of the test, one might add that the test “will help us understand your maturity (or your child’s), and explain that you would see things differently now than when you were much younger.” Clinical experience and research has indicated that such an introduction has hardly any effect on ego level scores (see the Psychometric Characteristics section).

When administering the WUSCT or SCT-Y in clinical practice or organizational settings, one should always consider the possibility that the client or his legal caretaker or representative might demand access to the item ratings, total protocol rating, or written interpretation of these scores.

COMPUTERIZATION

No computerized scoring system is available to rate the WUSCT or SCT-Y. The development of computerized scoring systems has been attempted by several investigators but to date no program has proved to be valid and reliable. Attempts to develop such programs have failed to make the subtle distinctions between scoring categories that sometimes distinguish one ego level from another. In other words, scoring the protocols has to be completed by trained raters (see the Test Description section). The final step in the scoring process, the conversion of the profile of 36 (or 32) item ratings into a total protocol rating by means of the ogive rules, could be computerized, because the ogive rules are “automatic” (see Hy, 1998).

CURRENT RESEARCH STATUS

By the early 1990s more than 300 empirical studies had employed the WUSCT (the SCT-Y was only recently developed and cross-validated and thus lacks a similarly large research base). The scope of this chapter does not permit a comprehensive review of these studies, and a few critical reviews have been published elsewhere (Cohn, 1991; Hauser, 1976, 1993; Loevinger, 1979, 1998; Manners & Durkin, 2000; see also Westenberg, Blasi, et al., 1998). Here we will only review the empirical evidence addressing a number of critical assumptions underlying the ego development construct.6

First, Loevinger, Westenberg, and other investigators maintain that ego development represents a single developmental continuum. Thus the sentence completion test should measure only one dimension. Several studies support this contention, including (1) the results of homogeneity and factor analyses (see Loevinger, 1998; Westenberg et al., 2000), and (2) the failure to identify separate subsets of items measuring distinct aspects of ego development (e.g., Blasi, 1971; Lambert, 1972).

Second, the current model of ego development also assumes sequentiality; that is, the model assumes that there is an invariant sequence of stages that individuals must traverse in the order proposed. Loevinger (1998) grouped the evidence for sequentiality into four categories: (1) evidence from cross-sectional studies in which ego level is correlated with age and educational grade (see Cohn, 1998); (2) evidence from longitudinal studies, in which individuals display significant gains between Time 1 and Time 2 testings (cf. Cohn, 1998); (3) evidence of asymmetry of comprehension (i.e., individuals can be prompted to produce SCT protocols that are scored lower than their own ego level, but individuals cannot be prompted to produce protocols that are scored higher than their own ego level; e.g., Redmore, 1976); and (4) evidence obtained from theory-relevant interventions that seek to foster development (see Cohn, 1998; Manners & Durkin, 2000).

A third assumption underlying the proposed model concerns individual differences: Specifically, within any cohort of individuals there should be a range of ego levels; likewise, within any age range (e.g., ages 10 to 20) there should be a variety of developmental trajectories, with some individuals displaying greater development than other individuals (perhaps due to different social experiences that may act as pacers for development). Cross-sectional studies have demonstrated the expected variety of ego level scores within age cohorts, and longitudinal studies have illustrated a range of ego level trajectories (e.g., Gfellner, 1986; Hauser, Powers, & Noam, 1991; Westenberg & Gjerde, 1999). For example, WUSCT data collected at age 14 and then again at age 23 years revealed that some precocious teenagers (as measured by scores on the WUSCT) do not mature into precocious adults, whereas some teens who displayed average or even below average ego levels achieved atypically high WUSCT scores at age 23 (several additional trajectories were also revealed; Westenberg & Gjerde, 1999). In their repeated measures longitudinal study tracking adolescents from ages 14 to 18, Hauser and colleagues operationally define specific ego development trajectories (e.g., progressive, regressive, profound arrest) based on ego stage scores shown by these adolescents over successive 3- and 4-year periods. They report, and describe in detail, individuals illustrating these varied ego development patterns (Hauser et al., 1991). Moreover, the various trajectories are not without consequence. It has been demonstrated, for example, that arrested ego development trajectories in adolescence are associated with insecure attachment in young adult years (Hauser, Gerber, & Allen, 1998).

A fourth assumption, directly related to the third one, is that ego levels represent a personality typology; that is, ego development represents a dimension of individual differences within a given age cohort. The bulk of the research addressing this issue provides strong support for this contention: Ego level is related to a host of other variables, even when controlling for age of respondent. Ego level was predictably related to personality types (John, Pals, & Westenberg, 1998), individual differences in personality (e.g., Westenberg & Block, 1993; for an overview, Pals & John, 1998), attachment representations (Hauser, Powers, et al., 1998), psychological maturity (Helson & Wink, 1987; McCrae & Costa, 1980), emotion expression (Hauser & Safyer, 1994), psychosocial and moral development (Adams & Fitch, 1983; Gfellner, 1986; Snarey, 1998), identity development (Adams & Fitch, 1983; Blasi, 1988), coping and ego resiliency (Hauser, Borman, Bowlds, et al., 1991; Hauser, Borman, Jacobson, Powers, & Noam, 1991), successful adaptation after divorce (Bursik, 1991), depression and suicide (Borst & Noam, 1993; Borst, Noam, & Bartok, 1991; Rierdan & Koff, 1991, 1993), social fears and anxiety disorders (Westenberg, Drewes, Goedhart, Siebelink, & Treffers, 2003; Westenberg et al., 1999; Westenberg, Siebelink, & Treffers, 2001), delinquency and antisocial behavior (Frank & Quinlan, 1976; Noam et al., 1984; Novy, Gaa, Frankiewicz, Liberman, & Amerikaner, 1992), interpersonal style and peer relations (Hansell, 1981; Hauser, 1978), risk taking (Kishton, Starrett, & Lucas, 1984; Philliber, Namerow, Kaye, & Kunkes, 1986), parental style and family contexts (e.g., Adams & Jones, 1981; Allen et al., 1994; Hauser et al., 1984), managerial performance (Fisher & Torbert, 1991; Torbert, 1989, 1994), self-image complexity (Hauser, Jacobson, Noam, & Powers, 1983), and various other variables (see Westenberg, Blasi, et al., 1998).

In addition to the many concurrent relations between level of ego development and a host of other variables, studies directed by Hauser, Allen, and Crowell have indicated several prospective relations. In their ongoing long-term longitudinal studies of normative and clinical (psychiatrically impaired) adolescents, ego development is studied in terms of specific stage-predictors as well as individual-based ego development trajectories (Hauser et al., 1998). In these studies ego development is viewed as an indicator of the adolescent’s capacity to integrate and balance needs of self and others in a way that permits the establishing of autonomy and maintaining relationships in interactions. Using this perspective and longitudinal approach, this research group finds clear threads of

continuity from adolescent ego development to functioning in social relationships over the following 20 years. Low levels of ego development in adolescence are related to individual’s hostility as reported by peers (e.g., failures of relatedness) and to lower levels of dating assertion (e.g., failures of autonomy) of our participants at age 25 (Allen, Hauser, & Borman-Spurrell, 1996). Similarly, ego development at age 14 predicted maturity in close relationships, over and above concurrent effects of attachment security, thus suggesting that each construct was making an independent contribution to close relationship functioning (Schultz & Selman, 1998). By age 35, low levels of adolescent ego development predict low levels of trust and greater distress in marital relationships and greater parenting stress in parents of 18-month-olds. In addition, these new studies reveal strikingly lower levels of perceived parenting competence when parenting 36- to 54 month-olds associated with lower levels of adolescent ego development (r ” .54; p # .01).

While the effects of adolescent psychopathology may in part be mediated by low levels of ego development, there is evidence that a number of the effects of adolescent ego development exist in addition to the predictive effects of severe adolescent psychopathology. For example, adolescent ego development is a stronger predictor of adult attachment coherence (a prime marker of secure adult attachment organization) at age 25 than is prior psychiatric hospitalization, suggesting an important continuity in underlying developmental processes that exists (at least in part) independently of manifest psychopathology (Hauser et al., 1998). Consistent with these predictions of ego development to later markers of social functioning, both ego development at age 14 and a lack of ego development from ages 14 to 25 were strongly predictive of marital distress at age 35, after accounting for the effects of adolescent-era psychopathology (beta for ego level ” !54; p # .001).

Finally, all psychological tools need to demonstrate incremental validity over already existing instruments and concepts. The WUSCT and SCT-Y have demonstrated incremental validity over age as the primary marker of development and have also displayed incremental validity over IQ and socioeconomic status (SES) in the prediction of other variables (e.g., Browning, 1987, Cohn, 1991; Westenberg & Block, 1993). A recent meta-analysis of more than 40 studies including approximately 4,700 participants reported that the weighted average correlation between ego level and (verbal) intelligence ranged from .20 to .34 (Cohn & Westenberg, 2003). The meta-analysis also included 16 studies (and 25 statistical tests) of the incremental validity of ego level scores over and above intellectual ability: 92% of the tests of incremental validity revealed significant relations between ego

608 Sentence Completion Measurement of Psychosocial Maturity

level and a host of criterion variables after statistically removing the influence of (verbal) intelligence. For example, in a sample of twins reared apart, Newman et al. (1998) showed that ego development scores have a considerable genetic basis even after controlling statistically for intelligence (the heritability estimate is about 50% after controlling for IQ).

USE IN CLINICAL AND ORGANIZATIONAL PRACTICE

The WUSCT and SCT-Y have demonstrated good psychometric properties when used with clinical populations (see the Psychometric Characteristics section). Studies also reveal that many variables with clinical relevance are related to ego level (see the Current Research Status section). Research also suggests that respondents cannot typically “fake” a higher ego level (unless they are well versed in the ego development concept; see Psychometric Characteristics). The research status of the WUSCT and SCT-Y allows for the question of how these instruments can be used in clinical and organizational practice. Three categories of practical uses are explored.

1. The client’s or employee’s interpersonal frame of reference. As was previously noted, the WUSCT and SCT-Y provide access to a client or employee’s core frame of reference or psychosocial “lens”; that is, these instruments provide insight into how individuals perceive themselves and other people in relationships and social interactions. Such insight can be quite important because it facilitates a better understanding of clients and employees and may contribute to an improved partnership and the pursuit of common goals (e.g., Dill & Noam, 1990; Young-Eisendrath & Foltz, 1998). Knowledge of a client’s ego level (frame of reference) is of particular relevance if the client or employee does not display the level of psychosocial maturity that is expected of a specific age group. A client or employee may score substantially above or below the modal ego level displayed by one’s age peers. This age-stage discrepancy might go unnoticed by the clinician or counselor, who might address a client erroneously on the basis of the intuitive average for that age group. Loevinger’s model of ego development and the accompanying assessment tools sought to provide a concept and measure of psychological maturity that was independent from age (see the Theoretical Basis section). Hence, the WUSCT and SCT-Y provide a yardstick of psychological maturity “that does not itself stretch with age” (Loevinger, 1990, p. 112). This age-independent measure of ego level maturity can be used to supplant or support the assessor’s intuitive impression of the client’s psychosocial maturity.

Knowledge about the client or employee’s ego level (frame of reference) can also help to organize other information that is available about the client, such as assessments of personality traits or psychopathology. For example, a high need for achievement will express itself differently at different levels of psychosocial maturity: Impulsive persons might have achievement fantasies but are unlikely to work with any consistency, Self-Protective persons might be very competitive and try to get others to do the dirty work, Conformist individuals may feel a strong sense of duty toward shared goals, Self-Aware persons might reflect on their needs for achievement, Conscientious people would feel the need to achieve in accordance with their own standards and strong sense of responsibility for others, and so on (see also Lasker, 1978). Therefore, employees of different ego levels may be motivated by different management approaches. In other words, the ego development perspective could be a useful tool for “situational leadership,” in which the manager’s management style matches the employee’s characteristic way of perceiving and responding to work-related tasks and social situations (for a similar argument see Graves, 1966).

Finally, knowledge about the client or employee’s ego level or frame of reference supports the selection of the most appropriate intervention method. Research indicates that clinical interventions should be tailored to the developmental level of the client, regardless of her or his chronological age (Borst & Noam, 1993; Kirshner, Hauser, & Genack, 1988; Noam, 1998). For example, Young-Eisendrath and Foltz (1998) found that clients at low ego levels perceive psychotherapy as a concrete service provided by therapists who are responsible for the outcome, whereas clients at higher ego levels perceive psychotherapy as a personal process of internal discovery for which the client is primarily responsible (see also Dill & Noam, 1990). A similar distinction was drawn by Lasker (1978) in a sample of industry workers in Curac¸ao. He noted that workers displayed different types of achievement needs—either low or high personal efficacy—and argued that screening for ego level was an effective way to tailor training content to increase its effectiveness in raising worker motivation. A mismatch between the services provided on the one hand and the client or employee’s ego level on the other hand is more likely to lead to failure and prematurely terminated interventions.

2. The link between ego development and problem behavior and psychopathology. A second way in which the ego development construct and measures might be of relevance for clinical practice concerns the relationship between ego level and psychopathology. Many clinicians, counselors, and lay people equate “low” with “bad,” and vice versa. Thus, low ego levels are often assumed to be associated with psychological problems, and clients with psychosocial difficulties are often assumed to be immature (see Noam, 1998). It is often proposed that clinical interventions should be aimed primarily at raising people’s level of ego development; likewise, it is sometimes assumed that personnel selection should be aimed at hiring high ego level personnel. But this presumed association between low ego level and psychosocial problems presents us with a logical problem: Children and adolescents who move through these low stages are not expected to always display psychosocial problems, and many adults who have outgrown the lowest ego levels might still have psychosocial problems. Therefore, the presence of psychosocial problems cannot intrinsically be a characteristic or product of low ego levels, and low ego levels cannot be a direct cause of psychosocial problems (Loevinger, 1968, 1976). Loevinger (1968) maintained that “every stage has its weaknesses, its problems, and its paradoxes, which provide both a potential for maladjustment and a potential for growth” (p. 169). She emphasized that to know more about the relationship between maturity and psychosocial problems or psychopathology, both need to be defined independently of each other. The initial purpose of the ego level construct was therefore to disentangle psychological maturity from psychological adjustment and to provide a measure of psychological maturity.

Even though psychosocial problems and ego development are not intrinsically related, research has revealed some meaningful relationships that suggest practical implications. Numerous studies over the past 30 years have illuminated the relationship between ego level and psychopathology and have, by and large, supported the contention that “every stage has its weaknesses.” Reviews of the literature by Rierdan (1998) and Noam (1998) indicate that the absence or presence of psychopathology is not related to ego development; however, the type of pathology displayed by individuals is related to level of ego development. This connection will be reviewed for the broad categories of externalizing and internalizing problems.

The general pattern is that behavior or externalizing problems (i.e., crime and conduct disorders, with or without internalizing problems) are mostly related to the Impulsive and Self-Protective ego levels, whereas the emotional or internalizing problems (without comorbid externalizing problems) are mostly related to the Conformist ego level or beyond. In other words, the higher one’s ego level, the smaller the likelihood of serious behavior problems. This connection is particularly pronounced in late adolescents and adults, suggesting that a developmental delay might be a risk factor for developing an externalizing problem (e.g., Noam, Paget, Valiant, Borst, & Bartok, 1994). The developmental delay itself might be partly responsible for the behavior problems: The mismatch between the person and the expectations others have about that person might yield disagreements and lead to escalating conflicts. In her work in elementary schools, Lamb (1996) observed that among children who had difficulties at recess was a group for whom a common factor was having a lower ego level while trying to participate in social play activities that demanded a higher level of functioning. Among the characteristics of ego development, those pertaining to interaction style and understanding of rules are most relevant to participation in social games. Social games were analyzed according to these factors (in addition to physical skills, leadership needs, and imaginative play requirements) and were paired with the different ego levels common among elementary school children. A small pilot study showed that when social games and leadership needs were matched to the ego level of young children with severe behavioral problems, aggression was minimized and children were able to participate successfully.

The association between the behavior difficulties and the developmental delay might not be an intrinsic one, but could both be due to other factors, such as genetic and environmental factors. In any case, the studies by Hauser and colleagues indicate that severe and chronic delays in the adolescent period are associated with a bad outcome in adulthood (e.g., greater hostility and poorer parenting skills; see the Current Research Status section). Whatever the causal relationship might be, the fact that most adolescents and adults with behavior problems are developmentally delayed, and that a chronic delay is associated with current and future interpersonal problems, has two implications for intervention: (1) they should not be overrated in terms of their expected maturity level and should be approached like younger persons, and (2) the intervention should not be restricted to the behavior problems themselves but should also be aimed at raising one’s ego level up to par with one’s age mates (see Cohn, 1998, and Manners & Durkin, 2000, for an overview of methods to stimulate ego development). For clients with externalizing problems, development to the Conformist level and above might serve as a protective factor that would ameliorate the detrimental effects of other factors.

Internalizing problems, such as anxiety and depression, can be present in people of all ages and ego levels, but are most noticeable at the Conformist level and beyond. From that level onward, the comorbid presence of externalizing problems declines, allowing the emotional difficulties to come to the fore. From this realization it can be concluded that emotional problems in children and older persons at relatively low ego levels might be obscured by their behavior problems. Borst and Noam (1993), for example, noted that depression in low-level adolescents often goes unnoticed (i.e., is not diagnosed) because of the much stronger presence of the behavior difficulties (e.g., impulsivity and acting-out behavior). The practical implications of this finding are obvious: In low-level individuals presenting with externalizing problems, diagnosticians should be extra careful not to miss emotional difficulties.

Internalizing problems may arise at any age or stage of development, but the specific type or manifestation of these difficulties appears to differ across the ego levels. Borst and Noam (1993) reported that suicidality in 14-year-old girls was equally prevalent at the different ego levels, but manifested itself in a strikingly different manner: The angry-defiant type was most prevalent at the Self-Protective ego level, whereas the self-blaming type was most prevalent at the Conformist level. Both types differed in terms of the reasons for and the methods by which they had attempted to commit suicide. The authors argued that the types had to be treated in a different way: The angry-defiant type would mostly benefit from a behavioral approach, the self-blaming type would best be served by psychotherapy.

Another example of the stage-dependent manifestation of emotional disorders is provided by the results of research on the relation between ego level and anxiety disorders in a population of children and adolescents referred to an outpatient psychiatric clinic (Westenberg et al., 1999). The two most prevalent and debilitating anxiety disorders in children and adolescents were empirically related to conceptually equivalent ego levels. The separation anxiety disorder (SAD) was related to the Impulsive ego level, and the generalized anxiety disorder (GAD) was related to the Conformist ego level with age, sex, IQ, and socioeconomic status statistically controlled for. This connection is thought to be due to common ground between specific anxiety disorders and certain ego levels: Vulnerability and dependency are common elements of SAD and the Impulsive ego level, the focus on social desirability and a self-blaming attitude are common elements of GAD and the Conformist ego level.

Apparently, anxiety disorders in children and adolescents are aggravated or pathological versions of normal concerns and preoccupations. This hypothesis is supported by research on the development of normal fears in a nonclinical sample: Impulsive individuals mostly report concrete-physical fears, Conformist individuals mostly report social-evaluative fears (Westenberg, Drewes, et al., 2003). Those connections suggest that the content of the anxiety disorder is not abnormal but is an integral part of that stage. Hence, the treatment of such anxiety disorders should not be aimed at the “removal” of the fear content but should be aimed at the reasons why the fear should be excessive and uncontrollable (e.g., by enhancing the child’s coping strategies).

3. The ego development level of clinicians, counselors, and managers. A third way in which the ego development construct and measure might be of relevance for clinical and organizational practice does not concern the client’s ego level, but concerns the ego level of the clinician, manager, or parent. From this perspective it appears that “higher” is “better.” Overall, high ego level counselors, managers, and parents do better than their counterparts at lower ego levels (e.g., Borders, 1998; Hauser, Borman, Jacobson, et al., 1991; Torbert, 1994). Borders (1989, 1998, Borders & Fong, 1989; Borders, Fong, & Neimeyer, 1986) examined how counselors’ ego level influences the perception of patients, in-session cognitions, and counseling ability. She observed that a relatively low ego level restricts the attainable level of counseling skills, although counseling training itself may actually promote ego development. Research findings reported by Torbert (1989, 1994; Fisher & Torbert, 1991) suggest that a relatively high ego level would be needed for the successful managing of organizational transformations. He argues that the Strategist level of managerial development—akin to Loevinger’s Individualistic ego level—is needed to bring about change in organizations. Managers at lower ego levels were less successful in bringing about change.

Why would higher ego levels promote success in counseling, management, and parenting? One possible answer is that higher ego levels embody—and have been shown to be related empirically to—properties needed to perform complex tasks, particularly in cases when the right outcome is not known in advance. Such situations require the ability to grasp the complexity in situations, the recognition of divergent perspectives, the awareness of one’s own role in such situations, and a strong sense of personal identity. As Rooke and Torbert (2001) noted: “CEOs whose cognitive-emotional structure recognizes that there are multiple ways of framing reality and that personal and organizational transformations of structure require mutual, voluntary initiatives—not just single-framed, hierarchical guidance—are more likely to succeed in leading organizational transformation” (p. 1).

A second answer might be that high ego level individuals are able to intuitively (and perhaps unwittingly) match their interaction style to the maturity level of their client, employee, or child, whereas low ego level individuals have fewer degrees of freedom in adapting to the various ego levels in their environment. Due to the asymmetry of comprehension embodied in any progressive developmental variable, people are able to understand and adjust to ego levels below their own level, but cannot understand and adjust to ego levels beyond their own level. Therefore, people of high ego level will be able to produce the desired match between self and many others, whereas low ego level individuals cannot match up with everyone else. High ego level parents are able to recognize and respect the different needs of their children at various stages in their development, and high ego level counselors and managers are able to adjust to the different frame of references of their clients and employees, thereby improving the desired match between the person and his or her environment.

In summary, the ego development construct should not be interpreted as a model of psychosocial health and the measure cannot be used as a screening tool for psychological or behavior problems. Many other measurement tools are available for the assessment of psychopathology; that is not what the WUSCT and SCT-Y are for. However, independence from pathology (and from age) makes the instrument highly relevant for clinical and organizational practice because it provides crucial information that is not obtained from other instruments.

FUTURE DEVELOPMENTS

One of the most important contributions of the ego development construct and measure is that it provides an ageindependent yardstick of psychosocial maturity. A drawback, however, is that a clear connection with age is lacking, leaving a basic question unanswered: What ego level is normative for each age group? Loevinger and her colleagues did not describe their samples as “representative,” nor did they claim that the prevalence of ego levels obtained during the measurement development process reflected ego level norms for the United States. It would, indeed, be a daunting task to collect precise norms for such a large country with a very heterogeneous population. Such norms would, however, be of interest from a developmental perspective, and also be of practical use from a clinical and organizational perspective because they would help one obtain a better sense of consistencies in relations between ego levels and age, and how these relations might vary in different contexts (e.g., social class, rural, or urban). Piecemeal information has been provided (e.g., Gfellner, 1986; Holt, 1980; Westenberg, Jonckheer, et al., 1998), but more systematic studies need to be mounted to obtain more comprehensive and reliable age data.

Related to questions about age and ego development is the topic of gender differences in ego development. One of the most intriguing findings arising from studies employing the WUSCT and SCT-Y has been the finding of gender differences during the adolescent period: Girls mature earlier than

boys, and the difference is notable. Yet by the mid-20s this gender difference disappears (see Cohn, 1991; Westenberg et al., 2000). This gender difference in the adolescent period needs to be validated against other, conceptually related gender differences. For example, the fear literature indicates that adolescent girls display more social fearfulness than boys (see Gullone, 2000). Is this greater sensitivity for social fearfulness related to the higher ego development of the girls? If validated, it needs to be explained why girls should be ahead of the boys and why they lose their lead in late adolescence. More research on the adolescent gender difference in ego development might be informative regarding (1) pacers for ego development (what makes people move from one ego level to the next?), (2) the nature of gender differences in the adolescent period (are some gender differences temporary?), and (3) a number of practical implications (e.g., teaching curricula might have to be tailored according to the ego level differences between boys and girls).

Another issue related to questions about age and ego development is the topic of ego development trajectories, referred to in the Current Research Status section. We already know that a given cohort of adolescents, when followed over several years, will show different paths of ego development (Hauser, Powers, et al., 1991). Some adolescents, for example, clearly progress from earlier (pre-Conformist) to either Conformist or post-Conformist stages. Other adolescents show no signs of change in their ego development, remaining either “profoundly arrested” (fixed at pre-Conformist stages) or “consistently Conformist,” expressing Conformist stage scores throughout each repeated testing over 3 or more adolescent years. Of interest is the fact that these varied ego change patterns are differentially distributed in special populations (e.g., profound arrests in adolescents previously hospitalized for nonpsychotic disorders; Hauser, Powers, et al., 1991).

These findings, emerging from a long-term longitudinal study of ego development in adolescents and young adults, have many theoretical and empirical implications. In terms of fruitful methods, they suggest the advantages of taking a person-centered approach in examining ego development over time. By conceptually defining theoretically meaningful “types” of ego development trajectories, one can derive operational definitions of these patterns, and then identify individuals fitting each profile. Once such types have been located, a host of intriguing questions can be addressed: How stable are these ego development paths from adolescence to adult years? Are they associated with specific adolescent family process, parental ego development paths, or other conceptually relevant aspects of adolescent contexts? Are such trajectories differentially found in different special populations (e.g., specific psychiatric diagnoses, chronic illness, se-

612 Sentence Completion Measurement of Psychosocial Maturity

vere poverty or other major familial or individual adversity)? Can such trajectories be modified through specially designed interventions? How does knowledge of an individual’s ego development trajectory (his or her “ego history”) add to our understanding of his or her concurrent or future behavior and/ or attitudes? One would suppose that knowing an individual’s history of ego stages and changes would provide a more comprehensive picture of that given person. But what do we gain in terms of predicting future behaviors or clarifying current phenomena, such as unexpected resilient outcomes? (Hauser, 1999). Finally, can discovering ego development trajectories clarify such matters as the nature of adult psychosocial development and intergenerational transmission of ego development? For instance, if we discern intergenerational continuities in ego development trajectories, then we would want to further examine associate familial patterns as candidates for underlying psychological mechanisms.

Finally, links between ego level and behavior, including psychopathology, require more attention. Recent ongoing longitudinal studies by Hauser and colleagues, briefly summarized previously, are pertinent to complex and important questions about the interplay (and underlying mechanisms) between individual ego development and concurrent as well as subsequent individual actions. Even though ego level scores derived from sentence completions have been shown to display external validity vis-a`-vis a host of behavioral measures (see the Current Research Status section), the mechanisms underlying these relations need to be specified. After all, ego development is not about behavior per se, but is about an internal frame of reference. One possibility is that ego level has an impact on the way social information is processed, thereby restricting and amplifying the behavioral options available to a given individual. For example, low ego level persons might be inclined to perceive other people’s behavior as threatening to the self and are therefore inclined to react with counteraggression or other self-protective measures. Another possibility is that ego level has an impact on the way one deals with situations. For example, along the line of expanding options, high ego level teens may not be less likely to engage in sexual behavior than low ego level teens, but they may be more likely to use contraceptives. Studies targeting mechanisms behind ego level behavioral links can lead to needed insights into the actual impact of ego development on daily functioning.

NOTES

1. The original scoring manual for the SCT-Y is in Dutch because it is based on research with Dutch children and adolescents. An English-language version, based on research with American, Australian, and New Zealand children and adolescents, is in preparation (please contact the first author of this chapter).

2. Several authors have mistakenly associated Loevinger’s model of ego development with Freud’s psychoanalytic model. The link is nonexistent, however, as Loevinger has repeatedly stated (Loevinger, 1976, 1993; see also Westen, 1998).

3. The first ego level, E-1, is not accessible by means of the sentence completion test (see Loevinger, 1976).

4. Examples are taken from the Hy and Loevinger (1996) scoring manual.

5. Loevinger and colleagues had constructed a form for children and adolescents, Form 77 of the WUSCT, but some of its items did not have a scoring manual and none of the item manuals had been validated for usage with children and adolescents. See Westenberg, Treffers, and Drewes (1998) for a comparison between the SCT-Y and Form 77 of the WUSCT.

6. A complete bibliography of ego development studies may be obtained from the first author of this chapter.

REFERENCES

• Adams, G.R., & Fitch, S.A. (1983). Psychological environments of university departments: Effects on college students’ identity status and ego stage development. Journal of Personality and Social Psychology, 44, 1266–1275.
• Adams, G.R., & Jones, R. (1981). Female adolescents’ ego development: Age comparisons and childrearing perceptions. Journal of Early Adolescence, 1, 423–426.
• Allen, J.P., Hauser, S.T., Bell, K.L., & O’Connor, T.G. (1994). Longitudinal assessment of autonomy and relatedness in adolescentfamily interactions as predictors of adolescent ego development and self-esteem. Child Development, 65, 179–194.
• Allen, J.P., Hauser, S.T., & Borman-Spurrell, E. (1996). Attachment theory as a framework for understanding sequelae of severe adolescent psychopathology: An eleven year follow-up study. Journal of Consulting and Clinical Psychology, 64, 254–263.
• Blasi, A. (1971). A developmental approach to responsibility training. Unpublished doctoral dissertation, Washington University, St. Louis, MO.
• Blasi, A. (1988). Identity and the development of the self. In D.K. Lapsley & F.C. Power (Eds.), Self, ego, and identity: Integrative approaches (pp. 226–242). New York: Springer-Verlag.
• Blasi, A. (1998). Loevinger’s theory of ego development and its relationship to the cognitive-developmental approach. In P.M. Westenberg, A. Blasi, & L.D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development (pp. 13–26). Mahwah, NJ: Erlbaum.
• Blumentritt, T., Novy, D.M., Gaa, J.P., & Liberman, D. (1996). Effects of maximum performance instructions on the Sentence

Completion Test of Ego Development. Journal of Personality Assessment, 67, 79–89.

• Borders, L.D. (1989). Developmental cognitions of first practicum supervisees. Journal of Counseling Psychology, 36, 163–169.
• Borders, L.D. (1998). Ego development and counselor development. In P.M. Westenberg, A. Blasi, & L.D. Cohn, L.D. (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development (pp. 331–346). Mahwah, NJ: Erlbaum.
• Borders, L.D., & Fong, M.L. (1989). Ego development and counseling ability during training. Counselor Education and Supervision, 29, 71–83.
• Borders, L.D., Fong, M.L., & Neimeyer, G.J. (1986). Counseling students’ level of ego development and perceptions of clients. Counselor Education and Supervision, 26, 36–49.
• Borst, S.R., & Noam, G.G. (1993). Developmental psychopathology in suicidal and non-suicidal adolescent girls. Journal of the American Academy of Child and Adolescent Psychiatry, 32, 501–508.
• Borst, S., Noam, G.G., & Bartok, J.A. (1991). Adolescent suicidality: A clinical-developmental approach. Journal of the American Academy of Child and Adolescent Psychiatry, 30, 796–803.
• Browning, D. (1987). Ego development, authoritarianism, and social status: An investigation of the incremental validity of Loevinger’s Sentence Completion Test (Short form). Journal of Personality and Social Psychology, 53, 113–118.
• Bursik, K. (1991). Adaptation to divorce and ego development in adult women. Journal of Personality and Social Psychology, 60, 300–306.
• Carlson, V.K., & Westenberg, P.M. (1998). Cross-cultural research with the WUSCT. In J. Loevinger (Ed.), Technical foundations for measuring ego development (pp. 57–75). Mahwah, NJ: Erlbaum.
• Cohn, L.D. (1984). Developmental changes in adolescent judgment under conditions of risk. Unpublished doctoral dissertation, Washington University, St. Louis, MO.
• Cohn, L.D. (1991). Sex differences in the course of personality development: A meta-analysis. Psychological Bulletin, 109, 252– 266.
• Cohn, L.D. (1998). Age trends in personality development: A quantitative review. In P.M. Westenberg, A. Blasi, & L.D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development (pp. 133–143). Mahwah, NJ: Erlbaum.
• Cohn, L.D., & Westenberg, P.M. (2003). Intelligence and character development: A quantitative review of the incremental validity of Loevinger’s measure of ego development. Manuscript submitted for publication.
• Dhruvaragan, V. (1981). Level of self differentiation: Caste status vs. mundane experience. Unpublished doctoral dissertation, Department of Sociology, University of Chicago.
• Dill, D.L. & Noam, G.G. (1990). Ego development and treatment requests. Psychiatry, 53, 85–91.
• Drewes, M.J., & Westenberg, P.M. (2001). The impact of modified instructions on ego level scores: A psychometric hazard or indication of optimal ego level? Journal of Personality Assessment, 76, 229–249.
• Fisher, D., & Torbert, W.R. (1991). Transforming managerial practice: Beyond the Achiever stage. In R.W. Woodman & W.A. Pasmore (Eds.), Research in organization change and development, Vol. 5 (pp. 143–173). Greenwich, CT: JAI Press.
• Frank, S.J. & Quinlan, D.M. (1976). Ego development and female delinquency: A cognitive-developmental approach. Journal of Abnormal Psychology, 85, 505–510.
• Gfellner, B.M. (1986). Changes in ego and moral development in adolescents: A longitudinal study. Journal of Youth and Adolescence, 15, 147–163.
• Graves, C.W. (1966). Deterioration of work standards. Harvard Business Review, 44, 117–128.
• Gullone, E. (2000). The development of normal fear: A century of research. Clinical Psychology Review, 20, 429–451.
• Hansell, S. (1981). Ego development and peer friendship networks. Sociology of Education, 54, 51–63.
• Hansell, S., Sparacino, J., Ronchi, D., & Strodtbeck, F.L. (1985). Ego development responses in written questionnaires and telephone interviews. Journal of Personality and Social Psychology, 47, 118–128.
• Hauser, S.T. (1976). Loevinger’s model and measure of ego development: A critical review. Psychological Bulletin, 33, 928–955.
• Hauser, S.T. (1978). Ego development and interpersonal style in adolescence. Journal of Youth and Adolescence, 7, 333–352.
• Hauser, S.T. (1993). Loevinger’s model and measure of ego development: A critical review II. Psychological Inquiry, 4, 33–40.
• Hauser, S.T. (1999). Understanding resilient outcomes: Adolescent lives across time and generations. Journal of Research on Adolescence, 9, 1–24.
• Hauser, S.T., Borman, E.H., Bowlds, M.K., Powers, S.I., Jacobson, A.M., Noam, G.G., et al. (1991). Understanding coping within adolescence: Ego development and coping strategies. In A.L. Greene, E.M. Cummings, & K. Karraker (Eds.), Life-span developmental psychology: Perspectives on stress and coping (pp. 273–295). Hillsdale, NJ: Erlbaum.
• Hauser, S.T., Borman, E.H., Jacobson, A.M., Powers, S.I., & Noam, G.G. (1991). Understanding family contexts of adolescent coping: A study of parental ego development and adolescent coping strategies. Journal of Early Adolescence, 11, 96–124.
• Hauser, S.T., Gerber, E.B., & Allen, J.P. (1998). Ego development and attachment: Converging platforms for understanding close relationships. In P.M. Westenberg, A. Blasi, & L.D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development (pp. 203–219). Mahwah, NJ: Erlbaum.

614 Sentence Completion Measurement of Psychosocial Maturity

• Hauser, S.T., Houlihan, J., Powers, S.I., Jacobson, A.M., Noam, G.G., Weiss-Perry, B., et al. (1991). Adolescent ego development within the family: Family styles and family sequences. International Journal of Behavioral Development, 14, 165–193.
• Hauser, S.T., Jacobson, A., Noam, G., & Powers, S. (1983). Ego development and self-image complexity in early adolescence: Longitudinal studies of diabetic and psychiatric patients. Archives of General Psychiatry, 40, 325–332.
• Hauser, S.T., Powers, S.I., & Noam, G. (1991). Adolescents and their families: Paths of ego development. New York: Free Press.
• Hauser, S.T., Powers, S., Noam, G., Jacobson, A., Weiss, B., & Follansbee, D. (1984). Familial contexts of adolescent ego development. Child Development, 55, 195–213.
• Hauser, S.T., & Safyer, A. (1994). Ego development and adolescent emotions. Journal of Research on Adolescence, 4, 487–502.
• Helson, R., & Wink, P. (1987). Two conceptions of maturity examined in the findings of a longitudinal study. Journal of Personality and Social Psychology, 53, 531–541.
• Holaday, M., Smith, D.A., & Sherry, A. (2000). Sentence completion tests: A review of the literature and results of a survey of members of the Society for Personality Assessment. Journal of Personality Assessment, 74, 371–383.
• Holt, R.R. (1980). Loevinger’s measure of ego development: Reliability and national norms for male and female short forms. Journal of Personality and Social Psychology, 39, 909–920.
• Hy, L.X. (1986). Cross-cultural measurement of ego development: Responses in different languages. Unpublished doctoral dissertation, Washington University, St. Louis, MO.
• Hy, L.X. (1998). Managing SCT data. In J. Loevinger (Ed.), Technical foundations for measuring ego development (pp. 41–48). Mahwah, NJ: Erlbaum.
• Hy, L.X., & Loevinger, J. (1996). Measuring ego development (Rev. ed.). Mahwah, NJ: Erlbaum.
• Jackson, D.N. (1993). Personality development and nonlinear measurement models. Psychological Inquiry, 4, 30–33.
• John, O.P., Pals, J.L., & Westenberg, P.M. (1998). Personality prototypes and ego development: Conceptual similarities and relations in adult women. Journal of Personality and Social Psychology, 74, 1093–1108.
• Jurich, J., & Holt, R.R. (1987). Effects of modified instructions on the Washington University Sentence Completion Test of Ego Development. Journal of Personality Assessment, 51, 186–193.
• Kegan, R. (1982). The evolving self: Problem and process in human development. Cambridge, MA: Harvard University Press.
• Kirshner, L., Hauser, S.T., & Genack, A. (1988). Effects of gender on short-term psychotherapy. Psychotherapy: Theory, Research, and Practice, 15, 158–167.
• Kishton, J., Starrett, R.H., & Lucas, J.L. (1984). Polar versus milestone variables in adolescent ego development. Journal of Early Adolescence, 4, 53–64.
• Kohlberg, L. (1969). Stage and sequence: The cognitivedevelopmental approach to socialization. In D.A. Goslin (Ed.),

Handbook of socialization theory and research (pp. 347–480). Chicago: Rand McNally.

• Kusatsu, O. (1977). Ego development and socio-cultural process in Japan. Keizagaku-Kiyp, 3, 41–109.
• Labouvie-Vief, G., Hakim-Larson, J., & Hobart, C.J. (1987). Age, ego, and the life-span development of coping and defense processes. Psychology and Aging, 2, 286–293.
• Lamb, N. (1996). Mastering the maze of social play, teacher’s resource manual: A developmental guide for social play among young children. Unpublished manuscript.
• Lambert, H.V. (1972). A comparison of Jane Loevinger’s theory of ego development and Lawrence Kohlberg’s theory of moral development. Unpublished doctoral dissertation, University of Chicago.
• Lamborn, S.D., Fischer, K.W., & Pipp, S. (1994). Constructive criticism and social lies: A developmental sequence for understanding honesty and kindness in social interactions. Developmental Psychology, 30, 495–508.
• Lasker, H.M. (1978). Interim summative evaluation report: An initial assessment of the Shell/Humanas OD program. Cambridge, MA: Harvard University Press.
• Lilienfeld, S.O., Wood, J.M., & Garb, H.N. (2000). The scientific status of projective techniques. Psychological Science in the Public Interest, 1, 27–66.
• Loevinger, J. (1968). The relation of adjustment to ego development. In S.B. Sells (Ed.), The definition and measurement of mental health (pp. 161–180). Washington, DC: U.S. Department of Health, Education, and Welfare.
• Loevinger, J. (1976). Ego development: Conceptions and theories. San Francisco: Jossey-Bass.
• Loevinger, J. (1979). Construct validity of the Sentence Completion Test for Ego Development. Applied Psychological Measurement, 3, 281–311.
• Loevinger, J. (1985). Revision of the Sentence Completion Test for Ego Development. Journal of Personality and Social Psychology, 48, 420–427.
• Loevinger, J. (1990). Ego development in adolescence. In R.E. Muuss (Ed.), Adolescent behavior and society (pp. 111–117). New York: McGraw-Hill.
• Loevinger, J. (1993). Measurement of personality: True or false. Psychological Inquiry, 4, 1–16.
• Loevinger, J. (1997). Stages of personality development. In R. Hogan, J. Johnson, & S. Briggs (Eds.), Handbook of personality psychology (pp. 199–208). San Diego, CA: Academic Press.
• Loevinger, J. (Ed.). (1998). Technical foundations for measuring ego development (pp. 49–55). Mahwah, NJ: Erlbaum.
• Loevinger, J., & Wessler, R. (1970). Measuring ego development: Construction and use of a Sentence Completion Test, Vol. 1. San Francisco: Jossey-Bass.
• Manners, J., & Durkin, K. (2000). Processes involved in adult ego development: A conceptual framework. Developmental Review, 20, 475–513.
• McCammon, E.P. (1981). Comparison of oral and written forms of the Sentence Completion Test for Ego Development. Developmental Psychology, 17, 233–235.
• McCrae, R.R., & Costa, P.T. (1980). Openness to experience and ego level in Loevinger’s Sentence Completion Test: Dispositional contributions to developmental models of personality. Journal of Personality and Social Psychology, 39, 1179–1190.
• Newman, D.L., Tellegen, A., & Bouchard Jr., Th. J. (1998). Individual differences in adult ego development: Sources of influence in twins reared apart. Journal of Personality and Social Psychology, 74, 985–995.
• Noam, G.G. (1998). Solving the ego development-mental health riddle. In P.M. Westenberg, A. Blasi, & L.D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development (pp. 271–296). Mahwah, NJ: Erlbaum.
• Noam, G.G., Hauser, S.T., Santostefano, S., Garrison, W., Jacobson, A.M., Powers, S.I., et al. (1984). Ego development and psychopathology: A study of hospitalized adolescents. Child Development, 55, 194–198.
• Noam, G.G., Paget, K., Valiant, G., Borst, S., & Bartok, J. (1994). Conduct and affective disorders in developmental perspective: A systematic study of adolescent psychopathology. Development and Psychopathology, 6, 519–532.
• Novy, D.M., & Francis, D.J. (1992). Psychometric properties of the Washington University Sentence Completion Test. Educational and Psychological Measurement, 52, 1029–1039.
• Novy, D.M., Gaa, J.P., Frankiewicz, D.L., Liberman, D., & Amerikaner, M. (1992). The association between patterns of family functioning and ego development of the juvenile offender. Adolescence, 27, 25–35.
• Pals, J.L., & John, O.P. (1998). How are dimensions of adult personality related to ego development? An application of the typological approach. In P.M. Westenberg, A. Blasi, & L.D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development (pp. 113–132). Mahwah, NJ: Erlbaum.
• Philliber, S., Namerow, P.B., Kaye, J.W., & Kunkes, C.H. (1986). Pregnancy risk taking among adolescents. Journal of Adolescent Research, 1, 463–481.
• Redmore, C. (1976). Susceptibility to faking of a Sentence Completion Test of Ego Development. Journal of Personality Assessment, 40, 607–616.
• Rierdan, J. (1998). Ego development, pubertal development, and depressive symptoms in adolescent girls. In P.M. Westenberg, A. Blasi, & L.D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development (pp. 253–269). Mahwah, NJ: Erlbaum.
• Rierdan, J., & Koff, E. (1991). Depressive symptomatology among very early maturing girls. Journal of Youth and Adolescence, 20, 415–425.
• Rierdan, J., & Koff, E. (1993). Developmental variables in relation to depressive symptoms in adolescent girls. Development and Psychopathology, 5, 485–496.
• Rooke, D., & Torbert, W.R. (2001). Organizational transformation as a function of CEO’s developmental stage. Retrieved October 15, 2001, from Boston College, Wallace E. Carroll School of Management web site: http://www.bc.edu/bc_org/avp/csom/ faculty
• Sasaki, M. (1981). Measuring ego development of female adolescents by sentence completions. Japanese Journal of Educational Psychology, 29, 147–151.
• Schultz, L.H., & Selman, R.L. (1998). Ego development and interpersonal development in young adulthood: A between-model comparison. In P.M. Westenberg, A. Blasi, & L.D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development (pp. 181–202). Mahwah, NJ: Erlbaum.
• Selman, R.L. (1980). The growth of interpersonal understanding: Developmental and clinical analyses. New York: Academic Press.
• Snarey, J. (1998). Ego development and the ethical voices of justice and care: An Eriksonian interpretation. In P.M. Westenberg, A. Blasi, & L.D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development (pp. 163–180). Mahwah, NJ: Erlbaum.
• Streich, D.D., & Swensen, C.H. (1985). Response to two presentations of the Sentence Completion Test of Ego Development. Journal of Personality Assessment, 49, 285–288.
• Sullivan, C., Grant, M.Q., & Grant, J.D. (1957). The development of interpersonal maturity: Applications to delinquency. Psychiatry, 20, 373–385.
• Torbert, W.R. (1989). Leading organizational transformation. In R. Woodman & W. Pasmore (Eds.), Research in organizational change and developments, Vol. 3. Greenwich, CT: JAI Press.
• Torbert, W.R. (1994). Cultivating post-formal adult development: Higher stages and contrasting interventions. In M. Miller & S. Cook-Greuter (Eds.), Transcendence and mature thought in adulthood: The further reaches of adult development (pp. 163– 179). Lanham, MD: Rowman & Littlefield.
• Vygotsky, L.S. (1978). Mind in society: The development of higher psychological processes. Cambridge, MA: Harvard University Press.
• Weiss, D.S., Zilberg, N.J., & Genevro, J.L. (1989). Psychometric properties of Loevinger’s Sentence Completion Test in an adult psychiatric outpatient sample. Journal of Personality Assessment, 53, 478–486.
• Westen, D. (1998). Loevinger’s theory of ego development in the context of contemporary psychoanalytic theory. In P.M. Westenberg, A. Blasi, & L.D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of

616 Sentence Completion Measurement of Psychosocial Maturity

Loevinger’s conception of ego development (pp. 59–70). Mahwah, NJ: Erlbaum.

• Westenberg, P.M., Blasi, A. & Cohn, L.D. (Eds.). (1998). Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development. Mahwah, NJ: Erlbaum.
• Westenberg, P.M., & Block, J. (1993). Ego development and individual differences in personality. Journal of Personality and Social Psychology, 65, 792–800.
• Westenberg, P.M., Drewes, M.J., Goedhart, A.W., Siebelink, B.M., & Treffers, P.D.A. (2003). A developmental analysis of selfreported fears in late childhood through mid-adolescence: Social-evaluative fears on the rise? Journal of Child Psychology and Psychiatry, 44, 1–15.
• Westenberg, P.M., Drewes, M.J., Siebelink, B.M., Treffers, P.D.A., Jonckheer, J., & Goedhart, A.W. (2000). Zinnenaanvullijst Curium (ZALC): Een instrument voor het meten van ego-ontwikkeling [A sentence completion test for measuring ego development in children and adolescents]. Lisse, The Netherlands: Swets Test Publishers.
• Westenberg, P.M., & Gjerde, P.F. (1999). Ego development during the transition from adolescence to young adulthood: A nineyear longitudinal study. Journal of Research in Personality, 33, 233–252.
• Westenberg, P.M., Jonckheer, J., Treffers, P.D.A., & Drewes, M.J. (1998). Ego development in children and adolescents: Another side of the Impulsive, Self-Protective, and Conformist ego levels. In P.M. Westenberg, A. Blasi, & L.D. Cohn (Eds.), Per-

sonality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development (pp. 89– 112). Mahwah, NJ: Erlbaum.

• Westenberg, P.M., Siebelink, B.M., & Treffers, P.D.A. (2001). Psychosocial developmental theory in relation to anxiety and its disorders. In W.K. Silverman & P.D.A. Treffers (Eds.), Anxiety disorders in children and adolescents: Research, assessment, and intervention (pp. 72–89). Cambridge: Cambridge University Press.
• Westenberg, P.M., Siebelink, B.M., Warmenhoven, N.J.C., & Treffers, P.D.A. (1999). Separation anxiety and overanxious disorders: Relations to age and level of psychosocial maturity. Journal of the American Academy of Child and Adolescent Psychiatry, 38, 1000–1007.
• Westenberg, P.M., Treffers, P.D.A., & Drewes, M.J. (1998). A new version of the WUSCT: The Sentence Completion Test for Children and youths (SCT-Y). In J. Loevinger (Ed.), Technical foundations for measuring ego development (pp. 81–89). Mahwah, NJ: Erlbaum.
• Westenberg, P.M., van Strien, S.D., & Drewes, M.J. (2001). Revised description and measurement of ego development in early adolescence: An artifact of the written procedure? Journal of Early Adolescence, 21, 469–491.
• Young-Eisendrath, P., & Foltz, C. (1998). Interpretive communities of self and psychotherapy. In P.M. Westenberg, A. Blasi, & L.D. Cohn (Eds.), Personality development: Theoretical, empirical, and clinical investigations of Loevinger’s conception of ego development (pp. 315–330). Mahwah, NJ: Erlbaum.

CHAPTER 46

Assessment of Object Representation in Children and Adolescents: Current Trends and Future Directions

FRANCIS D. KELLY

OBJECT-RELATIONS MEASURES 618

Rorschach-Based Measures: Mutuality of Autonomy Scale (MOAS) 618 TAT-Based Measures: Social Cognition and Object

Relations Scale (SCORS) 620 Measures Based on Early Memories: Comprehensive Early Memories Scoring System (CEMSS) 622 Interview-Related Material 623

CONCLUSIONS 623 APPENDIX: CHILD AND ADOLESCENT PERSONALITY ASSESSMENT 625 REFERENCES 625

Beginning in infancy, proceeding throughout childhood and adolescence, and continuing throughout the adult years the individual constantly creates, modifies, and recreates ever increasingly differentiated, multifaceted unconscious mental constructs and facsimiles relating to the self and to other persons in the phenomenological domain—and these are appropriately designated as object representations. This construct provides comprehensive information about the developmental level of personality organization and also the quality of external relations. In short, these representations or schemata are intervening variables that serve to guide, inform, and mediate one’s intra- and interpersonal dealings throughout the life cycle. How then is this construct assessed; or more specifically, what standardized procedures and measures are currently available to address clinical and developmental concerns in relation to a child’s object-relations capacity and functioning? Are these measures reliable, valid, easily implemented, applicable to clinical situations? These issues and concerns merit serious consideration and scrutiny if present projective measures should continue to be endorsed as appropriate devices to evaluate salient aspects and dimensions of self and other in children and adolescents.

A decade ago Stricker and Healey (1990) comprehensively reviewed the empirical literature on the projective assessment of object relations. Most of the cited literature referenced adults with only approximately 10% of the studies directly pertaining to children or adolescents. In the ensuing 10 years the situation has seen child clinicians and researchers gradually extend and expand their focus to consider more concerted projective assessment of object-relations phenomena in latency age children and adolescents. In the sections to follow, I will proceed in a manner that will allow for systematic examination and review of the literature regarding empirical approaches to the projective assessment of object representations in children and adolescents. The initial focus will be on outlining and evaluating Rorschach object-representation assessment measures, for example, a variety of content scales and indices. Next, the relevant literature involving objectrepresentation assessment employing the Thematic Apperception Test (TAT) will be presented. And finally, consideration of research discussing less frequently employed measures that is, early memories, interviews, and so forth—will be briefly reviewed. Thus, the intent of this chapter is to illustrate how it is increasingly possible to construct a reliable and valid multidimensional composite of a child’s objectrepresentational schemata via select consideration and utilization of projective test information—data that serves to illuminate and articulate critical and varied cognitive-affective personality attributes, traits, and dimensions of self and other as well as serving to inform and direct aspects of treatment.

OBJECT-RELATIONS MEASURES

In the sections to follow, the major empirical scales employed in the assessment of children’s object relations will be reviewed.

Rorschach-Based Measures: Mutuality of Autonomy Scale (MOAS)

Over the course of the past decade, the primary objectrelations scale employed with both latency age and adolescent children has been the Urist (1977, 1980) Mutuality of Autonomy Scale (MOAS). With theoretical underpinnings derived from various contemporary schools of psychoanalytic thought, that is, self-psychology, ego psychology, developmental psychoanalysis, and object-relations theory and primarily exemplified in the writings of Kernberg (1976); Kohut (1977); and Mahler (1968)—the scale is designed to provide, on one level, a sense of the individual’s objectrelations functioning and, on another level, an impression of the individual’s developmental status regarding the phenomena of separation-individuation.

The MOAS is a 7-point scale briefly outlined as follows in accord with definitions provided by Urist (1977). Rorschach responses involve either M, FM or m, with these represented in some type of active or passive exchange. Specifically, the scale points refer to (1) Level 1: Reciprocal individuality of figures is acknowledged; implicit or explicit reference is accorded regarding the separateness and distinct autonomy of the figures—there is, however, clear, expressed, and undeniable mutuality defining the relationship. (2) Level 2: Parallel activity clearly defines the activity noted in the response; mutuality is neither acknowledged nor denied. (3) Level 3: Figures are leaning on or supporting one another. There is the clear sense that one’s autonomy is defined by the need for external support, that is, “leaning on someone.” (4) Level 4: One figure is a reflection of the other. One of the figures can exist only if it represents an extension of the other, for example, shadow, reflection response, mirror image. (5) Level 5: Introduction of malevolent control along with the loss of capacity for separateness are components of this type of response. There is a severe and marked imbalance characterizing the mutuality of the relationship. Percepts emphasizing control, manipulation, and coercion of others are indicative of these responses. For example, in response to Card IX, “Two witches casting evil and controlling spells on those people below there.” (6) Level 6: The imbalance noted in Level 5 is intensified and now reflects decidedly malevolent, one-sided aggression and domination of the other object. Not only does the imbalance of mutuality between figures become unbalanced, it is defined with a distinct destructive and aggressive quality. Responses defined by parasitic relationships are also included herein. (7) Level 7: Figures are overpowered, swallowed up, annihilated, devoured or destroyed by forces completely beyond the individual’s control; that is, catastrophic events.

Urist and Shill (1982) later embellished their original material to include a theoretical perspective explicating the 7 scale points. Specifically, the scale points involve the following phenomena and constructs: Level 1: Reciprocity-Mutuality; Level 2: Collaboration-Cooperation; Level 3: Simple Interaction; Level 4: Anaclitic-Dependent; Level 5: Reflection-Mirroring; Level 6: Magical Control-Coercion; Level 7: Envelopment-Incorporation. Holaday and Sparks (2001), in an attempt to improve scoring guidelines and interrater reliability, have proposed a more parsimonious summarization of the scale points; one, however, that is quite compatible with Urist’s original rendering.

Following the scoring of the MOAS responses, five possible scores are derived from the protocol. First, the mean MOAS, reflecting the average or typical level objectrepresentational level or schema of the child; second, the modal MOAS score, with this being important where there may be preponderance of one specific score that could be overlooked if only the mean MOAS score were considered; third, the highest object-relations score (HOR) and, conversely, the lowest object-relations score (LOR) are obtained as these provide information in relation to the range, variability, and stability of the child’s object relations. Last, consideration to the percentage of responses to the last four scale points is indicated given that these cluster within a pathological range of object relatedness (Gacano & Meloy, 1994).

The application of the MOAS to children is largely attributable to the investigative efforts of Tuber (1989a, 1989b) and colleagues (Meyer & Tuber, 1989; Tuber & Coates, 1989; Tuber, Frank, & Santostefano, 1989). Employing both nomothetic and idiographic paradigms, Tuber utilized contemporary object relations as the theoretical approach to integrate his findings. His research has provided significant clinical and research findings in relation to assessment and treatment. In later sections, a comprehensive sampling of Rorschach research emphasizing the MOAS with latency age children is provided—attention is accorded to heuristic as well as psychometric concerns, including considerable material regarding different types of reliability and validity measures of the MOAS.

With respect to reliability measures, there is ample indication of sound interrater reliability in studies emphasizing nomothetic clinical group comparisons. In the investigations of Meyer and Tuber (1989); Ryan, Avery, and Grolnick, (1985); and Tuber (1983) the MOAS scale was easily extended to children’s Rorschachs, and consistently sound interrater reliability figures have been reported. In these studies exact agreement figures have ranged from .73 to .90, whereas agreement within 1 point consistently yielded results greater than 90%. More recently, Leifer, Shapiro, Martone, and Kassem (1991) used the MOAS as one measure to evaluate object relations in sexually abused latency females (mean age of 8.9) and reported exact agreement figures of 84%. Similarly, Kelly and Burch (2000) reported interrater MOAS reliability of .85 in a study that investigated object-relations functioning in latency age children who were subjected to chronic, complex trauma and abuse. Thus, it is possible to see that the MOAS scale has provided reasonable estimates of reliability in nomothetic investigations involving normal and clinical groups of children and adolescents.

MOAS results in children have been cited as cogent predictors of later object-relations functioning and healthy adaptation (Tuber, 1983), offering indication of both predictive and construct validity and confirming observations of researchers (Burns & Viglione, 1996) who concluded that Rorschach object-relations measures are strongly related to, and ultimately reflect, the quality of a person’s usual and characteristic interpersonal transactions. Additional support for the criterion validity of this object-relations measure is provided by other authors. Ryan et al. (1985) showed that more adaptive MOAS scores were associated and positively correlated with self-esteem ratings, cooperative working relationships with peers, and academic success. Goldberg (1989) demonstrated that depressed inner city girls had lower MOAS scores when compared to less depressed counterparts.

The MOAS has been utilized in research involving preadolescent boys with gender identity disorder (Coates & Tuber, 1988; Tuber & Coates, 1989). Boys with gender identity difficulties exhibited less autonomous self- and other object representations and significantly more responses referencing malevolent interpersonal interactions than did control subjects.

Employing an idiographic paradigm, Tuber (1989a, 1992) has shown how the MOAS has utility in the clinical situation. Findings suggested that MOAS scale responses paralleled actual interactions with the therapist, thus providing cogent diagnostic relational information and illustrating an obvious theoretical-clinical link between a construct (object representation) and manifest behavior, that is, transference phenomena.

Gacano and Meloy (1994) have used the MOAS to evaluate conduct-disordered (CD) children, comparing their results with Tuber’s (1989b) normative data. Their findings underscored the fact that CD children produced a predominance of Level 6 and 7 responses, whereas normal children generally produce Level 5 responses as their most extreme response. Maladaptive representations of self and other parallel and complement manifest behavioral characteristics, suggesting that “aggressive and antisocial behavior among children, defined by the conduct-disordered diagnosis, will be expressed representationally in a perceptual, associate, and judgmental task such as the Rorschach” (p. 32).

A limited number of efforts have used the MOAS to assess object relatedness in abused and maltreated children. Leifer et al. (1991) assessed object representations of sexually abused females between the ages of 5 and 16, finding that the maltreated group exhibited significantly more disturbed perceptions of interpersonal contact. LOR and median MOAS scores served to differentiate the sexually abused girls from their counterparts but this was not the case with HOR scores. The authors emphasized the role of the MOAS as a mediating variable that could be utilized to more accurately decipher and comprehend clinical states and manifest relatedness. In a case study of a 9-year-old conduct-disordered little boy with a history of chronic physical and sexual abuse occurring before the age of 6, Gacano & Meloy (1994) provided MOAS material indicative of variable and severely compromised object-relations functioning—mean MOAS score of 4.5; HOR, Level 2; LOR, Level 7. Of more concern is the fact that 83% of his responses fell within a pathological range (i.e., Levels 4, 5, 6, or 7).

Commentary

The results of the studies using the MOAS with a variety of children and adolescents are relatively clear—the scale is an appropriate and useful measure of object relations with both children and adolescents, demonstrating acceptable reliability and validity with an array of clinical and nonclinical populations. Its popularity appears to have waned over the course of the past several years and the reasons for this are somewhat unclear as alternative object-representation measures have not supplanted the MOAS, although other devices are available including the Developmental Analysis of the Concept of the Object Scale (DACOS) developed by Blatt and his colleagues (Blatt, Brenneis, Schimek, & Glick, 1976) and used sparingly with adolescents (Blatt & Lerner, 1983; Fritsch & Holmstrom, 1990; Kelly, 1986). As Stricker and Healey (1990) observed, the DACOS is more discriminating than the MOAS but probably limited in its use with younger children because it only utilizes human content responses. Content scales derived from the Comprehensive System (Exner, 2000)—for example, Good Human Responses (GHR) and Poor Human Responses (PHR); Cooperative (COP) and Aggressive (AG) movement responses—offer the possibility for additional means to assess object relations, albeit in a fairly global and limited way as the former are dichotomous measures, and the latter, up until this point, have not been investigated with children. At first glance it would appear that these do not provide the same degree of interpretive discrimination as the DACOS or the MOAS when assessing children and adolescents, but again, a final statement to this effect is quite premature given the virtual absence of research. A recent proposed object-representation index of malevolence—COP/ AG—has recently been described by Ornduff, Centeno, and Kelsey (1999) with results suggesting that sexually abused girls score significantly higher on this index than do nonabused girls. Findings suggest that object representations are defined by expectations of malevolence in the interpersonal domain.

Overall, it would appear reasonable to posit that the MOAS and DACOS should be considered when clinicians or researchers are intent on obtaining a more refined and differentiated look at a child or adolescent’s object-relations functioning the MOAS to be used with younger children and both to be employed with adolescents, with the DACOS probably providing more detailed information than the MOAS despite the paucity of research that would support this opinion. The COP/AG response is intriguing as it represents an array of clinical and heuristic possibilities but its incidence with children and adolescents is usually low, and thus the ultimate importance of this variable may be defined by its unique occurrence in a protocol.

TAT-Based Measures: Social Cognition and Object Relations Scale (SCORS)

Emphasizing that the TAT provided a rich source of material about an individual’s object relations because of the nature of the task—that is, reflecting on one’s internal representations to construct narratives—Westen and colleagues (Westen, Lohr, Silk, Kerber, & Goodrich, 1985) developed an objectrelations scale to measure four main areas of psychological functioning involving cognitive and affective representations of others; quality of affect valence defining relations; one’s capacity for emotional investment in relationships, moral standards, and values; and the individual’s capacity for understanding interpersonal motivation with respect to self and other. Based on these four areas, Westen devised the SCORS, in which each of the four domains may be characterized by five developmental levels, ranging from primitive to mature. Specifically, the four dimensions refer to Complexity of representations of people (CR); Affect-tone of relationship schemas (AT); Capacity for emotional investment in relationships (CEI); and Understanding of social causality (USC).

In a recent revision, Westen (1995) altered the scoring, adopted a Q-sort paradigm, and added four more dimensions: Emotional investment in values and moral standards; Experience and management of aggressive impulses; Self-esteem; and Identity and coherence of self.

Each of the four (or more recently eight) SCORS categories is coded for a specific developmental level for each TAT story told. With the exception of Affect-tone, the scales assess developmental dimensions along a continuum (i.e., Level 1 is primitive/undifferentiated, and Level 5 is mature/ differentiated). Level 1 scores are relatively infrequent and always highly pathological even in younger children (Ornduff, 1997). For example, an 11-year-old girl with a history of chronic sexual and physical abuse tells the following story to TAT Card 13 MF: “He feels sad that he killed someone . . . killed that woman . . . just before he got off bed, just killed her. I don’t know. He gets thrown in jail for murder. [Feel?] I don’t know.” Scores assigned to the original four SCORS categories would be: Complexity of representations of people—Level 1; Affect-tone—Level 1; Capacity for emotional investment—Level 1; Understanding of social causality— Level 2. Averaging scores of narratives for each SCORS category provides for, as Cramer (1999) observes, a systematic determination of developmental level, or relative degree of pathology, for different patients, based on one aspect of ego functioning, object relations.

Interrater reliability has been adequately demonstrated with children and adolescents. High levels of corrected interrater reliability generally in the range of .83 to .96 have been reported with children (Freedenfeld, Ornduff, & Kelsey, 1995; Ornduff, Freedenfeld, Kelsey, & Critelli, 1994; Westen et al., 1991) and also with adolescents (Westen, Ludolph, Block, Wixom, & Wiss, 1990; Westen, Ludolph, Lerner, Ruffins, & Wiss, 1990; Westen, Ludolph, Misle, Ruffins, & Block, 1990; Westen, Ludolph, Silk, et al., 1990).

Substantial research with the SCORS, noted in the following sections, has also clearly delineated various types of validity: concurrent, predictive, and construct. In an initial series of normative studies, Westen and colleagues (Westen et al., 1991) investigated object-representation functioning and development in 2nd-, 5th-, 9th-, and 12th-grade children and adolescents. The studies supported several conclusions: Object-relations capacity continues to develop through latency and adolescence; object representations are multidimensional in nature; interpersonal expectations are shaped and molded by past experiences and mediate the present processing of social information, and this is ultimately viewed via consideration of TAT stories and SCORS information.

In employing the SCORS to evaluate object-relations functioning in borderline female adolescents, Westen and associates (Westen, Ludolph, Silk, et al., 1990) found that borderlines could be discriminated from normal and nonborderline, psychiatrically disturbed female adolescents on the basis of TAT object representations. Results highlighted the clinical utility of the SCORS in underscoring the unique nature of the borderline group’s pathology: more pathological responses, poorly articulated representations of others, blurred boundaries, and eccentric thinking. For all scales, TAT data and interview material were highly related.

In another study involving female adolescent psychiatric inpatients, Westen and his colleagues (Westen, Ludolph, Block, et al., 1990) used the SCORS to illustrate the relationships between developmental history variables and different dimensions of these adolescents’ object relations. Results supported several conclusions. First, preoedipal experiences and the role of the mother have a cogent influence on the development of later psychopathology. Second, various types of maltreatment—for example, neglect, sexual abuse, as well as disruptive family experiences—occurring during the latency years also appear to have an important impact on objectrelations development. This finding provided additional confirmation to earlier research attesting to the fact that object relations continued to develop well into the childhood and adolescents years (Westen, 1991).

More recently, other researchers have turned their attention to employing the SCORS in the evaluation of children and adolescents who have been sexually and physically maltreated (Freedenfeld et al., 1995; Ornduff, 1997; Ornduff et al., 1994; Ornduff & Kelsey, 1996; Pistole & Ornduff, 1994).

Ornduff and associates (Ornduff et al., 1994) reported that sexually abused children, including adolescents, score significantly lower on all SCORS scales than do nonabused counterparts. Results were interpreted to support the impression that the abused children exhibited significantly impaired and lower levels of developmental functioning with respect to object-relations functions, and that SCORS material is useful in demonstrating a link between deleterious environmental events and the manifestation of impaired object relations; in addition, results were seen as integral in generating developmental diagnostic formulations and treatment interventions.

Freedenfeld and colleagues (Freedenfeld et al., 1995) demonstrated that another type of abuse, physical maltreatment, also has a deleterious impact on the child’s object-relations functions. Specifically, the abused group was noteworthy for significantly lower capacity for emotional investment in relationships, less accurate understanding of social causality, and more malevolent and negative affective attributions to interpersonal relationships. But on a scale involving the understanding of the complexity of representations (CR) of relationships, the groups did not differ. Results were interpreted to indicate that clinical interventions need to be directed to help abused children alter malevolent perceptions in relation to self- and other object representations and to assist them in recognizing the patterns and dynamics intrinsic to abusive relationships. SCORS results also provided clinicians information underscoring the need to be sensitive to transference phenomena in relation to both diagnostic and therapeutic process issues.

Ornduff and Kelsey (1996), in an effort to more comprehensively evaluate the effect of both sexual and physical abuse, evaluated the SCORS material of sexually and physically abused female children, comparing them to nonabused counterparts, and found that all object-relations scores were significantly lower in both of the abused groups. However, differential patterns of impairment were noted when the abuse groups were compared. While physically abused children exhibited more pervasive indication of impairment, sexually abused children created object representations defined by an affective tenor associated with profound expectations of malevolence and maltreatment.

Finally, Ornduff (1997) demonstrated the clinical utility of the SCORS as she compared three latency age girls—one sexually abused, another physically maltreated, and a third emotionally distressed. The intent was to offer a clinical analogue to the aforementioned empirical efforts. Results showed the diagnostic and therapeutic significance of SCORS responses in children manifesting different clinical presentations. Abused children produced more Level 1 scores and with more frequency, particularly on scales tapping more affective (i.e., AT, CEI) rather than cognitive (i.e., USC, CR) dimensions of object relations. Noting the importance of this is critical to those treating these children as they are at risk to develop templates and schemata that mediate and influence subsequent relationships—being exposed and subjected to abuse early in life makes it likely that the child will develop expectations and assumptions that harm and maltreatment will likely be an integral part of one’s relationships.

Commentary

The SCORS has been employed with a variety of children and adolescents manifesting an array of psychological presentations. To date, credible reliability and validity findings have been reported as Westen’s manuals (1995; Westen et al., 1985) have seen researchers use his approach and validate his initial findings. In addition, the bulk of the more recent research involving the SCORS has seen it employed with children and adolescents who have been subjected to varying types of abuse and maltreatment, and this has resulted in formulating and disseminating appropriate assessment and

treatment information for clinicians who work with these children. Recent efforts have not expanded to include additional child samples, which would further enhance validity and expand the clinical utility of the SCORS. Children and adolescents manifesting other types of psychological disorders are not represented either in nomothetic or idiographic material and this would be an important consideration as additional normative data would enhance validation.

At this juncture what can be said about the SCORS as a paradigm that organizes and interprets TAT narratives? For the most part, a mostly positive and unequivocal thumbs-up is warranted. The SCORS appears to offer the child clinician and researcher a tool and a paradigm that is both reliable and valid—refer to Cramer (1999) for more detailed thoughts about this—and should certainly be considered if there is the need to know more about the schemata or scripts that may serve to orient, advise, and direct the child’s transactions with others. Overall, this scoring and interpretive paradigm is unique in that it offers the opportunity to look at the child’s or adolescent’s object relations from a variety of cognitive and affective vantage points, thus acquiring information that is multidimensional and reflective of distinct aspects of objectrelations functioning. This provides for a composite that allows for a more definitive and meaningful appreciation of structural and dynamic personality attributes, with this ultimately informing clinicians and researchers as they seek to address assessment and treatment concerns.

Measures Based on Early Memories: Comprehensive Early Memories Scoring System (CEMSS)

Bruhn (1981, 1983) and colleagues (Bruhn & Davidow, 1983; Davidow & Bruhn, 1990; Last & Bruhn, 1983, 1985) conducted research on a comprehensive assessment device for evaluating children’s early memories. The CEMSS is comprised of 48 items across nine categories, one of which is object relations. The referenced theoretical paradigm involves contemporary psychoanalytic thought, ego psychology, and object relations, and the object-relations category of the CEMSS contains the following specific domains: Perception of Others, Perception of Self, Perception of the Environment, Individual Distinctiveness, and Degree of Interpersonal Contact. Scoring involves rating each subcategory from 1 to 3, with 3 being the most differentiated score. An object-relations score for the individual is obtained by summing all scores.

A limited number of studies have reported acceptable interrater reliability in the range of .92 to .96. In the studies reported here, the scale has been able to demonstrate concurrent and predictive validity by showing group differences between children with various psychological disorders (e.g., delinquent, schizoid, hyperactive). And the scale has also been able to show that structural personality factors (i.e., object relations) predict the degree of psychopathology—well adjusted, mildly disturbed, and severely maladjusted.

Last and Bruhn (1983, 1985), in several studies, examined the early memories of 8- to 12-year-old males with varying types of psychopathology. The CEMSS results provided a significantly effective means to distinguish among four diagnostic groups of children—delinquent, somatic complaints, and hyperactive; in addition, scoring system classifications of the CEMSS were more accurate than those arrived at by experienced clinicians regarding the designation of varying levels of psychopathology in preadolescent boys.

Perhaps the most impressive early childhood memory studies are exemplified by research conducted with delinquent adolescent boys (Bruhn & Davidow, 1983; Davidow & Bruhn, 1990). Results in both instances revealed that early memories successfully discriminated the two groups. In the latter study, 82% were accurately classified as delinquent, and 96% were designated nondelinquent, with this tentatively validating the use of this material as an object-relations measure. However, as Davidow and Bruhn (1990) caution, differences between the scoring approaches employed in the evaluation of early memories make it difficult to compare findings; also, adolescent middle-class males comprised the delinquent samples in the aforementioned research, serving to limit generalizability of findings.

Commentary

The CEMSS has not seen extensive investigation or clinical application with either young children or adolescents aside from the aforementioned research. Modifications for scoring and coding early memory material (Davidow & Bruhn, 1990) have suggested some potentially fruitful material with delinquent adolescents in relation to representations of self and others, but the lack of uniformity in scoring and coding procedures has hindered opportunities to compare findings. Expanded employment of other approaches to the investigation of early memories (e.g., Fowler, Hilsenroth, & Handler, 1998) may have considerable heuristic and clinical potential in the assessment of object-relations functioning with adolescents, likely less so with younger children. It would be important for researchers and clinicians to reconsider the systematic investigation of looking at early memory material in light of recent clinical efforts documenting the link between this object-relations representational data and manifest behavior in a variety of settings, including the treatment arena.

Interview-Related Material

A limited number of studies have described the use of a structured interview procedure (i.e., Attachment Story-Completion Task; Bretherton, Ridgeway, & Cassidy, 1990) to assess preschool children’s representations of mothers in play narratives. Oppenheim, Emde, & Warren (1997) reported that 4 and 5-year-old children’s representations of mothers were associated with their social-emotional adaptation as well as that of their mothers. Goodman and Pfeffer (1998) found that four latency age children with significant psychiatric problems exhibited marked signs of pathological relatedness and disorganized internal working models indicative of different and discrete types of attachment disorders. More recently, Waldinger, Toth, & Gerber (2001) also used the same diagnostic procedure to address the issues of whether objectrepresentations material is related to the experiencing of various types of maltreatment and neglect. Results suggested that physically abused and neglected children’s self-representations were defined by angry and oppositional attributes; neglected children also represented others as hurt, sad, or anxious more frequently than both abused and nonmaltreated children. However, perhaps somewhat unexpectedly, sexually abused children represented others in a positive light and also expressed more frequent wishes for closeness. The authors concluded that maltreated children’s internal representations vary according to the type of maltreatment.

Commentary

The concept of internal working models (Bowlby, 1979) represents another paradigm defining mental representations of self and other, and a growing body of research is now starting to accrue, mostly involving young children, addressing the relationship between this construct and attachment phenomena. The existing literature is diverse, with reliability and validity information now starting to accumulate, with tentative indication that the Attachment Story-Completion Task provides valuable information regarding the child’s representations of others (Oppenheim et al., 1997). But as Goodman and Pfeffer (1998) caution, the association between the child’s working models and psychiatric symptomatology has not been systematically explored—most researchers have not studied children with severe emotional disturbances. While it is beyond the scope of the present effort to thoroughly address the literature on the assessment of internal working models in children, it would be appropriate for researchers and clinicians to consider incorporating this device into assessment procedures, with both latency and preschool children, as the existent research certainly suggests that a more structured and

play-oriented device may be an appropriate way to assess the domains of self- and other object representations in younger children, particularly those in the preschool years.

CONCLUSIONS

The projective assessment of objective representation in children and adolescents has witnessed an important and exciting literature gradually accruing over the course of the past decade since Stricker and Healey (1990) first published a major review article on projective assessment of object relations. The Rorschach has been employed to assess object relatedness in both latency age and adolescent children with the MOAS proving its mettle as a reliable and valid measure of self- and object representation in child and adolescent clinical as well as nonclinical populations. The MOAS provides a brief, easily scored, and clinically rich measure that may be used in conjunction with more easily obtained initial information from the Rorschach protocol when considering initial results from the Comprehensive System—for example, Good Human Responses (GHR) and Poor Human Responses (PHR). These latter measures do not provide the degree of interpretive depth or perspective that is inherent in the MOAS paradigm and they are not specifically yoked to a theoretical perspective. The use of the MOAS in conjunction with other content measures appears indicated as there is a dearth of literature examining variables such as oral dependency, aggressive content, reflection responses, and so forth, with both latency age and adolescent clinical as well as nonclinical populations.

Consideration regarding the Rorschach assessment of object representations in either latency or adolescent populations sees the clinician having at least two options. First, employing the GHR and PHR in a global interpretive overview with limited opportunity to comment on a more multidimensional level, although it is certainly possible to integrate this with other content scores to embellish possible motivational factors contributing to the GHR/PHR composite. Or second, utilizing the MOAS, which certainly provides the clinician with the chance to move beyond the material obtained as this perspective allows for a more differentiated and multidimensional source of data, one providing for both structural as well as dynamic interpretive possibilities; information being integrated into a paradigm articulating where the child may be found along the continuum of separation-individuation with all that this implies in a greater developmental sense. The recently suggested scoring revision for the MOAS (Holaday & Sparks, 2001) should serve to clarify scoring issues and enhance interrater reliability.

Clinicians and researchers opting to employ the TAT in the assessment of a child’s or adolescent’s object relations are informed by the seminal works of Westen and his colleagues. The SCORS provides opportunity for assessing a more differentiated and clinically rich sample of a child’s object world than does the MOAS, providing as it does perspective on both affective and cognitive representational domains, with this differentiation proving to be clinically critical in the categorization and understanding of specific populations, for example, physically abused versus sexually abused children. The bulk of more recent research with the SCORS has been largely confined to the investigation of children who have experienced some type of maltreatment, but application and investigation of this object-relations scale has not seen wider utility with other populations, for example, adolescents, different psychiatric subgroups, and culturally diverse groups.

Nomothetic research findings provide one way of considering SCORS data, although this is still a somewhat sparse literature. An idiographic perspective may be employed, an approach that is often eschewed but one that seems both clinically responsible and indicated (Morgan & Morgan, 2001). From a psychometric vantage point, the research has more than attested to its reliability when scoring concerns are addressed; more importantly, the numerous studies with children and adolescents provide credence to observations that the SCORS manifests impressive concurrent, predictive, and construct validity measures, albeit with a fairly limited number of studies in the final consideration. Nonetheless, this is an instrument that deserves continued and expanded research and clinical application as it provides a relevant and useful interpretive paradigm for assessing TAT-based object-relations information taken from the child’s narratives.

Additional projective devices (i.e., early memory information and sentence stem completion material) have yielded interesting and potentially useful but as yet minimal information with both children and adolescents. The assessment of early memory information provides another source of potentially useful and important clinical material in relation to the child’s or adolescent’s object representations, but aside from the earlier referenced material, a dearth of recent research is available. It would be important for clinicians and researchers to consider utilizing this approach in concert with other projective devices, as there is the strong likelihood of convergent information. For example, a young adolescent, with possible periods of binge eating, reports early memory material relating to food being forced upon her, produces an elevated number of oral content responses on the Rorschach, and generates SCORS material repeatedly introducing themes of not having one’s needs met or satisfied. Curiously, this potentially rich source of information has languished even though interesting and impressive early memory clinical material has been demonstrated with adults (Fowler et al., 1998).

Finally, the efforts of attachment theorists have shown a recent interest in applying a quasi-projective device (Attachment Story-Completion Task) in assessing younger children’s object representations (i.e., internal working models). As noted earlier, most efforts have focused on children without emotional or psychiatric problems in an attempt to validate the use of this measure in identifying distinct types of attachment prototypes in preschool and latency age children. More recently, a limited number of studies have begun to evaluate how psychiatrically impaired children with differing types of attachment disorganization represent internal working models of significant others based on their responses to the aforementioned measure; results, while tentatively useful in discriminating between children manifesting discrete types of attachment disorders, obviously require that additional information be obtained before more serious consideration of this data is possible.

Clinicians and researchers who wish to deliberate whether the continued use of projective devices should be employed in assessing object-relations phenomena in children and adolescents are advised to proceed and expand their efforts given the status quo—a literature that is tightly knit, rich, and diversified, albeit still limited when one reviews child and adolescent clinical research endeavors. Nonetheless, these instruments, most specifically the Rorschach and the TAT, have much to offer and their continued and expanded usage is indicated with a range of clinical and nonclinical populations to strengthen and augment validation measures. It would also appear to be appropriate to argue for the expanded use of early memory devices as well as story completion assessment procedures with both younger children and adolescents, although a lack of uniformity in administration and scoring in both instances preclude more than tentative comments about the usefulness and applicability of findings at this time.

Overall, the literature regarding assessment of objectrelations phenomena in children and adolescents still sees a somewhat limited informational source, but one that certainly deserves the attention of clinicians and researchers given the increasing testimony for the role that projective devices play in the assessment and articulation of a specific personality construct—object-relations capacity—in children and adolescents. Using the TAT, Rorschach, or additional measures referenced in this chapter offers the opportunity to access and evaluate a source of vital information providing important diagnostic and developmental material often inaccessible if other diagnostic procedures such as rating scales, interviews, and self-report inventories are only employed in assessment. While it is clear that the projective assessment of object relations in children and adolescents provides reliable and valid diagnostic information, it is also apparent that the future literature will probably continue to reveal a rather limited sampling owing to the fact that children and adolescents constitute a smaller clinical population than adults, and child psychologists as well are a relatively small group. Nonetheless, there is reason to expect that the future will be a bright one: challenging, exciting, and ultimately productive for those involved in assessing the relational world of children and adolescents. But the need to replicate existing research still remains a priority, along with expanded efforts to further establish the reliability and validity of these object-relations measures in developmental and clinical undertakings.

APPENDIX: CHILD AND ADOLESCENT PERSONALITY ASSESSMENT

These sources should augment in-text reference material, providing additional clinical and research information pertaining to object-representation assessment of children and adolescents.

• Ames, L., Metraux, R., Rodell, J., & Walker, R. (1995). Child Rorschach responses: Developmental trends from 2 to 10 years (Rev. ed.). Northvale, NJ: Jason Aronson.
• Cramer, P. (1996). Storytelling, narrative, and the Thematic Apperception Test. New York: Guilford Press.
• Exner, J., & Weiner, I. (1995). The Rorschach: A comprehensive system: Volume 3. Assessment of children and adolescents (2nd ed.). New York: Wiley.
• Kelly, F. (1996). Object relations assessment in younger children: Rorschach and TAT measures. Springfield, IL: Charles C. Thomas.
• Kelly, F. (1997). The assessment of object relations phenomena in adolescents: TAT and Rorschach measures. Mahwah, NJ: Erlbaum.
• Kelly, F. (1999). The psychological assessment of abused and traumatized children. Mahwah, NJ: Erlbaum.
• Kissen, M. (1986). Assessing object relations phenomena. New York: International Universities Press.
• Leichtman, M. (1996). The Rorschach: A developmental perspective. Hillsdale, NJ: Analytic Press.

REFERENCES

Blatt, S., Brenneis, B., Schimek, J., & Glick, M. (1976). Normal development and psychopathological impairment of the concept of the object on the Rorschach. Journal of Abnormal Psychology, 85, 364–373.

• Blatt, S., & Lerner, H. (1983). The psychological assessment of object representation. Journal of Personality Assessment, 47, 7–28.
• Bowlby, J. (1979). The making and breaking of affectional bonds. New York: Tavistock.
• Bretherton, I., Ridgeway, D., & Cassidy., J. (1990). Assessing internal working models of the attachment relationship: An Attachment Story-Completion task for 3-year-olds. In M.T. Greenberg, D. Cicchetti, & E.M. Cummings (Eds.), Attachment in the preschool years: Theory, research, and intervention (pp. 273–308). Chicago: University of Chicago Press.
• Bruhn, A. (1981). Children’s earliest memories: Their use in clinical practice. Journal of Personality Assessment, 45, 258–262.
• Bruhn, A. (1983). The psychodiagnostic value of children’s earliest memories. Journal of Personality Assessment, 47, 597–603.
• Bruhn, A., & Davidow, S. (1983). Earliest memories and the dynamics of delinquency. Journal of Personality Assessment, 47, 476–482.
• Burns, B., & Viglione, D. (1996). The Rorschach human experience variable, interpersonal relatedness, and object representation in nonpatients. Psychological Assessment, 8, 92–99.
• Coates, S., & Tuber, S. (1988). The representation of object relations in the Rorschachs of extremely feminine boys. In H. Lerner & P. Lerner (Eds.), Primitive mental states and the Rorschach (pp. 647–654). New York: International Universities Press.
• Cramer, P. (1999). Future directions for the Thematic Apperception Test. Journal of Personality Assessment, 72, 74–92.
• Davidow, S., & Bruhn, A. (1990). Earliest memories and the dynamics of delinquency: A replication study. Journal of Personality Assessment, 54, 601–616.
• Exner, J. (2000). A primer for Rorschach interpretation. Asheville, NC: Rorschach Workshops.
• Fowler, C., Hilsenroth, M., & Handler, L. (1998). Assessing transitional phenomena with the transitional object memory probe. Bulletin of the Menninger Clinic, 62, 455–474.
• Freedenfeld, R., Ornduff, S., & Kelsey, R. (1995). Object relations and physical abuse: A TAT analysis. Journal of Personality Assessment, 64, 552–568.
• Fritsch, R., & Holmstrom, R. (1990). Assessing object representations as a continuous variable: A modification of the concept of the object on the Rorschach scale. Journal of Personality Assessment, 55, 319–334.
• Gacano, C., & Meloy, R. (1994). The Rorschach assessment of aggressive and psychopathic personalities. Hillsdale, NJ: Erlbaum.
• Goldberg, E. (1989). Severity of depression developmental levels of functioning in 8 to 16 year old girls. American Journal of Orthopsychiatry, 59, 167–178.
• Goodman, G., & Pfeffer, C. (1998). Attachment disorganization in prepubertal children with severe emotional disturbance. Bulletin of the Menninger Clinic, 62, 490–525.
• Holaday, M., & Sparks, C. (2001). Revised guidelines for Urist’s Mutuality of Autonomy Scale (MOAS). Assessment, 8, 145– 154.

626 Assessment of Object Representation in Children and Adolescents: Current Trends and Future Directions

• Kelly, F. (1986). Assessment of the borderline adolescent: Psychological measures of defensive structure and object representation. Journal of Child and Adolescent Psychotherapy, 3, 199– 206.
• Kelly, F., & Burch, C. (2000, March). The psychological sequelae of chronic, complex trauma in latency age children Rorschach indices. Paper presented at the annual meeting of the Society for Personality Assessment, Albuquerque, NM.
• Kernberg, O. (1976). Object relations theory and clinical psychoanalysis. New York: Jason Aronson.
• Kohut, H. (1977). The restoration of the self. New York: International Universities Press.
• Last, J., & Bruhn, A. (1983). The psychodiagnostic value of children’s earliest memories. Journal of Personality Assessment, 47, 476–482.
• Last, J., & Bruhn, A. (1985). Distinguishing child diagnostic types with early memories. Journal of Personality Assessment, 49, 187–192.
• Leifer, H., Shapiro, J., Martone, M., & Kassem, L. (1991). Rorschach assessment of psychological functioning in sexually abused girls. Journal of Personality Assessment, 56, 14–28.
• Mahler, M. (1968). On human symbiosis and the vicissitudes of individuation (Vol. 1). New York: International Universities Press.
• Meyer, J., & Tuber, S. (1989). Intrapsychic and behavioral correlates of the phenomena of imaginary companions in young children. Psychoanalytic Psychology, 6, 151–168.
• Morgan, D., & Morgan, R. (2001). Single-participant research design: Bringing science to managed care. American Psychologist, 56, 119–127.
• Oppenheim, D., Emde, R., & Warren, S. (1997). Children’s narrative representations of mothers: Their development and associations with child and mother adaptation. Child Development, 68, 127– 138.
• Ornduff, S. (1997). TAT assessment of object relations: Implications for child abuse. Bulletin of the Menninger Clinic, 61, 1–15.
• Ornduff, S., Centeno, L., & Kelsey, R. (1999). Rorschach assessment of malevolence in sexually abused girls. Journal of Personality Assessment, 73, 100–109.
• Ornduff, S., Freedenfeld, R., Kelsey, R., & Critelli, J. (1994). Object relations of sexually abused female subjects: A TAT analysis. Journal of Personality Assessment, 63, 223–238.
• Ornduff, S., & Kelsey, R. (1996). Object relations of sexually and physically abused female children: A TAT analysis. Journal of Personality Assessment, 66, 91–105.
• Pistole, D., & Ornduff, S. (1994). TAT assessment of sexually abused girls: An analysis of manifest content. Journal of Personality Assessment, 63, 211–222.
• Ryan, R., Avery, R., & Grolnick, W. (1985). A Rorschach assessment of children’s mutuality of autonomy. Journal of Personality Assessment, 49, 6–12.
• Stricker, G., & Healey, B. (1990). Projective assessment of object relations: A review of the empirical literature. Psychological Assessment, 2, 219–230.
• Tuber, S. (1983). Children’s Rorschach scores as predictors of later adjustment. Journal of Clinical and Consulting Psychology, 51, 379–385.
• Tuber, S. (1989a). Assessment of children’s object representations with the Rorschach. Bulletin of the Menninger Clinic, 53, 432– 441.
• Tuber, S. (1989b). Children’s object representations: Findings for a non-clinical sample. Psychological Assessment, 1, 146–149.
• Tuber, S. (1992). Empirical and clinical assessments of children’s object relations and object representations. Journal of Personality Assessment, 58, 179–197.
• Tuber, S., & Coates, S. (1989). Indices of psychopathology in the Rorschachs of boys with severe gender identity disorder: A comparison with control subjects. Journal of Personality Assessment, 53, 100–112.
• Tuber, S., Frank, M., & Santostefano, S. (1989). Children’s anticipation of impending surgery. Bulletin of the Menninger Clinic, 53, 501–511.
• Urist, J. (1977). The Rorschach test and the assessment of object relations. Journal of Personality Assessment, 41, 3–9.
• Urist, J. (1980). Object relations. In R.H. Woody (Ed.), Encyclopedia of clinical assessment (Vol. 1, pp. 821–833). San Francisco: Jossey-Bass.
• Urist, J., & Shill, M. (1982). Validity of the Rorschach Mutuality of Autonomy Scale: A replication using excerpted responses. Journal of Personality Assessment, 46, 450–454.
• Waldinger, R., Toth, S., & Gerber, A. (2001). Maltreatment and internal representations of relationships: Core relationship themes in the narratives of abused and neglected preschoolers. Social Development, 10, 41–58.
• Westen, D. (1991). Social cognition and object relations. Psychological Bulletin, 109, 429–455.
• Westen, D. (1995). Social Cognition and Object Relations Scale: Q-sort for projective stories (SCORS-Q), Unpublished manuscript, Department of Psychiatry, The Cambridge Hospital and Harvard Medical School.
• Westen, D., Klepser, J., Ruffins, S., Silverman, M., Lifton, N., & Boekamp, J. (1991). Object relations in childhood and adolescence: The development of working relationships. Journal of Clinical and Consulting Psychology, 59, 400–409.
• Westen, D., Lohr, N., Silk, K., Kerber, K., & Goodrich, S. (1985). Social Cognition and Object Relations Scale (SCORS): Manual for coding TAT data. University of Michigan, Ann Arbor.
• Westen, D., Ludolph, P., Block, J., Wixom, J., & Wiss, F. (1990). Developmental history and object relations in psychiatrically disturbed adolescent girls. American Journal of Psychiatry, 147, 1061–1068.
• Westen, D., Ludolph, P., Lerner, H., Ruffins, S., & Wiss, F. (1990). Object relations in borderline adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 29, 338– 348.
• Westen, D., Ludolph, P., Misle, B., Ruffins, S., & Block, J. (1990). Physical and sexual abuse in adolescents with borderline per-

sonality disorder. American Journal of Orthopsychiatry, 60, 55–66.

Westen, D., Ludolph, P., Silk, K., Kellam, A., Gold, L., & Lohr, N. (1990). Object relations in borderline adolescents and adults: Developmental differences. In S.C. Feinstein (Ed.), Adolescent Psychiatry: Developmental and Clinical Studies, 17, 360–384.

Projective Assessment of Affect in Children’s Play

SANDRA W. RUSS

TEST DESCRIPTION 628 The APS Play Task 628 The APS Rating Scale 629 THEORETICAL BASIS 629 APS as a Projective Test 630 TEST DEVELOPMENT 630 PSYCHOMETRIC CHARACTERISTICS 630 Reliability 630 Validity Studies 631 RANGE OF APPLICABILITY AND LIMITATIONS 637

The Affect in Play Scale (APS; Russ, 1987, 1993) was developed to meet the need for a standardized play task and for a reliable and valid scale that measures affective expression in children’s pretend play. The need for a measure of affective expression in children’s play was widely recognized (Howe & Silvern, 1981; Rubin, Fein, & Vandenberg, 1983). The lack of adequate measures of affect in play led to the “cognification” of play in that cognitive processes, not affective processes, were the focus in play research (Rubin et al., 1983). With the exciting developments in the research on emotion and on cognitive-affective interaction, it is essential that measures of affect in play be developed.

Pretend play is an arena in which cognitive and affective processes important in child development are expressed and developed. The APS is one of the very few standardized measures of affect in pretend play.

TEST DESCRIPTION

The APS play task and rating scale are described.

The APS Play Task

The Affect in Play Scale consists of a standardized play task and a criterion-based rating scale. The APS is appropriate for CROSS-CULTURAL FACTORS 638 ACCOMMODATION FOR POPULATIONS WITH DISABILITIES 638 LEGAL AND ETHICAL CONSIDERATIONS 638 COMPUTERIZATION 638 CURRENT RESEARCH STUDIES 638 USE IN CLINICAL PRACTICE 639 FUTURE DIRECTIONS 639 REFERENCES 639

children from 6 to 10 years of age, which includes children in Grades 1 through 3.

The play task consists of two human puppets, one boy and one girl, and three small blocks that are laid out on a table. The puppets have neutral facial expressions. The blocks are brightly colored and of different shapes. The play props and instructions are unstructured enough so that individual differences in play can emerge. The task is administered individually to the child and the play is videotaped. The instructions for the task are:

I’m here to learn about how children play. I have here two puppets and would like you to play with them any way you like for five minutes. For example, you can have the puppets do something together. I also have some blocks that you can use. Be sure to have the puppets talk out loud. The video camera will be on so that I can remember what you say and do. I’ll tell you when to stop.

The child is informed when there is 1 minute left. If the child stops playing during the 5-minute period, the prompt “You still have time left, keep going” is given. The task is discontinued if the child cannot play after a 2-minute period.

These are free-play instructions that leave much room for the child to structure the play and present themes and affects that are habitual to him or her. Although the instruction “For example, you could have the puppets do something together” does provide structure, we found that some structure was necessary for many children to be able to carry out the task. These instructions can be altered to elicit different types of affect. For example, to pull for aggression, the instructions would be “Play with them and have the puppets disagree about something,” rather than “Play with them any way you like.” The play task described here is appropriate for Grades 1 through 3 (6 to 10 years of age). In our experience, many kindergarten children have difficulty with the puppet task. However, the rating criteria could be used in a natural play observation situation for very young children. An adaptation of the APS for young children will be discussed later in this chapter.

The APS Rating Scale

The Affect in Play Scale measures the amount and types of affective expression in children’s fantasy play. The APS measures affect themes in the play narrative. Both emotion-laden content and expression of emotion in the play are coded. The APS also measures cognitive dimensions of the play, such as quality of fantasy and imagination.

Both Holt’s Scoring System for Primary Process on the Rorschach (1977) and J.L. Singer’s (1973) play scales were used as models for the development of the Affect in Play Scale. In addition, the work of Izard (1977) and Tomkins (1962, 1963) was consulted to ensure that the affect categories were comprehensive and covered all major types of emotion expressed by children in the 4 to 10 age group.

There are three major affect scores for the APS:

1. Total frequency of units of affective expression. A unit is defined as one scorable expression by an individual puppet. In a two-puppet dialogue, expressions of each puppet are scored separately. A unit can be the expression of an affect state, an affect theme, or a combination of the two. An example of an affect state would be “This is fun.” An example of an affect theme would be “Here is a bomb that is going to explode.” The expression can be verbal (“I hate you”) or nonverbal (one puppet punching the other).

The frequency of affect score is the total number of units of affect expressed in the 5-minute period.

2. Variety of affect categories. There are 11 possible affective categories. The categories are: Happiness/Pleasure; Anxiety/Fear; Sadness/Hurt; Frustration/Disappointment; Nurturance/Affection; Aggression; Oral; Oral Aggression; Anal; Sexual; Competition. The variety of affect score is the number of different categories of affect expressed in the 5-minute period.

3. Mean intensity of affective expression (1 to 5 rating). This rating measures the intensity of the feeling state or content theme. Each unit of affect is rated for intensity on a 1 to 5 scale.

Quality of fantasy and imagination is also scored. Although other scales (J.L. Singer, 1973) already tapped this dimension, it was important to include this aspect of pretend play in the scoring system so that the APS would be comprehensive in its assessment of fantasy play. The fantasy scores are

• Organization (1 to 5 global rating): This score measures the organization of the play and considers the quality of the plot and complexity of the story.
• Elaboration (1 to 5 global rating): This score measures the amount of embellishment in the play.
• Imagination (1 to 5 global rating): This score measures the novelty and uniqueness of the play and the ability to pretend.
• Quality of fantasy: This score is the mean of the previous three fantasy scores.

In addition, comfort in play is rated on a 1 to 5 scale. Comfort includes the involvement of the child in the play and his or her enjoyment of the play. Finally, an affect integration score is obtained by multiplying the quality of fantasy score by the frequency of affect score. The affect integration score is needed because it attempts to measure the construct of cognitive modulation of emotion. It taps how well the affect is integrated and controlled by cognitive processes.

To summarize, the nine major scores on the APS are total frequency of affect, variety of affect categories, mean intensity of affect, organization of fantasy, elaboration of fantasy, imagination, overall quality of fantasy, comfort, and affect integration.

Practically, the APS is easy to administer and takes only 5 minutes. The props, human puppets and blocks, are simple. The scoring system takes time to learn, but then takes about 15 to 20 minutes per child. Although the APS has not been used with clinical populations, we do have a number of studies with means for nonclinical populations. Usually, the mean frequency of affect expression is 11 to 13 units with a mean variety of categories of 3 to 4.

THEORETICAL BASIS

Before the APS could be developed, a classification system for types of affect in fantasy was necessary to provide conceptual guidelines for the instrument. In play, feeling states are expressed in fantasy in a variety of ways. Although no formal classification system had been developed, three dimensions of affective expression in fantasy emerged in the play and psychotherapy literature:

• Affect-states: Actual affective experiencing and expression of feeling states and emotions.
• Affect-laden thoughts: Expression of affect themes and content. Primary process content is a subtype of affectladen thoughts. Emotion may or may not accompany the content.
• Cognitive integration of affective material: The integration and modulation of affect into the fantasy.

Sarnoff (1976) stated that the effective use of fantasy reflects a cross-situational, cognitively based structure that results in the ability to modulate the expression and experience of otherwise disruptive emotions.

The affect scores in the APS measure both affect themes and affect states. The scores reflect a combination of those elements. The affect integration score measures cognitive integration of the affect.

The APS also measures cognitive dimensions of play such as organization of the story and imagination. Theoretically, the affect and cognitive scores should reflect related but separate processes with unique variance. Therefore, one would expect separate factors.

In order to ensure that the affect categories were comprehensive and covered all major types of emotion expressed by children in the 4 to 10 year group, the research on emotion by Izard (1977) and Tomkins (1962, 1963) was consulted. Holt’s Scoring System for Primary Process on the Rorschach (1977) was a major model for developing the scoring system for the APS. J.L. Singer’s (1973) play scales were also models, especially for the cognitive component of play.

APS as a Projective Test

One of the main characteristics of a projective test is that it be unstructured or ambiguous (Weiner, 1998). Where does the APS fall on the continuum of structured versus unstructured? The APS was designed to be sufficiently unstructured so that there was opportunity for the child to express the tendency to be open to affect and affect-laden thoughts. There is some structure in that the two puppets are human, although with neutral facial expressions. The blocks are unstructured and can be used to represent a number of different objects. The instructions are relatively unstructured. The child is told that he or she could have the puppets do something together, but the choice of theme and events is up to the child. The amount of structure can be altered by increasing the amount of direction in the instructions. For example, if we tell children to have the puppets disagree about something, the amount of aggressive themes in the play will significantly increase (Russ & Kaugars, 2000–2001). Structure can also be added by including play objects with more concrete properties. We did this with a preschool version of the scale, since young children need more structure to work with (Kaugars & Russ, 2000).

The APS gives quantitative information that tells us about the child’s tendency to express emotion in an unstructured fantasy play situation and the tendency to express a broad range of emotion. It also tells us about the child’s tendency to organize a story and generate a variety of ideas that are imaginative. The cognitive and affective processes that are tapped should have traitlike properties. The APS, as scored, does not tap personality dynamics. There is no interpretation of the affect themes. The APS measures personality structure (traitlike processes), not personality dynamics.

TEST DEVELOPMENT

Once the Affect in Play Scale was constructed, pilot studies were carried out to ensure that the task was appropriate for young children and would result in adequate individual differences among normal school populations (Russ, Grossman-McKee, & Rutkin, 1984). By 1984, the basics of the task and scoring system were in place. Early studies resulted in refinement of the scoring criteria and a shortening of the play period (from 10 minutes to 5 minutes). The next step was to develop reliability and build construct validity for the scale. To date, a substantial number of validity studies have been carried out with different populations and different examiners (see Table 47.1).

The current manual for scoring is basically the same as the 1993 version. We have eliminated a repetition scale and have clarified a number of scoring issues.

PSYCHOMETRIC CHARACTERISTICS

A number of studies have investigated the reliability and validity of the Affect in Play Scale.

Reliability

Interrater reliabilities in all of the studies have been consistently good. Because a detailed scoring manual was developed and raters were carefully trained, interrater reliabilities using a number of different raters have usually been in the .80s and .90s, with some in the .70s. For example, in a study

Authors	Date	Variables Investigated
Grossman-McKee	1989	Pain complaints
Peterson	1989	Self-esteem
Russ & Grossman-McKee	1990	Divergent thinking; primary process
Russ & Peterson	1990	Divergent thinking; coping in school
D’Angelo	1995	Adjustment; imagination
Niec & Russ	1996	Interpersonal themes in storytelling and interpersonal functioning
Christiano & Russ	1996	Coping and distress in dental visit
Russ, Robins, & Christiano	1999	Longitudinal prediction of creativity, coping, and older version of play
		task
Niec	1998	Internal representations and empathy
Perry & Russ	1998	Coping and adjustment in homeless children
Seja & Russ	1998	Parents’ reports of daily emotional expression
Seja & Russ	1999	Emotional understanding
Kaugars & Russ	2000	Creativity in preschoolers
Goldstein	2002	Imagination and anxiety
Russ & Cooperberg	2001	Longitudinal prediction of creativity, coping, and depression
Russ & Schafer	2001	Divergent thinking and emotion in memories
Goldstein & Russ	2000–2001	Coping

TABLE 47.1 Validity Studies of Affect in Play Scale

by Russ and Grossman-McKee (1990), based on 15 randomly chosen subjects, Pearson r correlation coefficients were as follows: total frequency of affect, r ” .90; variety of categories, r ” .82; intensity of affect, r ” .53; mean quality of fantasy, r ” .88; imagination, r ” .74; and comfort, r ” .89. With the exception of intensity of affect, which was therefore not included in the analysis, all of the interrater reliabilities were judged to be good.

In a study by Christiano and Russ (1996), interrater reliabilities for 20 participants were: total frequency of affect, r ” .91, variety of affect, r ” .90, quality of fantasy, r ” .85, and comfort, r ” .90. In a recent study by Seja and Russ (1999a), correlations were: frequency of affect,r ” .83, quality of fantasy, r ” .80, organization, r ” .72, elaboration, r ” .74, and imagination, r ” .78.

In two studies, we have investigated the internal consistency of the APS and found it to be good. We compared the 2nd and 4th minutes with the 3rd and 5th minutes of the play period for frequency of affect. In both studies we found, using the Spearman-Brown split-half reliability formula, good internal consistency of r ” .85 for frequency of affect (Russ & Peterson, 1990; Seja & Russ, 1999a).

Test-retest reliability for the APS needs to be determined. We are in the process of analyzing data for 50 children who were administered the task a second time after a 2- to 4-week interval.

Validity Studies

The development of construct validity for the APS has been carried out by investigating the relationships between the scores on the APS and criteria that should be related to the constructs of fantasy and affect in fantasy (Anastasia, 1988). By finding relationships between a measure and theoretically relevant criteria, conceptual validity is developed (Weiner, 1977). For the APS, validity studies have been carried out with four major types of theoretically relevant criteria: creativity, coping and adjustment, emotional understanding, and interpersonal functioning (see Table 47.1). Each area is briefly reviewed in the following sections.

APS and Creativity

One of the most robust findings in the literature is the relationship between pretend play and creativity. Russ (1993) postulated that pretend play is important in developing creativity because so many of the cognitive and affective processes involved in creativity occur in play. Russ’s (1993) model of affect and creativity identified the major cognitive and affective processes involved in creativity, and the relationships among them, based on the research literature.

Divergent thinking is one major cognitive process important in creativity and was a focus of several validity studies with the APS. As defined by Guilford (1968), divergent thinking is thinking that generates a variety of ideas and associations to a problem. Divergent thinking involves free association, broad scanning ability, and fluidity of thinking. It has been found to be relatively independent of intelligence (Runco, 1991).

Two affective processes important in creativity are access to affect-laden thoughts and the ability to experience affect states (Russ, 1993). Both the ability to think about affectladen fantasy and the capacity to experience emotion are important in creativity. In play, children express affect in

fantasy and experience emotion. For example, Fein (1987) concluded that play facilitated the development of an affective symbol system important in creativity. Waelder (1933) viewed play as a place in which primary process thinking can occur. Morrison (1988) conceptualized play as an arena in which children reconstruct past experiences and rework old metaphors.

Pretend play should facilitate the development of divergent thinking for several reasons. The expression of emotion and affect-laden fantasy in play could help develop a broad repertoire of affect-laden associations (Russ, 1993, 1996). This broad repertoire of associations and use of emotion to access these associations should facilitate divergent thinking because the involvement of emotion broadens the search process for associations (Isen, Daubman, & Nowicki, 1987). Play should also facilitate divergent thinking because in play children practice divergent thinking skills by using toys and objects to represent different things and by role playing different scenarios (D. Singer & Singer, 1990).

A growing body of research has found a relationship between play and creativity. Most of the research has been correlational in nature and has focused on cognitive processes. A substantial body of studies have found a relationship between play and divergent thinking (Clark, Griffing, & Johnson, 1989; Johnson, 1976; Peppler & Ross, 1981; D. Singer & Rummo, 1973). In addition, in experimental studies, play has been found to facilitate divergent thinking (Dansky, 1980; Dansky & Silverman, 1973; Feitelson & Ross, 1973; Hughes, 1987) and insight (Vandenberg, 1980). Flexibility in problem solving has also been related to play (Pelligrini, 1992).

A few studies have found a relationship between affective processes in play and creativity. Lieberman (1977) found a relationship between playfulness, which included affective components of spontaneity and joy, and divergent thinking in kindergarten children. Christie and Johnson (1983) also concluded that there was a relationship between playfulness and creativity. J.L. Singer and Singer (1981) found that preschoolers rated as high-imagination players showed significantly more themes of danger and power than children with low imagination.

In the first study with the APS, we were particularly interested in the relationship between the affect scores and creativity. Russ and Grossman-McKee (1990) investigated the relationships among the APS, primary process thinking on the Rorschach, and divergent thinking in first- and second-grade children. Sixty children individually received the Rorschach, APS, and Alternate Uses Test. A typical item on the Alternate Uses Test is “How many uses for a newspaper can you think of?” Holt’s Scoring System was the measure for the Rorschach (1977). Primary process thinking was included in the study because it is affect-laden ideation that has been found to be related to a number of creativity criteria (Russ, 1996; Suler, 1980). The version of the APS used in this study was different than the current version in that the play sessions were 10 minutes in length (rather than 5) and the play was audiotaped, not videotaped, with careful notetaking by the examiner as the play occurred.

A major finding of this study was that affective expression in play was predictive of divergent thinking. The predicted relationships between the play scores and the Alternate Uses Test were all significant for the total sample, except for the relationship between frequency of nonprimary process affect and divergent thinking. Divergent thinking was significantly related to frequency of affect [r(58) ” .42, p # .001], variety of affect categories [r(58) ” .38, p # .001], comfort [r(58) ” .23, p # .051], frequency of primary process affect [r(58) ” .41, p # .001], quality of fantasy [r(58) ” .30, p # .01], imagination [r(58) ” .35, p # .01], and integration of affect [r(58) ” .42, p # .001]. All correlations remained significant when IQ was partialed out, due to the fact that IQ had such low relationships with the play scores (e.g., r ” .09 with frequency of affect, r ” .01 with comfort, r ” .08 with quality, and r ” .12 with imagination). The fact that intelligence did not relate to any of the play measures is theoretically consistent with the model for the development of the scale and is similar to the results of J.L. Singer (1973). There were no gender differences in the pattern of correlations.

Also, as predicted, amount of primary process thinking on the Rorschach was significantly positively related to the amount of affect in play. Total frequency of primary process on the Rorschach was significantly positively related to the following play measures: frequency of affect [r(44) ” .34, p # .01]; variety of affective categories [r(44) ” .44, p # .001]; frequency of primary process affect [r(44) ” .30, p # .05]; frequency of nonprimary process affect [r(44) ” .26, p # .05]; comfort [r(44) ” .45, p # .001]; quality of fantasy [r(44) ” .48, p # .001]; imagination [r(44) ” .47, p # .01]; and the composite integration of affect score [r(44) ” .37, p # .01]. Percentage of primary process, which controls for general productivity, was also significantly related to most of the play variables, although the correlations were lower than those with total frequency. Percentage of primary process was significantly related to frequency of affect [r(44) ” .32, p # .05]; variety of affective categories [r(44) ” .28, p # .05]; frequency of primary process [r(44) ” .28, p # .05]; frequency of nonprimary process [r(44) ” .25, p # .05]; quality of fantasy [r(44) ” .27, p # .05]; imagination [r(44) ” .30, p # .05], and integration of affect [r(44) ” .32, p # .05].

Primary process thinking on the Rorschach was equally predictive for girls and for boys in the play situation. The relationships between the variables were not affected when intelligence was controlled for. The finding in this study, that primary process expression on the Rorschach was significantly related to affective expression in children’s play, is important because it shows that there is some consistency in the construct of affective expression across two different types of situations.

A study by Russ and Peterson (1990) investigated the relationships among the APS, divergent thinking, and coping in school in first- and second-grade children. The main purpose of this study was to obtain a large enough sample size (121 children) so that a sound factor analysis of the play scale could be carried out for the total sample and separately for boys and girls. A second purpose was to replicate the results of the Russ and Grossman-McKee (1990) study that found a positive relationship between affective expression in play and divergent thinking.

One hundred twenty-one children (64 boys and 57 girls) were individually administered the APS and a Coping in School scale. In a separate testing session, with a different examiner, they were administered the Alternate Uses Test. The APS used in this study and subsequent studies was the current version of the scale. The play period was 5 minutes instead of 10 minutes. A video camera was used rather than a tape recorder. Also, some of the affect categories were condensed, because of infrequent occurrence. Displeasure and frustration became one category and sadness and hurt became another category. A new category, competition, was added because of its prevalence in children’s play and because it is considered to be a derivative of aggressive content in Holt’s system. Finally, there were some minor adjustments in the intensity rating criteria.

The main finding in this study was that the APS was significantly positively related to divergent thinking. These results replicated the findings of the Russ and Grossman-McKee (1990) study with children of the same age. As in the previous study, there were no gender differences in the pattern of correlations. For the total sample, divergent thinking was significantly related to frequency of total affect [r(115) ” .26, p # .01]; variety of affect [r(115) ” .25, p # .01]; comfort [r(115) ” .37, p # .001]; quality of fantasy [r(115) ” .43, p # .001]; imagination [r(115) ” .42, p # .001]; primary process [r(115) ” .17, p # .05]; nonprimary process [r(115) ” .24, p # .01]; and integration of affect [r(115) ” .30, p # .001]. These relationships remained significant when IQ was partialed out. Based on this study, we can say with more confidence that affective expression in fantasy relates to divergent thinking, independent of the cognitive processes measured by intelligence tests.

It is important to note that in both the Russ and Grossman-McKee (1990) study and the Russ and Peterson (1990) study, the significant relationship between play and divergent thinking occurred in studies where the play task and the divergent thinking task were administered by different examiners. Given Smith and Whitney’s (1987) criticism that previous positive results that linked play and associative fluency were due to experimenter effects, these are important findings.

In a recent study with 47 first- and second-grade children, Russ and Schafer (2001) found hypothesized significant relationships between frequency of affect and variety of affect and divergent thinking. In this study we used three emotionladen objects and three non-emotion-laden objects for the Alternate Uses task. Frequency of affect related to total uses for emotion-laden objects (r ” .25, p # .05), and to originality of response for emotion-laden (r ” .32, p # .05) and nonemotion-laden (r ” .57, p # .01) objects. The relationship between variety of affect and divergent thinking followed a similar pattern. Most of these correlations did not remain significant when IQ was partialed out (with the exception of originality of response for non-emotion-laden objects and affect). IQ was related to the APS in this sample, an unusual occurrence in the play studies. So for this sample, play did relate to creativity, but usually not independent of intelligence.

Play also related to divergent thinking in a preschool sample of children, to be discussed later in this chapter (Kaugars & Russ, 2000). The affect scores especially were related to creativity and to teachers’ ratings of make-believe.

Using a different kind of imagination criterion, Goldstein (2002) found significant relationships between play and Singer’s imaginative predisposition interview (IPPI). Children with more fantasy and affect in play scored higher on this interview, which assessed preference for imaginative activities (frequency of affect and IPPI, r ” .42, p # .001; variety of affect and IPPI, r ” .35, p # .001; fantasy and IPPI, r ” .27, p # .05).

APS, Coping, and Adjustment

Theoretically, play ability should be related to coping ability and to broader measures of adjustment. This link should occur for several reasons. First, children use play to solve reallife problems and to resolve internal conflicts (Erikson, 1963; Freud, 1965). Children play out their problems in pretend play, express negative emotions in a controllable way, and practice with different behaviors. Second, the creative problem-solving skills developed in play should generalize to problem-solving skills in daily life. Creative problem solvers should be better copers because they bring their problem-solving skills to everyday problems. The ability to generate a variety of associations and alternative solutions should facilitate coping with daily stressors. There is some empirical support for this concept. Russ (1988) found a relationship between divergent thinking and teachers’ ratings of coping in fifth-grade boys. Similarly, Carsen, Bittner, Cameron, Brown, and Meyer (1994) found a significant relationship between figural divergent thinking and teachers’ ratings of coping.

Looking specifically at play on the APS and coping ability, Christiano and Russ (1996) found a positive relationship between play and coping and a negative relationship between play and distress in 7- to 9-year-olds. Children who were “good” players on the APS implemented a greater number and variety of cognitive coping strategies (correlations ranging from .52 to .55) during an invasive dental procedure. In addition, good players reported less distress during the procedure than children who expressed less affect and fantasy in play. Also, the Russ and Peterson study (1990) found a relationship between fantasy in play, self-report coping, and teachers’ ratings of coping.

Consistent with these findings, a recent study by Perry and Russ (1998) found that fantasy in play on the APS was positively related to frequency and variety of self-reported coping strategies on the Schoolagers Coping Strategies Semistructured Interview in a group of homeless children. This sample of homeless children was primarily African American (77%). Because videotaping was not permitted in the shelter, the play was transcribed as it occurred. In this sample, 61 children living in homeless shelters were administered the APS, Schoolagers Coping Strategies Interview, and adjustment measures. The coping measure was self-report. Quality of fantasy in play significantly positively correlated with the frequency of coping responses (r ” .42, p # .01) and with the variety of coping responses (r ” .35, p # .01). These relationships were independent of age and achievement. Thus, children who had good fantasy skills were able to report a greater number and variety of coping strategies to use when confronted with stressful events.

In a study with first-grade children, Goldstein and Russ (2000–2001) found a significant positive relationship between the imagination score on the APS and children’s selfreport of how they would cope with a specific situation. The imagination score related to total frequency of coping responses (r ” .43, p # .001) and total variety of strategies (r ” .40, p # .001). These relationships remained significant when IQ was controlled for. In a multiple regression analysis, the APS score accounted for 37% of the variance in the total number of coping attempts and 29% of the variance in types of strategies used after controlling for IQ.

These coping studies used different samples of children, different examiners, and different measures of coping. In these different studies, various scores on the APS were related to coping ability.

The APS has also been related to more global measures of children’s adjustment. Grossman-McKee (1989), using the APS with first- and second-grade boys, found that boys who expressed more affect in play had fewer pain complaints than boys with less affect in play. Good players were also less anxious on the State-Trait Anxiety Inventory for Children (Spielberger, 1973). The conclusion from this study was that the ability to express affect in play was associated with less anxiety and fewer psychosomatic complaints.

Peterson (1989) in a 1-year follow-up on a subsample of 50 of the original 121 children in the Russ and Peterson study, found that the APS predicted self-esteem on the Self-Perception Profile for Children (Harter, 1994). Also, D’Angelo (1995) found, in a group of inner-city first- and second-grade children, that ego resilient children had higher APS scores than less resilient children. Also, internalizing children on the Child Behavior Checklist (Edelbrock & Achenbach, 1980) had significantly lower fantasy and affect scores in play than did externalizing and ego resilient children. Externalizing children had significantly lower fantasy scores (but not affect) than did ego resilient children.

In the Perry and Russ (1998) study with homeless children, the APS was significantly related to depression in children, but not to anxiety. That is, better players had lower scores on the Children’s Depression Inventory. Goldstein (2001) found, similar to the Grossman-McKee (1989) finding, that good players were less anxious on the state anxiety component of the State-Trait Anxiety Inventory for Children (r ” .23, p # .05).

Longitudinal Prediction of the APS

A study by Russ, Robins, and Christiano (1999) followed up the first and second graders in the Russ and Peterson (1990) study that investigated the APS, divergent thinking, and coping. The children were now in the fifth and sixth grades. Thirty-one of the original 121 children participated. This was a longitudinal study that explored the ability of the APS to predict creativity and coping over a 4-year period. The Alternate Uses Test was the measure of divergent thinking and a self-report School Coping Scale was the measure of coping. In addition, a version of the APS for older children was administered. The same basic task was administered, but the children were instructed to put on a play. In essence, the puppet play task became a storytelling task in the form of a play. The stories were scored with the same criteria as for the APS.

As predicted, quality of fantasy and imagination in early play predicted divergent thinking over time, independent of IQ (see Table 47.2). Variety of affect categories and comfort showed low positive correlations with divergent thinking but did not reach significance. The APS also significantly predicted coping over time. The fantasy scores predicted the number of different responses generated on the coping measures.

In addition, the APS was predictive of the version of the scale for older children. Most of the APS scores were significantly related to the comparable score on the modified play task (see Table 47.3). The magnitude of the correlations is quite good for longitudinal data. The strongest correlations were for the affect scores—r ” .51, p # .01 for positive affect; r ” .38, p # .05 for variety of affect; and r ” .33, p # .05 for total frequency of affect expressed. These findings suggest that the cognitive and affective processes measured by the APS are stable over time and are important processes in divergent thinking. Although the APS also predicted coping over time, this finding should be interpreted with caution because the coping measure is a new measure

TABLE 47.2 Longitudinal Pearson Correlations of Play Scale Variables at First and Second Grade with Divergent Thinking and Coping Measures at Fifth and Sixth Grade

	Divergent Thinking		Coping
Affect in Play Scale	Fluency	Spontaneous Flexibility	Frequency	Quality
Frequency of Affect	.13	.11	.02	!.03
Variety of Affect	.25	.20	.26	.23
Comfort	.24	.17	.20	.22
Mean Quality	.34*	.25	.34*	.33*
Organization	.27	.16	.34*	.28
Imagination	.42**	!.35*	.42**	.45**

Note. N ” 30; *p # .05, **p # .01.

TABLE 47.3 Longitudinal Pearson Correlations of Affect in Play Scale Variables at First and Second Grade with Similar Variables of the Modified Affect in Fantasy Task at Sixth and Seventh Grade

Affect in Play Scale (First and Second Grade)	Affect in Fantasy Task (Six and Seventh Grade) r
Frequency of Affect	.33*
Positive Affect	.51**
Negative Affect	.21
Variety of Affect	.38*
Comfort	.29
Mean Quality of Fantasy	.27
Organization	.31*
Elaboration	.32*
Imagination	.08
Repetition	.08
Affective Integration	.40*

Note. N ” 30; *p # .05, **p # .01.

and has not been related to other measures of coping behavior. Also, it is a self-report measure not yet related to behavioral measures of coping. It measures how many different things the child can think of to do when real-life problems occur.

A recent study by Russ and Cooperberg (2001) followed these children into high school. We were able to recruit 49 of the original 121 children, who were now in the 11th and 12th grades. We administered in small groups the adult version of the Alternate Uses Test, two self-report coping measure (Ways of Coping-Revised; Adolescent Coping Orientation for Problem Experiences [ACOPE]), and the Beck Depression Inventory. The results were that quality of fantasy in early play was significantly related to high school divergent thinking (r ” .28, p # .05) and problem-focused coping on the ACOPE (r ” .34, p # .01). Variety of affect also related to problem-focused coping (r ” .24, p # .05). These relationships remained relatively unchanged when IQ was partialed out. These results support the stability of these play processes over time in that they relate to theoretically relevant criteria over a 10-year period. They also support the validity of the APS in measuring important processes in play.

One noteworthy finding in this study was that the frequency of negative affect in play significantly related to depression on the Beck Depression Inventory (r ” .31, p # .05). Children who had more negative affect in their early play had more symptoms of depression 10 years later. It is important to note that in this population, the mean on the Beck was 6.98. The relationship between negative affect in play and depression could be indicative of a tendency to feel mildly dysphoric rather than show symptoms of clinical depression. Nevertheless, these results suggest that play processes that can be adaptive in one area (creativity) can be maladaptive in another.

Factor Analysis of the APS

An important theoretical question is whether or not affect in fantasy and the cognitive components in fantasy are separate processes or are one process. The theoretical assumption underlying the play scale was that at least two separate processes are involved—one cognitive and one affective. On the other hand, it is possible that affect and fantasy are so intertwined in play that they cannot be measured separately. In the development of the scale, care was taken to make the scoring criteria of the affect scores separate from the cognitive dimensions. For example, the intensity rating of an aggressive expression should not be influenced by the amount of imagination in the play; the scoring of affective expressions themselves should be independent of the quality of the

fantasy. Also, the scoring of imagination should not be influenced by the amount of affect in the response. Thus, if only one underlying dimension were identified in a factor analysis, it would probably not be due to an artifact of the scoring system.

Factor analyses have been carried out with two separate samples of the APS with different examiners (D’Angelo, 1995; Russ & Peterson, 1990). Both studies had a large enough sample for a solid factor analysis to be carried out. In both studies, two separate factors were found to be the best model. These two factors appear to be a cognitive factor and an affect factor.

Looking at the Russ and Peterson (1990) data set of 121 children, a factor analysis of the total sample was carried out using the principal component analysis with oblique rotation (see Table 47.4). Seven major scores for the APS were included in the factor analysis. (Scores that involved statistical combinations of scores were not included in this particular factor analysis.) An oblique solution using the method default (Cattell & Jaspers, 1967) yielded two separate factors as the best solution. The first and dominant factor appears to be a cognitive factor. Imagination, organization, quality of fantasy, and comfort in play significantly loaded on this first factor. The second factor appears to be an affective factor. Frequency of affective expression, variety of affect categories, and intensity of affect loaded on this second factor. Although separate factors, there is a significant amount of shared variance (r ” .76), suggesting that the factors also overlap.

When factor analyses were carried out separately for girls and boys, similar factor structures emerged. For boys, the factor structure replicated that of the total sample. For girls, the only difference from the total sample was that intensity of affect loaded on the cognitive factor.

The important finding here is that affective expression and cognitive expression in fantasy play, though related, also have significant amounts of unique variance, which suggest that there are separate processes involved. This finding is consistent with the theoretical conceptualization underlying the scale.

TABLE 47.4 Oblique Factor Structure of the Affect in Play Scale for Total Sample

Play Scores	Cognitive	Affective
Frequency of Affect	!.27	.79
Variety of Affect	!.00	.60
Mean Intensity	.12	.40
Comfort	.55	.06
Quality of Fantasy	.62	.04
Organization	.69	!.09
Imagination	.65	!.08

Note. N ” 121.

APS and Emotional Understanding

Emotional understanding is the process by which people make inferences about their own and others’ feelings and behaviors that, in turn, influence their thoughts and actions (Nannis, 1988). Theoretical and empirical evidence suggest that there may be two reasons for a relationship between children’s play and emotional understanding. First, using imagination in play may relate to the cognitive ability to take the perspective of other people. Second, experiencing and expressing different emotions may be central to both fantasy play and emotional understanding.

The relationship between affect and cognitive processes in fantasy play and emotional understanding was examined in children in the first and second grades (Seja & Russ, 1999a). In this study, consistent, yet modest, relations were found between dimensions of fantasy play on the APS and emotional understanding as measured by the Kusche Affective Interview-Revised (Kusche, Greenberg, & Beilke, 1988). Cognitive dimensions of fantasy play, but not affect expression, were related to facets of emotional understanding. The children who were able to access and organize their fantasy and emotions in play were more likely to recall and organize memories related to emotional events and had a more sophisticated understanding of others’ emotions. These relationships remained significant when verbal ability was partialed out. The relationship between fantasy play and understanding others’ emotions supports Harris’s (1989) proposition that imaginative understanding may enable children to understand others’ mental states and affective experiences. A composite fantasy play score accounted for a significant amount of variance in a composite emotional understanding score (5%) when verbal ability was accounted for.

Contrary to initial hypotheses, frequency of affect expression was not related to emotional understanding of oneself and others. The results of this study have important implications for clinical work and suggest that the mere expression of emotion in play is not related to emotional understanding and may not be as useful as play therapists believe. Instead, the integration of affective and cognitive material may be more important in facilitating the development of emotional understanding.

APS and Interpersonal Functioning

Theories of development acknowledge that affect is linked to interpersonal functioning in multiple ways (Emde, 1989; Russ & Niec, 1993; Sroufe, 1989; Strayer, 1987). For example, affective sharing has been related to better quality of infant-parent attachment (Pederson & Moran, 1996; Waters, Wippman, & Sroufe, 1979); regulation of affect has been related to better peer relations and fewer behavior problems (Cole, Zahn-Waxler, Fox, Usher, & Welsh, 1996; Rubin, Coplan, Fox, & Calkins, 1995); openness to affect has been described as providing meaning to interpersonal experience (Sandler & Sandler, 1978) and has also been conceptualized as a key component of empathy (Feshbach, 1987). Given these associations between dimensions of affect and interpersonal functioning, two studies were conducted to investigate the relationship of the APS with children’s interpersonal functioning.

Niec and Russ (1996) investigated relationships among affect and fantasy in play, expression of interpersonal themes in projective stories, and peer and teacher ratings of interpersonal functioning in 49 first through third graders. Children were administered the APS, the Children’s Apperceptive Story Telling Test (CAST), and a brief IQ measure (Schneider, 1989). Teachers and peers rated subjects on their likability, disruptiveness, and withdrawal using the Pupil Evaluation Inventory (PEI; Pekarik, Prinz, Liebert, Weintraub, & Neale, 1976). Results found no relationship between the APS and interpersonal functioning. However, relationships were found between the APS and frequency of interpersonal themes on the CAST. Children who were better players in that they expressed a wide variety of affective categories, frequent positive affect, comfort in their play, and high-quality fantasy were more likely to project themes involving people and relationships in their stories.

In a study by Niec (1998), relationships among affect and fantasy in play, internal representations, and capacity for empathy were investigated. Eighty-six children in third and fourth grades completed the APS, the TAT, and the Bryant Index of Empathy for Children (Bryant, 1982; Murray, 1971). Teachers completed ratings of children’s empathy and helpfulness for each child. Thematic Apperception Test (TAT) stories were scored using Westen’s (1995) Social Cognition and Object Relations Scale (SCORS-Q).

As predicted, quality of fantasy on the APS was related to self-reported empathy. The finding supported the importance of imaginative ability in children’s empathic responding and is consistent with the previously discussed Seja and Russ finding (1999a). Children who were able to “put reality aside and imagine the feelings of someone else in a different (make-believe) situation” were likely to be self-described as more empathic to others (Harris, 1994, p. 19).

Access to affect in play did not relate to empathy, perhaps because the APS measures expression of affect-laden themes rather than the experience of emotion so important in empathic understanding.

Neither access to affect nor fantasy in play related to children’s representations of relationships on the TAT. This finding helped to answer the question posed by Niec (1994) as

to whether access to affect in play would be related to interpersonal representations when content rather than frequency is assessed. While in the Niec and Russ (1996) study, affect and fantasy in play were positively related to frequency of interpersonal themes in projective stories, Niec’s (1998) finding suggests that access to affect may not be related to the qualitative aspects of those representations. It may be that access to affect relates to access to interpersonal representations (i.e., frequency) regardless of the content of those representations (i.e., quality).

The two studies have refined the understanding of the constructs of affect and fantasy as measured by the APS. As expected, access to affect has related to access to interpersonal representations (Niec & Russ, 1996); however, it has not related to peer, teacher, or self-reported measures of interpersonal functioning including such dimensions as empathy, helpfulness, likability, disruptiveness, and withdrawal (Niec, 1998; Niec & Russ, 1996). Quality of fantasy on the APS has been related to both access to interpersonal representations (Niec & Russ, 1996) and to self-reported capacity for empathy (Niec, 1998). These findings and those of previous validity studies suggest that the APS may tap affective dimensions important in mental flexibility (e.g., creativity, role taking, problem solving), rather than the affective constructs important in communication and interpersonal behavior. This understanding is consistent with the theoretical conceptualization of the scale. Further studies that investigate both convergent and discriminant validity of the APS based on this conceptualization will enhance the usefulness of the scale.

In summary, the validity studies suggest that the affective and cognitive processes measured by the APS are predictive of theoretically relevant criteria. The affective processes are related to criteria of creativity, coping, and adjustment. They are not related to measures of emotional understanding, empathy, or interpersonal functioning. The cognitive fantasy processes are related to all criteria. Both cognitive and affective processes are stable over a 5-year period. A very important point is that, in this age group, the APS is independent of IQ. Thus, these processes are resources for children that are independent of intelligence. Finally, the factor analyses results suggest that the APS measures two processes, one cognitive and one affective. Thus, future studies should continue to use both sets of scores.

RANGE OF APPLICABILITY AND LIMITATIONS

The APS has been validated for normal, nonclinical schoolbased populations of children from first through third grade. The scale was developed for children ages 6 to 10. For children younger than 6, the preschool version of the APS should

638 Projective Assessment of Affect in Children’s Play

be used. A next step is to validate the APS with clinical and at-risk populations of children.

One limitation of the APS is that, for accurate coding, the play must be videotaped. This reality may deter clinicians from using the APS.

CROSS-CULTURAL FACTORS

The APS has been used with Caucasian, African American, and Asian children. To date, no differences have been found on the APS scores. In recent studies, an African American version of the puppets has been used with African American children.

The APS has been used in a variety of schools with children from diverse socioeconomic backgrounds. We find remarkable similarity in the means and standard deviations across studies. The mean frequency of affect in a 5-minute play period varies from 11 to 13 units. The mean variety of affect ranges from 2.5 to 3.5 in most studies. The fact that the affect scores do not relate to measures of intelligence suggests that the APS may be a relatively culture-free test.

Research is currently under way investigating the APS in a sample of Italian children. This study is being carried out by Adrienna Lis in Italy. Hopefully, more cross-cultural studies will be carried out.

ACCOMMODATION FOR POPULATIONS WITH DISABILITIES

No studies have been carried out with populations with disabilities. Because nonverbal expressions are coded, the APS could be used with children who have speech difficulties. A series of questions could be developed to help understand the play. However, the child’s scores could not be compared to the mean scores of other child populations.

The APS could be used with developmentally delayed children. If the APS was difficult for them, then the preschool version of the scale could be used.

LEGAL AND ETHICAL CONSIDERATIONS

The APS has not been validated with clinical populations and should not be used for diagnostic purposes at this time.

The main ethical consideration is that, for children who become uncomfortable and do not want to continue the task, the task should be discontinued. In the APS instructions, for children who cannot play at all after 2 minutes, the task is discontinued. If the child becomes uncomfortable or distressed before the 2-minute period, the task should be discontinued.

Another ethical consideration is that the APS should not be scored without a training period. Although criteria are spelled out in the manual, initial scoring samples should be reviewed by someone familiar with the system who has already achieved adequate interrater reliability.

COMPUTERIZATION

Although computerization of the coding system has not been developed, it could be. The narrative component of the play could be transcribed and analyzed by a computer program. Other narrative scoring systems have done this. For example, Pennebaker (2002) has developed a computer program that scores positive and negative emotion words. However, for the APS, the nonverbal expression of emotion is an important component of the scale. A number of affect expressions would not be included if only the words were analyzed. Also, affect tone influences the scoring of verbal expressions. Nevertheless, it would be interesting to investigate the validity of the verbal component of the scale.

CURRENT RESEARCH STUDIES

Current research studies are investigating the use of the scale with at-risk populations, the validity of a preschool version of the scale, the development of alternative scoring systems, and use of the APS as an outcome measure.

Kaugars, Russ, and Singer (2000) have investigated the APS and emotional expression in a cocaine-exposed population of 4-year-old children. Russ and Singer are continuing with a follow-up of these children at 6 years of age. These studies are being carried out in the research laboratory of Lynn Singer. In the study with 4-year-olds, the APS did not distinguish between the cocaine-exposed group and a matched control group. In the current study, we are investigating the validity of the APS with this at-risk population.

For the study with the 4-year-olds, we needed to develop a preschool version of the APS (APS-P). One of my graduate students, Astrida Seja Kaugars, took the lead in the development of the preschool version. Based on the understanding that puppets might be more difficult for young children to manipulate, we selected toys that would be easy to play with and that could elicit symbolic and fantasy play. The instructions for the preschool version are a little more structured than for the APS. Scoring is the same except that affect is scored as being present or absent in 10-second intervals.

We have completed one validity study with the APS-P. We looked at the relationship between play, creativity, social competence, and teacher ratings of play in 33 nursery school children from 4 to 5 years of age. Creativity was measured with the Multidimensional Stimulus Fluency Measure (Goodwin & Moran, 1990). Interrater reliability was good and correlations ranged from .82 to .97 and internal consistency (2nd and 4th minutes and 3rd and 5th minutes) for frequency of affect was good, r ” .88. The affect scores, frequency and variety, were significantly related to creativity and originality for the creativity measure. Comfort in play also was significantly related to creativity. Interestingly, in this study, the fantasy scores were not related to creativity. All play scores were significantly related to teachers’ ratings of daily play behavior. Finally, good players were rated by teachers as functioning well with little adult supervision in the classroom. The results of this study suggest that the APS-P is measuring constructs that relate to important criteria in this preschool age group. Theoretically, the results suggest that affective dimensions of play are related to important functions in child development.

The standardized play task could be used with other scoring systems. One of my students, Larissa Niec, has developed a coding system for the APS that measures interpersonal representations (Niec, Yopp, & Wells, 2001). The Interpersonal Themes in Play System (ITPS) has shown promising results. In one study the ITPS scores were related to the Westen SCORS-Q measure on the TAT and to various empathy measures. Especially promising is the Affect Tone in Play score, which measures the degree to which the play narrative reflects a safe, supportive, interpersonal world. Affect Tone in Play is significantly related to internal representations on the TAT and to several measures of empathy. These findings suggest that interpersonal schema can be assessed in play narratives. Future research in this area will be important.

Finally, I am investigating a variety of play interventions to improve play skills in children. The APS will be used as an outcome measure. This brings us to the next section in this chapter.

USE IN CLINICAL PRACTICE

One of the most appropriate uses of the APS is as a measure of outcome in child psychotherapy. The APS measures processes that clinicians are concerned about: constriction of affect, organization of thoughts, and integration or modulation of affect. Play behavior should change as psychotherapy progresses. The APS should be sensitive to these changing processes.

The APS has not been used for diagnostic purposes nor has research been carried out in this area. The APS could be valuable tool in determining whether a child could use play in therapy. The APS could be used to determine how a child compares to school-based populations in terms of affect expression and imagination in play. One limitation is that, for accurate scoring, a video camera should be used.

FUTURE DIRECTIONS

Next steps for the development of the APS are:

• 1. Determine validity with clinical populations and high-risk populations.
• 2. Continue to refine the scoring system. Of special importance is the addition of an emotion regulation score. Although the combination of the affect and fantasy scores attempt to measure integration of affect, a score that is coded separately may be more valid.
• 3. Use the APS as a measure of change in child therapy intervention studies.
• 4. Continue longitudinal studies with the APS.
• 5. Develop validity of the preschool version of the scale.
• 6. Develop a version of the APS that can be used without a video camera.

The growing body of validity studies to date suggests that the APS measures processes that are important in child development, predict adaptive functioning in children, and are separate from what traditional intelligence tests measure. The use of the APS in a variety of research programs and clinical settings with a variety of child populations will further the development of the measure. Research with the APS will also tell us about this important resource for children’s affect expression in fantasy play.

REFERENCES

• Anastasia, A. (1988). Psychological testing (6th ed.). New York: Macmillan.
• Bryant, B. (1982). An index of empathy for children and adolescents. Child Development, 53, 413–425.
• Carsen, D., Bittner, M., Cameron, B., Brown, D., & Meyer, S. (1994). Creative thinking as a predictor of school-aged children’s stress responses and coping abilities. Creativity Research Journal, 7, 145–158.
• Cattell, R.B. & Jaspers, J. (1967). A general plasmode for factor analytic exercises and research. Multivariate Behavior Research Monographs, 67(3), No. 30-10-5-2.
• Christiano, B. & Russ, S. (1996). Play as a predictor of coping and distress in children during an invasive dental procedure. Journal of Clinical Child Psychology, 25, 130–138.
• Christie, J., & Johnson, E. (1983). The role of play in socialintellectual development. Review of Educational Research, 53, 93–115.
• Clark, P., Griffing, P., & Johnson, L. (1989). Symbolic play and ideational fluency as aspects of the evolving divergent cognitive style in young children. Early Child Development and Care, 51, 77–88.
• Cole, P., Zahn-Waxler, C., Fox, N., Usher, B., & Welsh, J. (1996). Individual differences in emotion regulation and behavior problems in preschool children. Journal of Abnormal Psychology, 105, 518–529.
• D’Angelo, L. (1995). Child’s play: The relationship between the use of play and adjustment styles. Unpublished doctoral dissertation, Case Western Reserve University, Cleveland, OH.
• Dansky, J. (1980). Make-believe: A mediator of the relationship between play and associative fluency. Child Development, 51, 576–579.
• Dansky, J., & Silverman, F. (1973). Effects of play on associative fluency in preschool-aged children. Developmental Psychology, 9, 38–43.
• Edelbrock, C., & Achenbach, T.M. (1980). A typology of child behavior profile patterns: Distribution and correlates for disturbed children aged 6–16. Journal of Abnormal Psychology, 8, 441– 470.
• Emde, R. (1989). The infant’s relationship experience: Developmental and affective aspects. In A. Sameroff & R. Emde (Eds), Relationship disturbances in early childhood (pp. 33–51). New York: Basic Books.
• Erickson, E. (1963). Childhood and society. New York: Norton.
• Fein, G. (1987). Pretend play: Creativity and consciousness. In P. Gorlitz & J. Wohlwill (Eds.), Curiosity, imagination, and play (pp. 281–304). Hillsdale, NJ: Erlbaum.
• Feitelson, D., & Ross, G. (1973). The neglected factor—play. Human Development, 16, 202–223.
• Feshbach, N.D. (1987). Parental empathy and child adjustment/maladjustment. In N. Eisenberg & J. Strayer (Eds.), Empathy and its development (pp. 271–291). New York: Cambridge University Press.
• Freud, A. (1965). Normality and pathology in childhood: Assessment of development. New York: International Universities Press.
• Goldstein, A. (2002). The effect of affect-laden reading passages on children’s emotional expressivity in play. Unpublished doctoral dissertation, Case Western Reserve University, Cleveland, OH.
• Goldstein, A., & Russ, S. (2000–2001). Understanding children’s literature and its relationship to fantasy ability and coping. Imagination, Cognition, and Personality, 20, 105–126.
• Goodwin, L.S., & Moran, J.D. (1990). Psychometric characteristics of an instrument for measuring creative potential in preschool children. Psychology in the Schools, 27, 204–210.
• Grossman-McKee, A. (1989). The relationship between affective expression in fantasy play and pain complaints in first and second grade children. Dissertation Abstracts International, 50, 4219B.
• Guilford, J.P. (1968). Intelligence, creativity and their educational implications. San Diego, CA: Knapp.
• Harris, P.L. (1989). Children and emotion: The development of psychological understanding. Cambridge, MA: Blackwell Publishers.
• Harris, P. (1994). The child’s understanding of emotion: Developmental change and the family environment. Journal of Child Psychology and Psychiatry, 35, 3–28.
• Harter, P. (1994). Manual for the Self-Perception Profile for Children. Denver, CO: University of Denver.
• Holt, R.R. (1977). A method for assessing primary process manifestations and their control in Rorschach responses. In M. Rickers-Ovsiankina (Ed.), Rorschach psychology (pp. 375–420). New York: Krieger.
• Howe, P., & Silvern, L. (1981). Behavioral observation during play therapy: Preliminary development of a research instrument. Journal of Personality Assessment, 45, 168–182.
• Hughes, M. (1987). The relationship between symbolic and manipulative (object) play. In D. Gorlitz & J. Wohwill (Eds.), Curiosity, imagination, and play (pp. 247–257). Hillsdale, NJ: Erlbaum.
• Isen, A., Daubman, K., & Nowicki G. (1987). Positive affect facilitates creative problem solving. Journal of Personality and Social Psychology, 52, 1122–1131.
• Izard, E. (1977). Human emotions. New York: Plenum.
• Johnson, J. (1976). Relations of divergent thinking and intelligence test scores with social and nonsocial make-believe play of preschool children. Child Development, 47, 1200–1203.
• Kaugars, A., & Russ, S. (2000, March). Validity of the Affect In Play Scale–Preschool Version. Paper presented at the midwinter meeting of The Society for Personality Assessment, Albuquerque, NM.
• Kaugars, A., Russ, S., & Singer, L (2000, March). Behavioral and emotion regulation among children prenatally exposed to cocaine. Poster presented at Millenium Conference of Great Lakes Society of Pediatric Psychology, Cleveland, OH.
• Kusche, C.A., Greenberg, M.T., & Beilke, B. (1988). The Kusche Affective Interview. Unpublished manuscript, University of Washington, Seattle.
• Lieberman, J.N. (1977). Playfulness: Its relationship to imagination and creativity. New York: Academic Press.
• Morrison, D. (1988). The child’s first ways of knowing. In D. Morrison (Ed.), Organizing early experience: Imagination and cognition in childhood (pp. 3–14). Amityville, NY: Baywood.
• Murray, H.A. (1971). Thematic Apperception Test manual. Cambridge, MA: Harvard University Press.
• Nannis, E.D. (1988). Cognitive-developmental differences in emotional understanding. In E.D. Nannis & P.A. Cowan (Eds.), Developmental psychopathology and its treatment. New Directions for Child Development (No. 39, pp. 31–49). San Francisco: Jossey-Bass.
• Niec, L.N. (1994). Relationships among affect and interpersonal themes in children’s fantasy and interpersonal functioning. Unpublished master’s thesis, Case Western Reserve University, Cleveland, OH.
• Niec, L.N. (1998). Relationships among internal representations, affect in play, and interpersonal functioning. Unpublished doctoral dissertation, Case Western Reserve University, Cleveland, OH.
• Niec, L.N., & Russ, S.W. (1996). Relationships among affect in play, interpersonal themes in fantasy, and children’s interpersonal behavior. Journal of Personality Assessment, 66, 645–649.
• Niec, L.N., Yopp, J., & Russ, S. (2002). Children’s interpersonal themes in play and interpersonal functioning: Development of the Interpersonal Themes in Play System. Manuscript submitted for publication.
• Pederson, D., & Moran, G. (1996). Expressions of the attachment relationship outside of the Strange Situation. Child Development, 67, 915–927.
• Pekarik, E.R., Prinz, R., Liebert, D., Weintraub, S., & Neale, J. (1976). The Pupil Evaluation Inventory: A sociometric technique for assessing children’s social behavior. Journal of Abnormal Child Psychology, 4, 83–97.
• Pelligrini, A. (1992). Rough and tumble play and social problem solving flexibility. Creativity Research Journal, 5, 13–26.
• Pennebaker, J.W. (2002). What our words can say about us: Toward a broader language psychology. APA Monitor, January/February, pp. 8–9.
• Peppler, D., & Ross, H. (1981). The effects of play on convergent and divergent problem solving. Child Development, 52, 1202– 1210.
• Perry, D., & Russ, S. (1998, August). Coping strategies and adjustment in homeless children. Paper presented at the meeting of the American Psychological Association, San Francisco, CA.
• Peterson, N. (1989). The relationship between affective expression in fantasy play and self-esteem in third grade children. Unpublished master’s thesis, Case Western Reserve University, Cleveland, OH.
• Rubin, K., Coplan, R., Fox, N., & Calkins, S. (1995). Emotionality, emotion regulation, and preschoolers’ social adaptation. Development and Psychopathology, 7, 49–62.
• Rubin, K., Fein, G., & Vandenberg, B. (1983). Play. In P. Mussen (Ed.) Handbook of child psychology, Vol. 4 (pp. 693–774). New York: Wiley.
• Runco, M.A. (1991). Divergent thinking. Norwood, NJ: Ablex.
• Russ, S. (1987). Assessment of cognitive affective interaction in children: Creativity, fantasy, and play research. In J. Butcher &

C. Spielberger (Eds.), Advances in personality assessment, Vol. 6 (pp. 141–155). Hillsdale, NJ: Erlbaum.

• Russ, S. (1988). Primary process thinking on the Rorschach, divergent thinking, and coping in children. Journal of Personality Assessment, 52, 539–548.
• Russ, S. (1993). Affect and creativity: The role of affect and play in the creative process. Hillsdale, NJ: Erlbaum.
• Russ, S. (1996). Psychoanalytic theory and creativity: Cognition and affect revisited. In J. Masling & R. Borstein (Eds.), Psychoanalytic perspectives on developmental psychology (pp. 69– 103). Washington DC: APA Books.
• Russ, S., & Cooperberg, M. (2001). Longitudinal prediction of creativity, coping, and depression from pretend play. Manuscript in preparation.
• Russ, S., & Grossman-McKee, A. (1990). Affective expression in children’s fantasy play, primary process thinking on the Rorschach, and divergent thinking. Journal of Personality Assessment, 54, 756–771.
• Russ, S., Grossman-McKee, A., & Rutkin, Z. (1984). [Affect in Play Scale: Pilot project]. Unpublished raw data.
• Russ, S., & Kaugars, A. (2000–2001). Emotion in children’s play and creative problem solving. Creativity Research Journal, 13, 211–219.
• Russ, S., & Niec, L.N. (1993, April). Affective development and object relations: How much do we know? Paper presented at the meeting of the Society for Personality Assessment, Chicago, IL.
• Russ, S., & Peterson, N. (1990). The Affect in Play Scale: Predicting creativity and coping in children. Unpublished manuscript.
• Russ, S., Robins, D., & Christiano, A.B. (1999). Pretend play: Longitudinal prediction of creativity and affect in fantasy in children. Creativity Research Journal, 12, 129–139.
• Russ, S., & Schafer, E. (2001). Affect in play, emotional memories, and divergent thinking. Manuscript in preparation.
• Sandler, J., & Sandler, A.M. (1978). On the development of object relationships and affects. International Journal of Psycho-Analysis, 59, 285–296.
• Sarnoff, C. (1976). Latency. New York: Jason Aronson.
• Schneider, M. (1989). Children’s Apperceptive Story-Telling Test manual. Austin, TX: Pro-ed.
• Seja, A., & Russ, S. (1998, May). Children’s fantasy play, emotional understanding and parents’ reports of children’s daily behavior. Poster session presented at the Great Lakes Regional Conference on Child Health Psychology, Louisville, KY.
• Seja, A., & Russ, S. (1999a). Children’s fantasy play and emotional understanding. Journal of Clinical Child Psychology, 28, 269– 277.
• Seja, A., & Russ, S. (1999b, April). Development of the preschool Affect in Play Scale. Paper presented at a meeting of the Society for Research in Child Development, Albuquerque, NM.

642 Projective Assessment of Affect in Children’s Play

• Singer, D.L., & Rummo, J. (1973). Ideational creativity and behavioral style in kindergarten age children. Developmental Psychology, 8, 154–161.
• Singer, D.L., & Singer, J.L. (1990). The house of make-believe: Children’s play and the developing imagination. Cambridge, MA: Harvard University Press.
• Singer, J.L. (1973). The child’s world of make-believe. New York: Academic Press.
• Singer, J.L., & Singer, D.G. (1981). Television, imagination, and aggression. Hillsdale, NJ: Erlbaum.
• Smith, P.K., & Whitney, S. (1987). Play and associative fluency: Experimenter effects may be responsible for positive results. Developmental Psychology, 23, 49–53.
• Spielberger, C.D. (1973). State-Trait Anxiety Inventory for Children. Palo Alto, CA: Consulting Psychological Press.
• Sroufe, L. (1989). Relationships, self, and individual adaptation. In A. Sameroff & R. Emde (Eds.), Relationship disturbances in early childhood (pp. 70–96). New York: Basic Books.
• Strayer, J. (1987). Affective and cognitive perspectives on empathy. In N. Eisenberg & J. Strayer (Eds.), Empathy and its development (pp. 218–244). New York: Cambridge University Press.
• Suler, J. (1980). Primary process thinking and creativity. Psychological Bulletin, 88, 144–165.
• Tomkins, S.S. (1962). Affect, imagery, consciousness: Volume 1. The positive affects. New York: Springer.
• Tomkins, S.S. (1963). Affect, imagery, consciousness: Volume 2. The negative affects. New York: Springer.
• Vandenberg, B. (1980). Play, problem-solving, and creativity. New Directions for Child Development, 9, 49–68.
• Waelder, R. (1933). Psychoanalytic theory of play. Psychoanalytic Quarterly, 2, 208–224.
• Waters, E., Wippman, J., & Sroufe, L. (1979). Attachment, positive affect, and competence in the peer group: Two studies in construct validation. Child Development, 50, 821–829.
• Weiner, I. (1977). Approaches to Rorschach validation. In M. Rickers-Ovsiankina (Ed.), Rorschach psychology (pp. 575–608). New York: Krieger.
• Weiner, I. (1998). Principles of Rorschach interpretation. Mahwah, NJ: Erlbaum.
• Westen, D. (1995). Social Cognition and Object Relations Scale: Q-Sort for projective stories (SCORS-Q). Unpublished manual, Cambridge Hospital and Harvard Medical School, Cambridge, MA.