This article was downloaded by: [Nanyang Technological University] On: 28 April 2015, At: 15:19 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of Personality Assessment Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/hjpa20
Early Memories and the Criminally Dangerous Linda Hale Tobey & Arnold R. Bruhn Published online: 10 Jun 2010.
To cite this article: Linda Hale Tobey & Arnold R. Bruhn (1992) Early Memories and the Criminally Dangerous, Journal of Personality Assessment, 59:1, 137-152, DOI: 10.1207/s15327752jpa5901_11 To link to this article: http://dx.doi.org/10.1207/s15327752jpa5901_11
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions
JOURNAL OF PERSONALITY ASSESSMENT, 1992, 59(1), 137-152 Copyright o 1992, Lawrence Erlbaum Associates, Inc.
Early Memories and the Criminally Dangerous Linda Hale Tobey Washington, DC Downloaded by [Nanyang Technological University] at 15:19 28 April 2015
Arnold R. Bruhn Bethesda, MD
We demonstrated the efficacy of using early memory (EM) data to distinguish postdictively criminally dangerous from nondangerous psychiatric patients. Dangerous patients (n = 30) had been adjudicated not criminally responsible (NCR) for felonies, whereas nondangerous ~ s ~ c h i a t rpatients ic (n = 30) had no convictions for aggressive offenses. Groups were comparable in age, IQ, socioeconomic status (SES), marital status, and diagnoses. Group racial compositions, however, were significantly different. Statistical analyses were thus performed to clarify the potential effect of this racial confound. Overall, 73% of the dangerous and nondangerous patients were accurately classified via a newly developed EM scoring system, the Early Memory Aggressiveness Potential Score System (EMAPSS). Of those classified as dangerous, 15 of 16 (94%) were actually dangerous. The false-positive rate was an impressively low 6%, suggesting EM aggression is highly predictive of dangerousness postdictively.
It is commonly known that the re diction of rare events presents special challenges to researchers. Aggressive acting out is a n example of just such a rare event. Clinicians have attempted t o develop more accurate prediction instruments over the past 20 years, yet the prediction of dangerousness continues to be little better than chance (Monahan, 1981; Steadman, 1980). I n response to the dearth of successful results with standard assessmeint techniques, many researchers have developed actuarial equations as a n alternative method of predicting dangerousness (Klassen & O'Connor, 1988a, 1988b; Norton, 1988; Selby, 1984; Wenk, Robinson, & Smith, 1972). Although these actuarial procedures have provided a higher degree of accuracy than most clinical studies (Monahan, 1984), substantial false-positive and false-negative rates appear t o be the rule (Klassen & O'Connor, 1988b), and impressive actuarial results have not been replicated t o date.
Downloaded by [Nanyang Technological University] at 15:19 28 April 2015
138
TOBEY AND BRUHN
Given the preceding limitations, a need exists for assessment techniques that can accurately identify individuals at risk for dangerous acting out. We addressed this issue using the Early Memories Procedure (EMP)' (Bruhn, 1989) to identify psychological variables that underlie and appear to maintain dangerous behavior in a psychiatric population. The E M . has several advantages over other projective assessment procedures. Early memories (EMS) may be less threatening than the Rorschach and the Thematic Apperception Test (TAT; Bruhn, 1984, 1985, 1990), and they provide rich information at a cost of relatively little time (Mosak, 1958; Verger &Camp, 1970). Because EMS are not evoked by artificial stimuli, they constitute a " 'purer' projective method than . . . the more conventional approaches, such as the Rorschach and the TAT" (Bruhn, 1984, p. 112). Previous investigations have demonstrated that psychotic patients can be discriminated from less disturbed psychiatric patients on the basis of their EMS (Charry, 1959; Friedman, 1952; Friedman & Schiffman, 1962; Furlan, 1984; Hafner, Corotto, & Fakouri, 1980; Haher & Fakouri, 1978; Hafner, Fakouri, Ollendick, & Corotto, 1979; Langs, Rothenberg, Fishman, & Reiser, 1960; Plutchik, Platman, 6r Fieve, 1970). However, only two known studies have used EMS within a criminal population (Hankoff, 1987; Quinn, 1973). Quinn (1973) proposed EMS could be used to discriminate (a) prison recidivists from nonrecidivists and (b) inmates who committed crimes against people from inmates who committed property crimes. EM data, however, did not differentiate the groups. Quinn's scoring system may have been insufficiently sensitive to detect group differences. Alternatively, imprisonment may have altered the inmates' frame of reference, obliterating otherwise detectable differences that existed when the crimes were perpetrated. Hankoff (1987) compared the EMS of 32 incarcerated male criminals and 50 noncriminal controls screened for a history of delinquency and antisocial behavior. Criminal subjects' EMS were more dramatic and unpleasant than the noncriminal EMS. Criminals also ~ o r t r a ~ events ed involving more disturbed or aggressive interactions with others. No known studies to date have used EMS with ~ s ~ c h i a t rpatients ic who are also dangerous. The research described next was designed to identify EM variables that can postdictively distinguish dangerous psychiatric patients from nondangerous controls, thereby establishing the basis of an assessment procedure that might identify psychiatric patients at risk for engaging in dangerous behavior. 'The EMP is the only projective instrument available to assess the whole of lifetime autobiographical memory versus one or several EMS. The protocol in this study required only the first four memories from the EMP and the rating scales. Due to the populations studied, we adapted the EMP t o an oral format as the writing requirement would have made the task impossible for a chronic, psychotically inclined population. A copy ofthe EMP can be obtained from Arnold R. Bruhn, 4704 Hunt Avenue, Chevy Chase, MD 20815.
EARLY MEMORIES
139
METHOD
Downloaded by [Nanyang Technological University] at 15:19 28 April 2015
Subjects The dangerous group consisted of 30 male psychiatric patients, ranging in age from 21 to 48 years old (M = 33.5), who were confined to a maximum security forensic hospital. All subjects were classified as a danger to themselves and/or others. One subject had been transferred from a regional hospital due to uncontrollable aggressive behavior. The remaining 29 subjects had been adjudicated not criminally responsible (NCR) or not guilty by reason of insanity (NGBRI) for violent criminal offenses, including murder, rape, arson, assault, battery, and malicious destruction of property. The nondangerous group consisted of 30 males, ranging in age from 19 to 67 (M = 37.9), who were hospitalized at a state psychiatric hospital. Nondangerous subjects had never been adjudicated NCR/NGBRI or been convicted of a felony. Both groups were dominated by chronic psychiatric patients who had beien currently hospitalized for an average of 4 years and previously hospitalized an average of 5 to 6 times. There were no significant group differences in the distribution of diagnoses.' Almost all subjects (n = 59) were coded at the lowest SES levels (Level 4 and 5) on the Hollingshead (1957) Socioeconomic Scale. The remaining nondangerous subject was coded as middle class (Level 3). Gross extrapolation from the Wechsler Adult Intelligence Scale-Revised (WAIS-R) Vocabulary subtest suggested dangerous and nondangerous patients' mean verbal IQs wcre 81 and 84, respectively. Most subjects were not married (n = 56) and had only completed 2 years of high school (M = 10.5). In sum, both groups were composed of long-term psychiatric patients with psychotic and organic disorders. There were no group differences in,age, SES, intelligence, or marital status-all of which have been associated with EM differences(Davidow & Bruhn, 1990). There were no differences in educatiorial level or diagnosis. Significant differences were found, however, in group racial compositions. The dangerous group was composed of 17 Black patients and 13 White patients. In comparison, 6 nondangerous patients were Black and 24 were White. Given this significant difference, subsequent statistical analyses were performed to assess the potential effects of this racial confound. 'Both groups contained a large percentage of highly disturbed psychotic patients. In order to find
30 subjects in each group who were sufficiently integrated to cooperate with the research protocol, the first author was forced to screen a large pool of potential subjects. In addition, many of the dangerous subjects were too suspicious to volunteer their EMS and had to be eliminated. The subjects used in the study should be viewed as better integrated, less suspicious, and much less floridly psychotic than a randomly selected group of chronic patients.
140
TOBEY AND BRUHN
Downloaded by [Nanyang Technological University] at 15:19 28 April 2015
Measures During a 30-min semi-structured interview, four EMS were elicited from each subject using the format outlined in the EMP (Bruhn, 1989). The WAIS-R Vocabulary subtest was also administered. All interviews were audiotaped, with the exception of 2 subjects who refused to be taped and whose interviews, therefore, were transcribed by hand. The senior author interviewed all subjects. EMS were scored with the Early Memory Aggressiveness Potential Score System (EMAPSS; Bruhn & Tobey, 1991),~adapted from a pilot study of aggressive and nonaggressive criminals' EMS (Bruhn & Feigenbaum, 1987). Memories were also scored using the Comprehensive Early Memory Score System-Revised (CEMSS-R; Last & Bruhn, 1990; see footnote 3). The CEMSS-R is an updated version of the CEMSS (Last & Bruhn, 1983), a comprehensive, general purpose EM scoring system that assesses pathological and adaptive EM characteristics. The rationale for using two EM scoring systems was to test a hypothesis ~roferredby Bruhn and Schiffman (1982), who argued that carefully crafted, specific-use, pilot study tested EM scoring systems (e.g., the EMAPSS) should outperform general purpose EM scoring systems (e.g., the CEMSS-R).
Hypotheses Two hypotheses were tested: 1. Dangerous psychiatric patients will differ from nondangerous psychiatric patients by displaying more aggressiveness potential, reflected in higher scores on the EMAPSS. 2. Dangerous and nondangerous psychiatric patients will be more accurately classified using the EMAPSS than the CEMSS-R.
RESULTS Reliability Because raters demonstrated universal agreement in identifying the absence of aggression, reliability was defined as rater agreement when aggression was 3The EMAPSS is a specialized EM scoring system that attempts to assess aggressiveness potential. The CEMSS-R is a general purpose EM scoring system, analogous to such general purpose Rorschach scoring systems as Exner's, Klopfer's, and Beck's. Inquiries about the EMAPSS or CEMSS-R, which are both copyrighted, should be directed to Arnold R. Bmhn, 4704 Hunt Avenue, Chevy Chase, MD 20815.
Downloaded by [Nanyang Technological University] at 15:19 28 April 2015
present, a much more conservative standard for reliability. EMS with no aggression were excluded from reliability computations. Three naive judges reached 85.5% interrater agreement on the 12 EMAPSS variables; two out of three judges agreed 95.8% of the time. Several studies have demonstrated that the CEMSS is reliable (Acklin, Sauer, Alexander, & Dugoni, 1989; Last & Bruhn, 1983,1985), valid in differentiating diagnostic groups (Last & Bruhn, 1985), and capable of identifying the degree of psychopathology in a clinical population (Acklin et al., 1989; Last 6r Brutm, 1983). The CEMSS has not been previously used with aggressive psychiatric patients, and the reliability of the modified CEMSS system has not been assessed. CEMSS-R interrater reliability was therefore computed for this study. Given the large number of CEMSS-R variables (65 variables per EM), eight naive raters, divided into four pairs, scored the CEMSS-R. Categorical variables were considered reliable if at least one of the two raters' scores was identical with the ratings of the senior author. Continuous variables were considered 1). This resulted in an overall reliable if ratings fell within 1 point (i.e., CEMSS-R reliability of 94.5%, with individual items' interrater agreemeints ranging from 100% to 81.1%.
Reduction of EMAPSS and CEMSS-R Variables A preliminary analysis was conducted to eliminate EMAPSS and CEMSS-R variables with minimal discriminatory potential. Categorical variables endorsed by less than 10% of the subjects were dropped. Categorical and continuous variables that were not significantly different between the two groups at a univariate level of alpha equal to .05 were also eliminated. When these t.wo criteria were used, 7 EMAPSS (Table 1) and 13 CEMSS-R (Table 2) predictor variables were retained. TABLE 1 EMAPSS Predictor Variables Variable Physical abuse Someone is physically threatened* Psychological abuse** Patient is psychologically abused* Summary scores Number of aggressive incidents** Number of aggressive memories* Mean severity ratings* Maximum severity ratings*
*p
< .05. **p < .01.
142
TOBEY AND BRUHN TABLE 2 CEMSS-R Predictor Variables
Downloaded by [Nanyang Technological University] at 15:19 28 April 2015
Variable Setting Home* Sensory-motor Visual* Distal sensory-motor* Total sensory-motor* Content Trust* Affect Anxiety* Subject affect rating* Rater affect rating** Object relations Perception of sel!? Perception of environment* Individual distinctiveness* Level of object relations* Number of words**
Main Analyses
Hypothesis I. Dangerous psychiatric patients will differ from nondangerous psychiatric patients by displaying more aggressiveness potential, reflected in higher EMAPSS scores. This hypothesis was strongly supported. Significantly more dangerous patients recalled aggressive EMS (77%) than nondangerous patients (43%), XZ(l,N = 60) = 5.62, p < .05. More dangerous than nondangerous patients recalled physical abuse (60% vs. 42%), animal abuse (17% vs. 3%), and psychological abuse (37% vs. 3%), although psychological abuse was the only category where these differences were significant, x2(1,N = 60) = 8.44, g < .01. Dangerous patients recalled a greater number of aggressive EMS, ~ ' ( 2 ,N = 60) = 8.98, p < .01. Dangerous patients also reported significantly more aggressive incidents within their four EMS than the nondangerous patients, X2(2,N = 60) = 9.50, p < .01, indicating dangerous patients are more likely to recall EMS that include multiple aggressive incidents. Whenever aggression appeared in an EM, the seriousness of the aggression was rated on a 3-point scale ranging from minimal (1)to severe (3). As a group, the dangerous patients' mean severity rating (M = .09) was significantly higher than the nondangerous patient's mean severity rating (M = .04), t(51) = 2.30, P
Journal of Personality Assessment Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/hjpa20
Early Memories and the Criminally Dangerous Linda Hale Tobey & Arnold R. Bruhn Published online: 10 Jun 2010.
To cite this article: Linda Hale Tobey & Arnold R. Bruhn (1992) Early Memories and the Criminally Dangerous, Journal of Personality Assessment, 59:1, 137-152, DOI: 10.1207/s15327752jpa5901_11 To link to this article: http://dx.doi.org/10.1207/s15327752jpa5901_11
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions
JOURNAL OF PERSONALITY ASSESSMENT, 1992, 59(1), 137-152 Copyright o 1992, Lawrence Erlbaum Associates, Inc.
Early Memories and the Criminally Dangerous Linda Hale Tobey Washington, DC Downloaded by [Nanyang Technological University] at 15:19 28 April 2015
Arnold R. Bruhn Bethesda, MD
We demonstrated the efficacy of using early memory (EM) data to distinguish postdictively criminally dangerous from nondangerous psychiatric patients. Dangerous patients (n = 30) had been adjudicated not criminally responsible (NCR) for felonies, whereas nondangerous ~ s ~ c h i a t rpatients ic (n = 30) had no convictions for aggressive offenses. Groups were comparable in age, IQ, socioeconomic status (SES), marital status, and diagnoses. Group racial compositions, however, were significantly different. Statistical analyses were thus performed to clarify the potential effect of this racial confound. Overall, 73% of the dangerous and nondangerous patients were accurately classified via a newly developed EM scoring system, the Early Memory Aggressiveness Potential Score System (EMAPSS). Of those classified as dangerous, 15 of 16 (94%) were actually dangerous. The false-positive rate was an impressively low 6%, suggesting EM aggression is highly predictive of dangerousness postdictively.
It is commonly known that the re diction of rare events presents special challenges to researchers. Aggressive acting out is a n example of just such a rare event. Clinicians have attempted t o develop more accurate prediction instruments over the past 20 years, yet the prediction of dangerousness continues to be little better than chance (Monahan, 1981; Steadman, 1980). I n response to the dearth of successful results with standard assessmeint techniques, many researchers have developed actuarial equations as a n alternative method of predicting dangerousness (Klassen & O'Connor, 1988a, 1988b; Norton, 1988; Selby, 1984; Wenk, Robinson, & Smith, 1972). Although these actuarial procedures have provided a higher degree of accuracy than most clinical studies (Monahan, 1984), substantial false-positive and false-negative rates appear t o be the rule (Klassen & O'Connor, 1988b), and impressive actuarial results have not been replicated t o date.
Downloaded by [Nanyang Technological University] at 15:19 28 April 2015
138
TOBEY AND BRUHN
Given the preceding limitations, a need exists for assessment techniques that can accurately identify individuals at risk for dangerous acting out. We addressed this issue using the Early Memories Procedure (EMP)' (Bruhn, 1989) to identify psychological variables that underlie and appear to maintain dangerous behavior in a psychiatric population. The E M . has several advantages over other projective assessment procedures. Early memories (EMS) may be less threatening than the Rorschach and the Thematic Apperception Test (TAT; Bruhn, 1984, 1985, 1990), and they provide rich information at a cost of relatively little time (Mosak, 1958; Verger &Camp, 1970). Because EMS are not evoked by artificial stimuli, they constitute a " 'purer' projective method than . . . the more conventional approaches, such as the Rorschach and the TAT" (Bruhn, 1984, p. 112). Previous investigations have demonstrated that psychotic patients can be discriminated from less disturbed psychiatric patients on the basis of their EMS (Charry, 1959; Friedman, 1952; Friedman & Schiffman, 1962; Furlan, 1984; Hafner, Corotto, & Fakouri, 1980; Haher & Fakouri, 1978; Hafner, Fakouri, Ollendick, & Corotto, 1979; Langs, Rothenberg, Fishman, & Reiser, 1960; Plutchik, Platman, 6r Fieve, 1970). However, only two known studies have used EMS within a criminal population (Hankoff, 1987; Quinn, 1973). Quinn (1973) proposed EMS could be used to discriminate (a) prison recidivists from nonrecidivists and (b) inmates who committed crimes against people from inmates who committed property crimes. EM data, however, did not differentiate the groups. Quinn's scoring system may have been insufficiently sensitive to detect group differences. Alternatively, imprisonment may have altered the inmates' frame of reference, obliterating otherwise detectable differences that existed when the crimes were perpetrated. Hankoff (1987) compared the EMS of 32 incarcerated male criminals and 50 noncriminal controls screened for a history of delinquency and antisocial behavior. Criminal subjects' EMS were more dramatic and unpleasant than the noncriminal EMS. Criminals also ~ o r t r a ~ events ed involving more disturbed or aggressive interactions with others. No known studies to date have used EMS with ~ s ~ c h i a t rpatients ic who are also dangerous. The research described next was designed to identify EM variables that can postdictively distinguish dangerous psychiatric patients from nondangerous controls, thereby establishing the basis of an assessment procedure that might identify psychiatric patients at risk for engaging in dangerous behavior. 'The EMP is the only projective instrument available to assess the whole of lifetime autobiographical memory versus one or several EMS. The protocol in this study required only the first four memories from the EMP and the rating scales. Due to the populations studied, we adapted the EMP t o an oral format as the writing requirement would have made the task impossible for a chronic, psychotically inclined population. A copy ofthe EMP can be obtained from Arnold R. Bruhn, 4704 Hunt Avenue, Chevy Chase, MD 20815.
EARLY MEMORIES
139
METHOD
Downloaded by [Nanyang Technological University] at 15:19 28 April 2015
Subjects The dangerous group consisted of 30 male psychiatric patients, ranging in age from 21 to 48 years old (M = 33.5), who were confined to a maximum security forensic hospital. All subjects were classified as a danger to themselves and/or others. One subject had been transferred from a regional hospital due to uncontrollable aggressive behavior. The remaining 29 subjects had been adjudicated not criminally responsible (NCR) or not guilty by reason of insanity (NGBRI) for violent criminal offenses, including murder, rape, arson, assault, battery, and malicious destruction of property. The nondangerous group consisted of 30 males, ranging in age from 19 to 67 (M = 37.9), who were hospitalized at a state psychiatric hospital. Nondangerous subjects had never been adjudicated NCR/NGBRI or been convicted of a felony. Both groups were dominated by chronic psychiatric patients who had beien currently hospitalized for an average of 4 years and previously hospitalized an average of 5 to 6 times. There were no significant group differences in the distribution of diagnoses.' Almost all subjects (n = 59) were coded at the lowest SES levels (Level 4 and 5) on the Hollingshead (1957) Socioeconomic Scale. The remaining nondangerous subject was coded as middle class (Level 3). Gross extrapolation from the Wechsler Adult Intelligence Scale-Revised (WAIS-R) Vocabulary subtest suggested dangerous and nondangerous patients' mean verbal IQs wcre 81 and 84, respectively. Most subjects were not married (n = 56) and had only completed 2 years of high school (M = 10.5). In sum, both groups were composed of long-term psychiatric patients with psychotic and organic disorders. There were no group differences in,age, SES, intelligence, or marital status-all of which have been associated with EM differences(Davidow & Bruhn, 1990). There were no differences in educatiorial level or diagnosis. Significant differences were found, however, in group racial compositions. The dangerous group was composed of 17 Black patients and 13 White patients. In comparison, 6 nondangerous patients were Black and 24 were White. Given this significant difference, subsequent statistical analyses were performed to assess the potential effects of this racial confound. 'Both groups contained a large percentage of highly disturbed psychotic patients. In order to find
30 subjects in each group who were sufficiently integrated to cooperate with the research protocol, the first author was forced to screen a large pool of potential subjects. In addition, many of the dangerous subjects were too suspicious to volunteer their EMS and had to be eliminated. The subjects used in the study should be viewed as better integrated, less suspicious, and much less floridly psychotic than a randomly selected group of chronic patients.
140
TOBEY AND BRUHN
Downloaded by [Nanyang Technological University] at 15:19 28 April 2015
Measures During a 30-min semi-structured interview, four EMS were elicited from each subject using the format outlined in the EMP (Bruhn, 1989). The WAIS-R Vocabulary subtest was also administered. All interviews were audiotaped, with the exception of 2 subjects who refused to be taped and whose interviews, therefore, were transcribed by hand. The senior author interviewed all subjects. EMS were scored with the Early Memory Aggressiveness Potential Score System (EMAPSS; Bruhn & Tobey, 1991),~adapted from a pilot study of aggressive and nonaggressive criminals' EMS (Bruhn & Feigenbaum, 1987). Memories were also scored using the Comprehensive Early Memory Score System-Revised (CEMSS-R; Last & Bruhn, 1990; see footnote 3). The CEMSS-R is an updated version of the CEMSS (Last & Bruhn, 1983), a comprehensive, general purpose EM scoring system that assesses pathological and adaptive EM characteristics. The rationale for using two EM scoring systems was to test a hypothesis ~roferredby Bruhn and Schiffman (1982), who argued that carefully crafted, specific-use, pilot study tested EM scoring systems (e.g., the EMAPSS) should outperform general purpose EM scoring systems (e.g., the CEMSS-R).
Hypotheses Two hypotheses were tested: 1. Dangerous psychiatric patients will differ from nondangerous psychiatric patients by displaying more aggressiveness potential, reflected in higher scores on the EMAPSS. 2. Dangerous and nondangerous psychiatric patients will be more accurately classified using the EMAPSS than the CEMSS-R.
RESULTS Reliability Because raters demonstrated universal agreement in identifying the absence of aggression, reliability was defined as rater agreement when aggression was 3The EMAPSS is a specialized EM scoring system that attempts to assess aggressiveness potential. The CEMSS-R is a general purpose EM scoring system, analogous to such general purpose Rorschach scoring systems as Exner's, Klopfer's, and Beck's. Inquiries about the EMAPSS or CEMSS-R, which are both copyrighted, should be directed to Arnold R. Bmhn, 4704 Hunt Avenue, Chevy Chase, MD 20815.
Downloaded by [Nanyang Technological University] at 15:19 28 April 2015
present, a much more conservative standard for reliability. EMS with no aggression were excluded from reliability computations. Three naive judges reached 85.5% interrater agreement on the 12 EMAPSS variables; two out of three judges agreed 95.8% of the time. Several studies have demonstrated that the CEMSS is reliable (Acklin, Sauer, Alexander, & Dugoni, 1989; Last & Bruhn, 1983,1985), valid in differentiating diagnostic groups (Last & Bruhn, 1985), and capable of identifying the degree of psychopathology in a clinical population (Acklin et al., 1989; Last 6r Brutm, 1983). The CEMSS has not been previously used with aggressive psychiatric patients, and the reliability of the modified CEMSS system has not been assessed. CEMSS-R interrater reliability was therefore computed for this study. Given the large number of CEMSS-R variables (65 variables per EM), eight naive raters, divided into four pairs, scored the CEMSS-R. Categorical variables were considered reliable if at least one of the two raters' scores was identical with the ratings of the senior author. Continuous variables were considered 1). This resulted in an overall reliable if ratings fell within 1 point (i.e., CEMSS-R reliability of 94.5%, with individual items' interrater agreemeints ranging from 100% to 81.1%.
Reduction of EMAPSS and CEMSS-R Variables A preliminary analysis was conducted to eliminate EMAPSS and CEMSS-R variables with minimal discriminatory potential. Categorical variables endorsed by less than 10% of the subjects were dropped. Categorical and continuous variables that were not significantly different between the two groups at a univariate level of alpha equal to .05 were also eliminated. When these t.wo criteria were used, 7 EMAPSS (Table 1) and 13 CEMSS-R (Table 2) predictor variables were retained. TABLE 1 EMAPSS Predictor Variables Variable Physical abuse Someone is physically threatened* Psychological abuse** Patient is psychologically abused* Summary scores Number of aggressive incidents** Number of aggressive memories* Mean severity ratings* Maximum severity ratings*
*p
< .05. **p < .01.
142
TOBEY AND BRUHN TABLE 2 CEMSS-R Predictor Variables
Downloaded by [Nanyang Technological University] at 15:19 28 April 2015
Variable Setting Home* Sensory-motor Visual* Distal sensory-motor* Total sensory-motor* Content Trust* Affect Anxiety* Subject affect rating* Rater affect rating** Object relations Perception of sel!? Perception of environment* Individual distinctiveness* Level of object relations* Number of words**
Main Analyses
Hypothesis I. Dangerous psychiatric patients will differ from nondangerous psychiatric patients by displaying more aggressiveness potential, reflected in higher EMAPSS scores. This hypothesis was strongly supported. Significantly more dangerous patients recalled aggressive EMS (77%) than nondangerous patients (43%), XZ(l,N = 60) = 5.62, p < .05. More dangerous than nondangerous patients recalled physical abuse (60% vs. 42%), animal abuse (17% vs. 3%), and psychological abuse (37% vs. 3%), although psychological abuse was the only category where these differences were significant, x2(1,N = 60) = 8.44, g < .01. Dangerous patients recalled a greater number of aggressive EMS, ~ ' ( 2 ,N = 60) = 8.98, p < .01. Dangerous patients also reported significantly more aggressive incidents within their four EMS than the nondangerous patients, X2(2,N = 60) = 9.50, p < .01, indicating dangerous patients are more likely to recall EMS that include multiple aggressive incidents. Whenever aggression appeared in an EM, the seriousness of the aggression was rated on a 3-point scale ranging from minimal (1)to severe (3). As a group, the dangerous patients' mean severity rating (M = .09) was significantly higher than the nondangerous patient's mean severity rating (M = .04), t(51) = 2.30, P
