REVIEW OF FORENSIC ASSESSMENT INSTRUMENTS

There are only two known forensic assessment instruments designed specifically for use in evaluations for the question of insanity. Neither of these is a test in the strict sense; they are better seen as guides to help structure the assessment process.

Mental State at the Time of the Offense Screening Evaluation (MSE)

Authors

Slobogin, C., Melton, G., & Showalter, C.

Primary Author Affiliation

College of Law, University of Florida

Primary Reference

Slobogin, C., Melton, G., & Showalter, C. (1984). The feasibility of a brief evaluation of men­tal state at the time of the offense. Law and Human Behavior, 8, 305-320

Description

The Mental State at the Time of the Offense Screening Evaluation (MSB) is a semi-structured interview guide to assist in assessments of a defen­dant's mental/psychological condition at the time of the alleged offense. It was intended for use in screening out cases where an insanity defense would clearly not be applicable and for identifying the "obviously insane individual for whom a more comprehensive evaluation is unnecessary" (b, p. 235). When MSE results indicate that a significant mental abnormal­ity may have existed, there would typically be a recommendation for a more extensive evaluation.

The MSE has three sections that occur in the following sequence: I, Historical Information; II, Offense Information; and III, Present Mental Status Examination. Each section of the MSE outlines the interviewer's goals but does not specify the actual interview questions that the exam­iner should ask.

Section I (Historical Information) consists of five' subsections (A-E) and focuses on ruling out various "significant mental abnormalities" in the examinee's past. Subsection A is typical of the format of these subsections (f, p. 319):

A. Does the defendant have a history of prolonged bizarre behavior [i.e., delusions, hallucinations, looseness of association of ideas (thought processes incoherent and illogical), disturbance of affect (behavior disorganized, aggressive, intensely negativistic or with­drawal)]? If not, exclude:

1. Organic brain syndromes of progressive or chronic nature

a. Dementia

b. Organic personality syndrome

2. Psychoses

a. Schizophrenia

b. Paranoid disorders

c. Schizophreniform disorders

d. Affective disorders

Subsections B-E use a similar format to deal with convulsive disor­ders (B), bizarre behaviors suggestive of psychotic or various neuropsy­chological conditions (C), episodic bizarre behavior (D), and mental retardation (E). One always proceeds to Section II even if the above disor­ders are ruled out in the historical evaluation.

Section II (Offense Information) has two subsections for information from the defendant (A) and external sources (B). This section focuses on the examinee's mental state at the time of the offense. Categories of infor­mation to be obtained in each section are:

A. From the Defendant

1. Defendant's present general response to offense

2. Detailed account of offense

3. Events leading up to offense

4. Postoffenseresponse

B. From extrinsic sources

1. Indictment, information or complaint

2. Confessions, preliminary transcripts, statement of the police

3. Attorney's notes

4. Autopsy reports

5. Witness accounts

Section III (Present Mental Status Examination) is completed by employing any typical mental status examination, such as the Folstein Mini-Mental Status Examination (Folstein, Folstein, & McHugh, 1975), focusing on the person's mental state at the time of the interview.

After data collection, the examiner judges whether, on the basis of the data, "significant mental abnormality may have affected the defendant's actions at the time of the offense" (f, p. 311) or that there is "probably no evi­dence of 'significant mental abnormality' approaching legal relevance" (f, p. 319). No formulas are provided for arriving at these decisions from MSE data. The authors note, however, that the use of diagnoses in the MSE only serves as a tool for decision making; diagnoses are not dispositive of the evaluation question. For example, the mere presence of a significant mental abnormality at the time of the offense, without evidence that it impaired the examinee's functioning at that time, should lead to the recom­mendation that there is little likelihood of any insanity defense on the basis of mental state. Further, evidence of severe cognitive, affective, or volitional impairment at the time of the offense may warrant further evaluation, even when no signs of significant mental abnormality can be found.

Conceptual Basis

concept definition. The authors developed the concept of "signifi­cant mental abnormality." This refers to any mental disorder that could serve as a predicate for a legal insanity defense, based on a review of statutes defining legal insanity and diminished capacity.

operational definition. The authors extracted all diagnoses that they felt would constitute "significant mental abnormality" as defined above from DSM-III (a) (the version of the American Psychiatric Association's diagnostic manual that was being used at the time the instrument was developed). They prepared a logical argument for the choices, based on probable or possible effects of the diagnostic disorders on one's functioning in an aggressive, antisocial, or violent manner. The diagnoses chosen were:

• Dementia

• Organic personality syndrome

• Schizophrenias

• Paranoid disorders

• Schizophreniform disorders

• Affective disorders

• Epileptic disorders

• Brief reactive psychosis

• Intermittent and isolated explosive disorder

• Automatism (Post-concussion syndrome, Temporal lobe epilepsy, Cerebral anoxia)

• Dissociative disorders (Psychogenic future, Sleepwalking)

• Withdrawal, delirium, or hallucinations associated with psychoac­tive substance use

• Moderate to severe retardation

The format of the instrument operationalizes several concepts that frequently arise in courts' examination of evidence in insanity cases.

critique. To screen for "significant mental abnormality," the authors began with a review of relevant diagnoses in the DSM-III. Subsequently, there have been two new versions of that diagnostic manual (DSM-III-R and DSM-IV). More important, however, the current scheme does not include consideration of some diagnoses and clinical syndromes that have permissively served as the basis for successful insanity pleas, such as severe personality disorders and certain dissociative disorders (c, d, e). As a practical matter, the exclusion of these diagnoses would not necessarily negate the merit of the screening system. Diagnoses are merely a tool, they note, not a deciding factor.

Potentially more problematic from a conceptual view, the grouping of symptoms in some cases is inconsistent with DSM and standard diag­nostic nosology. Rogers and Shuman (g, p. 225) provide a detailed listing of these inconsistencies, which include:

• Hallucinations and delusions are listed as "bizarre behavior"

• Disorganized behavior is included as a "disturbance of affect"

• Disturbances of affect for assessment of mood disorders do not include depressed or elevated moods

• Mood disorders are limited arbitrarily to "prolonged' periods"

• Looseness of association is linked with incoherence

• Mood disorders are subsumed under psychotic disorders

• Sudden alterations of consciousness are subsumed under "bizarre behavior"

• Delirium can be excluded diagnostically without an assessment of consciousness.

Psychometric Development

standardization. As noted above, the MSE is not a test or psycho­metric measure. Because it is a semi-structured guide, even the adminis­tration and classification/interpretation of responses is not formalized or standardized.

reliability. Inter-rater reliability has not been established (or inves­tigated) either for the classification of symptoms and impairment or for ultimate opinions and conclusions regarding the defendant's MSO.

norms. (Not applicable)

critique. The absence of any empirical evidence relating to observa­tions or conclusions from the MSE does limit the confidence that one can place in the opinion derived from it. If the MSE is to remain viable as one of the two existing guides for assessment of insanity, studies demonstrat­ing both aspects of reliability should be a primary consideration.

Construct Validation

The single study of the MSE is best discussed under the following heading.

Predictive or Classificatory Utility

The authors of the MSE trained 24 psychiatrists and doctoral level clinical psychologists to use the instrument (f). Pairs of trainees each eval­uated three cases (36 cases), which were selected randomly from admis­sions to a hospital forensic unit. Trainees were provided little information on defendants' past histories. Because the MSE was intended for use as a screening device, examiners were instructed to use a low threshold for the possible presence of a significant mental abnormality (in other words, if necessary, to err on the side of overidentification). In addition, all defen­dants were also evaluated independently by hospital forensic teams (con­sisting of a psychiatrist, a psychologist, and a social worker) who typically perform the required insanity assessments for the court. Cases were then followed through the court dispositions process.

In 72% of the 36 cases, the MSE-informed conclusion agreed with the conclusion of the forensic team based on their comprehensive evaluation.

Of the 20 cases determined by the MSE pair to pass a screening threshold, the court found 2 cases insane, 6 cases had charges nol-prossed, 4 were convicted of a lesser charge, and 6 were convicted as charged. (The research report does not account for the remaining 2 cases.) Of the 10 cases identified by the hospital team as having potential merit for an insanity defense, 2 cases were found insane, 7 were nol-prossed and 1 was convicted as charged. For all 16 "screen out" cases, the court either con­victed as charged or gave mitigated dispositions (unrelated to mental abnormality). No validity data are reported with regard to MSE conclu­sions about specific symptoms, disorders, or the severity of impairment.

Critique. The authors of the study discussed the results in terms of classification error rates. It is interesting to translate these rates into hypo­thetical consequences if the trainees were to have constituted an actual screening process in the mental health criminal justice system. Sixteen of the 36 defendants would have been screened out instead of being com­mitted to a hospital for further evaluation. The results indicate that none of them would have been wrongly denied an insanity verdict. (That is, as defendants in this study, all of them were seen as ineligible by the hospital forensic team, and none of them received an insanity verdict in court.) The number of cases sent to the hospital for a costly inpatient evaluation, however, would have been reduced by about 45%, and the court would have avoided delays in these cases.

Of course, based on MSE screenings, several defendants who the inpa­tient team would deem ineligible for an insanity defense would still be sent to them for evaluation. Yet this type of error is of less concern than screen­ing out defendants who might otherwise be eligible for an insanity defense, from the point of view of fairness and defendants' fundamental rights.

While the agreement rate of 72% with no false negatives is promising, the limited sample size suggests that further research and replication is necessary. Furthermore, Rogers and Shuman (g) point out that although the MSE is intended to be used as a screening device, neither this nor any subsequent investigation has tested its effectiveness in a screening context or with a population that truly represents an outpatient forensic sample. Moreover, paired evaluation teams are not typical for screening or outpatient assessments.

A screening system of any type, though, might meet with opposition from some defense attorneys. One could argue that regardless of the empirical validity of the screening instrument, the principles of due process or equal protection are violated when the defendant is not pro­vided the benefit of a full or comprehensive evaluation, especially when it is provided for some defendants and not others.

On the other hand, the empiricist will ask whether comprehensive evaluations are any more valid than briefer methods such as the MSE, especially because the more comprehensive evaluation methods rarely have been examined for their validity. Indeed, there may be no way to examine any insanity evaluations for their true validity (see Chapter 3).

References

(a) American Psychiatric Association (1980). Diagnostic and Statistical Manual of Mental Disorders, Third Edition. Washington, DC: APA.

(b) Melton, G., Petrila, J., Poythress, N., & Slobogin, C. (1997). Psychological evaluations for the courts. New York: Guilford,

(c) Pasewark, R. (1981). Insanity plea: A review of the research literature. Journal of Psychiatry and Law, 9, 357-401.

(d) Pasewark, R., Pantle, M., & Steadman, H. (1979). Characteristics and dispositions of per­sons found not guilty by reason of insanity in New York State, 1971-1976. American Journal of Psychiatry, 136, 655-660.

(e) PhiUips, B. & Pasewark, R. (1980). Insanity plea in Connecticut. Bulletin of the American Academy of Psychiatry and Law, 8, 335-344.

(f) Slobogin, C., Melton, C., & Showalter, C. (1984). The feasibility of a brief evaluation of mental state at the time of the offense. Law and Human Behavior, 8, 305-320.

(g) Rogers, R. & Shuman, D. (2000). Conducting insanity evaluations (2ded.). New York: Guilford.

Rogers Criminal Responsibility Assessment Scales (R-CRAS)

Author

Rogers, R.

Author Affiliation

Department of Psychology, University of North Texas

Primary Reference

Rogers, R. (1984). Rogers Criminal Responsibility Assessment Scales. Odessa, FL: Psychological Assessment Resources

Description

The Rogers Criminal Responsibility Assessment Scales (R-CRAS) is designed to structure and quantify the decision making process in assess­ments of legal insanity. After the examiner conducts a thorough evalua­tion including relevant interviews and reviews of pertinent records, the R-CRAS presents 30 items called "Psychological and Situational Variables," which must be assigned a numerical rating. The examiner uses these ratings and the assessment information in a decision tree analysis, which leads to a conclusion that the defendant is either "sane" or "insane" according to the relevant legal standard.

Rogers (a) emphasizes that the R-CRAs can only be used by trained forensic examiners and does not substitute for clinical judgment. it does not provide a simple recipe for decision making or relegate the ultimate conclusion to a cutting score.

The examiner provides a numerical rating for each of the 30 items called "Psychological and Situational Variables" (hereafter, the "PSV items"), then uses ratings and other data to proceed through a decision tree to arrive at an opinion regarding legal insanity.

psychological and situational variables. There are 5 groups of PSV items in the instrument. These groupings have been used by Rogers as scales, especially for purposes of research with the R-CRAS:

A. Patient's Reliability (2 items)

1. Reliability of patient's self-report which is under voluntary control

2. involuntary interference with patient's report

B. Organicity (5 items)

3. Level of intoxication at time of crime

4. Evidence of brain damage or disease

5. Relation of brain damage to commission of alleged crime

6. Mental retardation

7. Relation of mental retardation to commission of alleged crime

C. Psychopathology (10 items)

8. Observable bizarre behavior

9. General level of anxiety

10. Amnesia about the alleged crime

11. Delusions

12. Hallucinations

13. Depressed mood

14. Elevated or expansive mood

15. Level of verbal incoherence

16. Intensity and appropriateness of affect

17. Evidence of formal thought disorder

D. Cognitive Control (4 items)

18. Planning and preparation

19. Awareness of the criminality of behavior

20. Focus of the crime (e.g., selective vs. random focus)

21. Level of activity in commission of alleged crime

E. Behavioral Control (7 items)

22. Responsible social behavior during week prior to alleged crime

23. Patient's reported self-control

24. Examiner estimate of patient's self-control

25. Relation of loss of control to psychosis

26. Impaired judgment

27. Impaired behavior

28. Impaired reality testing

Two additional PSV items do not contribute to any conceptual scale (29, Capacity for self-care; 30, Awareness of wrongfulness).

The manual describes each PSV item to help provide conceptual anchors for the rating. Each item is rated on a 5- or 6-point scale, with a 0 indicating that no information is available, a 1 that there is "no symptomatology or disorganization," and a 2 that the factor is clinically insignificant. Ratings from 2 to 5 or 6 designate increasing degrees of severity or symptom impairment as specified in the item description. The ratings may be summed to produce subscores for the five summary scales, or they may stand simply as 30 item ratings with which to approach the following process.

R-CRAS decision models. The R-CRAS was designed primarily to focus on insanity assessments using the American Law Institute (AlI) standard, but it may also be used with the McNaughtan standard and a Guilty but Mentally Ill (GBMI) standard. The analytic logic to apply the ratings is guided by separate decision trees. For example, the ALI model has six decision points (a, p. 33), each of which corresponds to a major interpretive issue (e.g., malingering; presence or absence of organic disor­der) or to a component of the ALI standard (e.g., loss of cognitive control). "Yes" and "no" answers by the examiner at each of these decision points provide branching routes for arriving, ultimately, at a clinical opinion concerning whether or not the defendant meets the psychological criteria for insanity according to the ALI standard.

The examiner considers the numerically summarized data from the PSV items when making each of the judgments in the decision model. Further, the manual provides paragraph supplements defining each of the decision points in the decision models, as well as several case studies demonstrating the PSV item scoring and use of the decision models with individual defendants.

Conceptual Basis

concept definition. By conducting a review of existing statutes, case law, and legal analyses, Rogers identified the specific cognitive and/or volitional elements for each of the major insanity standards. He then attempted to identify known psychological constructs that would correspond to the functional legal element. For example, the following constructs are proposed for the ALI standard: (a, pp. 34—35):

• Organicity: an organic mental disorder.

• Major Psychiatric Disorder: functional disorders as defined by DSM-III, excluding disorders that "by definition would not have significant impact on a legal standard" (a, p. 34). [Organicity and Major Psychiatric Disorder together define the ALI's "mental disease or defect."]

• Loss of Cognitive Control: loss of ability to recognize, at the time of the crime, that the conduct was criminal.

• Loss of Behavioral Control: loss of ability to choose and to withhold important behaviors (e.g., "delusional ideation may leave no recourse and 'demand' a particular behavior response") (a, p. 35).

operational definition. The above four psychological definitions, together with the question of malingering and the legally required causal element (mental disease or defect), form the basis of the six decision nodes on the R-CRAS for the ALI standard. (Other constructs and decision mod­els were developed for McNaughtan and for Guilty but Mentally Ill.) The logic of the legal standards also was translated into the "yes-no" path­ways between the decision nodes in the decision models.

The selection and development of the 30 PSV items were derived from consensus ratings of five experienced forensic psychologists and psychiatrists for the ALI standard, and from another panel of three experts for the McNaughtan standard and the Guilty But Mentally Ill stan­dard (GBMI criteria) (a). Rogers states that the structure of the R-CRAS rating system and derivation of decision models was heavily influenced by the Schedule of Affective Disorders and Schizophrenia (h, i).

critique. The general process of arriving at concepts to be assessed was systematically logical. The legal standard was analyzed as a construct with abstract elements, and Rogers translated these elements into psychi­atric or psychological concepts. These, in turn, were defined operationally as several subclasses of symptoms, behaviors, and (despite the manual's reference to "Loss of Cognitive Control" rather than the broader "cogni­tive-affective" concept of "appreciation") both cognitive and affective states that might be inferred regarding the defendant's state at the time of the crime.

The strength of the model is that it provides a structured and stan­dardized way to think about symptoms and levels of impairment. The limitation is that, while it creates constructs by relating legal elements to psychological functions, it does not—and arguably cannot—structure the application of those constructs to the ultimate threshold questions. For example, several items are related to the construct "Loss of Behavioral Control"; yet there is no specific guidance about how to conclude from these items whether the defendant had "no recourse" but to act in accord with his delusions. Some would argue that this is because those thres­holds mark the division between psychological and moral/normative judgments. Regardless, Rogers notes that some examiners might wish not to employ these aspects of the decision tree model (a, pp. 32—33).

Psychometric Development

standardization. Administration of the R-CRAS is not standardized in the formal sense. The relevant factors are specified and defined, but the R-CRAS does not require the examiner to ask any particular set of questions. Thus, the instrument specifies what type of information should be gathered, but does not require that it be gathered in any particular way. Similarly, PSV item scores are related to the decision nodes in a very structured way, but the ultimate conclusion is not standardized or driven by cutoff scores.

reliability. Rogers (a) simultaneously examined inter-examiner and test-retest reliability in a sample of 76 defendants in a forensic hospital for insanity evaluations. Each case was interviewed twice, on the average of 2.7 weeks between interviews, by different psychiatrists or psychologists trained in using the R-CRAS. Pairs of examiners demonstrated high agree­ment on the diagnostic variables, insanity variables, and ultimate conclu­sions in the ALI decision model. Rates of agreement ranged from 85% to 100%, with kappas between.48 and.94. Examiners agreed on the ultimate conclusion in 97% of cases (kappa.94). Interrater agreements on individual PSV items averaged r = 58, with lower correlations (below r =.40) on three items (7: Retardation and the crime; 16: Intensity of affect during crime; 21: Level of activity in committing the crime). Eighteen item correla­tions achieved very high statistical significance (p =.0001).

Rogers (a) reported alpha coefficients for internal consistency of the five summary scales of the R-CRAS: Patient's Reliability,.28; Organicity,.52; Psychopathology,.80; Cognitive Control,.64; and Behavioral Control,.77.

norms. The R-CRAS manual (a) provides summary scale means and 95% confidence levels for three sample sets (n = 73, n = 111, n = 76) of persons judged by forensic examiners (using the R-CRAS ALI decision model) to be legally insane or sane.

critique. The R-CRAS seems to provide considerably greater stan­dardization of the insanity assessment process than would be expected for assessments by examiners not employing the system. Yet it also pro­vides sufficient flexibility in data collection to meet the requirements of specific cases.

Although interrater agreement on individual PSV items was less than one would hope, measures of interrater reliability for the components of insanity are strong (.75 for cognitive control and.80 for behavioral con­trol) and those for the ultimate conclusions (sane/insane) are impressive (.94). It is possible that greater specification of item rating criteria and the inclusion of additional rating examples for each item could improve scor­ing standardization and inter-examiner reliability of item ratings.

The normative samples were drawn from forensic assessment centers in Chicago and Toledo, with secondary (smaller) samples drawn from several centers in other geographic areas. Therefore, the information on norms may be generalizable to various settings where insanity assess­ments are performed.

The low internal consistency of the Patient's Reliability summary scale probably is due to the presence of only two PSV items in this scale. Further, one would not expect these two items to be related. The first item refers to the veracity or unreliability of the patient's report as a function of voluntary truthfulness or distortion, whereas the second refers to involuntary inter­ference with the patient's report (e.g., because of delusional thought).

Construct Validation

Factor analysis of a set of R-CRAS protocols produced three factors (c):

• Bizarre behavior: lack of awareness; self-reported low control over crime; delusions; and final R-CRAS insanity decision.

• High activity: inappropriate, intense affect; self-reported low con­trol over crime; absence of intoxication; and final R-CRAS insanity opinion.

• High anxiety: malingering; depressed feelings; lack of reliability of self-report; and self-reported low control over crime.

Rogers (a) reported a discriminant analysis using the five summary scales as predictors, and final R-CRAS opinions of the examiners (does or does not meet the ALI standard using the R-CRAS model) to define criterion groups. In the original and cross-validation attempts respectively, false posi­tive rates were 5.7% and 16.7%, and false negative rates were 1.3% and 3.6%.

Rogers and Sewell (e) conducted a further examination of the con- struct/structural validity of the R-CRAS using discriminant analysis on each component (e.g., loss of cognitive control, loss of volitional control). As predicted, they found patterns of items that distinguished the constructs (average hit rates of 94%), and the results indicated that the items helped to "explain" the constructs (explaining, on average, 64% of the variance).

No significant relations have been found between final R-CRAS insan­ity opinions and age, race, gender, education, work history, competency to stand trial, prior felony arrests, and several other legal variables (a).

critique. Factor analysis of the R-CRAS items did not produce fac­tors similar to the five summary scales, but rather factors that might describe three prototypic insanity assessment cases. One wonders, there­fore, whether the summary scales should be conceptualized as scales as such, rather than simply as groups of PSV items. More research on the structure of the R-CRAs is needed to address this question.

One should note that the discriminant analysis described previously does not demonstrate the R-CRAS' external validity, but rather the degree of consistency between its summary scale ratings and the final opinions of insanity made by examiners using the same ratings along with the ALI decision model. Interestingly, the results suggest that use of the statistical analysis of summary scale ratings to make the final insanity classification usually will produce the same result (in 93-97% of the cases) as if the examiners had made the final decisions. This result speaks well for the internal validity of the R-CRAS system. Examiners, however, should not take these results to suggest that the R-CRAS is a valid indicator of insan­ity in the sense of external validity.

Finally, the study's use of examiners' final insanity decisions as a dependent variable or criterion should not be taken as a justification for tes­tifying to these decisions in actual practice; arguably these opinions about the ultimate legal question are beyond the proper scope of expert testimony.

Predictive or Classificatory Utility

To examine the external validity of the R-CRAS, Rogers (a) examined the rate of agreement between R-CRAS-based opinions about sanity or insanity and court dispositions for a sample of 93 defendants. To avoid criterion contamination, courts were not advised of R-CRAS results for these defendants. The overall hit rate of the R-CRAS-based opinions was 88.3%, with 4.8% false negatives and 26.7% false positives. (Base rate of insanity verdicts was 32%). The chance-corrected kappa coefficient aver­aged.72. The rate of agreement for cases the evaluators assessed as sane was higher (95%) than for those assessed as insane (73%). Further analysis revealed that cases in which there was a disagreement between the R-CRAS opinion and the verdict were primarily cases in which the severity of psychopathology was moderate. Greater agreement was reached when the severity rating was either extremely high or extremely low.

critique. The false positive rate in this study seems higher than would be desirable. On the other hand, Rogers (a) correctly notes that comparison of R-CRAS results to court determinations is an equivocal test of the instrument. One does not know (nor can one test) the validity or accuracy of judicial decisions, because there is no appropriate external cri­terion. At a minimum, at least the R-CRAS provides some evidence for known error rates in relation to some criterion, and this is more than cur­rently exists for unaided or unstructured insanity assessments.

References

(a) Rogers, R. (1984). Rogers Criminal Responsibility Assessment Scales. Odessa, FL: Psychological Assessment Resources.

(b) Rogers, R. & Cavanaugh, J. (1981). The Rogers Criminal Responsibility Assessment Scales. Illinois Medical Journal, 160, 164-169.

(c) Rogers, R., Dolmetsch, R., & Cavanaugh, J. (1981). An empirical approach to insanity evaluations. Journal of Clinical Psychology, 37, 683-687.

(d) Rogers, R., Seman, W., & Wasyliw, O. (1983). The R-CRAS and insanity: A cross-validation study. Journal of Clinical Psychology, 39, 554-559.

(e) Rogers, R. & Sewell, K. (1999). The R-CRAS and insanity evaluations: A reexamination of construct validity. Behavioral Sciences and the Law, 17, 181-194.

(f) Rogers, R. & Shuman, D. (2000). Conducting insanity evaluations (2d Ed.). New York: Guilford.

(g) Rogers, R., Wasyliw, O., & Cavanaugh, J. (1984). Evaluating insanity: A study of con­struct validity. Law and Human Behavior, 8, 293-303.

(h) Spitzer, R, & Endicott, J. (1978). Schedule of affective disorders and schizophrenia. New York: Biometrics Research.

(i) Spitzer, R., Endicott, J., & Rollins, E. (1975). Clinical criteria and DSM-III. American Journal of Psychiatry, 132, 1187-1192.