REVIEW OF FORENSIC ASSESSMENT INSTRUMENTS

Mental health examiners may improve their assessments related to competence to stand trial by evaluating functional abilities that are espe­cially relevant for the legal definition of the competence.

Several instruments that have been developed for use in competence to stand trial assessments were not selected for review. Three are from an earlier era, providing "checklists" of information to be obtained in compe­tence evaluations but no quantitative scores or ratings (Bukatman, Foy, & De Grazia, 1971; Lawrence, 1981; Robey, 1965). Another (the Computer- Assisted Competence Assessment Tool) was not included because no manual was available (for a review of this instrument, see Melton et al., 1997, pp. 143-144). The Interdisciplinary Fitness Interview (IFI), which was reviewed in the first edition of EC, was not re-reviewed because no further work on the IFI has been published since the review in the first edition.

Finally, at the time this chapter was written, a manual for the Evaluation for Competency to Stand Trial-Revised (ECST-R) by Rogers, Tillbrook, and Sewell (undated) was in preparation but not yet in final form. The instrument's constructs and semi-structured interview items include the full range of abilities typically associated with competence to stand trial, a format for identifying the relation of psychiatric symptoms to defendants' deficits on those abilities, a system of rating the items, and an index for detecting feigning of symptoms. Initial promising information on the instrument has been published (Rogers, 2001; Rogers, Grandjean, Tilbrook, Vitacco, & Sewell, 2001; Rogers, Sewell, Grandjean, & Vitacco, in press; Tillbrook, 2001).

In each of the following reviews, citations are made to lettered refer­ences (for example, a, b, c) that appear at the end of the instrument's review.

MacArthur Competence Assessment Tool-Criminal Adjudication (MacCAT-CA)

Authors

Poythress, N., Nicholson, R., Otto, R., Edens, J., Bonnie, R., Monahan, J., & Hoge, S.

Primary Author Affiliation

Florida Mental Health Institute, University of South Florida, Tampa FL

Primary Reference

Poythress et al. (1999). The MacArthur Competence Assessment Tool-Criminal Adjudication: Professional manual. Odessa, FL: Psychological Assessment Resources.

Description

The MacArthur Competence Assessment Tool-Criminal Adjudication (MacCAT-CA) was developed to assess abilities associated with "adju­dicative competence." As discussed later in this review, adjudicative com­petence is a concept that recognizes the broad range of factors that may be involved in legal determinations of defendants' "competence to proceed (assist counsel)" and their "decisional competence." (b) The adjudicative competence concept was one of the guiding structures for an effort during the 1990s by the MacArthur Research Network on Mental Health and Law to develop improved measures of capacities associated with compe­tence to stand trial.

The MacCAT-CA is a structured and highly standardized interview that consists of 22 items organized in three parts called Understanding, Reasoning and Appreciation. (k) The eight Understanding items and eight Reasoning items are based on a brief vignette, introduced at the beginning of the interview, regarding two men who get in a fight in a bar while play­ing pool. One of the men is arrested and faces charges and a trial process.

In this hypothetical context, the Understanding items ask questions of the examinee to determine the examinee's comprehension of eight things:

• roles of defense and prosecution attorneys

• the elements of an offense with which the defendant could be charged

• elements of a lesser included offense

• role of the judge

• role of the jury

• the consequences of conviction

• consequences of pleading guilty

• rights waived when one pleads guilty.

For 6 of these items, an examinee's failure to give an adequate response results in the examiner employing a brief, standardized "teaching" process regarding the element that is being tested, with a subsequent repeat of the question to determine whether the defendant can now provide an adequate response. For the 2 items about nature of the offense, the items begin with a "teaching" disclosure without a "pre-test" of the defendant's knowledge.

The examinee's responses are examined with objective scoring crite­ria in the manual, classifying them as 2, 1, or 0 (adequate, questionable, inadequate).

The eight Reasoning items are of two types. Five of the items offer two "facts" about the incident and ask the examinee which of the two would be more important for the defendant to tell his lawyer. The other three items are based on a description of two pleading choices: pleading guilty under the conditions of a plea agreement with the prosecutor, and plead­ing not guilty and going to trial. The items focus on the decision process, not the choice: whether the examinee seeks more information before deciding, offers both an advantage and a disadvantage for the chosen option, and manifests evidence of having compared the chosen option to the rejected option. Responses are scored (2, 1, 0) according to objective criteria provided in the manual.

The six Appreciation items are not based on the hypothetical case used in Understanding and Reasoning, but on the examinee's own legal situa­tion. The examinee is asked whether, "compared to other people who are in trouble with the law," the examinee thinks that he or she is "more likely, less likely, or just as likely" to:

• be treated fairly in the legal process

• be assisted by defense counsel

• fully disclose case information to his or her defense attorney

• be found guilty

• get the same punishment as others if found guilty

• plead guilty

Examinees are then asked for an explanation for their choice, and queries are made to determine whether the examinee's explanation contains evi­dence of "unrealistic or idiosyncratic beliefs that defendants have about themselves or their situations" such that they are "clearly implausible and colored by symptoms of mental illness (e.g., delusions)" (k, p. 13).

The process yields Understanding, Reasoning, and Appreciation scores, but the manual does not provide or encourage a "total MacCAT-CA score" such as a sum of the three scale scores. The examinee's scores can be compared to norms, provided in the manual, for three groups of adult defendants involved in research during the instrument's development: non-selected jail detainees, defendants with mental illness but presumed competent, and defendants who had recently been found incompetent to stand trial. Tables offer percentages of persons in each of these groups who obtained each of the possible scores on the three scales. The tables also offer recommended cut-off points for levels of performance desig­nated "minimal or no impairment," "mild impairment," and "clinically significant impairment." The authors explain that these are not meant to define competence or incompetence, but to assist in describing the degree of deficits or strengths that the defendant manifests.

Conceptual Basis

concept definition. Two sets of concepts played central roles in the development of the MacCAT-CA: adjudicative competence as conceptual­ized by Bonnie (b), and the three elements of Understanding, Reasoning, and Appreciation.

Bonnie's theory of adjudicative competence employs two compo­nents: competence to assist counsel and decisional competence. Competence to assist counsel includes a basic understanding of matters pertaining to the trial process, as well as the ability to work with one's attorney and to assist the attorney by relating pertinent factors that may assist the defense. Decisional competence refers to abilities needed for "autonomous decision making with respect to strategic issues that arise in the course of prosecution" (k, p.

The elements of Understanding, Reasoning, and Appreciation were derived from earlier work in the structuring of decision making abilities associated with competence to consent to or refuse treatment, as identi­fied through case law reviews (a, n) and employed in other studies of assessment of abilities associated with competence to consent (d). According to Grisso and Appelbaum (d), Understanding refers to basic comprehension of relevant facts. Reasoning refers to the ability to process information (e.g., seek information, weigh advantages and disadvantages of options) in the course of decisionmaking. Describing the MacCAT-CA concepts, Otto et al. also include in Reasoning the "ability to discern the potential legal relevance of information" (j, p. 436). Appreciation refers to "rational awareness of the meaning and consequences of the proceedings in one's own case" (j, p. 436). This element focuses specifically on dis­torted application of information to one's own situation due to delusional beliefs associated with mental disorder.

As conceptualized by Grisso and Appelbaum (d), these elements are inter-related but sufficiently discrete to warrant separate concepts. For example, one might fully understand the facts of trials, yet have diffi­culties manipulating the information in a decision process. Or one might have the cognitive capacities to process information quite well in arriving at a decision using simple logic, while also being at risk of making a deci­sion based on a delusional belief (e.g., regarding one's immortality or invulnerability) that does not allow one to realistically appreciate the significance of the trial for oneself.

The specific content to be employed in each of these three sections of the MacCAT-CA derived from various sources. The content of the Understanding section borrowed from an evolving consensus in law and clinical practice (as manifested in many other psycholegal instruments) regarding the specific things that defendants are expected to know about charges and trials when questions of their competence are raised. The content domains of the Reasoning and Appreciation sections are more novel and were determined by the researchers in consultation with mem­bers of the MacArthur research network that sponsored the development of the instrument.

operational definition. The various item strategies employed in the MacCAT-CA (such as the use of a hypothetical case, the teaching process of the Understanding items, the Reasoning section's method for assessing ability to perceive relevant facts and to process information, and the Appreciation section's format) had been used in the MacSAC-CD, which had demonstrated the viability and value of these strategies in ear­lier studies with criminal defendants (g). They were refined in the course of a group process of discussion and revision that involved a broader community of researchers and law scholars (the MacArthur Research Network on Mental Health and Law). Example responses for use in the scoring criteria were also derived from the earlier MacSAC-CD and from piloting of the prototype MacCAT-CA.

critique. The use of two sets of concepts to guide the development of the MacCAT-CA anchors it in both law and existing psycholegal con­cepts for structuring the assessment of relevant abilities for legal compe­tence. It is unique, for example, in focusing part of the assessment on decision making abilities associated with competence to stand trial, which have achieved greater attention in practice as a result of their recognition in the U.S. Supreme Court decision in Godinez v. Moran. The decision to use the three elements that have been derived in other areas of legal com­petence also increases its potential to contribute to a more unified concep­tualization of legal competencies.

The absence of separate competence to assist counsel and decisional com­petence sections of the MacCAT-CA seems logical. The authors explain that the Understanding items and some of the Reasoning items are especially associated with the competence to assist counsel component, and the Appreciation items and some of the Reasoning items are especially associ­ated with decisional competence. But these two concepts, while capturing somewhat different aspects of competence to stand trial, are not completely discrete. Understanding of essential facts is necessarily required for meaningful decision making, and delusional beliefs can impair one's abil­ity to assist counsel. Thus, the choice not to try to structure the instrument too rigidly along the lines of these two concepts seems wise.

As acknowledged by the MacCAT-CA authors (k), the MacCAT-CA does not collect information on all of the abilities and deficits that might be relevant for competence to stand trial evaluations. Among those that are not included, for example, are the defendant's ability to engage in log­ical communications with a defense attorney, the ability to testify, and the ability to manage the stress of trial. This is not a weakness unless a mea­sure's items insufficiently represent the primary psycholegal constructs that have guided the development of the instrument. In this sense, the MacCAT-CA items seem to have been chosen meaningfully.

The high degree of structure employed in the MacCAT-CA offers excellent potential for the instrument to be used not only in clinical assess­ments but also in research, and it provides the opportunity for construct­ing meaningful norms and demonstration of its psychometric properties. The use of a hypothetical case, as well as the rigid standardization of ques­tions and inquiries, reduces the examiner's ability to explore issues that may arise specific to individual defendants (e.g., probing the nature of their misunderstanding of concepts, determining whether they under­stand their own charges and their consequences). Rogers (1) criticized the MacCAT-CA's reliance on defendants' responses to hypothetical offenses and offenders, pointing out that such data were insufficient to address case-specific issues in actual forensic assessment practice. The MacCAT-CA authors were clearly aware of this. They acknowledged (j, k) that the instrument was not intended to provide all of the information that a clini­cian will need when assessing defendants' abilities; it should be supple­mented with mental status examination and inquiries designed to assess the defendant's grasp of information specific to his or her own circum­stances. Use of those more individualized methods alone, of course, would provide no way to compare defendants to norms regarding degrees of abil­ity related to competence to stand trial. This is the value of using multiple methods, some highly structured like the MacCAT-CA and others with greater capacity to capture nuanced information specific to the individual.

The use of a teaching mechanism in the Understanding items is unique. It incorporates the notion, suggested in various texts on psycholegal assessment (c, h), that it is the capacity to understand, not merely the amount one currently understands, that is most relevant for competence to stand trial. When defendants at first appear not to understand certain aspects of trials, examiners typically are encouraged to attempt to "teach" them and to assess their capacity to grasp the matters with instruction. The MacCAT-CA makes this a standard part of the competence assess­ment process. It is not clear why the authors did not construct a subscale of Understanding that expresses the defendant's original level of under­standing, in contrast to the final Understanding score based on their performance after having been taught. This could be done by examiners, but the MacCAT-CA manual and current research provide neither a pro­cedure with which to do this nor normative data to which to compare the defendant's initial understanding.

Similarly, one should be aware that the Reasoning scale contains two relatively distinct sets of items. Five items pertain to the defendant's abil­ity to make inferences about the relevance or irrelevance of facts in the case for purposes of building a defense. Three other items focus on the individual's processing of a decision (e.g., whether they attend to both advantages and disadvantages of options). While all of those items can be conceptualized as related to a concept called Reasoning, the fact that they are of two different types raises the possibility that they might create a scale with less internal consistency than one might wish.

A review of the scoring criteria reveals scoring distinctions, explana­tions, and examples that generally will not be difficult for clinicians to comprehend or to explain to courts that inquire about the basis for their scores. One exception is the scoring for the Appreciation items. The con­cept of Appreciation refers specifically to evidence of the presence or absence of "implausible" explanations based on delusional beliefs. The Appreciation scoring system, however, gives "0" credit for two types of responses. One is for "reasons that are clearly implausible... based on a delusional premise or a serious distortion of reality." The other is for cases in which the defendant "offers no reason or fails to answer the question." Due to the second of these criteria, a low summary Appreciation score cannot be interpreted as signifying the concept that Appreciation repre- sents—that is, "implausibility due to delusional content"—because in some cases it will represent mere failure to respond. In clinical practice, one can easily interpret what a defendant's Appreciation score means by examining the specific responses of the defendant. But this creates greater problems for researchers, who will not be able to make meaningful com­parisons of Appreciation scores across studies (unless they examine every research subject's "0" responses on Appreciation to discover and report the proportion that are delusional responses and the proportion that are "I don't know" responses).

Psychometric Development

standardization. Instructions for administering the MacCAT-CA are clear, highly structured, and aided by a standardized form for admin­istration, response recording, and scoring. Scoring criteria are explicitly defined and supplemented with example responses.

reliability. Interrater reliability was established by examining responses on 48 protocols in the sample on which the MacCAT-CA norms were developed, with eight of the study's research assistants scoring all protocols (other than the six that a given research assistant had originally scored) (j, k). Intraclass correlations for scale scores were.90 for Under­standing,.85 for Reasoning, and.75 for Appreciation. All items but two in Reasoning and three in Appreciation manifested intraclass correlations above.60. Test-retest reliability has not been examined.

norms. Currently the original validation study for the MacCAT-CA is the source for normative data on the MacCAT-CA (j, k). The study involved administration of the MacCAT-CA and other clinical measures to:

• 283 defendants admitted to forensic psychiatric units after being adjudicated incompetent to proceed (HI: hospitalized incompetent),

• 249 defendants in jail who were receiving treatment for mental health problems but who were presumed competent (JT: jail treated), and

• 197 randomly selected jail inmates who were presumed competent (JU: jail unscreened).

The sample was obtained from data collection in six states in a study that managed consistency across sites in test administration and the data col­lection process. Defendants were between the ages of 18 and 65. Males made up about 90% of the sample, with about one-half being non-Hispanic white defendants. Gender, race, age, and educational level were not markedly different across the three comparison groups. The MacCAT-CA manual provides the percentile rank for scores on each of the three MacCAT-CA scales separately, for the HI group and for the combined JU/JT groups. Certain scores are designated "mild impairment" and "clinically significant impairment" based on their departure below the mean of the JU/JT group by 1.0 and 1.5 standard deviations, respectively.

critique. The degree of standardization of administration provided by the MacCAT-CA, as well as evidence for interscorer reliability, inspires confidence in the norms and their application in clinical cases. These psy­chometric features also provide excellent potential for the instrument's use in research. This psycholegal assessment instrument is the first in its area to offer meaningful norms to which to compare scores in individual cases involving defendants whose competence to stand trial is questioned and evaluated. Moreover, data were obtained from a number of sites nationally, which increases confidence in the generalizability of the norms.

One unanswered question is whether the norms can be applied to women. Only small proportions of the validation samples were women, and reports of the instrument's development did not indicate whether any differences were found (or whether the small female sample could have provided such information with confidence).

Although the authors devised cut-off scores, they have been very careful (k) to discourage users from employing these criteria as indicators of "competence" or "incompetence." As long as they are not misused in that way, the cut-off scores offer a useful tool for making comparative statements about the performance of defendants in individual assessment cases.

Construct Validation

In the original study (i, j), Cronbach's alpha was.85 for Understanding,.81 for Reasoning, and.88 for Appreciation. The mean inter-item correla­tions within each of these scales was.42 for Understanding,.36 for Reasoning, and.54 for Appreciation. Overall these results suggest that items within each scale are assessing something in common associated with the scale's construct.

Reports of the original study did not describe the degree to which the three scales of the MacCAT-CA are measuring distinctly different domains of ability (e.g., interscale correlations or factor analytic struc­ture). Rogers et al. (m) performed a factor analysis on a new sample of 149 defendants, achieving both two-factor and three-factor solutions with factor eigenvalues over 1.0. The three-factor solution, which accounted for more of the variance, manifested a distinct Appreciation factor and two other factors that corresponded roughly to the Understanding and Reasoning constructs. However, the Reasoning factor contained the five items (out of eight) that focused on defendants' capacities to infer rele­vance and irrelevance of information for a defense. Concerning the other three Reasoning items that focus on the reasoning process, two loaded on a factor with the majority of the Understanding items and did not load significantly on any of the factors.

The study producing the norms for the MacCAT-CA (j, k) found significant correlations between the MacCAT-CA scales and a number of clinical variables, the most significant being Full Scale IQ (with Understanding,.41; Reasoning,.34; Appreciation,.14), the Psychoticism scale of the Brief Psychiatric Rating Scale (r = -.40 to -.52), and Psychoticism on the Minnesota Multiphasic Personality Inventory-2 (r = -.21 to -.33).

critique. Available evidence suggests good internal consistency for the MacCAT-CA scales. More evidence is needed to examine empirically the degree to which the three scales are sufficiently distinct to warrant the basic three-scale structure of the MacCAT-CA. Preliminary factor-analytic evidence from Rogers et al. (m), however, suggest a rough correspon­dence with the intentions of the instrument's authors.

Note that Understanding and Reasoning were related more substan­tially to Full Scale IQ than was Appreciation. This is consistent with find­ings of Grisso et al. (f) for another psycholegal measure (the MacArthur Competence Assessment Tool—Treatment) that used the same three con­structs for that instrument's structure. The Understanding and Reasoning constructs would be expected to be related more closely to measures of general cognitive ability, while the Appreciation concept focuses on indi­vidual's beliefs and affective reactions to their own circumstances.

The finding that the MacCAT-CA scales are most consistently related to psychoticism is in agreement with findings in many other studies (e.g., i) that psychotic disorders constitute the most frequent clinical con­dition found in persons who are considered incompetent to stand trial in actual practice.

Predictive or Classilicalorv Utility

The original MacCAT-CA norming study (j, k) compared the jail unscreened, jail treated, and hospitalized incompetent samples on the MacCAT-CA scales. Moreover, because many of the hospitalized incompe­tent defendants were residing in hospital programs for the restoration of incompetence, clinicians were asked to make independent evaluations and decisions (without exposure to MacCAT-CA data) for each of the defen­dants hospitalized for incompetence, indicating whether they appeared to be competent or incompetent to stand trial at the time of the MacCAT-CA evaluation. Significant differences were found on all three MacCAT-CA scales between the HI defendants who were judged competent and those judged incompetent by clinicians. Scale scores were correlated with these competence judgments for all three scales (Understanding,.36; Reasoning,.42; Appreciation,.49). Moreover, both of the HI groups were significantly lower than the jail unscreened and jail treated groups (presumed compe­tent) on all three of the MacCAT-CA scales.

The MacCAT-CA manual includes tables that provide "hit rates" and various indicators of predictive utility (e.g., sensitivity, specificity, false negative rate, positive and negative predictive value) for each score on all three of the MacCAT-CA subscales.

critique. The evidence regarding the relation between MacCAT-CA scores and independent criteria for competence to stand trial is very encouraging. The findings need to be supported in other studies, but the likelihood that these results will be replicated is increased by the fact that the original sample was obtained from a number of sites nationally.

Potential for Expressing Person-Situation Congruency

The MacCAT-CA is based largely on stimuli that use a hypothetical case. It does not seek to assess the individual's understanding and reason­ing abilities in relation to the individual's own circumstances. Therefore, it does not provide a means for comparing individuals' abilities to actual demands of their situations.

References

(a) Appelbaum, P., & Grisso, T. (1988). Assessing patients' capacities to consent to treat­ment. New England Journal of Medicine, 319, 1635-1638.

(b) Bonnie, R. (1992). The competence of criminal defendants: A theoretical reformulation. Behavioral Sciences and the Law, 10, 291-316.

(c) Grisso, T. (1988). Competence to stand trial evaluations: A manual for practice. Sarasota, FL: Professional Resource Press.

(d) Grisso, T., & Appelbaum, P. (1998). Assessing competence to consent to treatment: A guide for physicians and other health care professionals. New York: Oxford University Press.

(e) Grisso, T., & Appelbaum, P.S. (1995). The MacArthur Treatment Competence Study, III: Abilities of patients to consent to psychiatric and medical treatment. Law and Human Behavior, 19, 149-174.

(f) Grisso, T., Appelbaum, P.S., Mulvey, E., & Fletcher, K. (1995). The MacArthur Treatment Competence Study, II: Measures of abilities related to competence to consent to treat­ment. Law andHuman Behavior, 19, 127-148.

(g) Hoge, S., Bonnie, R., Poythress, N., Monahan, J., Feucht-Haviar, T., & Eisenberg, M. (1997). The MacArthur adjudicative competence study: Development and validation of a research instrument. Law and Human Behavior, 21, 141-179.

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

(i) Nicholson, R., & Kugler, K. (1991). Competent and incompetent criminal defendants: A quantitative review of comparative research. Psychological Bulletin, 109, 355-370.

(j) Otto, R., Poythress, N., Edens, J., Nicholson, R., Monahan, J., Bonnie, R., Hoge, S., & Eisenberg, M. (1998). Psychometric properties of the MacArthur Competence Assessment Tool-Criminal Adjudication. Psychological Assessment, 10, 435-443.

(k) Poythress, N., Nicholson, R., Otto, R., Edens, J., Bonnie, R., Monahan, J., & Hoge, S. (1999). The MacArthur Competence Assessment Tool—Criminal Adjudication: Professional manual. Odessa, FL: Psychological Assessment Resources.

(l) Rogers, R. (2001). Focused forensic interviews. In R. Rogers (Ed.), Handbook of diagnostic and structured interviewing (pp. 296-357). New York: Guilford.

(m) Rogers, R., Grandjean, N., Tillbrook, C., Vitacco, M., & Sewell, K. (2001). Recent inter­view-based measures of competence to stand trial: A critical review augmented with research data. Behavioral Sciences and the Law, 19, 503-518.

(n) Roth, L., Meisel, A., & Lidz, C. (1977). Tests of competence to consent to treatment. American Journal of Psychiatry, 134, 279-284.

Fitness Interview Test—Revised (FIT-R)

Authors

Roesch, R., Zapf, P., Eaves, D., & Webster, C.

Primary Author Affiliation

Mental Health, Law and Policy Institute, Simon Fraser University

Primary Reference

Roesch, R., Zapf, P., Eaves, D., & Webster, C. (1998). Fitness Interview Test (Revised Edition). Burnaby, British Columbia, Canada: Mental Health, Law and Policy Institute, Simon Fraser University.

Description

The original 1984 version of the Fitness Interview Test (FIT) (d) was based on the Competence Assessment Interview (CAI) (see review later in this chapter). It included all of the functional ability items of the CAI as well as additional items related to understanding of trial procedures. Further, a separate section of 12 items guided examiners' evaluations of defendants' relevant clinical mental status. Research in the 1980s reported the structure, reliability, validity, and utility of the FIT (a, b, c). The revi­sion of the instrument, resulting in the Fitness Interview Test-Revised Edition (FIT-R) (e), was begun soon after Canada's 1992 revision of its Criminal Code (called Bill C-30) that offered a statutory definition of "unfit to stand trial" replacing earlier Canadian reliance on case law definitions (as described later).

The revisions in the FIT were extensive. Although some of the items were retained in the FIT-R, some were eliminated and other new ones added and regrouped, the FIT's clinical mental status items were deleted, and the scoring system was completely revised. Thus research evidence on the orig­inal FIT cannot be extrapolated to address the properties of the FIT-R.

The FIT-R is a structured interview that was designed to assist clini­cians in obtaining information on "all important aspects of fitness to stand trial" (e, p. 17), while recognizing that clinical opinions about fitness would require combining this information with clinical information obtained from an additional process for determining the presence and symptoms of mental disorder. The interview begins with four "back­ground" questions (e.g., Do you have a lawyer at this point?"). This is followed by 70 questions, grouped according to 16 items that are clustered in 3 sections. The 3 sections, their items and number of questions are:

• Section I: Understanding the Nature or Object of the Proceedings: Factual Knowledge of Criminal Procedure

1. Understanding of Arrest Process (5 questions)

2. Understanding of the Nature and Severity of Current Charges (5 questions)

3. Understanding of the Role of Key Participants (9 questions)

4. Understanding of the Legal Process (4 questions)

5. Understanding of Pleas (8 questions)

6. Understanding of Court Procedure (5 questions)

• Section II: Understanding the Possible Consequences of the Pro­ceedings: Appreciation of Personal Involvement in and Importance of the Proceedings

7. Appreciation of the Range and Nature of Possible Penalties (4 questions)

8. Appraisal of Available Legal Defenses (3 questions)

9. Appraisal of Likely Outcome (3 questions)

• Section Ø: Communicate with Counsel: Ability to Participant in Defense

10. Capacity to Communicate Facts to Lawyer (3 questions)

11. Capacity to Relate to Lawyer (5 questions)

12. Capacity to Plan Legal Strategy (6 questions)

13. Capacity to Engage in Own Defense (3 questions)

14. Capacity to Challenge Prosecution Witnesses (2 questions)

15. Capacity to Testify Relevantly (2 questions)

16. Capacity to Manage Courtroom Behavior (3 questions)

Although the questions should be posed in the sequence and wording provided in the manual, it is appropriate for examiners to probe regarding examinees' initial answers if they need clarification. When the interview is completed, examiners rate each item (2, 1, 0) based on a consideration of the examinee's responses to all of the questions associated with that item. The manual urges examiners to rate according to their opinions about the examinee's abilities, not simply with regard to the specific words used in the examinee's answers. General paragraph explanations are provided concerning the meaning of each item, but specific examples or scoring cri­teria are not provided. A 2 rating indicates "severe impairment of ability," 1 indicates "moderate impairment of ability," and 0 indicates "little or no impairment of ability" to meet the legal criterion.

The manual offers a coding sheet for recording these ratings, as well as assigning a single summary rating (2, 1, 0) for each of the three Sections. The examiner then indicates on the coding sheet an "overall judgement of fitness" (Fit, Questionable, Unfit). The FIT-R system does not provide any "formulas" for combining the item ratings to arrive at this final judgment, because not only the functional ability ratings, but also other evidence related to the presence and nature of mental disorder, must be taken into account to reach a final opinion about fitness. In addi­tion, no specific level of ability is necessarily required for all types of defendant trials and circumstances.

Conceptual Basis

conceptual definition. As noted earlier, the FIT-R was constructed to be compatible with 1992 changes in the Criminal Code of Canada (known as Bill C-30). The three sections of the FIT-R, therefore, followed the new definition of unfitness:

Unable on account of mental disorder to conduct a defence at any stage of the proceedings before a verdict is rendered or to instruct counsel to do so, and, in particular, unable on account of mental disorder to (a) understand the nature or object of the proceedings, (b) understand the possible consequences of the proceedings, or (c) communicate with counsel (Criminal Code of Canada, Section 2, 1992).

operational definition. The items for each section were partly drawn from the FIT (which had been influenced by the content of the CAI, reviewed later in this chapter) with a few additions and modifications of content. The specific questions, however, are largely original (not bor­rowed from other instruments) and were developed by the authors as tools for probing the examinee's knowledge and inferential processes asso­ciated with the items in question. For example, Item 6, "Understanding of Court Procedure," is assessed with the following questions:

• Who is the only one at your trial who can call on you to testify?

• After your lawyer finishes asking you questions on the stand, who can ask you questions next?

• Why does the Crown counsel (prosecutor) ask you questions?

• What will the judge do if you plead guilty?

• What questions would you ask your lawyer before you decide whether or not to plead guilty?

critique. The logic for structuring the FIT-R according to Canadian law is quite clear. Moreover, it does not limit the FIT-R to application in Canada. Most of the concepts assessed in the interview are relevant for U.S. definitions of competence to stand trial. As noted in the third item above, the wording of questions in the FIT-R manual recognizes the potential application of the instrument in other countries.

It is instructive to compare the content of the FIT-R to that of the MacArthur Competence Assessment Tool-Criminal Adjudication (MacCAT-CA), reviewed earlier in this chapter. The first sections of both instruments focus on "understanding" of information about charges, penalties, and legal process, and they have considerable content overlap. The second section of the MacCAT-CA, called "Reasoning," focuses on the defen­dant's abilities to make inferences about the importance of facts that might be relevant to communicate to counsel. This is most similar to Section III of the FIT-R, although the inquiry for this section in the FIT-R is much broader, including a wider range of "ability to participate in defence" items and questions than is found in the MacCAT-CA (e.g., questions about ability to testify at one's own trial, capacity to manage behavior in the courtroom).

The title of the FIT-R's Section II (Understand the Possible Consequences of the Proceedings) uses the term understanding. But the criterion phrases describing the three items in this section clearly indicate that their purpose is similar to that of the "Appreciation" section of the MacCAT-CA. That is, Section II is intended to identify whether the defendant has a "realistic" perception of the penalties, possibilities for defense, and outcome. The focus, therefore, is not merely on understand­ing of facts, but whether or not defendants have beliefs about how those matters apply to their own circumstances that are distorted by delusional or idiosyncratic beliefs. As such, the content of this section of the FIl-R seems to depart from the plain wording of Canada's Criminal Code defi­nition, which says nothing specifically about distorted beliefs ("apprecia­tion," or in Dusky v. U.S., "rational understanding" as contrasted with "factual understanding"). Thus whether (and, if so, how) Section II is rel­evant for application to Canadian fitness hearings requires further expla­nation, although inclusion of this "appreciation" factor clearly increases the instrument's potential application to U.S. definitions of competence to stand trial.

Other significant differences in format between the FIT-R and the MacCAT-CA include the FIT-R's focus on the defendant's own circum­stances (compared to the MacCAT-CA's reliance on questions about a hypothetical criminal case), as well as the FIT-R's far more diverse set of questions to explore the nature of defendants' abilities. In contrast, the MacCAT-CA's more narrow and more highly standardized format allowed for the development of highly objective scoring criteria. These are sacrificed in the FIT-R, which uses more general opinion rating criteria, in favorof allowing theinfluence ofclinicaljudgmentthatthe MacCAT-CA's strict "correct" or "incorrect" scoring criteria do not allow. Neither of these approaches is necessarily better than the other, but each has its own advantages and limitations. For example, in principle, the MacCAT-CA's more highly standardized format offers greater potential for inter-examiner reliability, while the FIT-R provides greater opportu­nities for individualized deficits (and their explanations) to arise in the fitness interview.

Psychometric Development

standardization. Questions are asked in a standardized sequence, although probing of examinees' answers are at the examiner's discretion. Rating criteria for the various items are in the form of a paragraph description of the concept covered by each item, but without specific guides for 2,1, or 0 ratings.

reliability. In the only study of interrater reliability of the FIT-R (f), 49 professionals (psychiatrists, psychologists, graduate students, and nurses) were provided training in the FIT-R, then asked to perform FIT-R ratings for defendants observed in two FIT-R videotaped interviews. Intraclass correlation coefficients were calculated separately for each profession and showed only occasional differences between professions. Averaged across professions, correlations for most items were in the.80 to.95 range, with a few in the.70s. Only one item fell below this average ("Appraisal of Likely Outcome,".67). Average correlations for the three Section ratings were not as good (I =.70, II =.54, III =.59), but the aver­age correlation for the overall judgment of fitness was.98.

norms. The FIT-R manual provides no norms, and none have been published to date in other sources.

critique. The lack of specific scoring criteria apparently decreases the degree of interrater reliability found on the MacCAT-CA for some items. Most are satisfactory, however. More troublesome are the relatively low agreements on the Section II and III summary ratings. It is also difficult to understand why the overall judgments of fitness with the FIT-R manifests such high agreement, when there is only modest agree­ment between raters on the sections that represent the major components examiners are supposed to take into account when making a final fitness rating. Further research might profitably seek an explanation for this anomaly.

Construct Validity

In a study in British Columbia involving 76 male forensic inpatients remanded for fitness evaluations, FIT-R scores were moderately (but sig­nificantly) related to scores on two other measures of "Competence to Confess" and "Competence to Plead Guilty" (g). Fitness/Unfitness on the FIT-R corresponded with competence/incompetence conclusions based on the competence to confess measure 85% of the time, and with compe- tence/incompetence conclusions based on the competence to plead guilty measure 81% of the time.

Both the MacCAT-CA and the FIT-R were administered to 100 male forensic inpatients remanded for fitness evaluations (i). There was con­siderable overlap in results, in that 28% were found to be "impaired" on both instruments and 48% were found "unimpaired" on both instru­ments (76% agreement). But more patients were found impaired on the MacCAT-CA (48%) than on the FIT-R (32%). The report did not specifi­cally describe how a conclusion of impairment on either instrument was defined. But by extrapolation from tables in the report, it appears that patients were considered impaired on the FIT-R if they were rated as impaired on any one of the three Sections of the FIT-R, and impaired on the MacCAT-CA if they fell below the impairment cut-off scores (as defined in the MacCAT-CA: see previous MacCAT-CA review) on any one of the three main sections of the MacCAT-CA. This same report indi­cated that almost all of the patients were either unimpaired on all three sections of the FIT-R or impaired on all three sections; only 9% were impaired "selectively" on some section(s) and not others. In contrast, on the MacCAT-CA, 35% of the patients were "selectively impaired" on some section(s) but not others.

critique. Theoretically, there should be some degree of consistency in findings across various measures of fitness or competence in the criminal process. The results of these studies suggest that the FIT-R produces find­ings that are consistent with this theoretical notion. However, the differences between the FIT-R and the MacCAT-CA in the comparative study could have been found for either of two reasons. First, the FIT-R uses examiners' judgments to make ratings of abilities, while the MacCAT-CA uses more highly structured scoring criteria. A tendency for greater similarity in a defendant's ratings across the three sections of the FIT-R might arise if raters form general impressions of a defendant's fitness that then influence their judgments in their ratings for the three sections. Second, it is possible that the three sections of the FIT-R simply measure abilities that are less distinct from each other than are the sections of the MacCAT-CA. Factor analysis of the FIT-R might help to address the likelihood of this second interpretation.

Predictive of Classificatory Utility

Although two reports have compared the FIT-R to clinicians' inde­pendentjudgments about patients' fitness, the earlier of these reports (h) used a sample that appears to have been incorporated later into a larger sample in the second report (j). Therefore, only the results of the second report are described here.

Two samples were included, obtained from the same Canadian foren­sic inpatient facility where men were remanded for fitness evaluations (during two different years). The first included 250 inpatients, and the second included 100 who were diagnosed with psychotic disorders. The FIT-R and the Structured Clinical Interview for DSM-III-R—Patient Edition (SCID-P) were administered to the first sample, while the FIT-R and the BPRS were administered to the second. Fitness evaluations and opinions on these patients were also obtained from forensic examiners at the facilities who had no knowledge of the results of the above measures. The independent examiners reached unfitness conclusions for only 4% of the first sample and 10% of the second sample, while the FIT-R produced unfitness conclusions for 13% of the first sample and 32% of the second sample. For both samples, only 2% of patients were identified as fit by the FIT-R and unfit by the institution's examiners.

critique. The authors of the FIT-R have pointed out that while the instrument can serve as part of a broader fitness evaluation, it may also be used as a screening tool. As with all screening instruments, it would be expected to identify more patients as impaired—as it did in this study— than might be found unfit in comprehensive evaluations. This same result could be found, however, if the FIT-R is more sensitive to the full range of abilities associated with the legal construct of fitness than are forensic clinicians. In other words, there is no guarantee regarding the integrity or superiority of the standard to which the FIT-R was being compared.

Potential for Expressing Person-Situation Congruency

The FIT-R manual does not offer specific instructions to consider the degree of the defendant's abilities in relation to the demands of the defen­dant's specific legal situation. However, the flexibility of the rating system would allow for such judgments to be made, given that the examiner has some knowledge of the defendant's trial circumstances.

References

(a) Bagby, R., Nicholson, R., Rogers, R., & Nussbaum, D. (1992). Domains of competence to stand trial: A factor analytic study. Law and Human Behavior, 16, 491-506.

(b) McDonald, D., Nussbaum, D., & Bagby, R. (1991). Reliability, validity and utility of the Fitness Interview Test. Canadian Journal of Psychiatry, 36, 480-484.

(c) Roesch, R., Jackson, M., Sollner, R., Eaves, D., Glackman, W., & Webster, C. (1984). The Fitness to Stand Trial Interview Test: How four professions rate videotaped fitness inter­views. International Journal of Law and Psychiatry, 7, 115-131.

(d) Roesch, R., Webster, C., & Eaves, D. (1984). The Fitness Interview Test: A method for exam­ining fitness to stand trial. Toronto, Ontario, Canada: Research Report of the Centre of Criminology, University of Toronto.

(e) Roesch, R., Zapf, P., Eaves, D., & Webster, C. (1998). Fitness Interview Test (Revised Edition). Burnaby, British Columbia, Canada: Mental Health, Law and Policy Institute, Simon Fraser University.

(f) Viljoen, J., Roesch, R., & Zapf, P. (in preparation). Interrater reliability of the Fitness Interview Test across four professional groups. Burnaby, British Columbia, Canada: Simon Fraser University.

(g) Whittemore, K., Ogloff, J., & Roesch, R. (1997). An investigation of competence to par­ticipate in legal proceedings in Canada. Canadian Journal of Psychiatry, 42, 869-875.

(h) Zapf, P., & Roesch, R. (1997). Assessing fitness to stand trial: A comparison of institu­tion-based evaluations and a brief screening interview. Canadian Journal of Community Mental Health, 16, 53-66.

(i) Zapf, P., & Roesch, R. (2001). A comparison of the MacCAT-CA and the FIT for making determinations of competence to stand trial. InternationalJournal of Law and Psychiatry, 24, 81-92.

(j) Zapf, P., Roesch, R., & Viljoen, J. (2001). Assessing fitness to stand trial: The utility of the Fitness Interview Test (Revised Version). Canadian Journal of Psychiatry, 46, 426-432.

Competence Assessment for Standing Trial for

Defendants with Mental Retardation (CAST-MR)

Authors

Everington, C., & Luckasson, R.

Primary Author Affiliation

Department of Psychology, Miami University (Ohio).

Primary Reference

Everington, C., & Luckasson, R. (1992). Competence Assessment for Standing Trial for Defendants with Mental Retardation: Test manual. Worthington, OH: IDS Publishing Corp.

Description

The Competence Assessment for Standing Trial for Defendants with Mental Retardation (the CAST-MR) (c) was developed to respond to needs that were considered unique for defendants with mental retardation. The instrument forms the basis of a structured interview. It consists of 50 ques­tions organized in three sections:

• I: Basic Legal Concepts

• II: Skills to Assist Defense

• III: Understanding Case Events

Section I includes 25 questions that focus on assessing defendants' "knowledge of the criminal justice process" (c, p. 3). In general these ques­tions use the words "What is ______________ " or "What does_______ mean" or

"What does it mean to-------," and they focus on the following:

• The roles of several of the persons in trials

• The meanings of words and concepts (hearing, sentence, maximum/ minimum sentence, crime, guilty, innocent, acquitted, felony, mis­demeanor, plead guilty, penitentiary, time served, probation, plea bargain, fine)

• The meanings of certain functions (What happens when you go to court, in a trial, when a prosecutor asks you questions, when you go to a penitentiary)

Section II consists of 15 questions addressing defendants' "under­standing of the client-attorney relationship" (c, p. 3). The items are ques­tions that pose circumstances ("What if------------------------------- " or "Let's pretend------- ") to

which the defendant is asked to respond with an appropriate answer: for example, "What if your lawyer asks you to do something you don't want to do like getting a haircut? What would you do?" "Let's pretend that you took something from the store and you got arrested for it. You didn't mean to do it, and you felt really bad about it. When your lawyer asks you if you did it, what would you do?" Some of the questions focus on attorney-client interactions, some on events outside the courtroom (e.g., talking to other detainees about one's situation), and many focus on potential responses to events in the courtroom (including witnessing events or testifying at one's own trial).

Each of the questions in Sections I and II has three multiple choice response options. For each question, the defendant is asked to listen to all three optional answers, then select "the one that is right." In contrast, Section III consists of 10 open-ended, very brief questions focusing on defendants' ability to "discuss the facts concerning the incident in a coher­ent manner and to understand the relationship between the alleged facts in the case and the subsequent arrest and charges" (c, p. 3). The questions are more or less chronological in orientation, beginning with "What were you doing that caused you to be arrested," continuing through questions focusing on where and when the events occurred, what happened when the police arrived, and the charges that were filed.

Administration instructions include practice questions to give exam­inees prior to each section, in order to orient examinees to the task and to determine whether they can manage the response format. The manual provides instructions for how to manage cases in which examinees have difficulty grasping the nature of the task. The materials include an Examiner Form that has the questions, response options, and indicators for the correct answers, as well as a Subject Form that is given to the examinee during the evaluation. The Subject Form provides the questions and the multiple choice options visually (without correct answers indi­cated) while the examiner reads the same material aloud to the examinee, to which the examinee responds verbally. Questions may be read to examinees up to three times for Sections I and II, and up to two times in Section III, and certain allowable non-leading prompts are described in the manual.

Scoring of responses in Sections I and II simply requires assigning 1 point for each correct answer. Thus subscores range from 0 to 25 on Section I and 0 to 15 on Section II. Responses to Section III questions are scored according to a more detailed set of scoring criteria provided in the manual. Each Section III item may receive 0,.5, or 1 credit, allowing for a subscale III range of 0 to 10. Total scores may be calculated (the sum of points on all 50 items), ranging from 0 to 50.

Conceptual Basis

concept definition. The manual indicates that the areas of knowl­edge and ability related to the three sections were chosen in light of case law and psychological literature on competence to stand trial (c), but further detail is not provided.

Operational definition. Decisions about operationalizing the three concepts were based substantially on the authors' intentions to construct a competence to stand trial instrument that would be especially suitable for administration to persons with mental retardation. A multiple-choice format, for example, was considered superior to an open-ended format because it was expected to place less of a demand on the respondent to answer independently. In addition, vocabulary and syntax were kept to the simplest possible level so that persons with mental retardation would not be unfairly penalized because of their lower level of linguistic ability.

To refine the format and content of the instrument, the authors asked 10 experts on criminal disability law to rate various parts of the CAST-MR for legal appropriateness of content, format with reference to the abilities of persons with mental retardation, and usability (e.g., clarity) (a). In gen­eral, ratings on the various features of the instrument were primarily "good" or "excellent," although some respondents felt that the format might be too difficult for persons with mental retardation. Readability analyses placed the reading level of the text at about grade 2 to grade 6.

critique. The types of abilities that directed the content development of the CAST-MR are much like those of other instruments for use in compe­tence to stand trial evaluations. Like most such instruments, it was not developed to assess all relevant abilities. But the variety of the questions, especially in Section II (how defendants would respond to hypothetical situations from arrest to sentencing), is somewhat greater than in other instruments of its type.

The notion to develop an instrument specifically for defendants with mental retardation is novel and raises two general points. First, the logic for using a multiple choice format seems clear. Persons with mental retar­dation often have language deficits, making it difficult for them to express answers in open-ended response formats despite the fact that they might actually comprehend the information or concept for which they are being tested. A multiple choice format removes the necessity for constructing one's answers. This is not to say that the multiple choice format, however, makes for an easier response format. Choosing from among three answers, after they are presented to the individual verbally, requires audi­tory reception and retention, as well as a subsequent review and compar­ison of the three options, in the process of selecting the correct answer. This process itself is relatively more difficult for persons with mental retardation than for persons of average cognitive and functional capacity.

Second, given that the multiple choice format is successful in remov­ing the effect of expressive deficits when assessing defendants' knowl­edge and functional responses to trial circumstances, one must recognize that this response format is a bit further removed from the types of func­tioning that are actually needed during trial participation. During the pretrial and trial process, defendants frequently will not be offered alternative interpretations of what is happening to them. In addition, they may sometimes be required to employ those same receptive and expressive abilities that the format of the CAST-MR was designed to avoid.

This being said, the logic of the multiple choice format is sound in theory, in that it is likely to help avoid penalizing persons with mental retardation who may grasp the nature of the trial and its participants yet have difficulty expressing what they know. Moreover, Section III does provide the examiner the opportunity to test whether the defendant has the expressive capacities necessary to communicate essential facts and defendant opinions to counsel in construction of a defense.

Psychometric Development

standardization. Administration of the CAST-MR is described very clearly, including the invariant sequence of the items and the allow­able probes (c). Scoring is absolute on Sections I and II, requiring no exam­iner judgment. The scoring criteria for Section III (involving open-ended responses that require scoring judgments on the part of the examiner) are clearly defined in the manual, which, for each Section III item, offers general and specific criteria as well as scoring examples for 0,.5, and 1 point responses.

reliability. Interscorer agreement for Section III items (Section I and II provide scores requiring no examiner judgment) was examined in two studies. In one (a), observations between pairs of scorers on 10 cases manifested 80% agreement on individual items and 88% agreement on total Section III score. Another study (b) compared 11 scorers (who had obtained half-day training) to the CAST-MR authors' scores on at least three cases per scorer. Mean level of agreement was reported as 87%. Concerning test-retest reliability, a Pearson correlation of.90 was found between two administrations of the CAST-MR (two week interval) for 23 persons in group homes for persons with mental retardation (a).

norms. The manual (c) provides means and standard deviations for groups of individuals with and without mental retardation, as described later in "Construct Validation."

critique. Standardization of the CAST-MR is quite good, allowing for adequate interscorer reliability and reasonable confidence in the abil­ity of different examiners to produce similar results with a given exami­nee. Normative data currently are based on relatively small samples (see "Construct Validation").

Construct Validation

Concerning internal consistency, two studies involving persons with and without mental retardation reported Cronbach's alpha (a) and Kuder- Richardson coefficients (b) for the various sections of the CAST-MR. These were reported as follows:.93 and.92 for Section I;.76 and.73 for Section II; and.83 and.84 for Section III.

In one study (b), IQ was correlated substantially with scores on CAST-MR Sections I (.64), II (.54) and III (.59), and with total CAST-MR scores (.66). Correlations with other defendant characteristics have not been reported.

The first validation study (a) for the CAST-MR compared four groups of defendants who were expected to manifest progressively lower CAST-MR scores:

• A: defendants with no mental retardation (n = 46),

• B: defendants with mental retardation but not referred for compe­tence evaluations (n = 24),

• C: defendants with mental retardation who were referred and found competent (n = 12), and

• D: defendants with mental retardation who were referred and found incompetent (n = 11).

A significant group effect was found for all 3 Section subscores and the total CAST-MR scored. The mean for Group D (mental retardation and found incompetent) was significantly different from those of all other groups on all four of the comparisons, and Group A's means (no mental retardation) were significantly different from those of Groups B and C except for Section III scores (A was not significantly different from B).

The second CAST-MR validation study (b) compared CAST-MR means for defendants with mental retardation whom clinicians recom­mended as competent (n = 15) and incompetent (n = 20). Clinicians mak­ing the competence judgments were specialized forensic examiners who did not have knowledge of the CAST-MR scores. Statistically significant differences were found between the two groups, with the "incompetent" group having lower mean scores for each of the sections of the CAST-MR and for total CAST-MR score. Seven items (3 from Section I and 3 from Section III) manifested significant differences between the two groups at or above p <.01.

critique. The sample sizes of the validation studies are somewhat small. But within those limits, internal consistency appears to be satisfac­tory, and CAST-MR scores appear to be related to intellectual functioning (and mental retardation status) in a manner that would be expected based on theoretical inference. An examination of the absolute differences in mean scores, however, indicates that mean differences between compe­tent and incompetent groups were not substantial, and it is interesting that only 7 of the 50 items appeared to contribute to the mean differences on the subscales. The stability of those findings is worth determining in future research.

Predictive or Classificatory Utility

In the first validation study (a), a discriminant function analysis correctly classified 9 of the 16 defendants judged competent by independ­ent examiners and 7 of the 11 defendants judged incompetent. In the second validation study (b), better agreement between cut-off scores and independent examiner judgments were produced by Section I and Section II scores than by Section II or total CAST-MR scores. On Sections I and II, scores above 13 on Section I, and scores above 9 on Section II, produced 71% agreement. The pattern for both of these sections indicated that the CAST-MR was able to correctly classify about two-thirds of defendants judged competent and two-thirds of the defendants judged incompetent.

critique. These results may not appear encouraging. On the other hand, they are similar to results found with other competence to stand trial screening instruments (reported elsewhere in this chapter). Moreover, the authors (b) correctly note that clinicians' judgments about competence among persons with mental retardation are probably less reliable than for defendants with mental illnesses. Thus one does not know whether the less-than-satisfying results of the CAST-MR's classifi­cation validity is due to error in the instrument or error in the criterion to which it was being compared.

Future research might examine the relation of the CAST-MR to competent and incompetent groups of persons with mental retardation in interaction with other variables. For example, persons with mental retardation might differ in capacities relevant for defendant participation as a function of age and legal experience. Perhaps the CAST-MR might play a more effective role in discriminating between competent and incompetent defendants when variance related to these other variables is controlled.

Potential for Expressing Person-Situation Congruency

The CAST-MR offers no quantitative method for comparing specific types of deficits of defendants in relation to the specific demands posed by their own trial circumstances. However, one of the interesting features of the CAST-MR is its range of content and its potential for qualitative interpretation. For example, Section II uses items that focus on a number of specific circumstances that are more or less likely to arise depending on the nature of the defendant's case, such as being questioned by the prose­cutor or having to testify in a manner consistent with previous counsel between attorney and client. Defendants' responses to those items may be of greater importance in cases where defendant testimony is especially likely to be needed, offering information for interactive interpretation apart from CAST-MR scores.

References

(a) Everington, C. (1990). The Competence Assessment for Standing Trial for Defendants with Mental Retardation (CAST-MR): A validation study. Criminal Justice and Behavior, 17, 147-168.

(b) Everington, C., & Dunn, C. (1995). A second validation study of the Competence Assessment for Standing Trial for Defendants with Mental Retardation (CAST-MR). Criminal Justice and Behavior, 22, 44-59.

(c) Everington, C., & Luckasson, R. (1992). Competence Assessmentfor Standing Trialfor Defendants with Mental Retardation (CAST*MR): Test manual. Worthington, OH: IDS Publishing Corporation.

Georgia Court Competence Test-Mississippi State

Hospital (GCCT-MSH)

Authors

Wildman, R., Batchelor, E., Thompson, L., Nelson, F., Moore, J., Patterson, M., & deLaosa, M.

Primary Author Aifiliaiion

Central State Hospital, Milledgeville, GA

Primary Reference

For the original GCCT: Wildman et al. (1978). The Georgia Court Competence Test: An attempt to develop a rapid, quantitative measure of fitness for trial. Unpublished manuscript: Forensic Services Division, Central State Hospital, Milledgeville, GA.

For the GCCT-MSH: Nicholson, R., Briggs, S., & Robertson, H. (1988). Instruments for assessing competence to stand trial: How do they work? Professional Psychology: Research and Practice, 19, 383-394. (GCCT-MSH published as an Appendix to the article.)

Description

The Georgia Court Competence Test-Mississippi State Hospital (GCCT- MSH) (b) is a revision of the original Georgia Court Competence Test (GCCT) by Wildman et al. (h). The original GCCT was intended to provide "a rapid, quantitative measure of the knowledge and skills necessary for competence for trial" (h, p. 2). It was developed at the Forensic Services Division of Central State Hospital in Millidgeville, Georgia and first described in 1978 (h). According to Wildman (personal communication), the GCCT was not developed with the intention of widespread dissemi­nation and use; it evolved from experience and local need, rather than from a research project. Nevertheless, it came into use by other mental health professionals, primarily in southern states. At least two revisions of the GCCT evolved in the 1980s, one of them by psychologists at Mississippi State Hospital. That revision, the GCCT-MSH, has been most often cited in subsequent research and practice. This review examines only studies performed with the GCCT-MSH, not the GCCT.

The original GCCT consists of 17 questions that are asked of the exam­inee in the context of a pretrial competence assessment by the examiner, usually requiring about 10 minutes for its administration. The questions are grouped in the following six categories and sequence:

1. Picture of Courf. Seven questions focus on the examinee's descrip­tion of the location of participants in a courtroom (e.g., "Where does the Judge sit?" "Where will the witness sit to testify?"). The questions are accompanied by a drawing of a courtroom that does not contain depictions of participants themselves. The exami­nee points to the area of the drawing to signify answers to the questions.

2. Functions. Five questions inquire about the functions of the Judge, jury, defendant's lawyer, prosecutor, and witnesses (e.g., "What does the jury do?").

3. Charge. Two questions, 'What are you charged with?" and "What does that mean?"

4. Helping the lawyer. One question asks how the examinee plans to help the examinee's lawyer in a defense.

5. Alleged crime. One question asks the examinee to describe "what actually happened about the charge you are here on?"

6. Consequences. One question asks what the examinee expects will be done to him/her if a guilty verdict is returned.

The GCCT-MSH differs from the GCCT in only two ways. First, 4 new questions were added in order to provide additional information about defendants' knowledge of courtroom procedures and ability to assist counsel. Specifically, 2 questions were added to #2 above, asking what is done by people watching the trial, and what "you will do during the trial." Two other questions were added to #4 above ("What is your attorney's name?" and "How can you contact him/her?"). Second, some of the maximum possible scores for each item were changed, and scoring criteria for questions about charges and possible penalties were redefined for improved clarity.

The maximum points possible vary from item to item, with the total possible being 50. Brief scoring criteria are provided for each question that does not have an obvious correct or incorrect answer. Points are multi­plied by 2 to obtain a total score between 0 to 100. The original GCCT used 69 or below to suggest possible incompetence, and the GCCT-MSH has retained this cut-off (b).

Conceptual Basis

concept definition. The authors' examination of competence to stand trial standards, and the opinions of forensic experts in several states, led them to the following conceptual definitions for abilities related to legal competence to stand trial: knowledge of the charge; knowledge of the possible penalties; some understanding of courtroom procedure; and the ability to communicate rationally with an attorney in the preparation of a defense.

operational definition. Test items were developed logically to relate to the four conceptual definitions. Scoring decisions allowed some items to contribute a greater proportion to the total score than others, based on the authors' perceptions of the relative importance of each ability when one assumes the presence of a competent and helpful attorney (h).

critique. The logical derivation of the concepts to be measured has not been described fully by the authors. The concepts themselves, however, seem to grasp some minimum of abilities relevant to pretrial competence.

Differences of opinion might arise concerning the criteria for ade­quacy or inadequacy of certain examinee responses. On the item, "What do witnesses do?" for example, some test users might disagree that the response, "Answers questions about the case," demonstrates adequate understanding of the witness function, although this receives full credit by the test's scoring criteria.

The choice of a picture stimulus (courtroom) for one set of items is unique among competence to stand trial instruments. The authors did not explain the reasoning for this choice; one possible rationale for a picture stimulus is the structure it provides early on in the test process, as well as establishing a concrete visual image that can facilitate exploration of the examinee's overall knowledge of the trial process.

Psychometric Development

standardization. Specific questions are asked in a given sequence. Scoring criteria (right and wrong answers) are briefly characterized, with example responses offered for some items.

reliability. Two studies have examined inter-rater reliability. One reported an intraclass correlation of.95 between two trained raters (c) and the other.82 (e). No test-retest reliability figures have been published for the MSH version. Test-retest reliability for the original GCCT was reported as.79 when the instrument was first developed (h). But it is uncertain whether this can be generalized to the GCCT-MSH, due to some changes in its scoring criteria.

norms. No norms as such have been published. But mean GCCT- MSH scores for a group of defendants in a competence restoration pro­gram were 81,8 (SD = 8.8) for defendants considered restored and 43.2 (SD = 18.4) for defendants continuing to be seen as incompetent (g). Another study involving defendants referred for CST evaluations reported mean scores of 81.2 (SD = 18.3) for defendants considered com­petent and 51.3 (SD = 24.8) for those considered incompetent (d).

critique. Overall the GCCT-MSH appears to have adequate inter­rater reliability and internal consistency. It should be noted that the study (e) reporting lower alpha coefficients also reported lower inter-rater relia­bility, and the latter would typically result in lower alpha coefficients. Alternatively, the lower alpha for that study might be due to sample differences, over half of the subjects having not been drawn from the CST evaluation or restoration processes that characterized samples in the comparison studies.

Construct Validity

Internal consistency was demonstrated by alpha coefficients of.88 and.89 in two studies involving defendants in the CST evaluation or restoration process (b, g). But a marginal alpha of.70 was found in a study for which half of the subjects were mentally ill offenders not involved in the CST process (e). An average item-to-total score correlation has been reported as ranging from.47 to.67 (b), and inter-item correlations have averaged.28 (g) and.36 (b).

Two factor analyses of the GCCT-MSH, with defendants referred for CST evaluations, independently arrived at three-factor solutions with factors representing (I) most of the courtroom location items, (II) items involving the functions of participants, and (III) items involving the circumstances of one's own trial, charges and penalties (a, b). Two other factor analyses with different types of samples, however, claimed that a two-factor solution better fit the data (f, g). One of these studies found that the two-factor solution outperformed the aforementioned three- factor solution in a confirmatory factor analysis applied to a sample of defendants in a CST restoration program (f).

The GCCT-MSH was identified in one study (of defendants referred for CST evaluations) as correlating significantly (negatively) with mental retardation and with race (being non-white) but not with psychosis, and positively with years of education (b). In a sample of persons involved in CST restoration, multiple analysis of variance found no relation between GCCT-MSH competent and incompetent patients on any of the diagnoses provided by the SCL-90-R (g). In the same study, however, low IQ, psychotic disorder, and non-psychotic affective disorder did account for significant portions of the variance in GCCT-MSH competent and incompetent classifications.

Finally, the relation of the GCCT-MSH to the Competence Screening Test (CST) (see review in this chapter) was examined in two studies. One found a correlation of.76 between continuous scores on the two measures (b), while the other found a correlation of.47. In the latter study, the two instruments agreed on competence and incompetence classification (using their recommended cut-off scores) in only 63% of 85 cases.

critique. The results of the factor analyses of the GCCT-MSH are very difficult to interpret. This is in part due to reporting problems; for example, one study (g) attempted to compare its factor analysis to that of Nicholson et al. (b), but the factor loadings from Nicholson et al. that were "reproduced" in the report were not the same as those shown in the Nicholson et al. article itself. Difficulty in interpreting the results is also due to differences in samples. The two studies involving defendants referred for CST evaluations are in agreement on a three-factor structure, while the two studies using other types of samples (Ustad et al.'s restora­tion patients, and Rogers et al.'s mentally ill offenders with no referral for CST) claim the superiority of a different solution. For examiners concerned about the use of the GCCT-MSH with defendants referred for CST evaluations, it is reasonable to give greater weight to the results of the two studies that employed samples drawn from that population.

Similarly, the GCCT-MSH correlated substantially with another com­petence instrument (the CST) in a study involving defendants referred for competence evaluations, but not in a study involving patients who had all been found incompetent at one time and were now in various stages of restoration. These conflicting results either render the issue equivocal, or one accepts the results as population-specific. At any rate, if GCCT-MSH scores correspond only very imperfectly to classifications provided by the CST instrument, this does not argue against the concurrent validity of the GCCT-MSH, since the validity and utility of the CST itself has been ques­tioned independent of these results (see review of the CST instrument in this chapter).

Predictive or Classificatory Utility

In one study (resulting in two journal reports), the GCCT-MSH cut­off score produced classifications of 140 defendants that correlated.44 with independent forensic professional staff decisions about competence and incompetence (b), identifying as incompetent 10 out of 14 persons (71%) who were classified incompetent by staff (b). The GCCT-MSH over­identified incompetence, however, so that the instrument was "right" only 3 in 10 times when it classified defendants as incompetent.

In a meta-analytic study, Nicholson and Kugler (c) reported that GCCT scores correlated on average.42 with independent clinical judg­ments of competence, based on results of four reports: the two identified above (which actually reported a single study), as well as the original Wildman et al. study of the GCCT and an unpublished study.

critique. The correspondence of the GCCT-MSH to clinical judg­ments about competence is relatively uncertain, in that there has been only one study examining the present MSH version of the GCCT. That result suggests that the GCCT-MSH over-predicts incompetence, a result that is not of particularly great consequence in that the instrument is intended as a screening tool, resulting simply in further evaluation for examinees scoring below the cut-off score. One can tolerate a significant number of false positives under those circumstances.

Potentially more troubling is the fact that the study suggested a sen­sitivity rate of only 71%. This would mean that if the instrument were used to determine who would receive more comprehensive evaluations, it would fail to "screen in" more than one-quarter of the incompetent defendants in the population (who, hypothetically, would proceed to trial because their potential for incompetence had failed to be identified). Concern about this figure is mitigated, however, by two things. First, the baserate of incompetence was very low in this study sample, thus making the predictive task very difficult for any instrument. Second, there is no reason to believe that the professional judgments about incompe­tence, the criterion to which the GCCT-MSH was being compared, were themselves reliably made or were valid as indicators of defendants' actual capacities.

Potential for Expressing Person-Situation Congruency

The GCCT-MSH includes no assessment of the trial situation facing the examinee, and the primary reference does not mention examination of responses in light of the specific demands and circumstances of the examinee's future legal situation.

References

(a) Bagby, R., Nicholson, R., Rogers, R., & Nussbaum, D. (1992). Domains of competence to stand trial: A factor analytic study. Law and Human Behavior, 16, 491-507.

(b) Nicholson, R., Briggs, S., & Robertson, H. (1988). Instruments for assessing competence to stand trial: How do they work? Professional Psychology: Research and Practice, 19, 383-394.

(c) Nicholson, R., & Kugler, K. (1991). Competent and incompetent criminal defendants: A quantitative review of comparative research. Psychological Bulletin, 109, 355-370.

(d) Nicholson, R., Robertson, H., Johnson, W., & Jensen, G. (1988). A comparison of instru­ments for assessing competence to stand trial. Law and Human Behavior, 12, 313-321.

(e) Rogers, R., Grandjean, N., Tillbrook, C., Vitacco, M., & Sewell, K. (2001). Recent inter­view-based measures of competence to stand trial: A critical review augmented with research data. Behavioral Sciences and the Law, 19, 503-518.

(f) Rogers, R., Ustad, K., Sewell, K., & Reinhardt, V. (1996). Dimensions of incompetence: A factor analytic study of the Georgia Court Competence Test. Behavioral Sciences and the Law, 14, 323-330.

(g) Ustad, I., Rogers, R., Sewell, K., & Guarnaccia, C. (1996). Restoration of competence to stand trial: Assessment with the Georgia Court Competence Test and the Competence Screening Test. Law and Human Behavior, 20, 131-146.

(h) Wildman, R., Batchelor, E., Thompson, L., Nelson, F., Moore, J., Patterson, M., & deLaosa, M. (1980). The Georgia Court Competence Test: An attempt to develop a rapid, quan­titative measure for fitness for trial. Unpublished manuscript, Forensic Services Division, Central State Hospital, Milledgeville, GA.

Competence to Stand Trial Assessment Instrument (CAI)

Authors

McGarry, A.L. and Associates, Laboratory of Community Psychiatry, Harvard Medical School

Primary Author Affiliation

Harvard Medical School

Primary Reference

Laboratory of Community Psychiatry, Harvard Medical School (1973). Competence to stand trial and mental illness. Rockville, MD: NIMH, Department of Health, Education, and Welfare: DHEW Publication No: (ABM) 77-103:

note on second edition review. Since 1986, only one research article has published results using the CAI. Thus the present review offers little more than did the first edition with regard to empirical support for the instrument. The review is included here primarily for its historical value, because the CAI influenced the development of almost all other instru­ments for competence to stand trial evaluations. The DHEW monograph that constituted the CAI manual is no longer available from NIMH, although it is still found in some university libraries.

Description

The Competence to Stand Trial Assessment Instrument (CAI) was devel­oped to deliver clinical opinion to the court in language, form, and sub­stance sufficiently common to the disciplines involved to provide a basis for adequate and relevant communication. The purpose of the instrument was to standardize, objectify, and quantify, the relevant criteria for compe­tence to stand trial (d, p. 99). An interdisciplinary team of psychiatrists, psychologists, and lawyers developed the instrument at Harvard's Laboratory of Community Psychiatry during a project funded by a research grant from the Center for Studies of Crime and Delinquency, NIMH. It was intended as a companion instrument for the Competence Screening Test (described later in this chapter).

The CAI describes 13 functions related to a defendant's "ability to cope with the trial process in an adequately self-protective fashion" (d, p. 99). These functions are examined in an interview procedure that employs the 13 functions as items, with two or three recommended inter­view questions for each item. The following are brief descriptions of the functions and their content, as well as a sample question for each of them (worded as in d, pp. 101-114):

1. Appraisal of available legal defenses: the accused's awareness of his possible legal defenses and how consistent these are with the real­ity of his particular circumstances ("How do you think you can be defended against these charges?").

2. Unmanageable behavior: the appropriateness of the current motor and verbal behavior of the defendant and the degree to which this behavior would disrupt the conduct of a trial ("What do you think would happen if you spoke out or moved around in the courtroom without permission?").

3. Quality of relating to attorney: interpersonal capacity of the accused to relate to the average attorney ("Do you have confidence in your lawyer?").

4. Planning of legal strategy including guilty pleas to lesser charges where pertinent: degree to which the accused can understand, partici­pate, and cooperate with his counsel in planning a strategy for the defense that is consistent with the reality of his circumstances ("Is there anything that you disagree with in the way your lawyer is going to handle your case, and if so, what do you plan to do about it?").

5. Appraisal of role of: (a) defense counsel, (b) prosecuting attorney, (c) judge, (d) jury, (e) defendant, (f) witnesses ("In the courtroom, during a trial, what is the job of...").

6. Understanding of court procedure: degree to which the defendant understands the basic sequence of events in a trial and their import for him ("After your lawyer is finished asking you ques­tions on the stand, who then can ask you questions?").

7. Appreciation of charges: the accused's understanding of the charges against him and, to a lesser extent, the seriousness of the charges ("Do you think people in general would regard you with some fear on the basis of such a charge?").

8. Appreciation of range and nature of possible penalties: the accused's concrete understanding and appreciation of the conditions and restrictions which could be imposed on him and their possible duration ("If you're found guilty as charged what are the possi­ble sentences the judge could give you?").

9. Appraisal of likely outcome: how realistically the accused perceives the likely outcome and the degree to which impaired understand­ing contributes to a less adequate or inadequate participation in his defense ("How strong a case do they have against you?").

10. Capacity to disclose to attorney available pertinent facts surrounding the offense: the accused's capacity to give a basically consistent, rational, and relevant account of the motivational and external facts ("Tell us what actually happened, what you saw and did and heard and thought before, during, and after you are sup­posed to have committed this offense").

11. Capacity to realistically challenge prosecution witnesses: the accused's capacity to recognize distortions in prosecution testimony ("Suppose a witness against you told a lie in the courtroom. What would you do?").

12. Capacity to testify relevantly: the accused's ability to testify with coherence, relevance, and independence of judgment (no exam­ple questions are provided in the primary reference).

13. Self-defeating versus self-serving motivation (legal sense): the accused's motivation to adequately protect himself and appropriately utilize legal safeguards to this end ("Suppose the District Attorney made some legal errors and your lawyer wants to appeal a guilty finding in your case-would you accept that?").

The interview is structured by the 13 functions and recommended questions. Instructions, however, urge clinicians to conduct the interview with appropriate clinical flexibility rather than as a recitation of questions. The authors suggest that examination and scoring with a defendant who is in relatively good contact with reality will require no more than one hour.

Each item is given a rating by the examiner, ranging from 1 (total lack of capacity to function) to 5 (no impairment, no question that defendant can function adequately), or 6 if the available data permit no rating within reasonable clinical certainty. Each rating pertains to the degree of impair­ment or functioning on that item (function) only; a rating of 1 on any sin­gle item does not necessarily constitute a clinical decision about overall incapacity to function in a trial setting. Rating proceeds on the assump­tion that the defendant will be adequately assisted by counsel, and that the professional who has administered the assessment interview has a basic understanding of realities of the criminal justice system. No specific rating criteria are provided, but clinical case examples are given for cer­tain ratings on each of the 13 functions/items.

The 13 scores are neither summed nor weighted. They are left to stand individually for use in forming testimony when combined with information from other sources. The authors note that the CAI is not intended to predict a defendant's performance in a future trial, because "with the passage of time and variations in clinical status, even from day to day, a given defendant will vary in the scores attained" (d, p. 100).

Conceptual Basis

concept definition. During the first year of a four-year project, the interdisciplinary team (including lawyers, psychiatrists, and psycholo­gists) performed reviews of appellate cases and legal literature, observed pretrial competence hearings, and interviewed attorneys and judges. This allowed them to arrive at a three-part definition of the competence to stand trial standard: an ability to cooperate with one's attorney in one's own defense; an awareness and understanding of the nature and object of the proceedings; and an understanding of the consequences of the proceedings.

The functions (items) and their definitions were derived logically from the same review and observation process. Early versions of the CAI used 15 functions, but 2 were eliminated because they seemed duplicative of other functions.

operational definition. The recommended interview questions for eliciting defendant information relevant for the 13 functions were devised and refined in a process involving administration of early drafts of the instrument to nonhospitalized defendants preparing for trial and com­mitted patients at Bridgewater State Hospital (Massachusetts). Further refinement occurred with administration of the CAI, at a later stage of its development, to Bridgewater examinees who had been committed to determine their competence to stand trial. Case examples to assist inter­viewers in assigning ratings on each function were drawn from experi­ences with the latter group of defendants.

critique. The method for arriving at the 13 functions of the CAI offers no empirical assurance that the instrument covers the range of potentially relevant functions associated with pretrial competence. Nevertheless, the listing seems relatively complete in its coverage on a rational basis, and each of the functions focuses the examiner on elements that seem to have considerable content relevance for the issues in compe­tence to stand trial determinations.

Controversy arose very soon after the publication of the CAI regard­ing the definitions of certain functions and their case examples. Brakel (a) argued that the rating criteria for certain functions contained presump­tions and prejudices that would penalize some defendants unfairly. For example, Brakel noted that one of the case examples under "Quality of Relating to Attorney" was of a black defendant who was rated a 4 ("mild degree of incapacity") because he said that he did not yet know whether he had confidence in his public defender, but that he did not think the attorney was very interested in his case. Other functions (e.g., "Appraisal of Available Legal Defenses") require examiners to use their own pre­sumptions concerning when legal strategies such as plea bargaining gen­erally are realistic or desirable. Brakel's argument, then, charged the CAI with bias of two types: bias against persons who, for political or personal reasons, simply do not have confidence in the criminal trial process; or bias due to examiner assumptions about the general nature of trial cir­cumstances and attorney performances, to which a defendant's response about specific attorneys or trial circumstances may be compared unfairly.

In defense of the CAI against these charges, its authors clearly speci­fied their intention to evaluate defendants' perceptions of attorneys, judges, and trial processes under the assumption that the defendant will be represented by competent counsel and tried by a legal standard of fair­ness. As long as this is made clear to the court during testimony, there would seem to be no need for concern about the presumptiveness of the examiner because the court may weigh the matter based on its better knowledge of the quality of counsel in the immediate case. In addition, in evaluating the defendant's perceptions of legal defenses, the examples provided in the CAI manual indicate that the examiner should pay special attention to the logical process employed by the defendant, more than the defendant's final choice of defense strategy.

It is unfortunate that the CAI was not developed with more specific guidelines concerning what questions to ask and which aspects of exami­nees' responses are critical for ratings. For example, the function "Planning for Legal Strategy" focuses the examiner on the degree to which the accused can understand and cooperate with counsel in planning a realistic strategy of defense. The related text focuses the examiner on the content area of plea bargaining, and the questions to be asked direct one toward discovering which defenses the examinee would prefer or accept. Only by extrapolation from the case examples does it become apparent that it is the defendant's reasoning, not simply the choice of defense, that contributes to the rating.

Psychometric Development

standardization. Few instructions are provided for administration of the CAI. For example, the authors provided no instructions in the pri­mary reference (d) for introducing the CAI interview to defendants, for the sequence with which the various functions should be assessed, or for the extent of additional inquiry that may be necessary beyond the basic questions to be asked. The standard questions themselves are described as "sample questions," and many of them require only a "yes" or "no" answer by the examinee. Undoubtedly, examiners were expected to probe beyond the responses to these questions. Consequently, it is likely that different examinees receive somewhat different stimulus questions from different examiners.

No specific criteria are provided for each of the 1 to 5 ratings on each function. Standardization of rating depends on paragraph definitions of the function and three case vignettes per function, for which the authors' ratings are provided.

reliability. Interrater reliability was established by examining intra­class correlations for a group of three experienced raters and a group of three inexperienced raters (d). In all cases, raters apparently observed (but did not themselves conduct) the interviews to be rated. Across CAI func­tions, correlations between the experienced raters ranged from r =.84 to.97 (average.92), compared to r =.43 to.96 (average.87) between inexpe­rienced raters. Correlations may have been attenuated by the fact that raters discussed the ratings they had given after viewing each interview. Currently there are no data on temporal consistency of CAI ratings or on reliability as a function of different interviewers administering the CAI.

norms. Neither norms nor mean scores for relevant samples is available.

critique. The CAI offers consistency and reliability for pretrial com­petence evaluations in the sense that it assures that each CAI-assisted evaluation will provide data on 13 functions. Beyond this, however, very little is known about the consistency or stability of CAI ratings. The rating system produced adequate interrater agreement for very experienced raters in the setting in which the CAI was developed. Yet without a description of standard instructions for CAI administration, one does not know whether one's own way of administering the CAI in everyday practice is similar to the interview procedure on which reliability ratings were performed. Published details of standard administration would be necessary before meaningful studies of interrater reliability could be performed.

Construct Validation

Schreiber (f) compared CAI-based recommendations for compe- tence/incompetence to those based on two other instruments designed to be relevant for the construct of competence to stand trial. The study employed actual defendants to whom the CAI was administered in con­junction with the Competence Screening Test and/or the Interdisciplinary Fitness Interview. Statistically significant agreement was found between the CAI and both of the other instruments, although the CAI tended to find a somewhat greater number of defendants incompetent than did the Interdisciplinary Fitness Interview.

Schreiber, Roesch, and Golding (c, g) reported that competence judg­ments based on the CAI corresponded to the consensus of expert panels in about 78% of cases. Schreiber (e) found that CAI-based evaluations rec­ommended findings of competence in all cases in which defendants had not received a psychotic diagnosis, and for incompetence in one-half of the psychotic cases. Finally, in one study, CAI scores were significantly related to behavioral observation ratings of defendants while they per­formed in the role of defendant in a mock trial setting (b).

critique. The existing research lends only meager evidence for the construct validity of the CAI. Nevertheless, it does appear to correlate with other instruments designed to assess the same construct underlying the development of the CAI. Further, it does not automatically produce incompetence findings for individuals diagnosed psychotic, although it does produce competence findings for defendants without significant psychosis. These results generally are consistent with legal presumptions about pretrial competence as a construct. The finding that CAI scores were related to observations of defendants' behaviors during mock trials is intriguing and may offer a method that can be used in future research on other measures of abilities related to competence to stand trial.

Predictive or Classificatory Utility

McGarry (d) reported that in a series of 15 cases in which the CAI formed the basis for expert testimony in Massachusetts Superior Court, judicial determinations of pretrial competence/incompetence matched the CAI-based recommendations in all but one case. The judicial deci­sions, however, were made with judicial knowledge of the CAI data. In another study in which CAI-based recommendations were not available to the court, those recommendations matched with the competence or incompetence decisions of the courts in 82% of the cases (c, f, g).

critique. The existing evidence suggests that the CAI produces results similar to independent court opinions about competence or incom­petence to stand trial. This is based on only one study, however, and it is not known whether CAI-based evaluations are any more effective in this regard than are traditional assessments for competence to stand trial. Courts' acceptance of CAI results in expert testimony, and legislative or administrative encouragement of its use, clearly has not been based on demonstrated empirical validity of the instrument.

There is some evidence, in fact, that it is the conceptual structure of the CAI rather than its operational or metric features that contributed to its acceptance in many legal and mental health systems. Schreiber (e) examined usage of the CAI and its companion instrument (the Competence Screening Test) in four states. He found that the CAI's 13 functions were being used to structure evaluators' observations and reports, but that the CAI's rating system itself either was actively opposed or simply was not being used. Further, in the survey inquiry that was performed for the first edition review of instruments in this book, many respondents reported using their own "modified McGarry interview," involving slightly altered versions of the CAI's 13-function outline without employment of the rating system.

The lack of research on the empirical validity of CAI ratings, there­fore, was probably a consequence of users' lack of interest in, or their judgments about the questionable value of, its quantification. The CAI continues to be used as a conceptual tool—a way to think about functional abilities—with apparently very little use of the instrument as a scorable tool. Indeed, the CAI does not really have the status of an "instrument." Without the use of the quantified rating system, and with no standardized description of the interview process, the CAI's sole remaining contribution is a list of 13 legally relevant concepts and definitions. Valuable as this may be, it neither calls for nor can receive empirical support without greater attention to the quantitative aspect of the original CAI.

Potential for Expressing Person-Situation Congruency

The CAI includes no systematic evaluation of the trial circumstances or attorney relations facing the examinee. Three of the functions/items, however, instruct the examiner to compare the examinee's knowledge, perceptions or capacities to the "reality of his particular circumstances" (Function 1, Appraisal of Available Legal Defenses; Function 4, Planning of Legal Strategy; Function 9, Appraisal of Likely Outcome). Thus for some functions, the CAI asks the examiner to evaluate the congruency or incongruency between the examinee's ability and these aspects of the demands or reality of the situation. Yet the CAI provides no dimensions or guidelines for characterizing the environmental situation itself. Examiners might be greatly assisted by the development of such guidelines.

References

(a) Brakel, S. (1974). Presumption, bias, and incompetence in the criminal process. Wisconsin Law Review, pp. 1105-1130.

(b) Gannon, J. (1990). Validation of the Competence Assessment Instrument and elements of competence to stand trial. Dissertation Abstracts International, 50-B, 3875.

(c) Golding, S., Roesch, R., & Schreiber, J. (1984). Assessment and conceptualization of com­petence to stand trial: Preliminary data on the Interdisciplinary Fitness Interview. Law and Human Behavior, 8, 321-334.

(d) Laboratory of Community Psychiatry, Harvard Medical School. (1973). Competence to stand trial and mental illness (DHEW Pub. No. ADM-77-103). Rockville, MD: Department of Health, Education and Welfare.

(e) Schreiber, J. (1978). Assessing competence to stand trial: A case study of technology dif­fusion in four states. Bulletin of the American Academy of Psychiatry and the Law, 6, 439-457.

(f) Schreiber, J. (1983). Evaluation of procedures for assessing competence to stand trial. (Final Report of NIMH Research Grant No. ROI-MH33669.) Rockville, MD: Center for Studies of Antisocial and Violent Behavior, National Institute of Mental Health.

(g) Schreiber, J., Roesch, R., & Golding, S. (1987). An evaluation of procedures for assessing competence to stand trial. Bulletin of the American Academy of Psychiatry and the Law, 15, 143-150.

Competence Screening Test (CST)

Authors

Lipsitt, P., & Lelos, D., with McGarry, A. L. and Associates, Laboratory of Community Psychiatry, Harvard Medical School

Primary Author Affiliation

Massachusetts General Hospital, Harvard Medical School (Community Mental Health-Law Program)

Primary Reference

Laboratory of Community Psychiatry, Harvard Medical School (1973). Competence to stand trial and mental Illness (DHEW Publication No. ADM 77-103). Rockville, MD: NIMH, Department of Health, Education, and Welfare.

note on second edition review. This review has changed little since the first edition, because only a few studies have been performed with the instrument since that time. The DHEW monograph that constituted the CAI manual is no longer available from NIMH, although it is still found in some university libraries.

Description

The Competence Screening Test (CST) was intended as a brief, psychome­tric instrument that would help to decide whether more extensive assess­ment for competence to stand trial was needed. The authors hoped that the CST, in conjunction with a brief psychiatric interview, might serve a screen­ing function to divert clearly competent defendants from the lengthy hospi­talization that was typically required (in past decades) for full competence assessments. Thus the CST sought to identify persons who clearly were competent to stand trial. The identification of incompetent defendants apparently was not a formal goal in the development of the CST; defen­dants' low scores "raise the issue of (in)competence" (c, p. 74), but do not provide evidence for incompetence itself. The CST, therefore, can be viewed as a psychometric definition of a threshold for recommending or ruling out further evaluation related to legal decisions about pretrial incompetence.

The CST was developed by an interdisciplinary team of psychiatrists, psychologists, and lawyers, funded by a research grant from the Center for Studies of Crime and Delinquency, NIMH. The instrument was intended as a companion to the Competence To Stand Trial Assessment instrument (CAI) reviewed earlier in this chapter.

The CST consists of 22 items, each of which is the beginning of an incomplete sentence. The sentence stems are

1. The lawyer told Bill that

2. When I go to court the lawyer will

3. Jack felt that the judge

4. When Phil was accused of the crime, he

5. When I prepare to go to court with my lawyer

6. If the jury finds me guilty, I

7. The way a court trial is decided

8. When the evidence in George's case was presented to the jury

9. When the lawyer questioned his client in court, the client said

10. If Jack has to try his own case, he

11. Each time the D.A. asked me a question, I

12. While listening to the witnesses testify against me, I

13. When the witness testifying against Harry gave incorrect evidence, he

14. When Bob disagreed with his lawyer on his defense, he

15. When I was formally accused of the crime, I thought to myself

16. If Ed's lawyer suggests that he plead guilty, he

17. What concerns Fred most about his lawyer

18. When they say a man is innocent until proven guilty

19. When I think of being sent to prison, I

20. When Phil thinks of what he is accused of, he

21. When the jury hears my case, they will

22. If I had a chance to speak to the judge, I

The examinee's sentence completions are scored according to the 22 sets of definitions and examples provided in the CST manual. For example, the scoring criteria for Item 1 are as follows:

1. The lawyer told Bill that

(a) Legal criteria: ability to cooperate in own defense, communicate, relate

(b) Psychological criteria: ability to relate or trust

Score 2: includes obtaining and/or accepting advice or guidance

Examples: "he should plead not guilty," "he was free," "he should plead nolo," "he should plead guilty," "he would take his case," "he would need to know all the facts concerning the case," "he should turn himself in," "the outlook was good," "he will try to help him."

Score 1:

Examples: "he is innocent," "everything is all right," "be truthful," "he will be going to court soon," "he is competent to stand trial," "it will be filled."

Score 0: includes regarding lawyer as accusing or judgmental

Examples: "he was wrong in doing what he did," "he is guilty," "he is going to be put away," "no comment."

Each response is scored 2, 1, or 0, and scores are summed to arrive at a total CST score (range, 0-44). In the manual, a total score below 20 was used as a signal to raise the question of incompetence (c, p. 74). Another reference by the authors (d, p. 138), however, designates "low scorers" as those obtaining CST scores of 20 or below, and this rule has been used most often in subsequent research with the CST.

Conceptual Basis

concept definition. During the first year of the 4-year project in which the CST was developed, the interdisciplinary team performed reviews of appellate cases and legal literature, and made observations in many pretrial competence hearings and in interviews with attorneys and judges. This allowed the team to arrive at a three-part definition of the competence to stand trial standard: an ability to cooperate with one's attorney in one's own defense; an awareness and understanding of the nature and object of the proceedings; and an understanding of the conse­quences of the proceedings.

These legal concepts were modified to provide a structure for devel­oping the CST items. The three measurement constructs used in writing items were:

• the potential for a constructive relationship between the client and his lawyer,

• the client's understanding of the court process, and

• ability to deal emotionally with the criminal process (c, p. 27).

operational definition. The three measurement constructs were used to structure the writing of 50 sentence stems. Pilot work reduced these to 22 items, refined the "legal criteria" and "psychological criteria" statements for each item, and developed definitions and examples for 2, 1, and 0 scores for each item. Each of the three measurement constructs is represented by about one-third of the CST items (although scores are summed only for total CST, not for construct dimensions).

critique. Questions about the content relevance of the CST can be addressed at three points in the translation process: the derivation of measurement constructs, the development of test items, and the formula­tion of scoring criteria.

Concerning the first, Brakel (b) severely criticized the CST authors for what he believes were unwarranted and dangerous presumptions when measurement constructs were developed to represent the Dusky standards. For example, where Dusky calls for an examination of a defendant's "ability to assist in one's defense," the CST construct inquires about "con­structive relationship with one's lawyer." Brakel claims that these are not the same constructs, because a good relationship with a good lawyer would promote an adequate defense, but a good relationship with a poor lawyer would not. The issue of assisting in one's defense, he claims, is broader than the nature of one's ability to relate to a lawyer. Similarly, where legal standards call for "an understanding of the consequences of the proceedings," the CST counterpart focuses on "an ability to deal emotionally with the criminal process." Thus the legal construct is cogni­tive, whereas the measurement construct refers to an affective component.

Whether or not Brakel's criticism is fair depends on the CST authors' purposes in translating competence standards into measurement con­structs. If they intended to produce three psychiatric constructs each of which would parallel the three Dusky (legal) components, then Brakel's criticisms are sound. At face value, the three measurement constructs seem essentially different from the three legal components derived more directly from Dusky.

On the other hand, it would seem likely that the three measurement constructs were intended to capture three psychological factors, each of which may contribute to any of the Dusky components: a social or inter­personal construct ("relationship"), a cognitive construct ("understand­ing"), and an affective or coping construct ("deal emotionally"). From this perspective, the CST attempts to measure three psychological functions, but using content specific to trial contexts (lawyers, trial events, criminal court outcomes) rather than assessing them in the abstract as in more general psychological tests of relating to others, intellectual abilities, and affective coping mechanisms. If that is the case, however, it is still true that the concepts themselves are considerably different than those that are used in modern interpretations of the constructs associated with competence to stand trial.

The second question of content relevance concerns translation of the measurement constructs into test items. The decision to develop approxi­mately an equal number of items for each of the three constructs seems benign. Yet this means that CST total scores give equal weight to all three measurement constructs in all cases. It should be recognized when inter­preting CST scores that low scores may be achieved because of marked deficits in only some areas of the CST, or because of moderate deficit across all areas.

At face value, several of the items (sentence stems) seem to have been worded in a manner that might result in interpretive error, in relation to the constructs that were to be assessed. For example, the first two items are intended to assess one's relationship capacities in interaction with one's lawyer. Yet the items do not specify a defense lawyer or one's own lawyer, but rather a lawyer of undefined role ("The lawyer told Bill that...," "When I go to court the lawyer will..."). This leaves open the possibility that the examinee may complete the sentence to express per­ceptions of prosecutors rather than defense attorneys.

The third question concerns the development of scoring criteria. Apparently the CST authors made no concerted attempt to determine whether professionals in law or in mental health outside of the project's research team would perceive the 2, 1, and 0 credit criteria or examples as relevant and appropriate to the legal purposes of the instrument. This is of concern because, at face value, it is often difficult to trace the logical connection between many of the scoring examples and the criteria that were written for the corresponding items. For example, for "The lawyer told Bill that...," it is not clear why the completion "... he should plead guilty" should receive 2 points, indicating fulfillment of the item criterion "Ability to relate or trust."

In other instances it is difficult to distinguish conceptually between 2 points and 1 or 0 points credit. For example, on item 6 ("If the jury finds me guilty, I..."), the response "... will serve my sentence" receives 2 points, whereas "... will take the punishment" receives 1 point. (The item intends to assess the defendant's realistic assessment of consequences.) A 0-point credit on the same item is assigned if the defendant responds, "... will be sentenced to the maximum penalty"; yet this may be a very realistic assessment of consequences in certain circumstances and for certain defendants.

Psychometric Development

standardization. The CST was designed as a paper and pencil test, although researchers (k, n) have administered it orally (with oral responding) when examinees lacked sufficient reading abilities. A stan­dard set of instructions (c, p. 74) informs the examinee that the items deal with courtrooms and the law, that there are no right or wrong answers, and that one is to complete the sentences with first impressions. Scoring is standardized by way of brief general criteria for each item and numerous specific item responses for each scoring category on each item.

reliability. There are five published reports of interscorer reliability for total CST scores, using scorers with various degrees of training in the method (c, e, j, k, n). Different statistical tests of reliability were used in the various studies, and coefficients of correlation have been uniformly high, generally.93 or higher. There have been no reports of temporal consis­tency (test-retest reliability) of CST scores.

NORMS. No norms are provided as such, but Lipsitt, Lelos, and McGarry (d) described mean CST scores for several hospitalized and non­hospitalized samples employed in initial studies with the CST: compe­tence examinees, 18.4; college students, 25.9; mental hospital control group, 24.6; men's club, 24.3; state hospital civil commitments, 23.8. Ustad et al. (p) provided means for competent (25.4) and incompetent (11.92) defendants in a competence restoration program.

critique. Interscorer agreement for the total CST scores generally has been quite good, despite observations that several of the examples provided in the manual for 2-point and 1-point responses appear to be quite similar to each other. No studies have examined interscorer agree­ment on individual items, however; it may be that CST total scores are highly comparable between scorers despite scoring disagreements on some of the individual items. Thus criticisms of CST scoring criteria may be important when making inferences about individual item responses, but relatively unimportant when employing total CST scores to make clinical decisions.

Construct Validation

The internal consistency (Cronbach's alpha) of the CST was identi­fied as.85 in one study (e) and.79 in another (p). Inter-item correlation in those two studies was.28 (e) and.14 (p). Differences between the two studies might be related to sample differences, the second study having been performed with defendants who were in a competence restoration program.

In the original study of the CST (c), factor analyses were performed with two samples of CST data: 91 patients representing a mix of the general hospital population and competence assessment defendants at Bridgewater State Hospital, and 83 persons who were not patients. The first analysis (hospitalized examinees) used an orthogonal rotation that produced six rotated factors accounting for 56.3% of the total vari­ance on the test. The authors described these six factors (b, p. 28) as reflecting: (I) relationship of defendant to attorney, (II) understanding of court process, (III) responsiveness and reaction to accusation and guilt, (IV) a second factor for understanding of court process, (V) trust and con­fidence in lawyer, and (VI) future orientation. The factor analysis with the second ("normal") sample produced markedly different factors; in the authors' opinion, these factors did not seem to represent any concepts related to pretrial competence definitions.

A factor analysis performed by Nicholson et al. (e) produced three factors that were quite different from those in the original study and were essentially uninterpretable when examined for the content of items load­ing strongly on the various factors. A re-analysis of the Nicholson et al. data by Bagby et al. (a) using a six-factor solution (like the original valida­tion study) produced results only slightly more like those found in the original study. Ustad et al. (p) also reported efforts to factor analyze the CST, having failed because a large number of subjects (in a competence restoration program) simply could not complete the instrument, either being unable to respond to some sentence stems or simply choosing to discontinue.

Schreiber (m) found that the CST produced competence/incompe- tence findings in proportions significantly similar to those based on interviews with the Competence Assessment Instrument (CAI: see previous review). Disagreements between the two instruments were all in the direction of CST incompetence findings in cases considered competent by CAI interviewers. Ustad et al. (p) found that scores on the CST correlated only.45 with those on the Georgia Court Competence Test-Mississippi State Hospital (GCCT-MSH: see accompanying review), and the measure agreed with competence/incompetence classification on the GCCT-MSH in only 61% of the cases.

One would expect pretrial competence to be related at least moder­ately to general intellectual ability. More specifically, individuals with very low levels of intellectual ability might be expected to perform poorly on the CST. Among patients with adequate intellectual functioning, how­ever, variance due to psychopathology might obscure any effects related to intellectual capacity. Correlations between CST scores and IQ measures in three studies have been very low (r =.10) (p), low (r =.24) (o), and moderate (r =.42) (l). A fourth study (h) reported a moderately greater proportion of mental retardation diagnoses among low CST scorers (30.7%) than among high CST scorers (13.5%), and a fifth (g) reported a correlation of -.33 between mental retardation and CST scores.

Within mental hospital samples, one would expect type or severity of diagnoses to be related to CST scores. Several studies have reported that a diagnosis of schizophrenia accounts for a greater proportion of low CST scorers than of high CST scorers. In three studies, the proportions of schizophrenic diagnoses in low and high CST groups, respectively, were 37.5% and 4.5% (h), 81.8% and 60.0% (p), and 73.9% and 50.0% (d). One study (g) reported a correlation of -.12 between psychosis and CST scores.

Studies examining the criterion validity of the CST, of course, pro­vide further empirical information concerning construct validity. These will be examined in the next section.

critique. The factor analysis of CST scores indicates that the CST's content is somewhat more complex conceptually than would be suggested by the three measurement concepts with which the authors began. Moreover, factor analyses subsequent to the original report cast considerable doubt on the stability of conceptual meaning for the CST across different samples.

Predictive or Classificatory Utility

Eleven studies have compared total CST scores to either judicial deci­sions about competence or concurrent findings of forensic clinicians who had performed comprehensive competence assessments and made rec­ommendations (d-2 studies, e, g, h, i, j, k, l, n, o).

All studies employed a CST score of 20 or lower as the cutoff score. Most of the studies used consecutive or randomly selected admissions of competence examinees to forensic assessment units of hospitals. In all cases, criterion decisions (judicial decisions or conclusions of forensic eval­uations) were made without knowledge of the defendants' CST scores.

Nicholson and Kugler (f) reviewed all of these studies and calculated effect sizes to summarize the relation of independent competence judg­ments to scores above and below the cut-off on the CST (and those of sev­eral other competence assessment tools). The meta-analytic finding of r = -.37 was poorer than for the other instruments reviewed (-.42 to -.52).

Typically, studies have found that the CST identifies far more defendants as potentially incompetent than are found incompetent by independent clinical judgments. For example, in their meta-analysis, Nicholson et al. (g) found that about 78% of criterion incompetent defen­dants made scores below the cut-off on the CST (sensitivity), while about 70% of criterion competent defendants scored above the cut-off (specificity). Had the CST been used as a screening tool in that sample, it would have identified about 34% of the sample as in need of further eval­uation, and about 23% of those who were further evaluated would have been found incompetent. Only 3% of those above the cut-off on the CST were criterion incompetent. However, those defendants accounted for nearly one-quarter of the incompetent defendants in the study. Some other studies have found somewhat better results with the CST, especially with samples that have baserates of criterion competence above 20% (the baserate in the Nicholson et al. meta-analysis was about 10%). Overall, however, the Nicholson et al. results are representative of other studies with baserates of competence that approximate the 10-15% baserate reported in many forensic services in the U.S.

critique. The available data suggest two answers to the question of the CST's utility in identifying "clearly competent" defendants, as judged against criteria such as final court decisions and the outcome of more extensive psychiatric evaluations.

In jurisdictions with competence examinee populations that have a greater than 20% base rate of incompetence, about 9 out of 10 high CST scorers would meet judicial or comprehensive evaluation standards for pretrial competence. Stated another way, if the CST were the primary fac­tor for decisions to proceed to trial, 1 out of every 10 defendants who met the CST criterion (above 20 cutoff score) would be tried unfairly, that is, would be falsely identified as competent. If low CST scores were the deciding factor in commitment for extensive competence evaluation because of "questionable competence," about one-third of those defen­dants would eventually be found competent. In contrast, in jurisdictions with very low base rates of pretrial incompetence (e.g., less than 15%), it is likely that the use of the CST as a screening device would produce a hit rate poorer than would be attained if one simply concluded that all defendants are competent.

Potential for Expressing Person-Situation Congruency

The CST makes no attempt to examine the defendant's knowledge or ability in relation to the idiosyncratic demands of the defendant's own spe­cific trial circumstances. The conceptual development of the CST suggests three dimensions of the environment corresponding to the three measure­ment construct originally associated with the CST: interpersonal qualities and skills of the defense attorney; the complexity of the anticipated trial; and anticipated characteristics of the trial and its probable outcomes, espe­cially its potential for affect arousal and induction of stress reactions.

References

(a) Bagby, R., Nicholson, R., Rogers, R., & Nussbaum, D. (1992). Domains of competence to stand trial: Afactoranalytic study. Law and Human Behavior, 16, 491-507

(b) Brakel, S. (1974). Presumption, bias, and incompetence in the criminal process. Wisconsin Law Review, pp. 1105-1130.

(c) Laboratory of Community Psychiatry, Harvard Medical School (1973). Competence to stand trial and mental illness (DHEW publication No. ADM 77-103). Rockville, MD: NIMH, Department of Health, Education, and Welfare.

(d) Lipsitt, P., Lelos, D., & McGarry, A. L. (1971). Competence for trial: A screening instru­ment. American Journal of Psychiatry, 128, 105-109.

(e) Nicholson, R., Briggs, S., & Robertson, H. (1988). Instruments for assessing competence to stand trial: How do they work? Professional Psychology, 19, 383-394.

(f) Nicholson, R., & Kugler, K. (1991). Competent and incompetent criminal defendants: A quantitative review of comparative research. Psychological Bulletin, 109, 355-370.

(g) Nicholson, R., Robertson, H., Johnson, W., & Jensen, G. (1988). A comparison of instru­ments for assessing competence to stand trial. Law and Human Behavior, 12, 313-321.

(h) Nottingham, E., & Mattson, R. (1981). A validation study of the Competence Screening Test. Law and Human Behavior, 5, 329-335.

(i) Pope, B., & Scott, W. (1967). Psychological diagnosis in clinical practice. New York: Oxford University Press.

(j) Randolph, J., Hicks, T., & Mason, D. (1981). The Competence Screening Test: A replication and extension. Criminal Justice and Behavior, 8, 471-481.

(k) Randolph, J., Hicks, T, Mason, D., & Cuneo, D. (1982). The Competence Screening Test: A validation in Cook County, Illinois. Criminal Justice and Behavior, 9, 495-500.

(l) Roesch, R., & Golding, S. (1980). Competence to stand trial. Champaign, IL: University of Illinois Press.

(m) Schreiber, J. (1983). Evaluation of procedures for assessing competence to stand trial (Final Report for Grant No. R0I-MH33669). Rockville, MD: Center for Antisocial and Violent Behavior, National Institute of Mental Health.

(n) Shatin, L. (1979). Brief form of the Competence Screening Test for mental competence to stand trial. Journal of Clinical Psychology, 34, 464-467.

(o) Shatin, L., & Brodsky, S. (1979). Competence for trial: The Competence Screening Test in an urban hospital forensic unit. Mt. Sinai Journal of Medicine, 46, 131-134.

(p) Ustad, I., Rogers, R., Sewell, K., & Guarnaccia, C. (1996). Restoration of competence to stand trial: Assessment with the Georgia Court Competence Test and the Competence Screening Test. Law and Human Behavior, 20, 131-146.