handbook of psychology phần 7 pot

374 Psychological Assessment in Correctional Settings two well-defined factors. The first reflects an egocentric, self- ish interpersonal style with its principle loadings from such items as glibness/superficial charm (.86), grandiose sense of self-worth (.76), pathological lying (.62), conning/manipulative (.59), shallow affect (.57), lack of remorse or guilt (.53), and callous/lack of empathy (.53). The items loading on the second factor suggest the chronic antisocial behavior associated with psychopathy: impulsivity (.66), juvenile delin- quency (.59), and need for stimulation, parasitic life style, early behavior problems, and lack of realistic goals (all loading .56; Hare et al.). Some use the PCL-R to identify psychopaths; although the conventional cutting score is 30, Meloy and Gacono (1995) recommend a cutting score of 33 for clinical purposes. Others treat the PCL-R as a scale and enter PCL-R scores into predictive equations. These differing practices reflect a funda- mental disagreement about the nature of psychopathy; that is, is psychopathy a dimension of deviance, or are psychopaths qualitatively different from other offenders? A number of studies have shown that PCL-R scores corre- late with recidivism in general and violent recidivism in par- ticular. In their follow-up of 618 men discharged from a maximum security psychiatric institution, Harris et al. (1993) reported that, of all the variables they studied, the PCL-R had the highest correlation (+ .35) with violent recidivism, and they included psychopathy, as defined by PCL-R scores greater than 25, as a predictor in their VRAG. Rice and Harris (1997) reported the PCL-R was also associated with sexual reoffending by child molesters and rapists. Reviewing a number of empirical investigations, both retrospective and prospective, Hart (1996) reported that psychopaths as diag- nosed by the PCL-R had higher rates of violence in the community and in institutions than nonpsychopaths, and that psychopathy, as measured by the PCL-R, was predictive of violence after admission to a hospital ward and also after conditional release from a hospital or correctional institution. He estimated that the average correlation of psychopathy with violence in these studies was about .35. In their meta- analysis of 18 studies relating the original and revised PCLs to violent and nonviolent recidivism, Salekin, Rogers, and Sewell (1996) found 29 reports of effect sizes ranging from 0.42 to 1.92, with a mean of 0.79. They reported, “We found that the PCL and PCL-R had moderate to strong effect sizes and appear to be good predictors of violence and general recidivism” (p. 203). Hart summarized it best when he concluded, “predictions of violence using the PCL-R are considerably better than chance, albeit far from perfect” (1996, p. 64). As is the case with many risk assessment instruments, PCL-R scores in the clinical range are meaningful but those below the cutoff have no clear relation to behavior. Specifi- cally, low PCL-R scores do not guarantee that an offender will never recidivate or be violent. Although the PCL-R has been used most often for risk assessment, it also has implications for treatment planning. Suedfeld and Landon (1978, p. 369) summarized the results of attempting to treat psychopaths as “not much to show for the amount of time, effort, and money spent.” In correctional facilities where treatment resources are scarce and access must be limited to those most likely to profit from interventions, such findings suggest that psychopaths should have lower priority than other offenders. The PCL-R has shown rather good generalizability, being associated with recidivism and violence among male offenders in the United States and Sweden (Grann, Längström, Tengström, & Kellgren, 1999), as well as those in Canada. There is some question, however, about its applicability to minorities. BlackAmerican menscore higherthan theirWhite American counterparts, and there is insufficient research on the PCL-R with large samples ofAsians, Hispanics, or Native Americans or with women (Meloy & Gacono, 1995). To obtain reliable and valid PCL-R ratings, it is important to have good case histories and interviewer-raters who are trained in Hare’s technique. Such records and personnel are more likely to be found in correctional mental health facilities and neuropsychiatric hospitals than in prisons, and it is not surprising that the PCL-R has been used most successfully in those settings. In ordinary correctional institutions and jails, it would probably not be practical to use the PCL-R for mass screening, although it may be feasible to administer it to select groups, such as previously violent offenders being considered for parole. Evaluating Risk Assessment Instruments It is impossible to evaluate the predictive validity of risk assessment instruments accurately. Consider a parole prediction instrument. To evaluate it properly, one must first predict which prisoners eligible for parole are most likely to succeed or fail according to whatever criteria one selected. Then they must all be paroled, regardless of the predicted risk. After a year or so, a follow-up should be conducted that will enable the researcher to calculate whether those predicted to fail actually were more violent, committed more new crimes, or vi- olated the conditions of parole more than those predicted to succeed. If not all applicants were released, it is impossible to determine how many of those who were predicted to fail and denied parole actually would have succeeded had they been released (i.e., the false-positive rate; Megargee, 1976). Un- fortunately for researchers, parole boards are understandably Risk Assessment and External Classification 375 reluctant to release all eligible applicants in order to test their predictive devices. Similar considerations apply to security- and custody- level assignments. To properly assess their accuracy, it would be necessary to assign offenders randomly to different facilities without regard for their estimated risk levels. Otherwise, we cannot know whether a high-risk offender who failed to act out was a classification error or was simply deterred from misconduct by being assigned to a maximum security setting with stringent external controls. Base rates are another vital concern. The closer the incidence of the behavior in question isto 50%, the greater the potential contribution that a predictive tool can make. The more infrequent the behavior, the greater the number of false positives that can be expected (Brennan, 1993; Finn & Kamphuis, 1995; Meehl & Rosen, 1955; Megargee, 1976, 1981). Since violence is still a rare event, even in prisons, the number of false positives is likely to be high. For this reason it is important to consider the consequences of incorrect classifications (Megargee, 1976). If the risk assessment merely influences the dormitory to which offenders are assigned and has no impact on their programming or other conditions of confinement, the results of being misclassified are relatively benign. On the other hand, if the outcome is involuntary commitment or preventive deten- tion, the consequences for false positives are quite serious. Campbell (2000) recently argued that the schemes suggested for assessing the likelihood that sexual predators will reof- fend are, at best, experimental. Likening them to phrenology, he maintained that, at this stage of their development, using them to decide whether a sex offender should be kept in custody beyond the expiration of his prison term is contrary to the American Psychological Association’s (1992) ethical standards governing the use of psychological tests. The generality of predictive instruments is another concern. In order to economize, predictive devices derived in one setting have frequently been applied in other jurisdictions. For example, the National Institute of Corrections (1981) en- couraged other states to adopt the Wisconsin method of risk assessment for probation and parole decisions rather than going to the time and expense of developing their own instruments. However, when Wright, Clear, and Dickson (1984) tested the Wisconsin system in New York, they discovered that “a number of variables in the Wisconsin model were found to be unrelated to outcome” in their sample (p. 117). They advised practitioners to test the generality of prediction models in their settings before using them in actual decision making. Although the emphasis in risk assessment is on diagnosing the most dangerous offenders, the greatest contribution of these classification tools has been to identify low-risk prisoners who could safely be assigned to less secure correctional programs or placed in the community (Austin, 1993; Glaser, 1987; Solomon & Camp, 1993). When making subjective predictions of violence, classifications personnel are often overly conservative, placing many offenders in higher-than- necessary risk categories (Heilbrun & Heilbrun, 1995; Monahan, 1981, 1996; Proctor, 1994; Solomon & Camp). This is not surprising. The public is rarely incensed if low-risk offenders are retained in more restrictive settings than necessary, but clinicians can expect to be castigated if someone they approved for minimum security or early release goes out to rape, pillage, and plunder the community. Reducing the extent of overclassification has three important benefits. First, it is the correct thing to do; as noted previously, the courts have consistently ruled that offenders have the right to be maintained in the least restrictive settings con- sistent with maintaining safety, order, and discipline. Second, less restrictive settings are more economical; confining an offender in a maximum security institution costs $3,000 a year more than a minimum security facility and $7,000 more than a community setting. Third, the residents benefit because more programming is possible in less restrictive settings, and the deleterious effects of crowding are diminished (Proctor, 1994). Internal Classification After external classification and risk assessment have determined offenders’ custody and security levels and assigned offenders to the most appropriate correctional facilities, internal classification is used to further subdivide the institutional population into homogenous subgroups for housing and management. According to Levinson, Internal classification is the final stage in the classification process. It is a systematic method that identifies homogeneous prisoner subgroups within a single institution’s population. Although the degree of variation among one facility’s inhabi- tants is smaller than that found in the total prison system, every institution has a range of inmates—from the predators at one extreme to their prey at the other end of the continuum. Various labels are used to define these individuals: thugs, toughs, wolves, agitators, con-artists, in contrast to weak sisters, sheep, depen- dents, victims, and other expressions less acceptable in polite society. (1988, p. 27) The goal of internal classification is to separate these groups in order to reduce the incidence of problematic and disruptive behavior within the institution. 376 Psychological Assessment in Correctional Settings Other factors that influence management and housing decisions are the amount of supervision each offender is likely to need, his or her sense of responsibility and response to supervision, the approach correctional officers should take in working with him or her, and whether he or she will respond better to strict discipline or a more casual correctional atmos- phere (Wright, 1986, 1988). In many BOP facilities, Quay’s (1984) Adult Internal Management System (AIMS) is used for internal classification. The Adult Internal Management System Based on extensive factor analytic research with juvenile (Jenkins, 1943; Jenkins & Glickman, 1947; Hewitt & Jenkins, 1946; Quay, 1965) and adult offenders, Quay (1973, 1974, 1984) defined five adult-offender types: • Type I (aggressive-psychopathic) offenders are the most antisocial and have the most trouble with authorities. Eas- ily bored, and having little concern for others, they are the ones who are most apt to exploit others and cause difficulties and disturbances in an institution. • Type II (manipulative) offenders are less aggressive and confrontational but no less untrustworthy, unreliable, and hostile to authority. They may organize inmate gangs and manipulate others for their own ends. • Type III (moderate) inmates are neither very aggressive nor very weak. Often situational offenders, they have less extensive criminal histories than the first two types and are more responsible and trustworthy. • Type IV (inadequate-dependent) offenders are weak, im- mature, and indecisive. Rarely involved in disciplinary in- fractions, they are seen by staff as emotionally dependent and clinging. • Type V (neurotic-anxious) offenders are anxious, worried, and easily upset. They are apt to be exploited or victimized by other offenders. The primary goal of the AIMS system is to separate the heavy (Types I and II) from the light (Types IV and V) offenders by assigning them to separate living units and arranging their programs so they have minimal contact with one another (Levinson, 1988). However, Quay (1984) also provides differential programming guidelines for the heavy, moderate, and light offenders with regard to educational programming, work assignments, counseling, and staff approach. For example, correctional staff are advised to adopt a no-nonsense, by- the-book approach for the heavies, to supervise moderates only as needed, and to be highly verbal and supportive with the lights. Categorization into the Quay types is based on two rating forms, the Correctional Adjustment Checklist (CACL) and the Checklist for the Analysis of Life History Records of Adult Offenders (CALH). The CACL is filled out by trained correctional officers on the basis of their observations of the inmates’ behavior during the first 2 to 4 weeks after admission (Quay, 1984). Each of the 41 items, such as “Easily upset” or “Has a quick temper,” is scored as 0 (not observed) or 1 (observed). Each item is indicative of a different Quay type, and the number of items checked determines the raw score on each of the five scales. The 27-item CALH is filled out by a trained caseworker on the basis of the information contained in the presentence investigation report. It contains such behavioral items as “Has few, if any, friends” or “Thrill-seeking,” and, as with the CACL, each is scored as present or absent. Offenders are classified into the category on which they receive the highest score. Quay (1984) did not provide interrater reliability data. In addition to the factor analytic research that guided the development of the AIMS system, Quay (1984) cites five sources of evidence for the validity and utility of the AIMS system: (a) significant reductions in the number of assaults at penitentiaries where it was adopted, (b) significant reductions in misconduct where it was adopted, (c) testimonials from wardens and administrators, (d) convergence between AIMS classifications and parole board Salient Factor Scores, and (e) convergence between AIMS classifications and BOP custody- and security-level ratings. One drawback to the AIMS system is the time required to obtain valid CACL ratings. Staff should have 2 to 2 weeks to observe behavior before completing the CACL, and some correctional facilities demand quicker results. Quay (1984) himself acknowledges that AIMS has limited utility in jails that have rapid turnover and sparse case history records. Another concern is the availability of adequate life history information. Attempting to implement the AIMS system in Scotland’s largest prison, Cooke, Walker, and Gardiner (1990) found it was difficult to obtain the biographical information needed to complete the CALH. In some settings, staff members resist spending the time and effort required to observe inmates, review case files, and fill out the rating forms. In Van Voorhis’s (1994) comparison of five psychological classification systems for adult male offenders, she reported that the AIMS was the most difficult to complete because of the lack of staff cooperation. Some staff sabotaged the administration by checking every item for every inmate. She eventually had to hire additional personnel in order to get the CACL and CALH forms completed properly. However, she reported, “Despite these difficulties, we observe numer- ous significant relationships between this typology and Risk Assessment and External Classification 377 important institutional behaviors” (1994, p. 126). Correc- tional psychologists using the AIMS system should be prepared to devote the time and effort required to working with and motivating the staff members who are responsible for making the assessments on which the system depends. Needs Assessment Sooner or later, almost all of the nearly 2 million incarcerated adult offenders will be released to return to their communities and to the approximately 2.6 million children they left behind. The goal of treatment is to maximize the chances that former offenders will become productive citizens and responsible parents instead of continuing to prey on society until they are once more arrested and returned to prison. If the correctional system is to reform or rehabilitate inmates, each offender’s educational, vocational, emotional, and mental health needs must be appraised and individual management and treatment programs formulated. Treatment planning requires both of the following: 1. Psychological assessments in order to identify offenders in need of mental health interventions. These individuals include those who are depressed, psychotic, emotionally disturbed, and prone to self-injurious behavior, as well as those with problems centering around alcohol and substance abuse. In addition to assessing offenders’ needs for treatment, program planning involves estimating each inmate’s likely response to and ability to benefit from various types of intervention. In systems with limited mental health resources, each inmate’s priority for treatment, based on diagnosis and prognosis, needs to be determined. (Priority for various programs is also likely to be influ- enced by other factors such as the offender’s security level and behavior in the institution. In the 1970s, the Bureau of Prisons used a formula based on the offender’s age, prior sentences, and the length of the present sentence, as well as the caseworker’s rating, to determine priorities. Case- worker ratings being equal, younger offenders with few priors and short sentences were given priority for programming over older offenders with long records and considerable time left to serve.) 2. Cognitive appraisals to evaluate each offender’s need for and ability to profit from educational programming. These decisions can be based in part on the educational history; there is no need to place college graduates in a general equivalency diploma (GED) program. However, given the extent of social promotion, a high school diploma does not necessarily guarantee literacy, so intelligence and achievement tests are often needed. As with mental health treatment, when educational resources are limited, it may be necessary to determine offenders’priorities for education based on their ability and motivation. Intake Screening Inmates who have just arrived at a jail or prison must be screened for serious mental illness, suicide potential, and retardation before they are placed in a double cell or mingled with the general population (Anno, 1991). In jails, the burden of this screening typically falls on the correctional staff who receive new arrivals. In prisons, the receiving evaluation should include a screening for mental illness and suicide potential by a qualified health care professional who may or may not be part of the mental health staff. The NCCHC provides intake and mental health evaluation forms that appropriately trained reception personnel can use to screen new admissions to jails (NCCHC, 1996) and prisons (NCCHC, 1997), while the ACA has developed a self- instructional course designed to train correctional officers to recognize the signs of suicide and intervene appropriately (Rowan, 1998). This author has been unable to locate any published reports evaluating the reliability or validity of these screening instruments. Mental Health Assessment Many prisoners require mental health treatment and care. Reviewing a number of studies, Anno (1991) estimated that 5% to 7% of the adult prison population suffers from serious mental disorders, not including personality disorders or substance abuse problems, and an additional 10% may be considered mentally retarded. During the course of confinement, emotional problems will naturally arise. Anno (1991) estimated that, in addition to those suffering from serious psychiatric disorders, another 15% to 20% of a prison’s inmates require mental health services or interventions at some time during their incarceration. As noted earlier, all new inmates should receive a mental health assessment within the first week (AACP, 2000) or two (NCCHC, 1997) after admission. This should include an interview and screening with group tests of intellectual and personality functioning, followed by more extensive evaluations of those who appear to show signs of mental illness or retardation or who appear at risk for self injury or suicide (AACP; NCCHC). Intake Interview The NCCHC’s prison standards (1997, p. 47) stipulate that the mental health assessment should include a structured 378 Psychological Assessment in Correctional Settings interview by a member of the mental health staff who in- quires into the offender’s (a) current psychotropic medications, (b) suicidal ideation and history, and (c) emotional response to incarceration, as well as his or her history of (d) psychiatric hospitalizations and treatments, (e) drug and alcohol use, (f) sex offenses, (g) expressive (angry) aggres- sion or violence, (h) victimization, (i) special education placement, and (j) cerebral trauma or seizures. In addition to the interview, a group personality test and a brief group or individual test of intelligence should be administered (AACP, 2000; NCCHC). If the initial screening or the subsequent mental health assessment indicates mental illness or retardation or suggests the possibility of suicidal or self-injurious behavior, the inmate should be referred for further evaluation by a qualified mental health professional. (The NCCHC’s (1997) prison standards state, “The mental health staff includes those qualified health professionals who may not have had formal training in working with the mentally ill or retarded, but who have received instruction in identifying and interacting with individuals in need of mental health services. Qualified mental health professionals include psychiatrists, psychologists, psychiatric social workers, psychiatric nurses, and others who by virtue of their education, credentials, and experience are permitted by law to evaluate and care for the mental health needs of patients” (p. 47, italics in the original).) Minnesota Multiphasic Personality Inventory–2 The MMPI-2 is the most widely used and thoroughly researched personality assessment device in the world (Butcher, 1999). The MMPI-2 and the original MMPI have been used in corrections for almost 60 years. There are well- established correctional norms and cutting scores available (Megargee, Mercer, & Carbonell, 1999) and the correlates of the scales among criminal offenders have been thoroughly studied over the years (Megargee, 2000). Megargee (2000) has provided detailed instructions for administration in correctional settings.Although MMPI-2 administration is not difficult, it must be done properly to achieve optimal results. A sixth-grade reading level is needed to complete MMPI-2, so it is best to administer it after reading ability has been assessed. Audiotaped forms are available for poor readers. For inmates who are not proficient in English, the MMPI-2 is available in a number of other languages. There are three levels of analysis available for correctional psychologists using the MMPI-2. The first is to interpret the scores on the various MMPI-2 scales and indices using correctional norms and cutting scores (Megargee, 2000; Megargee et al., 1999). The second is to use Megargee’s MMPI-2-based offender classification system (Megargee, Carbonell, Bohn, & Sliger, 2001). The third is to consult the interpretative scales and possible problem areas identified by Megargee’s (2000) recently developed interpretive scheme. Each will be discussed in turn. The MMPI-2 has four types of scales: validity, basic, supplementary, and content. The eight validity scales enable the user to identify offenders who are (a) answering nonre- sponsively, (b) malingering (faking bad), or (c) dissembling (faking good). In assessing MMPI-2 validity in correctional settings, it is important to consult appropriate offender norms (Megargee, 2000; Megargee et al., 1999). For example, criminal offenders answering honestly may get elevations on the Infrequency (F) scale that would be regarded as invalid in free-world settings. The basic, supplementary, and content scales assess a broad array of traits and behaviors, many of which are relevant to mental health assessment and treatment planning in correctional settings. For example, elevations on MMPI-2 Scales 1 (Hs, Hypochondriasis), 3 (Hy, Hysteria), and HEA (Health Concerns) identify offenders who are likely to use sick-call frequently. Scales 2 (D, Depression) and DEP (Depression) identify those who are depressed, and Scales 7 (Pt, Psychasthenia) and ANX (Anxiety) are associated with anxiety. Scales 4 (Pd, Psychopathic Deviate), 9 (Ma, Hypomania), and ASP (Antiso- cial Practices) reflect authority problems, antisocial behavior, and acting-out. Scale ANG (Anger) indicates problems with anger control, and Scales 4, MDS (Marital Distress), and FA M (Family Problems) identify offenders who may be alienated or estranged from their families. The MAC-R (MacAndrewAlco- holism Scale–Revised) and AAS (Addiction Admission Scale) suggest alcohol or substance abuse. Scales 6 (Pa, Paranoia), 8 (Sc, Schizophrenia) and BIZ (Bizarre Mentation) identify those who might have mental disorders that require further assessment. Scales 5 (Mf, Masculinity-Femininity), 0 (Si; Social Introversion), and SOD (Social Discomfort) are associated with passivity, introversion, and awkward interpersonal relations that may lead to exploitation by more predatory inmates in prison settings (Butcher & Williams, 1992; Graham, 2000; Megargee, 2000). Whereas most measures used in correctional settings assess only negative characteristics, the MMPI-2 can also indicate positive attributes. Offenders with moderate elevations on Scale 0 are unlikely to be defiant or cause problems for those in authority, those high on Scale Re (Responsibility) should be more mature and cooperative than most, and those with elevations on Scale Do (Dominance) should be leaders. The second level of analysis is to classify MMPI-2 profiles according to Megargee’s empirically derived offender classifications system (Megargee & Bohn with Meyer & Sink, 1979; Megargee et al., 2001). Derived from cluster analyses Risk Assessment and External Classification 379 of criminal offenders’ original MMPIs, the system is com- posed of 10 types labeled with neutral, alphabetic names. In- dependent studies applying similar clustering procedures to the MMPIs of male and female offenders in various settings have demonstrated the reliability of the typology, consistently replicating most of the 10 groups (Goeke, Tosi, & Eshbaugh, 1993; Mrad, Kabacoff, & Duckro, 1983; Nichols, 1979/1980; Shaffer, Pettigrew, Blouin, & Edwards, 1983). Independent investigators have reported the successful application of the MMPI-based system among male and female offenders in probation, parole, and correctional settings. Within correctional institutions, it has been utilized in federal, state, military, and local facilities with security levels ranging from minimum to maximum. It has also been applied in halfway houses, community restitution centers, forensic mental health units, and local jails. The specialized populations to which the system has been applied include death row inmates, presidential threateners, and mentally disor- dered sex offenders (MDSOs; Megargee, 1994; Sliger, 1992; Zager, 1988). Gearing (1981, pp. 106–107) wrote that “this new MMPI system unquestionably defines the present state of the art in correctional classification.” A unique aspect of the MMPI-2-based system is the fact that the characteristics of the 10 types were determined en- tirely through empirical research in offender populations. The original MMPI research delineating the attributes of male offenders has recently been replicated with the MMPI- 2 (Megargee, 1994; Megargee et al., 2001) and a number of new studies have extended the system to female offenders (Megargee, 1997; Megargee et al.; Sliger, 1997). In addition, almost 100 independent investigations have further explored the attributes and behaviors of the 10 types in various criminal justice settings (Megargee et al.). Based on the patterns of empirically observed differences, individual descriptions of each of the 10 MMPI-2-based types were written that discussed their modal family back- grounds; social and demographic characteristics; patterns of childhood and adult adjustment; and educational, vocational, and criminal histories. In addition, a number of studies have examined how the types differ in their adjustment to prison— which ones are most likely to be disruptive or cause trouble, which are most likely to do well or poorly in educational or vocational programming, and which are most likely to succeed or fail on parole. Strategies for management and treatment have been formulated that address the optimal setting, change agent, and treatment program for each type (Megargee & Bohn, 1977; Megargee, Bohn, et al., 1979; Megargee et al., 2001). Although the system is designed pri- marily for needs assessment, Bohn (1979) obtained a 46% reduction in serious assaults over a 2-year period when he used it for internal classification to separate the more predatory inmates from those most likely to be victimized. One of the advantages of an MMPI-2-based system is that it can reflect changes in offenders over time in a way that systems based on the criminal history or current offense systems cannot. Studies have shown that many offenders’ classifications do change over the course of their sentences. Doren and Megargee’s (1980) research indicated that these differences reflect changes in the client rather than unreliability in the system. If a year or more has passed since an offender’s last MMPI-2, it is advisable to readminister the MMPI-2 and re- classify him or her if important programming or treatment decisions are to be made. A third level of analysis for evaluating MMPI-2’s in correctional settings involves consulting a series of interpretive statements recently devised by Megargee (2000). Unlike risk assessment instruments, these ratings include positive as well as negative aspects of offender behavior. Using algorithms based on the Megargee system of classification and cutting scores on selected MMPI-2 scales, offenders are evaluated as being high, medium, or low, relative to other criminal offenders on nine behavioral dimensions that are especially relevant to corrections: (a) apparent need for mental health assessment or programming; (b) indications of socially deviant behavior or attitudes; (c) extraversion and need for social participation; (d) leadership ability or dominance; (e) likelihood of hostile or antagonistic peer relations; (f) indications of conflicts with or resentment of authorities; (g) likelihood of mature, responsible behavior and positive response to supervision; (h) likelihood of positive or favorable response to academic programming; and (i) likelihood of positive or favorable response to vocational programming. In addition to these nine bipolar scales, Megargee (2000) has also developed a list of nine red flags, or warnings of possible problem areas, including the possibility of (a) difficulties with alcohol or substance abuse, (b) thought disorder, (c) depressive affect or mood disorder, (d) extensive use of sick call, (e) overcontrolled hostility, (f) manipulation or exploitation others, (g) problems with anger control, (h) awkward or difficult interpersonal relationships, passivity, and submis- siveness, and (i) family conflict or alienation from family. The purpose of these warning statements is to raise hypothe- ses for clinicians to evaluate using case history data, inter- views, staff observations, and other psychological tests. These interpretive scales and statements are contained in Megargee’s (2000) computerized MMPI-2 Criminal Justice and Corrections Report, which also provides MMPI-2 profiles, scores, and indices on all the validity, basic, and content scales as well as selected supplementary scales and the offender’s Megargee system classification. 380 Psychological Assessment in Correctional Settings Although Megargee’s (2000) interpretive scales and warnings of problem areas are based on well-established correlates of the MMPI-2 scales and offender types, the interpretations themselves have not yet been empirically validated, and as yet they apply only to male offenders. As with any computerized assessments, they should be used only by qualified correctional psychologists in conjunction with other sources of information. Intelligence Screening As noted earlier, the NCCHC’s (1997) Prison Standards stipulate that a brief intellectual assessment should be part of the postadmission mental health evaluation. The primary purpose of this assessment is to identify developmentally disabled inmates who may be victimized or exploited by predatory inmates. However, a more thorough intellectual evaluation should also be conducted as part of offenders’ needs assessment, to determine their need for educational programming and their ability to profit from instruction. Two brief screening instruments often used in corrections, one verbal and the other nonverbal, will be described. Shipley Institute of Living Scale The Shipley Institute of Living Scale (SILS; Shipley, 1940; Zachary, 1994) is a brief, self-administered verbal test of intellectual functioning in adults aged 16 to 64 that is designed for group or computer-based administration. It has two, timed 10-min subtests. The Vocabularysubtest contains 40 multiple- choice items of increasing difficulty on which the respondent selects which of four terms best conveys the meaning of the stimulus word. It thus involves reading and recognition of vocabulary words. The Abstraction subtest consists of 44 increasingly difficult sequences of letters, words, and numbers. The respondent’s task is to deduce the logical principle governing each sequence and to use it to produce the next symbols in the sequence. It thus involves reading, abstract reasoning, and pro- duction (as opposed to recognition) of the correct answer. Age-specific T scores can be computed on each of the subtests and on the total of both subtests (Zachary, 1994). Shipley (1940) originally designed the SILS as a test of intellectual deterioration or impairment based on the now- discredited notion that deterioration is evidenced by the discrepancy between scores on hold tests, such as Vocabu- lary, and don’t-hold tests, such as Abstraction. Today the SILS is used as a brief screening instrument for intellectual appraisals. The SILS manual (Zachary, 1994) reports split-half internal consistency coefficients, corrected by the Spearman-Brown formula, of .87 for Vocabulary, .89 forAbstraction, and .92 for the Total score. Test-retest stability coefficients over the course of 8 to 16 weeks ranged from .62 to .82 with a median of .79. Correlations between the SILS Total score and Wechsler Adult Intelligence Scale (WAIS) Full Scale IQs in 11 samples of psychiatric patients ranged from .74 to .90 with a median of .78; correlations with Wechsler Adult Intelligence Scale–Revised (WAIS-R) Full Scale IQs in two samples of psychiatric patients were .74 and .85 (Zachary). The manual (Zachary) provides a procedure for estimating WAIS-R IQs from SILS Total scores; the estimated IQs correlated .85 with actual WAIS-R Full Scale IQs. Although the SILS manual has been revised and the norms updated, the SILS items have not been changed since 1940. Perhaps because Shipley (1940) derived the test using high school and college students, the SILS works best in young adults; until age-specific T-score tables became available, it tended to underestimate the IQs of older respondents. Wood, Conn, and Harrison (1977) administered the SILS and the WAIS to prisoners at a county penal farm and reported that the SILS was an adequate predictor of WAIS scores, but cautioned that the estimates were better for White than for Black offenders. Bowers and Pantle (1998) administered the SILS and the Kaufman Brief Intelligence Test (KBIT; Kaufman & Kaufman, 1990) to 52 female inmates. They reported that the SILS correlated .83 with the KBIT IQ and that there were no significant mean differences between the offenders’ mean scores on the two measures. The SILS manual (Zachary, 1994, p. 2) warns that, “Be- cause the scale is self-administered, it is not recommended for individuals who are either unable or unwilling to cooper- ate,” and notes (p. 3), “While the Shipley may be used to obtain a quick estimate of intellectual functioning, it is not a substitute for more detailed assessment procedures.” In corrections, the SILS is best used as a brief screening device for estimating verbal intelligence. If offenders obtain scores in the average range of intellectual functioning or higher, it can be presumed that their intellectual ability is adequate for the educational programming afforded at most correctional institutions. Those obtaining below average scores should receive a more comprehensive individual intellectual assessment with an instrument such as the Wechsler Adult Intelligence Scale–Third Edition (WAIS-III; Wechsler, 1997), especially if their scores suggest possible retardation. General Ability Measure for Adults The General Ability Measure for Adults (GAMA; Naglieri & Bardos, 1997) provides a brief nonverbal measure of general intellectual ability for adults aged 18 and older. The GAMA has 66 items consisting of attractive blue and yellow diagrams, Risk Assessment and External Classification 381 each of which has six possible multiple-choice responses. There are four scales: 1. The Matching scale items present the respondent with a stimulus diagram. From an array of six similar diagrams, he or she must select the one that matches the stimulus item in color, shape, and configuration. 2. The Analogies subtest presents respondents with logic problems of the nature “A is to B as C is to (?),” but diagrams are used instead of letters. Respondents must choose the correct answer from six possible diagrams. 3. Sequences presents test takers with an interrupted sequence of five diagrams showing a figure that is being rotated or otherwise moved through space. In each sequence the middle (third) diagram is missing and test takers must select from an array of six possibilities the one design that correctly completes the sequence. 4. Construction presents respondents with fragments of shapes; from an array of six completed figures, they must choose the one diagram that could be built with the fragments. The GAMA can be scored by hand or by computer, and tables are available for converting raw scores to scaled scores for each of 11 age levels ranging from 18 to 19 years at the lower end to 80 or older at the upper. Although the tasks are nonverbal, a third-grade reading level is needed to follow the directions. (A Spanish version is available for those who should be tested in that language.) Respondents have 25 min to complete the 66 GAMA items. The authors took great pains in selecting the 2,360 partic- ipants in the national normative group. Each of the 11 age groups was stratified on the basis of the 1990 U.S. Census into the two usual genders, five racial or ethnic groups (African American, American Indian, Asian–Pacific Islander, Hispanic, or White), five education levels (grade school, attended high school, graduated high school or GED, attended college, or completed bachelor’s degree or more), and four geographic regions of the United States. Detailed tables in the GAMA manual (Naglieri & Bardos, 1997) provide complete comparisons with the 1990 Census data. Split-half internal consistency coefficients, averaged over the 11 age groups and corrected by the Spearman Brown formula, were .66 for the Matching subtest, .81 for Analogies, .79 for Sequences, .65 for Construction, and .90 for the overall IQ. Test-retest coefficients over the course of 2 to 6 weeks were .55 for the Matching subtest, .65 for Analogies, .74 for Sequences, .38 for Construction, and .67 for the overall IQ. Practice effects were evident on all of the retest means except Matching. The magnitudes of these reliability coefficients suggest that psychologists should discuss the confidence lim- its when reporting GAMA scores. Naglieri and Bardos (1997) reported that GAMA IQs correlated .65 with WAIS-R Verbal IQs, .74 with Performance, and .75 with Full Scale IQs. They also obtained correlations of .72 with the SILS and .70 with the KBIT. Given the multiplicity of ethnicities and the low reading levels typically encountered among criminal offenders, the GAMA appears to have considerable potential as a brief, nonverbal intellectual screening device for correctional settings, and it is currently being marketed for that purpose. Ad- ditional data on the GAMA’s use in corrections are needed. As with the SILS, its best use appears to be as an indicator of possible intellectual deficiency, with low-scoring offenders being referred for a more complete individual examination with WAIS-III. Wechsler Adult Intelligence Scale–Third Edition Offenders who are suspected of being developmentally disabled or for whom a more definitive appraisal of intelligence is needed should be tested with WAIS-III (Wechsler, 1997) by a qualified administrator (NCCHC, 1997). The gold standard (so to speak) for the appraisal of adult intelligence, the WAIS-III has been updated and undergone several modifications that make it more appropriate for correctional use than its predecessor, the WAIS-R. In addition to updating the 11 familiar WAIS subtests, three new supplementary scales have been added. On the new Verbal scale, Letter-Number Se- quencing, the examiner reads a series of randomly ordered letters and numbers that the respondent must recall, reorder, and recite back in ascending order, numbers first. One of the new Performance scales, Symbol Search, is a true-false test on which the respondent indicates whether either of two tar- get stimuli, such as stars or crosses, appears in an array of seven similar stimuli. The other Performance scale, Matrix Reasoning, consists of a series of pictures, each of which shows five geometric shapes that the respondent must identify. The new Performance scales should improve the assessment of intelligence among linguistically challenged inmates and reduce the importance of perceptual speed in assessing Performance IQs (Cohen & Swerdlik, 1999). Correctional assessment will also be improved by the downward extension of the floor for most subtests, making them more suitable for testing intellectually challenged clients. Despite this, the overall administration time is less for the WAIS-III than it was for the WAIS-R (Aiken, 2000). Several modifications make the WAIS-III more suitable for older respondents than its predecessors were. They include making some of the stimuli larger so they can be seen better by older clients, and extending the norms to adults aged 382 Psychological Assessment in Correctional Settings 74 to 89. Unlike with the WAIS-R, scaled scores are computed based on age-specific norms (Cohen & Swerdlik, 1999). The WAIS-III was standardized on a national sample of 2,450 adults. Within each of 13 age bands, ranging from 16 to 17 at the lower end and from 85 to 89 at the upper, the sample was stratified according to race or ethnicity (White, Black, Hispanic, other), gender, educational level, and geographic region. In addition to the familiar Verbal, Performance, and Full Scale IQs and the scaled scores on the various subtests, the WAIS-III also provides four new factor scores, Verbal Comprehension, Working Memory, Perceptual Organization, and Processing Speed. Educational Screening Although most correctional psychologists are trained in clinical psychology, in correctional settings they may also have to undertake some assessments that would fall to counseling or school psychologists in the free world. One such task is assessing offenders’ needs for educational programming. Intelligence tests, especially nonverbal and performance measures, are supposed to reflect intellectual ability rather than achievement. On an individual test such as the WAIS- III, it is possible to obtain an above average IQ without being able to read. In assessing offenders’ needs for educational programming, it is essential to evaluate their present educational level and skills. Obviously, the best way to determine how many years of formal education an offender has completed is to check the presentence investigation report. Unfortunately, the number of grades attended may not reflect adults’ actual skills in reading, mathematics, or language. Aiken (2000, p. 118) recently reported that “at least one out of every four employees is functionally illiterate and must ‘bluff it out’ in performing a job requiring reading skills.” Undoubtedly, the illiteracy rate is higher behind bars than in the free world. Therefore offenders’ educational skills should be independently assessed. Test of Adult Basic Education The Test ofAdult Basic Education (TABE; CTB/McGraw Hill, 1987) is a flexible test of basic adult educational skills that is used in a number of correctional settings. It comes in two forms, 5/6 and 7/8, and five levels: L(Literacy; grades 0.0–1.9), E (Easy; grades 1.6–3.9), M (Medium; 3.6–6.9), D (Difficult; (6.6–8.9), and A (Advanced; 8.6–12.9). Relatively brief Loca- tor tests are used to diagnose what level is appropriate for an offender in each content areas. Form 5/6 covers seven content areas (Reading Vocabulary, Reading Comprehension, Mathe- matics Computation, Mathematics Concepts and Applications, Language Expression, Language Mechanics, and Spelling). Form 7/8 covers Reading, Mathematics Computation,Applied Mathematics, Language, and Spelling. Any subtest can be administered independently. For basic screening, Form 7/8’s Reading and Mathematics subtests can be administered in less than an hour. The full TABE battery takes about 3 hr; a con- densed TABE Survey requires 90 min, and the Locator takes about 35 min (CTB/McGraw-Hill). The choice of instrument depends on how detailed an educational evaluation is needed. The test materials were prepared by teachers and drawn from adult basic education texts from around the country. The TABE is often administered to minorities, so great pains were taken to eliminate ethnic biases (Rogers, 1998). The basic evidence of validity is how the test was constructed and its man- ifest content; correlations with external criteria such as grades or GED scores are not provided (M. D. Beck, 1998; Rogers). Although more technical data are needed, the TABE provides correctional users with a broad array of testing options. In concept and design, it reflects current educational practices (Lissitz, 1992). An especially attractive feature of Form 7/8 for corrections use is that norms are provided based on 1,500 adult and juvenile offenders (M. D. Beck, 1998). Malingering on Intellectual and Achievement Measures The basic assumption in most ability and achievement testing is that those being evaluated are motivated to perform at their best. Unfortunately, this is not always the case in assessing criminal offenders, so correctional psychologists must be alert to possible malingering. Unlike personality assessment devices, intelligence and achievement tests do not have validity scales that reflect fake- bad tendencies, so appraisal of malingering must be based on other criteria. Correctional psychologists should keep the purpose of any assessment in mind, and ask themselves whether poorly performing offenders might think it is advisable to appear intellectually challenged. Although forensic assessment is beyond the scope of this chapter, correctional psychologists might find themselves evaluating offenders who are trying to establish a basis for a challenge to their criminal responsibility or legal competency. To take an extreme example, a death row inmate has an obvious incentive for being evaluated as not competent for execution (Small & Otto, 1991). A marked discrepancy between the intellectual level indicated by the case history and the results of intelligence testing is another red flag. Although there has been relatively little research on criminal offenders’ malingering on intelligence and achievement tests, researchers in other settings have examined the factors associated with deliberately poor performance on these Concluding Comments 383 measures. Some of the earliest studies were designed to detect draftees trying to evade induction into the armed services by feigning mental illness. More recent research has focused on patients feigning neurological disorders and memory deficits in conjunction with damage suits. Individual Intelligence Tests Schretelen (1988) reviewed 11 studies, many of which used individual intelligence tests such as the WAIS. He reported that the most reliable signs of malingering were absurd or grossly illogical answers, approximations, and inconsistent performance across tests or subtests. He concluded that, “At this point, examination of response ‘scatter’ appears to be the most powerful and well validated detection strategy. It is based on the finding that fakers tend to fail items genuine patients pass, and pass items genuine patients fail” (p. 458). However, he noted that this guideline is difficult to apply on brief scales and those on which the items are arranged hierar- chically in order of difficulty. Schretelen (1988) also noted that it was easier to detect malingering from a battery of tests than it was from any single measure. If, for example, an intelligence test is administered in conjunction with MMPI-2, and the MMPI-2’s validity scales suggest malingering, it would be prudent to question the intelligence test results as well. Symptom Validity Testing Originally developed to assist in the diagnosis of conversion reactions (Pankratz, 1979) and later applied to those feigning neurological and memory impairment (Rees, Tombaugh, Gansler, & Moczynski, 1998; Tombaugh, 1997), symptom validity testing (SVT) has recently been applied to correctional assessment by Hiscock and her associates (Hiscock, Laymen, & Hiscock, 1994; Hiscock, Rustemier, & Hiscock, 1993). In SVT, suspected malingerers are administered a forced-choice, two-alternative test that may appear challeng- ing but that is actually very easy. Hiscock employed two very easy 72-item tests, one of General Knowledge and the other of Moral Reasoning. A typical item on the General Knowledge test was, “Salt water is found in: (a) lakes or (b) oceans.” On two-choice tests, a person answering randomly should get half the items correct merely on the basis of chance. On SVT instruments, malingering is indicated by a score that is significantly lower than chance performance. Hiscock et al. (1994) found that when male prisoners were instructed to take her tests as if they were poorly educated and could not tell the difference between right and wrong, 71% scored below chance on the General Knowledge test and 60% were below chance on the Moral Reasoning measure, whereas none of a control sample of offenders who took the tests under standard instructions scored this low. Coaching inmates on how to fake successfully reduced the hit rates to 60% on General Knowledge and 43% on Moral Reasoning, showing that the SVT technique works best on unsophisti- cated offenders. CONCLUDING COMMENTS Corrections is a growth industry. Scholars at Brown Univer- sity have projected that, if current trends continue, by 2053 the United States will have more people in prison than out (Alter, 2001; given current ratios, everyone else will probably be on probation or parole). As the correctional population grows, so does the need for reliable, valid, cost-effective assessments. The standards issued by professional organiza- tions concerned with correctional health care are an important first step in encouraging correctional agencies to provide offenders with access to mental health care, including objec- tive, reliable, and valid psychological assessment. Few psychologists are trained to deliver psychological services, including assessment, in correctional settings, and few psychological tests and instruments have been developed in correctional settings to address correctional issues. In- stead, correctional assessment has had to rely on personnel and methods from other settings. Psychologists entering the correctional field should be aware that assessment is different in correctional settings. The clients differ, the issues differ, and the situational factors differ. Therefore, they should seek out instruments developed in or empirically adapted for use in correctional settings, and be prepared to determine the norms, patterns, and cutting scores appropriate in their par- ticular settings. Those instruments that have been developed or adapted for use in corrections need to be continually reassessed. Risk- assessment devices need to be cross-validated before they are applied in new settings or to new problems. Studies reviewed in the present chapter showed that models developed in one state did not always work in another, and factors related to one criterion, such as general recidivism, did not necessarily apply to another, such as sexual reoffending. Predictors may also change over time; not long ago, having a tattoo was an item on Walters, White, and Denney’s (1991) Lifetime Crim- inality Screening Form. It is questionable whether that item would be valid today. Despite the difficulties in validating risk-assessment devices, they at least have the advantage of having correction- ally relevant criterion measures against which they can be [...]... classification of female offender MMPI profiles Journal of Crime and Justice, 6, 57 66 Shipley, W C (1940) A self-administering scale for measuring intellectual impairment and deterioration The Journal of Psychology, 9, 371 – 377 Sliger, G L (1992) The MMPI-based classification system for adult criminal offenders: A critical review Unpublished manuscript, Florida State University, Department of Psychology, ... criminal offenders with personality disorders in Sweden Law and Human Behavior, 23, 205–2 17 386 Psychological Assessment in Correctional Settings Hanson, R K (19 97) Development of a brief scale for sexual offender recidivism Ottawa, ONT: Public Works and Government Services of Canada Hanson, R K., & Bussiere, M T (1998) Predicting relapse: A meta-analysis of sexual offender recidivism studies Journal of. .. in the United States, 19 97 Washington, DC: Bureau of Justice Statistics Beck, M D (1998) Review of Tests of Adult Basic Education, Forms 7, & 8 In J C Impara & B S Plake (Eds.), The thirteenth mental measurements yearbook (pp 1080–1083) Lincoln: Buros Institute of Mental Measurements of the University of Nebraska, Lincoln Boer, D P., Hart, S D., Kropp, P R., & Webster, C D (19 97) Manual for the Sexual... performance also correspond to subjective ratings of peak and off-peak times of the day (e.g., Horne & Osterberg, 1 976 ) For example, approximately 40% of college students (aged 18–25) tend to experience peak performance in the evening, whereas most (approximately 70 %) older adults (aged 60 75 ) tend to peak in the morning (Yoon, May, & Hasher, 19 97) Yoon et al (19 97) , as cited in Ishihara, Miyake, Miyasita,... (Eds.), Principles of geriatric medicine and gerontology (2nd ed., pp 640–644) New York: McGraw-Hill Goga, J A., & Hambacher, W O (1 977 ) Psychologic and behavioral assessment of geriatric patients: A review Journal of the American Geriatrics Society, 25, 232–2 37 Gottfries, C G (19 97) Recognition and management of depression in the elderly International Clinical Psychopharmacology, 12(Suppl 7) , 31–36 Grayson,... Chicago: University of Chicago Press Proctor, J L (1994) Evaluating a modified version of the federal prison system’s inmate classification model: An assessment of objectivity and predictive validity Criminal Justice and Behavior, 21, 256– 272 Schretelen, D (1988) The use of psychological tests to identify malingered symptoms of mental disorder Clinical Psychology Review, 8, 451– 476 Shaffer, C E., Pettigrew,... Minneapolis: University of Minnesota Press Quay, H C (1 973 , November) An empirical approach to the differential behavioral assessment of the adult offender Paper presented at the meeting of the American Society of Criminology, New York City Megargee, E I., Mercer, S J., & Carbonell, J L (1999) MMPI-2 with male and female state and federal prison inmates Psychological Assessment, 11, 177 –185 Meloy, J R.,... of health and mental health problems are often accompanied by the administration of medications The combination of health problems, mental health problems, and medication effects and side effects offers a unique array of challenges for the clinician—particularly the clinician who is unaccustomed to the provision of services to older adults Although these challenges are sufficiently daunting in and of. .. discussions of the assessment of culturally diverse young and older adults The kinship systems of older adult ethnic groups are often an important element of their culture Such systems are collections of social relationship that often define group life (Morales, 1999) The system governs the individual’s relationships and status within the culture Older adults tend to rely more on members of their kinship... Justice Standards and Goals (1 973 ) Report on corrections Washington, DC: U.S Department of Justice, Law Enforcement Assistance Administration Rogers, B G (1998) Review of Tests of Adult Basic Education, Forms 7, & 8 In J C Impara & B S Plake (Eds.), The thirteenth mental measurements yearbook (pp 1083–1085) Lincoln: Buros Institute of Mental Measurements of the University of Nebraska, Lincoln National . & Glickman, 19 47; Hewitt & Jenkins, 1946; Quay, 1965) and adult offenders, Quay (1 973 , 1 974 , 1984) defined five adult-offender types: • Type I (aggressive-psychopathic) offenders are the. Checklist for the Analysis of Life History Records of Adult Offenders (CALH). The CACL is filled out by trained correctional of cers on the basis of their observations of the inmates’ behavior during. Bardos (19 97) reported that GAMA IQs correlated .65 with WAIS-R Verbal IQs, .74 with Performance, and .75 with Full Scale IQs. They also obtained correlations of .72 with the SILS and .70 with

Định dạng
Số trang	66
Dung lượng	680,22 KB