Assignment: Clinical Personality Assessments

Chapter 9

PERSONALITY ASSESSMENT INVENTORY

The Personality Assessment Inventory (PAI: Morey, 1991) is a broadband measure of the major dimensions of psychopathology found in Axis I disorders and some Axis II disorders of the DSM-IV-TR (American Psychiatric Association, 2000). The PAI consists of 4 validity, 11 clinical, 5 treatment consideration, and 2 interpersonal scales (see Table 9.1 ). There also are three or four subscales for 9 of the 11 clinical scales and for one treatment consideration scale. Finally, a PAI Structural Summary provides the tables for scoring and profiles for plotting supplemental indices. Table 9.2 provides the general information on the PAI.

HISTORY

The PAI (Morey, 1991) was developed following a sequential, construct-validation strategy. The underlying construct for most of the clinical syndrome scales based on the extant research is multidimensional, and so the scale to measure each clinical syndrome was to be composed of several subscales. Once these component subscales were identified, items were written so that the content was directly relevant for each one. Each item in the original item pool of over 2,200 items then was rated by four individuals for its appropriateness for the specific subscale. Then four experts were asked to assign times to the appropriate scale, and items that did not reach 75% agreement either were dropped or rewritten. These items then were reviewed by a bias-review panel as to whether they could be perceived as being offensive on the basis of gender, race, religion, or ethnic-group membership. Any item that was perceived as being offensive or could inappropriately identify a normal behavior as psychopathology was deleted.

Expert judges, who were nationally recognized within the content area of each scale, then were used to sort the remaining items to ensure that each item was related to its actual construct for each scale on the PAI. The overall agreement was 94.3% among these judges for the 776 items that were retained for the alpha version of the PAI.

Groups of college students then completed the alpha-version of the PAI in one of three conditions: (1) standard, in which students were asked to respond frankly and honestly; (2) positive-impression management, in which the students were asked to respond as if they were trying to impress a potential employer; and (3) malingering, in which the students were asked to simulate the responses of a person with a mental disorder. Items for the beta

283

284 Self-Report Inventories

Table 9.1 Personality Assessment Inventory (PAI) scales

Validity Scales ICN

INF

NIM

PIM

Clinical Scales SOM

SOM-C

SOM-S

SOM-H

ANX ANX-C ANX-A

ANX-P

ARD

ARD-0

ARD-P

ARD-T

DEP

DEP-C

DEP-A

DEP-P

MAN

MAN-A

MAN-G

MAN-I

PAR

PAR-R

PAR-H

PAR-P

scz SCZ-P

SCZ-S

SCZ-T

BOR

BOR-A

BOR-I

BOR-N

BOR-S

ANT

ANT-A

ANT-E

ANT-S

ALC DRG

Inconsistency

Infrequency

Negative Impression Management

Positive Impression Management

Somatic Complaints Conversion

Somatization

Health Concerns Anxiety

Cognitive

Affective

Physiological Anxiety-Related Disorders

Obsessive-Compulsive Phobias Traumatic Stress

Depression

Cognitive Affective Physiological

Mania Activity Level Grandiosity

Irritability

Paranoia Resentment Hypervigilance Persecution

Schizophrenia Psychotic Experience Social Detachment Thought Disorder

Borderline Features

Affective Instability Identity Problems Negative Relationships Self-Harm

Antisocial Features Antisocial Behaviors Egocentricity Stimulus-Seeking

Alcohol Problems Drug Problems

Personality Assessment Inventory 285

Table 9.1 (Continued)

Treatment Consideration Scales AGG Aggression

AGG-A Aggressive Attitude

AGG-V Verbal Aggression

AGG-P Physical Aggression

SUI Suicidal Ideation STR Stress NON Nonsupport RXR Treatment Rejection

Interpersonal Scales DOM Dominance WRM Warmth

version of the PAI were selected on six bases: (I) reasonable variability across the construct, essentially an item-difficulty parameter; (2) a positive, corrected part-whole correlation of the item with the total score of the other items on the scale; (3) the corrected part-whole correlation was higher than the correlation with measures of social desirability and positive and negative impression management; (4) a higher correlation with their own scale than other scales; (5) less face valid or "transparent" measures of the construct embodied in the

Table 9.2 Personality Assessment Inventory (PAI)

Authors:

Published: Edition: Publisher:

Website: Age range: Reading level: Administration formats:

Additional languages:

Number of items: Response format: Administration time: Primary scales:

Additional scales: Hand scoring: General texts:

Computer interpretation:

Morey

1991

1st Psychological Assessment Resources www.parinc.com 18+ 4th grade paper/pencil, computer, CD, cassette

Arabic, French Canadian, Korean, Norwegian, Serbian, Slovene

and Swedish 344 False/Not at all True, Slightly True, Mainly True, Very True 40--50 minutes 4 Validity, 11 Clinical, 5 Treatment Considerations, 2 Interpersonal

Subscales for 9 clinical scales and 1 Treatment Consideration scale Self-scoring answer sheet Morey (2003), Morey (2007a)

Psychological Assessment Resources (Clinical: Morey; Corrections: Morey & Edens)

286 Self-Report Inventories

scale; and (6) absence of gender differences. Using these criteria, a total of 597 items were retained for the beta-version of the PAL

The beta-version of the PAI was administered to three groups of individuals: (1) com­ munity adults; (2) clinical patients; and (3) college students with either positive impression or malingering instructions. Similar item characteristics were assessed for the beta ver­ sion of the PAI as were assessed with the alpha version. The final 344 items on the PAI represented the best balance of all these item characteristics, including the requirement that no item could be scored on more than one scale-there are no overlapping items on the PAI.

Normative data for the PAI were collected from three groups: (1) 1,462 community­ dwelling adults from which a subsample of 1,000 were selected who were census-matched; (2) 1,265 clinical patients from 69 clinical sites; and (3) 1,051 college students. The norms for the PAI are based on 1,000 individuals from the census-matched sample. The skyline profile on the standard profile form demarcates two standard deviations above the mean in the clinical sample allowing the clinician to compare the individual simultaneously with both the census-matched and clinical samples (see Figure 9.1).

PAI Scales - Side A 8PAI" 10 11 A C D E y z

110 - ,o: ..,: - 110

70-=------- -~= _36=_-_- 40=------oo: ~=-70 ....

I - 0060

30: 25- 20- ,,_- ,0- 5-

20 : - 25: 20= 20: 1!5: - 5-

- =-20- 15-_- - 15------5--- H'i=------5-----,..,~---=- 5050------ 15- - - - - - - ,o: 5-

15-

0- ,: 10- ,0- 10- 10-

:. 4040 0-

100 15-

80.: 10- 15-

66-

OS- ":ss: ,o: .,_

.,_ 30-

35: 315- 3J-

,.,_

20-

- 100

:_ 80

3!5-

30.: :_ 30

5- 10-

20 _: :_ 20

Raw /CN INF NIM PIM

1 SOM

2 ANX

3 ARD

4 DEP

5 MAN

6 PAR

7 SCZ

8 BOR

9 ANT

10 A.LC

11 DRG

A A.GG

B SUI

C STR

D NON

E RXR

Y DOM

Z WRM

Raw

Tscore Tscore

Figure 9.1 PAI profile form.

Personality Assessment Inventory 287

Short Form of the PAI

The first 160 items of the PAI can be used to provide a reasonable estimate of 20 of the 22 clinical scales for all scales but Inconsistency (JCN) and Stress (STR). These estimates are possible because the items with the largest item-scale correlations were located at the beginning of the test when the final version of the PAI was developed. Table 11.1 in the Personality Assessment Inventory Professional Manual (Morey, 1991, p. 142) provides the descriptive characteristics for these 160 items. The short form only should be used in the most unusual circumstances, and the estimated scores must be considered as generating only the most tentative interpretive hypotheses. Frazier, Naugle, and Haggerty (2006) found that agreement between the short- and full-form of the PAI was affected adversely when the validity scales were elevated. They also noted that individuals with lower levels of ability were more likely to leave items missing and produce invalid protocols. These individuals are the very ones for whom the short form was designed. The hope was that it would provide information about the presence of psychopathology that otherwise might not be available from a self-report inventory.

PAI-A (Adolescent)

As a result of interest by professionals in using the PAI with adolescents in clinical settings, work was begun in 1999 on piloting an adolescent version of the inventory (Morey, 2007b ). The intent of this work was to explore the applicability of an adolescent version that would closely parallel the adult version of the PAL It would retain the structure and, as much as possible, the items of the adult version rather than be an entirely new version targeted specifically at an adolescent population. The development of the PAI-A involved an adaptation of the items of the adult PAI so that the content was meaningful when applied to adolescents. The approach taken was a conservative one-the question was not whether the item was optimized to capture the experience of an adolescent, but rather whether the item would retain its original meaning when read by the adolescent. This conservative approach was merited in that the items on the adult PAI had been selected on the basis of numerous criteria, and the rewording or replacement of items could have significant and unanticipated effects on the final properties of the adolescent version and its interpretability as parallel to the adult version. Thus, these revisions included rewordings of relatively few items and involved close equivalents of the original wording.

The next stage in development involved collecting a diverse and representative sample of adolescent patients, and determining the psychometric comparability of items on the adolescent and adult versions. A relatively small number of items were identified that appeared to have different characteristics in adolescent patients than in adult patients, and the decision was made to explore the impact of elimination of these items. On the basis of these analyses, items were removed in an effort to eliminate the most problematic items and yield an item distribution pattern that would closely parallel the adult instrument. On the basis of this strategy, the final PAI-A included 264 items. The PAI-A was then standardized using a census-matched normative sample of 707 adolescents aged 12 to 18, as well as a diverse clinical sample of 1,160 patients in the same age range. The average internal consistency for the 22 clinical scales was .79 in the community sample and .80 in the

288 Self-Report Inventories

clinical sample, while the average test-retest reliability for these scales was .78 over an

interval of approximately 18 days.

ADMINISTRATION

The first issue in the administration of the PAI is ensuring that the individual is invested in the process. Taking a few extra minutes to answer any questions the individual may have about why the PAI is being administered and how the results will be used will pay excellent dividends. The clinician should work diligently to make the assessment process

a collaborative activity with the individual to obtain the desired information. This issue of therapeutic assessment (Finn, 1996; Fischer, 1994) was covered in depth in Chapter 2 (pp. 43-44).

Reading level is a crucial factor in determining whether a person can complete the PAI; inadequate reading ability (to be discussed) is a major cause of inconsistent patterns of item endorsement. Morey (1991) suggests that most individuals who can read at the fourth­ grade level can take the PAI with little or no difficulty because the items are written on an fourth-grade level or less. The PAI has the easiest reading level of any of the self-report inventories reviewed in this Handbook. As such, one reason for selecting the PAI is the larger number of clients who can complete it successfully compared with the MMPI-2 (Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989) and the MCMI-III (Millon, Davis, & Millon, 1997), both of which are written at the eighth-grade level.

SCORING

Scoring the PAI is relatively straightforward either by hand or computer. A different answer sheet is used for hand scoring (Form HS Answer Sheet) and optical scanning (Form SS Answer Sheet), so the proper answer sheet must be selected for the method of scoring. If the PAI is administered by computer, the computer automatically scores it. If the individual's responses to the items have been placed on an answer sheet, these responses can be entered into the computer by the clinician for scoring or they can be hand scored. If the clinician enters the item responses into the computer for scoring, they should be double entered so

that any data entry errors can be identified. The first step in hand scoring is to examine the answer sheet carefully and indicate

omitted items and double-marked items by drawing a line through all four responses to these items with brightly colored ink. Also, cleaning up the answer sheet is helpful and facilitates scoring. Responses that were changed need to be erased completely if possible, or clearly marked with an "X" so that the clinician is aware that this response has not been

endorsed by the client. The PAI (Morey, 1991) and the NEO PI-R (Costa & McCrae, 1992) are the only self­

report inventories reviewed in this Handbook that do not use "true/false" items. Both of these inventories have the same publisher (Psychological Assessment Resources), which may account for not using "true/false" items. The PAI uses a four-point Likert scale ranging from "false, not at all true," "slightly true," "mainly true," to "very true." These potential response options always are presented in this same order on the answer sheet. When "very

Personality Assessment Inventory 289

true" is the scored direction for a specific item, the response options are scored as 0, 1, 2, or 3 ("very true"). When "false, not at all true" is the scored direction, the preceding four response options are scored as 3 ("false, not at all true"), 2, 1, or 0. Thus, the total raw score on an eight-item scale, which is the characteristic number of items on each subscale of the clinical scales, can range from Oto 24. It is imperative that the clinician realize that the total score is the sum of the response options for each scale, not the total number of items endorsed on the scale, which is the method for scoring the MCMI-III, MMPI-2, and MMPI-A.

The PAI is easier to score than other self-report inventories because no templates are required. The answer sheet, on which the person records his or her responses, is self-scoring. The items on each scale are designated by ruled and shaded boxes that are identified by scale abbreviations. The total raw score for each scale or subscale is entered in the corresponding box with the same abbreviation on Side B of the profile form. The subscales for the various scales on the PAI are plotted on Side B of the profile form. The total scores, which are the sum of the scores on the subscales, for all scales are entered on Side A of the profile form.

Although this process of hand scoring may sound somewhat complex, it is straightfor­ ward and can be carried out in 10 to 15 minutes. It is advisable to have another person double-check all the scoring and transferring of numbers to catch any scoring or transcrip­ tion errors before the interpretive process begins.

ASSESSING VALIDITY

Figure 9.2 provides the flowchart for assessing the validity of this specific administration of the PAI and the criteria for using this flowchart are provided in Table 9.3. The clinician is reminded that the criteria provided in Table 9.3 are continuous, yet ultimately the decisions that must be made in implementation of the flowchart in Figure 9.2 are dichotomous. General guidelines will be provided for translating these continuous data into dichotomous decisions on the PAI, but these guidelines need to be considered within the constraints of this specific client and the circumstances for the evaluation.

Item Omissions

Morey (1991) recommends that more than 95% of the items should be endorsed if the PAI is to be interpreted; that is, no more than 17 (.05 x 344) items should be omitted. Table 9.3 shows that omitting 17 items is somewhere between the 93rd and 98th percentile in both normal and clinical samples. Morey also recommends that more than 80% of the items should be endorsed for any individual scale to be interpreted. The subscales of the clinical scales all have six to eight items, so the omission of two items from one of these subscales (6/8 = 75%) would mean that subscale should not be interpreted, although the entire scale could be interpreted if more than 80% of its items were endorsed.

Consistency of Item Endorsement

Consistency of item endorsement on the PAI is assessed by the Inconsistency Scale (/CN) and Infrequency Scale (INF). The Inconsistency Scale (/CN) scale consists of 10 pairs of

310 Self-Report Inventories

shown on this form also allows the clinician to compare the individual's scores on each scale with the clinical sample.

APPLICATIONS

As a self-report inventory, the PAI is easily administered in a wide variety of settings and for a variety of purposes. Although the PAI was developed as a broadband measure of psychopathology in clinical settings, its use has gradually been extended to forensic and

criminal settings, neuropsychological settings, and medical settings. One of the primary reasons for its rising popularity in these settings is that it is shorter and easier to read than the other self-report inventories.

Somewhat different issues must be considered in the administration of the PAI in per­ sonnel selection and forensic settings compared with the more usual clinical setting. These general issues were covered in Chapter 6 with the MMPI-2 (pp. 197-198) and will not be repeated here, but they should be consulted by anyone who is using the PAI in personnel selection or forensic settings for the first time.

One of the considerations in the use of any assessment test or technique in forensic settings is whether it will meet the legal standards for admissibility. These considerations were raised in Chapter 8 with the MCMI-III (pp. 276-277) because various authors have opined that the MCMI-III does or does not meet these legal standards. Morey, Warner, and Hopwood (2007) have described how the PAI meets the legal standards for admissibility. In a survey of forensic psychologists, Lally (2003) reported that the PAI was rated as being acceptable for the evaluation of mental status at the time of the offense, risk for violence, risk for sexual violence, competency to stand trial, and malingering.

The PAI is increasingly being used in correctional settings because it is shorter and easier to read than other self-report inventories. Edens, Cruise, and Buffington-Vollum (2001) have provided a general overview of the issues involved in using the PAI in forensic and correctional settings. Edens and Ruiz (2006) reported that elevated scores on the Positive Impression Management (PIM > T56) scale in conjunction with elevated scores on the Antisocial Features (ANT > T59) scale predicted institutional misconduct among male inmates. Caperton, Edens, and Johnson (2004) found that elevated scores on the Antisocial Features (ANT > T69) scale identified sex offenders who were more likely to be management risks while in prison. Finally, Kucharski, Duncan, Egan, and Falkenback (2006) found that three levels of psychopathy as measured by the PCL-R were not related to scores on Negative Impression Management (NIM) scale, the Malingering Index (MAL), or Rogers' discriminant function (RDF), and that the criminal defendants with higher levels of psychopathy were not more likely to malinger as measured by the PAI scales.

Finally, the PAI is being used in neuropsychological settings to evaluate whether

the effects of brain injury have produced any psychological sequelae. Demakis et al. (2007) found that 34.7% of their sample of 95 individuals who had suffered a traumatic brain injury did not elevate any clinical scale on the PAI above a T score of 69. This number of unelevated profiles in individuals with brain injury is commonly found (cf. Warriner, Rourke, Velikonja, & Metham, 2003). The most common two-point codetypes were: SCZ/BDL-(Schizophrenia/Borderline Features)-18.9%; DEP/SCZ-(Depression/ Schizophrenia)-12.6%; and SOM/ANX-(Somatic Complaints/Anxiety)-10.5%.

Personality Assessment Inventory 311

PSYCHOMETRIC FOUNDATIONS

Demographic Variables

Age

Morey (1996a) reported age has minimal impact on the PAI scale scores. Individuals who

were 18 to 29 years of age elevated the Paranoia (PAR) scale 5 T points, the Borderline Features (BOR) scale 6 T points, the Antisocial Features (ANT) scale 7 T points, the Aggression (AGG) scale 5 T points, and the Stress (STR) scale 4 T points higher than other age groups. The primary subscale impacted by this elevation in score was Paranoia­ Persecution (PAR-P), Borderline Features-Identity Problems (BOR-1), Antisocial Features­ Stimulus Seeking (ANT-S), and Aggression-Verbal Aggression (AGG-V). There are no subscales for Stress (STR). Individuals who were 60+ years of age lower these same five scales 4 T points. The primary subscale lowered by this elevation was Paranoia-Resentment (PAR-R), Borderline Features-Identity Problems (BOR-1), Antisocial Features-Antisocial Behavior (ANT-A), and Aggression-Physical Aggression (AGG-P).

Gender

Gender does not create any general issues in PAI interpretation because the items were selected to eliminate gender bias. Men elevated the Antisocial Features (ANT) scale by 3 T points more than women (Morey, 1996a). This elevation primarily impacted the Antisocial Features-Antisocial Behavior (ANT-A) subscale.

Education

The potential effects of education have not been investigated in any systematic manner on the PAI, although such research clearly is needed.

Ethnicity

The effects of ethnicity on PAI performance also have not been investigated in any system­ atic manner. Morey (1996a) reported that nonwhite individuals elevated the Paranoid (PAR) scale 6 T points compared with White individuals. This elevation primarily impacted the Paranoid-Hypervigilance (PAR-H) subscale.

Reliability

The PAI Professional Manual (Morey, 1991, Appendix E) reported the reliability data for 75 community-dwelling adults who were retested after an average of 24 days and 80 undergraduate students who were retested at 28 days. The test-retest correlations ranged from .85 to .94 in the adult sample and ranged from .66 to .90 in the student sample across the 11 clinical scales. The standard error of measurement ranges from 2.8 to 4.6 T points for these 11 clinical scales, that is, the individual's "true" score on the clinical scales will be within ±3 to 5 T points two-thirds of the time.

Codetype Stability

There are limited empirical data that indicate how consistently individuals will obtain the same two highest clinical scales on two successive administrations of the PAL Codetype

312 Self-Report Inventories

stability was examined in all 155 individuals who were part of the examination of retest reliability just described. When only the single highest scale was examined across the two administrations, 57.4% had the same high-point scale. When this analysis was limited only to those individuals with significant elevations (20/155), 76.9% had the same high-point scale. These data should only be considered to be an estimate of the actual codetype stability of the PAI. Because only a single high-point scale was considered, there has to be a lower

rate of stability when the two highest scales are required to be the same. On the other hand, clinical samples would produce higher elevations on the PAI clinical scales than these normal individuals and the preceding data suggest that concordance rates would be higher for more elevated profiles.

CONCLUDING COMMENTS

The PAI (Morey, 1991) is the newest of the self-report inventories reviewed in this Hand­ book. The PAI is gradually gaining a wide base of usage because it is shorter than all other self-report inventories except the MCMI-III and it has the lowest reading level of any of them. There has been a substantial increase in research with the PAI in each ensuing year that continues to validate its use in a number of different settings.

REFERENCES

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author.

Bagby, R. M., Nicholson, R. A., Bacchiochi, J. R., Ryder, A.G., & Bury, A. S. (2002). The predictive capacity of the MMPI-2 and PAI validity scales and indexes to detect coached and uncoached feigning. Journal ofPersonality Assessment, 78, 69-86.

Baity, M. R., Siefert, C. J., Chambers, A., & Blais, M.A. (2007). Deceptiveness with the PAI: A study of nai"ve faking with psychiatric inpatients. Journal ofPersonality Assessment, 88, 16-24.

Butcher, J. N., Dahlstrom, W. G., Graham, J. R., Tellegen, A. M., & Kaemmer, B. (1989). MMPI-2: Manual for administration and scoring. Minneapolis: University of Minnesota Press.

Caperton, J. D., Edens, J. F., & Johnson, J. K. (2004). Predicting sex offender institutional adjustment and treatment compliance using the PAI. Psychological Assessment, I 6, 187-191.

Cashel, M. L., Rogers, R., & Sewell, K. (1995). The PAI and the detection of defensiveness. Assess­ ment, 2, 333-342.

Clark, M. E., Gironda, R. J., & Young, R. W. (2003). Detection of back random responding: Effec­ tiveness of MMPI-2 and PAI validity indices. Psychological Assessment, 15, 223-234.

Costa, P. T., Jr., & McCrae, R.R. (1992). Revised NEO Personality Inventory (NEO PI-R) and NEO Five-Factor Inventory (NEO-FFI) professional manual. Odessa, FL: Psychological Assessment Resources.

Demakis, G. J., Hammond, F., Knotts, A., Cooper, D. B., Clement, P., Kennedy, J., et al. (2007). The PAI in individuals with traumatic brain injury. Archives of Clinical Neuropsychology, 22, 123-130.

Edens, J. F., Cruise, K. R., & Buffington-Vollum, J. K. (2001). Forensic and correctional applications of the PAI. Behavioral Sciences and the Law, 19, 519-543.

Edens, J. F., Poythress, N. G., & Watkins-Clay, M. M. (2007). Detection of malingering in psychiatric unit and general population prison inmates: A comparison of the PAI, SIMS, and SIRS. Journal ofPersonality Assessment, 88, 33-42.

Personality Assessment Inventory 313

Edens, J. F., & Ruiz, M. A. (2006). On the validity of validity scales: The importance of defensive responding in the prediction of institutional misconduct. Psychological Assessment, 18, 220--224.

Finn, S. (1996). Using the MMPI-2 as a therapeutic intervention. Minneapolis: University of Min­ nesota Press.

Fischer, C. T. (1994). Individualizing psychological assessment. Hillsdale, NJ: Erlbaum.

Frazier, T. W., Naugle, R. L, & Haggerty, K. A. (2006). Psychometric adequacy and comparability of the short and full forms of the PAL Psychological Assessment, I 8, 324-333.

Hopwood, C. J., Morey, L. C., Rogers, R., & Sewell, K. (2007). Malingering on the PAI: Identification of specific feigned disorders. Journal of Personality Assessment, 88, 43-48.

Kucharski, L. T., Duncan, S., Egan, S.S., & Falkenbach, D. M. (2006). Psychopathy and malingering of psychiatric disorder in criminal defendants. Behavioral Sciences and the Law, 24, 633---644.

Kucharski, L. T., Toomey, J. P., Fila, K., & Duncan, S. (2007). Detection of malingering of psy­ chiatric disorder with the PAI: An investigation of criminal defendants. Journal of Personality Assessment, 88, 25-32.

Lally, S. J. (2003). What tests are acceptable for use in forensic evaluations? A survey of experts. Professional Psychology: Research and Practice, 34, 491-498.

Millon, T., Davis, R., & Millon, C. (1997). MCMI-III manual (2nd ed.). Minneapolis, MN: National Computer Systems.

Morey, L. C. (1991). Personality Assessment Inventory professional manual. Odessa, FL: Psycho­ logical Assessment Resources.

Morey, L. C. (1996a). An interpretive guide to the PAI. Odessa, FL: Psychological Assessment Resources.

Morey, L. C. (1996b). PAI structural summary. Odessa, FL: Psychological Assessment Resources.

Morey, L. C. (1999). PAI interpretive explorer module manual. Odessa, FL: Psychological Assessment Resources.

Morey, L. C. (2003). Essentials of PAI assessment. Hoboken, NJ: Wiley.

Morey, L. C. (2007a). An interpretive guide to the PAI. Odessa, FL: Psychological Assessment Resources.

Morey, L. C. (2007b). Personality Assessment Inventory---A.dolescent professional manual. Odessa, FL: Psychological Assessment Resources.

Morey, L. C., & Hopwood, C. J. (2004). Efficiency of a strategy for detecting back random responding on the PAL Psychological Assessment, 16, 197-200.

Morey, L. C., Warner, M. B., & Hopwood, C. J. (2007). Personality Assessment Inventory: Issues in legal and forensic settings. In A. M. Goldstein (Ed.), Forensic psychology: Emerging topics and expanding roles (pp. 97-126). Hoboken, NJ: Wiley.

Peebles, J., & Moore, R. J. (1998). Detecting socially desirable responding with the PAI: The Positive Impression Management scale and the Defensiveness Index. Journal ofClinical Psychology, 54, 621---628.

Rogers, R., Sewell, K. W., Morey, L. C., & Ustad, K. L. (1996). Detection of feigned mental disor­ ders on the Personality Assessment Inventory: A discriminant analysis. Journal of Personality Assessment, 67, 629-640.

Warriner, E. M., Rourke, B. P., Velikonja, D., & Metham, L. (2003). Subtypes of emotional and behavioral sequelae in patients with traumatic brain injury. Journal ofClinical and Experimental Neuropsychology, 25, 904-917.