The Personality Assessment Inventory is mostly for psychologists whom test for psychopathology. This test is used in psychotherapy, crisis evaluation, employment selection, forensic, child custody, and pain/medical evaluation. This personality test is one of the best assessments available and is more reliable and valid than most others. A thorough description of the test including the structure and scales is given. By evaluating this assessment from a practical and technical perspective, mental health professionals can see the many benefits of this tool. Practical and concise, it gives an overview of psychopathology. By examining the reliability and validity, one can see that it is a trusted test. However, the self-report aspect may pose problems since it requires honesty. The strengths and weakness are examined that may warrant future research.
Personality Assessment Inventory Critique
The title of the test is Personality Assessment Inventory. The author is Leslie C. Morey, the publisher is Psychological Assessment Resources, and the publishing date was 1991. The cost of the Examination Kit is $315.00, which includes the manual, two reusable testing books, two administration books, 25 hand scorable answer sheets, 25 adult profile and critical item forms. The test manual alone is $74.00. An online version is not available. The Personality Assessment Inventory (PAI) was first developed in 1991 for psychologists who tested personality and psychopathology in psychotherapy, crisis/evaluation, employment selection, forensic, child custody, and pain/medical evaluation (Morey, 1997). Corrections and forensics use this test routinely.
This test is an objective rather than projective measure and is a personality test. An alternate form was created for adolescents in 2007, which is called the PAI-A. This is used for adolescents 12 to 18 and only requires only a fourth-grade reading level and 45 minutes to finish. The PAI-A also test for borderline personality characteristics, which other tests do not. The Personality Assessment Screener, developed in 1997, a shorter test version, reveals the likelihood that a clinical disorder will appear on the full test (Morey, 1997).
A brief history of the PAI entails describing the need for a method to measure abstract psychological concepts. Psychologists wanted a statistically strong test that could determine personality pathology. The PAI was supposed to be more practical than other assessments because it was more concise, which meant it took less time to complete. This was due to less reading required, modern diagnostic concepts were emphasized, and the attention to clinical management problems (Kurtz & Blais, 2007). In addition, the assumptions of the PAI are that if an individual has a psychopathology, it can be detected quickly.
The PAI has 22 non-overlapping scales that provide a total overview of psychopathology (Morey, 1997). The scales are grouped in Neurotic, Personality Disorders, Psychotic, and Behavioral Disorders. The test is also used to detect Malingering and assessing possible aggression, suicide, and motivation for treatment.
The test measures 11 constructs which are somatization, anxiety, drug and alcohol problems, suicidal ideation, anxiety-related disorders, depression, mania, paranoia, schizophrenia, borderline, and antisocial (Slavin-Mulford). In addition, the constructs are multidimensional, and the theoretical and empirical foundations of the test were based upon previous importance in categorization of mental disorder and modern diagnostic practice (Morey, 1997).
The purpose of the PAI is the contribution of information in assisting diagnosis, treatment and screen for mental disorders that include DSM-IV classifications. Again, this test was designed for psychologists who needed a way to screen for personality and psychopathology in psychotherapy, crisis/evaluation, employment selection, forensic, child custody, and pain/medical evaluation (Morey, 1997).
The structure of the test includes 344-item with 10 scales and subscales that are conceptually derived. In addition, it has four different scales such as validity scales that measure the test taker’s approach to the test. This includes whether they are faking good or bad, defensive, or exaggerating. It also has clinical scales, which are similar to psychiatric diagnostic groups, and treatment consideration scales, that assess factors relating to treatment. Risk factors that may not be captured in diagnoses such as suicidal ideation are detected. Finally, it contains interpersonal scales that contribute information concerning personality functioning. They were developed to detect a warm friendly or cold rejecting and a controlling or submissive style.
The treatment scales provide indication of possible complications in treatment that may not be detected from diagnostic results. These five scales detect potential harm to others or self, two evaluations of the test taker’s circumstances, and one indicator of treatment motivation (Morey, 1997). A Borderline Features scale has several subscales due to the complexity of the constructs. Finally, the Antisocial Features has three dimensions that evaluate antisocial behaviors and traits.
The test has been developed in computerized format and also a shortened form. The scale used a cluster analysis instead of a two point code so scales would be practical for many applications. Interpretation of results can be made as a two-point code but may question the reliability (Morey, 1997). In addition, the test uses a 4-point item response format that ranges from false to very true.
Test administration procedures include explaining instructions to respondents so they clearly understand. Special attention should be given to marking the correct sex in the gender specific forms. The necessary administration qualifications are a doctorate degree in psychology or related field such as counseling, education social work, human resources, etc. However, an individual who is directly supervised by the previous qualifications is able to administer it (Morey, 1997).
Few special testing conditions must be considered. The test requires a sixth grade reading level and takes approximately 40-50 minutes to finish. Four choices per answer range from False to Very True, which reduces forced choices. It does not provide a comprehensive evaluation of normal personality (Morey, 1997). In addition, when circumstances do not provide for a desk or table, the handy PAI Administration Folio can hold the booklet and answer sheet and still has a hard surface for completing the inventory.
Test scoring is relatively simple because no scoring keys are needed and responses are entered on a two-part carbonless answer sheet that contains scores for all items. The test is hand-scored so there is no waiting for scores to be mailed. Also, software is available for administering, scoring, and interpretation. PAI and subscale raw scores are changed to T scores, which have an average of 50 and a standard deviation of 10. Thus, T-score values greater than 50 are above average in the standard sample. In addition, T scores equal or more than 70 detect an abnormal response compared to the normative sample (Morey, 1997).
The standardization sample was 1,000 community-dwelling adults who were matched based on gender, age, and race. Also, a sample of 1,265 patients and 1,051 college students were used in the assessment. This variety of settings makes the profiles comparable with clinical and normal populations (Wise, Streiner, & Walfish, 2010).
The procedure followed in obtaining the sample was very adequate (Wise et al., 2010). In order to empirically evaluate the test, the test developers administered two versions of the PAI, one to a group of college students and then to a normative sample. The sample was selected carefully to match the U.S. Census for race, gender, and age. Thus, the standardization sample is very adequate.
The norms are very adequate because the first sample of 1000 individuals with similar demographics is useful in detecting the severity of mental disorders (Wise et al., 2010). The second sample of 1,246 psychiatric patients facilitates evaluation of severity of psychopathology relative to other patients.
The evidence of reliability is adequate (Wise et al., 2010). Using several criteria, they assessed the internal consistency of the scales and the ability to fake good or bad. This is adequate reliability evidence to support potential uses of the test.
Reliability studies show that the PAI is very internally consistent, which means that results are stable over 2-4 weeks. The split-half/Cronbach’s alpha correlations exceed .80 for all scales. Test-retest reliability reveals that do also, which means that one can be almost certain that an individual’s true score is within 1.96 standard error of measurement. This is a very good confidence interval (Wise et al., 2010). Thus, reliability evidence is adequate to support potential uses of the test.
Validity studies show convergent and discriminant validity with 50 plus measures of psychopathology such as the Minnesota Multiphasic Personality Inventory (MMPI) assessment (Kurtz & Blais, 2007). In addition, the construct validation was based on theory. Items were based on empirical support and conceptual nature. The PAI has predictive validity, too, because it uses two scales to evaluate deviations from conscientious responding and detect impression management by respondents (Kurtz & Blais, 2007). Malingering and defensiveness can be predicted, too. Therefore, validity evidence is adequate to support potential uses of the test.
The quality of test materials is very good and the reading level is appropriate. However, some low-functioning individuals may experience difficulty. The graphic design and presentation of test material are of good quality, also. In addition, the PAI is very easy to administrate, score and interpret. All constructs in the PAI are commonly used by psychologists so they can be easily understood (Morey, 1997). The instructions are clear and comprehensive. All it requires to administer the test are the manual, test booklet, answer and profile sheet, and template. The test can be taken in 50-60 minutes. They can be hand scored or computerized within 15-20 minutes so they do not have to wait to be mailed back. Interpreting the scores is quick due to simple translation of scores to T scores. The scale names are easy to understand, and the interpretation and diagnostic information provided from the computer report facilitate assessment (Morey, 1997). In addition, unlimited-use interpretive software is available.
Summary Evaluation and Critique
The PAI has much strength. For example, the four-point scale prevents respondents from being forced to choose an answer that does not truly reflect them. The scale is also economical with only 344 items and is easy to understand. In addition, the results are compared to two large samples, which insure that The PAI has much statistical strength shown by the content and discriminant validity (Kurtz & Blais, 2007).
In contrast, the test has weaknesses. Because the test is a self-report questionnaire, it must rely on the honesty of the individual. Accuracy is also an issue because some individuals are not self-aware and do not know what their true feelings are. Thus, it may be important to supplement the PAI with other tests. Moreover, when interpreting the results, non-English speaking administrators must be cautious. Finally, researchers have shown concerns about reliability and validity. For example, a study revealed that the PAI-A scales did not directly correspond with categories in the DSM (Boyle and Lennon, 1994).
The recommendations for uses of the test include assessment for disorders of personality, assessing aggression and violence, and detecting institutional adjustment of female prisoners (Kurtz & Blais, 2007). The revisons should include using psychiatric patients instead of university students to role-play the faking bad or good (Baity, Siefert, Chambers, & Blais, 2007). Psychiatric patients would better understand how to act out these behaviors. In addition, counselors need to be aware of their client in regards to choosing assessments. Some assessments are more geared towards college students than toward clients with mental disorders. Normal data cannot be generalized to clinical samples (Wise, 2010). Problems related to diversity have to be considered. Ethnic and socioeconomic differences matter. Further study should be focused on reliability data from clinically representative populations. In addition, future research should focus on the validity and practicality of the Clinical and Treatment scale. This would increase confidence in the construct validity of the PAI (Kurtz & Blais, 2007). Another concern is the validity scales’ efficacy in detecting response distortion. For example, detecting carelessness to malingering may warrant more study (Kurtz & Blais, 2007).