Personality Assessment Inventory (PAI).

Devised by:  The PAI was authored by Lesley C. Morey and made copyright in 1991 by Psychological Assessment Resources.

Type of Instrument:  The PAI is a self-administered, objective inventory of adult personality and psychopathology.  The PAI contains 344 items comprising 22 nonoverlapping full scales:  validity scales, clinical scales, 5 treatment scales and 2 interpersonal scales (Morey 1991).  The  PAI has been developed in several computerised forms and can be used in a shortened form.

Description:  The PAI measures manifestation of clinical syndromes which were selected based upon their historical importance in classification of mental disorder and their significance in contemporary diagnostic practice (Morey 1991).  The PAI provides information to assist diagnosis, treatment and screening for psychopathology which parallels DSM-IV categories.

Clinical scales are clustered in Neurotic, Psychotic, Personality Disorders and Behavioural Disorders.  In addition to measurement of clinical constructs, interpretation of results also provides measures for detecting Malingering; evaluating potential for Aggression and Suicide; and motivation for Treatment.

The development of the scale used a cluster analysis rather than a two point code type so that scales would be useful across a number of different applications.  Profile interpretation can be made as a two-point code but the author warns against this method of assessment, “...the reliability of the small differences that can determine a two-point code on any psychological instrument is often suspect” (Morey 1996).

The PAI requires a Grade 6 reading level and takes about 40-50 minutes to complete.  The four choice per answer, from False to Very True reduces resistance to forced choice.  Low functioning clients may experience difficulties.  It is not designed to provide a comprehensive assessment of normal personality.  


Internal consistency estimates for PAI full scales across different demographic strata in census-matched normative sample ranged from .75 to .79 (Morey 1991).

Split-half/Cronbach’s Alpha: The full scales of the PAI have median alphas of .81, .86 and .82 for normative, clinical and college samples respectively.  Internal consistency estimates ICN and INF tend to be lower than those for other scales because they measure the care with which the test was completed rather than a theoretical construct (Morey 1991).


Test-Retest Reliability: The test-retest evaluation indicates that one can be 95% certain that a subject’s true score on a PAI full scale falls within 1.96 standard errors of measurement of the observed score.  This confidence interval is useful when attempting to evaluation the meaning of changes in scores over time.  Changes that do not exceed this interval are not likely to be clinically meaningful (Morey 1991)


Alternate Form Reliability:  The PAI was designed to maximize the utility of information gathered in the first part of the test.  The ordering of the items allows a sampling of item content on nearly all scales and subscales within the first 160 items.  This short form has optimal stability and internal consistency characteristics.  The median internal consistency (coefficient alpha) for the 20 scorable scales is .76, whereas the median test-retest reliability is .79.  The median correlation between the short form and full scale scores is .91.  However, for most applications the short form scores should not be considered an acceptable substitute for administration of the complete inventory (Morey 1991).


Validity & Factor Analysis

Construct Validity:  The construct validation framework was theoretically based.  Items were included in the final inventory based on conceptual nature and empirical adequacy as part of a sequential construct validation strategy..  Test development intended that no single quantitative item parameter should be the sole criterion for item selection.

Predictive Validity:  The PAI uses two scales to assess deviations from conscientious responding (INF and ICN) and two to provide information on impression management by the respondent (NIM and PIM).  A Malingering Index is generated by an actuarial measure of tendency to distort the profile negatively in a systematic manner.  A Defensiveness Index is generated through the instrument to indicate the probability of profile distortion to disguise psychopathology.

Convergent and Discriminant Validity:  The comparatively short history of the PAI limits the accumulation of data required for convergent and discriminant validity.  However, A number of correlational studies have measured the PAI validity scales against other scales measuring similar constructs (Ban, Fjetland, Kutcher, & Morey, 1993;  Costa & McCrae, 1992;  Morey, 1991).  The NIM scale correlated significantly (r=.54) with the MMPI (MMPI; Hathaway & McKinley, 1967) F scale;  the PIM was associated with the Marlowe-Crowne (Crowne & Marlowe, 1957) Social Desirability scale (r=.56) and the MMPI K (r=.47) and L (r=.41) scales (Morey, 1991).  As expected the PAI scales INF and ICN produce negligible correlations with other measures since these are designed as indicators of measurement error.

Scoring Method:

Scoring can be done manually or by computer.  Scores are presented in the form of linear T scores with a mean of 50 and standard deviation of 10.  Interpretive hypotheses may be generated at four different levels: item, subscale, full scale and configuration level.  The Professional Manual (Morey 1991) provides detailed guidelines on the interpretation of scale scores and profile configurations.


The PAI was standardised for clinical assessment of individuals aged 18 and over.  The PAI scales and subscales are transformed to T scores to provide comparison with a standardised sample of 1,000 community-dwelling adults.  The sample was carefully selected to match projections of 1995 U.S. Census for race, age and gender.


Ban, T. A., Fjetland, O.K., Kutcher, M., & Morey, L.C. (1993). CODE-DD:  Development of a diagnostic scale for depressive disorders.  In I. Hindmarch & P. Stonier (Eds.), Human pscyhopharmacology: Measures and methods. Vol. 4, pp. 73-86). Chichester, England: Wiley

Costa, P.T., Jr., & McCrae R. R. (1992).  Normal personality in clinical practice:  The NEO Personality Inventory.  Psychological Assessment: A Journal of Consulting and Clinical Psychology.

Crowne, D.P., & Marlowe, D. (1964).  The approval motive:  Studies in evaluative dependence. New York: Wiley.

Hathaway, S. R., & McKinley, J.C. (1967). MMPI manual (rev.ed.). New York: Psychological Corporation

Morey, L.C. (1991). Personality Assessment Inventory - Professional Manual. Florida, USA: Psychological Assessment Resources, Inc.

Morey, L.C. (1996) An interpretive Guide to the Personality Assessment Inventory (PAI). USA: Psychological Assessment Resources, Inc.

Morey, L. C., Waugh, M.H., & Blashfield, R.K. (1985).  MMPI scales for DSM-III personality disorders:  Their derivation and correlates.  Journal of Personality Assessment, 49, 245-251.


Above written by: Ms. Kate Earl (September 2001)

Reviewed, edited and approved by: Dr. Grant J. Devilly