Predicting employee theft: A quantitative review of the validity of a standardized measure of dishonesty

One tool used in the preemployment setting for the prediction of theft is thePersonnel Selection Inventory (PSI) (London House, 1980). The PSI is a paper and pencil measure containing three scales: honesty, violence/emotional instability, and drugs. This paper focussed on the usefulness of the PSI dishonesty scale in the prediction of employee theft. Criterion-related validity data from 1,806 persons across 23 samples were quantitatively summarized. All samples met predetermined decision rules. The data indicated that the validity of the PSI dishonesty scale was dependent on a severity of consequences moderator. Validity coefficients are highest when the subject's responses are anonymous, when the theft criteria are self-report measures, when the subjects are students, and when the subjects are aware that the investigator has another source of information on their honesty. In all analyses, the validity of the PSI scale was found to be in useful ranges and showed little variance across studies.The opinions expressed in the report are those of the authors, and do not necessarily represent those of any organizations with which the authors are associated. This paper was neither sponsored nor endorsed by the Defense Personnel Security Research and Education Center. Reprints may be obtained from Michael A. McDaniel, Defense Personnel Security Research and Education Center, 99 Pacific Street, Building 455, Suite E, Monterey, CA 93940.