Assessing the Validity of a Theoretically-Based Substance Abuse Scale for Personnel Selection

This paper describes the development of a theoretically-based substance abuse scale to be used for personnel selection. Twenty-eight studies were conducted to validate the predictor. A meta-analysis was conducted using all 28 of the studies (N = 4,118). The mean validity coefficients for bare bones and fully corrected analyses were .25 and .43, respectively. The studies were analyzed according to the type of subject (students, applicants, employees), method used to collect the criteria (self-report, not self-report), and job relevance of the criteria (job-related, not job-related). The method used to collect the criteria was a moderating variable: mean validity coefficients for studies using self-report and nonself-report criteria were .64 and .33, respectively.