Nontraditional Validity Calculations and Differential Interviewer Experience: Implications for Selection Interviews

Abstract

The traditional way to calculate the criterion-related validity of selection interviews was compared to two nontraditional techniques. College admissions interview ratings were the predictors, and college GPAs were the criteria. Traditional interview-GPA correlations usually were not significant, whereas statistical techniques set forth by Dreher, Ash, and Hancock (1988) and Kenny and La Voie (1985) usually yielded significant validities. The differences were more pronounced when separate analyses were conducted for groups of experienced and inexperienced interviewers. Only the latter ratings provided valid predictors of GPA, and other analyses suggested that the underlying judgments were more complex.