Personality Test Validity Differs Between Low-Stakes and High-Stakes Employment Settings

Abstract

The impact of applicant faking on personality test validity in high-stakes settings remains debated in personnel selection research, with some arguing it distorts scores while others suggest minimal effects on validity. This meta-analysis compares personality test validity across low-stakes (e.g., employee assessments) and high-stakes (e.g., applicant testing) settings. Results show validity was consistently higher in low-stakes settings across both unmatched and matched samples. In unmatched studies, personality test validity was higher in low-stakes settings (r' = 0.17, k = 20, N = 8883) than in high-stakes settings (r' = 0.13, k = 215, N = 68,372). Matched studies showed a substantial difference, where low-stakes validity (r' = 0.27) was 125% larger than high-stakes validity (r' = 0.12). These findings provide strong empirical evidence that faking substantially reduces personality test validity in selection contexts. We recommend organizations treat low-stakes validity evidence as provisional and use it only for interim hiring decisions until high-stakes validation data is available. To improve selection accuracy, organizations should prioritize validation studies in motivated samples, apply statistical corrections for faking, and implement faking-resistant measures (e.g., forced-choice formats).