Personality testing and response distortion in selection: A randomized, longitudinal field experiment to test the effects of selection, control, and warning instructions on job performance and career development.
Øyvind Lund Martinsen et al.
Abstract
Response distortion (faking) on personality tests in high-stakes selection contexts, and the use of warnings to mitigate it, remain critical issues in industrial-organizational psychology. Research on these issues, however, has often been limited by practical and methodological constraints, leaving notable gaps in our understanding of the processes involved. To address these gaps, we conducted a study where 1,123 participants in a military selection setting were randomly assigned to three experimental groups: low-stakes control, high-stakes selection, and high-stakes warning. Performance indicators included admission to officer training, future officer rank, and job performance. The results showed that group-level trait means changed in the expected directions under high-stakes conditions, while the factorial invariance and predictive validity of the personality measure were preserved. The incremental validity of personality traits over general mental ability remained consistent across groups, supporting the robustness of personality assessments. While warnings reduced the mean score inflation, they failed to enhance predictive validity and negatively impacted convergent validity with an unaffected premeasure of personality. Warnings also resulted in other distortions, particularly among high-ability candidates. Overall, these findings indicate that personality assessments retain practical utility in high-stakes settings when excluding warnings, as such interventions may introduce unintended biases without improving selection outcomes. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
Evidence weight
Balanced mode · F 0.40 / M 0.15 / V 0.05 / R 0.40
| F · citation impact | 0.50 × 0.4 = 0.20 |
| M · momentum | 0.50 × 0.15 = 0.07 |
| V · venue signal | 0.50 × 0.05 = 0.03 |
| R · text relevance † | 0.50 × 0.4 = 0.20 |
† Text relevance is estimated at 0.50 on the detail page — for your query’s actual relevance score, open this paper from a search result.