Testing personality in student selection: dispositional but not situational characteristics predict faking

The use of personality tests in selection procedures is controversial because of their susceptibility to faking. The purpose of this study was to examine the extent and prediction of faking, including dispositional and situational characteristics in a within-subjects design. We compared Big Five scores from a low-stakes and a high-stakes situation in candidates applying for initial teacher education. Across all Big Five traits, participants scored significantly higher in the high-stakes situation than in the low-stakes situation. We found that the extent of faking depended on the personality trait, with high effect sizes in emotional stability (<i>d</i> = 0.94) and low effect sizes in extraversion (<i>d</i> = 0.29) and agreeableness (<i>d</i> = 0.19). Results from hierarchical regression analyses indicated that male gender and intelligence positively predicted faking in certain personality traits. Unexpectedly, emotional stability and conscientiousness also positively predicted faking. Situational factors could not explain incremental variance in the criteria over and above the dispositional factors. Overall, the amount of variance explained in faking was low, stressing that we still know too little about the interindividual differences in faking and thus to address faking as a systematic process.