When does random responding distort self-report personality assessment? An example with the NEO PI-R

Recent meta-analyses suggest that, notwithstanding almost 100 years of study, concerns about the impact of response biases on the validity of self-report personality scales remain to be justified. This study addresses the topic by demonstrating that random responding is indeed an issue for self-report personality scales, and that limitations associated with base rates exist that affect the ability of moderated multiple regression to evaluate fully the effects of a response bias. Overall, it is suggested that dismissing response biases as a concern for self-report personality assessment is premature when based only on a lack of significant regression moderator effects.

► Random responding moderates personality scale validity only when random responding is non-extreme. ► When incapable of moderating validity, random responding scales can identify invalid data. ► In combination with faking research, findings apply generally to response biases. ► Emphasizing only scale validity fails to address individual respondent protocol validity. ► Dismissing response biases based only on non-significant moderator effects is premature.