Underestimating the effects of faking on the validity of self-report personality scales

The influence of faking on the validity of self-reported personality is examined in two experimental studies. Although separate instructional group analyses and moderated multiple regression analyses using instructional group as a moderator indicated statistically significant effects for faking on validity, moderated multiple regression analyses with social desirability scales generally did not. Effect size estimates for faking were much larger for separate instructional groups analyses and for instructional group as a regression moderator than for socially desirability scales as regression moderators. It is concluded that using social desirability scales in moderated multiple regression can substantially underestimate the effects of experimental faking on validity. (C) 2007 Elsevier Ltd. All rights reserved.