THE IMPACT OF FAKING AND CURVILINEARITY ON THE ORGANIZATIONAL UTILITY OF PERSONALITY TESTS

The current study reports the findings from a Monte Carlo investigation examining the potential impact on the organizational utility of two independent factors, socially desirable responding (faking) and a curvilinear relationship between personality and job performance, for organizations choosing to use personality tests in a selection setting. This study explores the impact of faking on the observed predictive validity () and utility of the test, as well as the impact on the composition and job performance of the selected group () when using non-faked scores as opposed to faked scores. In addition, this study hypothesizes a curvilinear relationship between personality and performance, and examines the impact of this non-linear relationship on , utility and . The study simulates several variables that are likely to differ across selection scenarios, including the true predictive validity of the test ( ), the correlation between faking and job performance (), selection ratio (), average tenure (T) and various levels of the standard deviation of performance in dollars (). Consistent with previous research, results show a relatively minor impact on caused by faking, though more significant changes in utility and . This relationship is greatly impacted by the correlation between faking and performance (), where results show a detrimental impact on the outcome variables when this relationship is negative. However, when this relationship is positive, faking serves to augment the quality of selection systems. Results also indicate that a curvilinear relationship between personality and performance results in much more significant changes to , utility and the . Implications and directions for future research are also discussed.M.S. in Psychology, December 201