Applying generalized analysis of covariance for accommodating individual differences: A study effect of faking on personality test

Abstract

Previous studies suggest that there are individual differences that affect the way that people fake their responses on personality measures to varying degrees. These factors should be taken into account to obtain more accurate information regarding faking behavior. This study demonstrated an analysis technique that can accommodate individual differences in an experimentally induced faking study. This mixed-design experiment used two randomized groups (honest vs. faking condition), who each completed a five-factor personality measure twice. This study analyzed data using the generalized ANCOVA (g-ANCOVA) as an alternative to the traditional ANCOVA, since the g-ANCOVA can accommodate both individual differences in prior manipulation (covariates) and interaction, estimating the effects of inducement to fake. We also demonstrated the use of EffectLite, a program for the univariate and multivariate analysis of unconditional, conditional, and average mean differences between groups, and which supported the present study by providing analysis using g-ANCOVA.