The Effects of Faking on the Construct Validity of Personality Questionnaires: A Direct Faking Measure Approach
Authors
Maša Tonković Grabovac
Department of Psychology, Faculty of Humanities and Social Sciences,
University of Zagreb, Croatia
Željko Jerneić
Department of Psychology, Faculty of Humanities and Social Sciences,
University of Zagreb, Croatia
Zvonimir Galić
Department of Psychology, Faculty of Humanities and Social Sciences,
University of Zagreb, Croatia
Keywords:
faking, personality questionnaire, construct validity, direct measure of faking
Abstract
Some authors clearly showed that faking reduces the construct validity of personality questionnaires, whilst many others found no such effect. A possible explanation for mixed results could be searched for in a variety of methodological strategies in forming comparison groups supposed to differ in the level of faking: candidates vs. non-candidates; groups of individuals with "high" vs. "low" social desirability score; and groups given instructions to respond honestly vs. instructions to "fake good". All three strategies may be criticized for addressing the faking problem indirectly – assuming that comparison groups really differ in the level of response distortion, which might not be true. Therefore, in a within-subject design study we examined how faking affects the construct validity of personality inventories using a direct measure of faking. The results suggest that faking reduces the construct validity of personality questionnaires gradually – the effect was stronger in the subsample of participants who distorted their responses to a greater extent.
