Faking Good and Bad on Self-Report Measures: A Comparison of 15 Detection Methods

• In Event: Measuring and Modeling Test-Taker Behaviors and Experiences

Abstract

We compared fifteen methods of scoring the Balanced Inventory of Desirable Responding (BIDR; Paulhus, 1991) for accuracy in detecting faked responses. In contrast to previous findings in which 7-point polytomous scoring provided better psychometric results for score validation, statistical control, and outcome assessment, the present results indicate that alternative scoring methods are more effective for detecting both faking good and faking bad. These alternative methods include dichotomous procedures from the research literature as well as newly developed methods described in this paper. Classification accuracy of honest and faked responses for the most effective methods ranged from 88% to 96%. Overall, faking bad was easier to detect than faking good, and impression management were easier to spot than self-deceptive enhancement.

Authors

• Walter Peter Vispoel, University of Iowa

• Murat Kilinc, University of Iowa

• Carrie A. Morris, The University of Iowa