The Marlowe-Crowne affair: short forms, psychometric structure, and social desirability

The Marlowe-Crowne Social Desirability scale (Crowne & Marlowe, 1960) is widely used to assess and control for response bias in self-report research. Several abbreviated versions of the Marlowe-Crowne scale have been proposed and adopted in psychology and medicine. In this article I evaluate the adequacy of 9 short forms using confirmatory factor analysis across 2 samples (combined N = 867). There was some evidence for the adequacy of different short forms, but model adequacy was not consistent across samples. Supplementary analyses revealed a multidimensional structure to the full Marlowe-Crowne scale and indicated that the apparent adequacy of model fit for some short forms might be a statistical artifact. Using the Marlowe-Crowne scale or its various short forms as a control for response bias is discouraged on empirical and conceptual grounds.