Confirmatory factor analyses of scores from full and short versions of the Marlowe-Crowne social desirability scale

Self-report measures are much used in social psychology. However, such measures are susceptible to response bias. The 33-item Marlowe-Crowne Social Desirability scale (Crowne & Marlowe, 1960) is widely used to detect social desirability in responding. This study used confirmatory factor analyses and item and scale analyses to evaluate different, time-saving, short versions of the scale. The results from 633 management undergraduates showed that all the short versions and the 2-factor model are a significant improvement in fit over the full scale. There were no gender differences on the 14 measures. It is recommended that when the full scale is used, the separate attribution and denial scores are also used. Researchers who decide to use a short version should seriously consider Ballard's (1992) Scale I or composite versions because the present study identified these as the best short versions.