Easier When Done Than Said

Evidence for the criterion validity of indirect self-esteem measures is still limited, with only some studies finding effects of implicit (ISE) independent of explicit (ESE) self-esteem. This may be due to the fact that studies predicting actual behavior are particularly rare. The present study contributes evidence to the predictive validity of the Implicit Association Test and a recently developed self-judgment task under cognitive load. We used criteria beyond self-report: experimenter ratings of anxiety, linguistic aspects of anxiety, and spontaneous self-confident behaviors. Using paired criteria, we tested a double dissociation of ISE and ESE. Results supported our hypothesis: ISE predicted self-confident behaviors or aspects of anxiety that ESE was not able to predict, and vice versa. Thus, differential predictive validity of both measures of self-esteem was demonstrated. With behavioral criteria that tapped into impulsive processes, ISE was a better predictor than ESE was.