Key points are not available for this paper at this time.
A cross-sectional study of dating partners and a longitudinal study of college roommates revealed that the confidence and accuracy of their impressions were often dissociated. For example, relationship length and degree of involvement tended to increase the confidence of people's impressions, but neither variable consistently increased the accuracy of their impressions of their partners' sexual histories, activity preferences, and so on. A third study showed that relationship length and involvement increased the richness of impressions, and richness fostered confidence. The authors conclude that although confidence-accuracy dissociations are surely problematic in some instances, their apparent pervasiveness raises the possibility that confidence may sometimes contribute to relationship quality even when it is unrelated to accuracy.
Swann et al. (Wed,) studied this question.