Capturing the biases of socially anxious people by addressing partner effects and situational parameters.

To expose biases in self-perceptions of people high in social anxiety, information is needed on actual and perceived informant reports following social situations. We measured trait social anxiety (SA) in 90 college students arranged in pairs for "getting acquainted" conversations. Half participated in a small-talk task, where they took turns answering superficial questions; half participated in a closeness-generating task, where questions required gradual increases in self-disclosure. Afterward, students rated themselves and their partner on positive and negative attributes and how they think their partner viewed them. People with high SA judged themselves more negatively and less positively than their partner did (accuracy); when interacting with a partner endorsing low SA, they possessed enhanced negativity biases about how they expected to be viewed (meta-accuracy), and believed their partner's judgments were less positive than their own low self-judgments (perceived dissent). Conversely, people with low SA showed evidence of a self-enhancement bias about the impression they made on low SA strangers. Other moderators of the social cognitions of people with high SA included gender and the social situation (distortions being amplified in men and small-talk conversations). Our findings suggest that the study of SA cannot be understood using decontextualized approaches, instead requiring consideration of the synergy among the person, partner, and situation.