The authors investigate the relative importance of actor and interaction partner as determinants of dyadic behavior. Using the social relations model (D. A. Kenny, 1994a; D. A. Kenny & L. La Voie, 1984), the authors estimate the variance attributable to each determinant plus the reciprocity of behavioral responses from 7 studies. The authors find evidence for moderate behavioral consistency in a person's behavior across interaction partners, little or no evidence that people consistently engender the same behavioral response from others, and preliminary evidence of unique responding to particular partners. They also consider several methodological issues concerning behavioral measurement as well as the implications of the results for the study of accuracy.

The accuracy of behavioral predictions based on minimal information was assessed using the Social Relations Model (D. A. Kenny & L. Albright, 1987). Eighty women in unacquainted groups of 4 completed round robin trait ratings and predictions about extraversion-related behaviors. Each S then had a 5-min videotaped 1-on-l interaction with each of the other group members. Behaviors corresponding to those predicted were coded from the videotapes. Significant consensus was obtained for ratings of Extraversion and for behavioral predictions. Cross-partner consistency emerged for all coded behaviors. Although predictions of behavior were not accurate at the dyadic level, significant generalized accuracy did emerge. Thus, if a person was consensually predicted to be talkative, for example, that person tended to be talkative across partners. Results are discussed in relation to the accuracy of interpersonal perception and Swann's (1984) theory of pragmatic accuracy.

No abstract

Meta-accuracy, knowing how others view the self, was examined using the Social Relations Model. Fifteen groups of 4-6 acquainted individuals gave self-ratings, perceptions of other group members, and estimated others' perceptions of self (metaperceptions) on the Big Five and Interesting. Individuals also rated liking and metaperceptions of liking. Trait perceptions were consensual, and self-other agreement emerged for most traits. Affect judgments were entirely relational; individuals differentiated among targets. Trait metaperceptions were dominated by perceiver variance. Individuals differed in the impression they believed others generally held about them. Affect metaperceptions, however, were relational in nature. Correlations between perceptions and metaperceptions assessed 2 types of meta-accuracy. Generalised meta-accuracy was obtained for some trait ratings. Affect judgments revealed significant dyadic meta-accuracy.

This study focuses on the relationship between sport volunteerism and social capital, defined here as a resource that stems from participation in certain social networks. A position generator and a resources generator were used to measure the social capital of respondents. Results from this pilot study survey, exploring several aspects of volunteerism in sport in two Canadian communities (one in Québec, the other in Ontario), show a strong relationship between volunteerism in sport and social capital but do not allow a precise measure of the direction of this relationship. Results also show stronger relationships between sport volunteerism and social capital when we control for gender, language, and age.