Consensus and meta-accuracy in self-managing work groups: A social relations analysis

Date of Completion

January 1996


Psychology, Industrial




This research investigated peer appraisal within simulated self-managing work groups. Over repeated occasions across time, groups of undergraduate students were asked to work on a cooperative group task, to evaluate one another's performance, and to estimate how others in the group evaluated them. Social Relations Model (SRM, Kenny & La Voie, 1984) analyses explored the degree and effects of consensus and meta-accuracy in evaluation. Analyses found evidence for both consensus and meta-accuracy within a short time (p $<$.05). Therefore, consensus in and meta-accuracy of evaluation emerges quickly (after only thirty minutes of cooperative work). However, across time, evaluations were more directly related to immediate performance information than to previous evaluations.^ Further analysis found that evidence for generalized meta-accuracy was greater than for dyadic meta-accuracy, suggesting that people were more aware of their general impression than particular peers' evaluations. And, aspects of performance differed with respect to consensus and meta-accuracy. Although consensus quickly reached significance for ratings of concrete variables (esp., individual contribution), similar increases were not found for variables more subject to interpretation (e.g., accuracy). Finally, impression management had both direct and indirect effects (interactions with previous peer evaluations) predicting future performance (p $<$.10). In the early stages of acquaintance, some people, impression managers, are sensitive to how others evaluate them and may actively change performance output to affect future evaluations.^ Conclusions suggest, first, that peer appraisal systems are a viable measure of individual performance. However, careful thought should be given to the aspects of performance evaluated, both to ensure adequate samples of behavior and a common understanding of performance standards. Results also suggest that some people are sensitive to others' evaluations and may, early on, change behavior to affect those evaluations. More generally, because they estimate the rater, ratee, and relationship effects on social interaction data, SRM analyses provide a useful apparatus for appraisal research. Researchers should be searching for analytic techniques that take advantage of the power of their data to aid organization in designing successful evaluation systems. The Social Relations Model represents one such valuable tool. ^