A model for the comparison of individual and group performance scores

Many investigators have compared the average performance of individuals and of groups working on a standard task. Of additional interest is the question whether groups perform more effectively than their most competent (or least competent) members would have performed if they had worked alone. Methods of answering such questions have been reported in the literature, but practical difficulties have limited their usage. This paper reports an alternative procedure which will sometimes be appropriate when others are not, and which is based upon the logic of sampling theory as it concerns percentile scores.