Three different approaches to the creation of popularity indices from sociometric data are described. One involves a factor analysis of the sociometric matrix and the other two are different approaches to the weighting of sociometric choices. All turn out to have the same mathematical solution when the relationships are symmetric; it is certain eigenvectors of the sociometric matrix. This technique has the additional benefit of giving the clique structure at a glance. The technique is compared to Hubbell's (1965) method for clique identification. The method is then illustratively applied to structural data on the pattern of overlap in membership among a set of high school activities. A measure of "centrality" in this structure, analogous to individual popularity in sociometric structures, is calculated for each of the activities and the results are compared to common sense expectations about high school activities. (Author)