Prominence indices can be used to classify actors into subgroups according to the roles that they play within a social network. Traditional prominence indices, such as centrality and prestige measures, do not provide an adequate criterion for the classification of actors into clusters. Faust and Wasserman (1992) provide an integrative methodology designed to measure the prominence of network actors based on the dyadic interaction models of Holland and Leinhardt (1981). The raw stochastic centrality and prestige parameters from these models, though meaningful in their interpretation as log odds ratios, are not easily comparable to the traditional measures used to assess actor importance. Standardization of the model prominence parameters will enable one to compare indices across methods. Furthermore, due to the statistical nature of the centrality and prestige indices, standard errors can be calculated and parameter fit evaluated. These standard errors provide a logical criterion for the classification of network actors into overlapping clusters, which represent varying levels of prominence within the network.