The study of networks is a growth area in public management. This article argues that small studies of networks need to be supplemented with large n studies that permit one to include more theoretically relevant control variables and to deal with issues of causality. Using survey data from several hundred agency heads, this article presents a reliable measure of management network activities that has demonstrated substantial empirical import. If the right network nodes are selected, contact information on only a limited number of nodes is needed. Who initiates contacts within the network is also shown to be important.