FIXED LIST VERSUS SNOWBALL SELECTION OF SOCIAL NETWORKS

Two ways of collecting network data are discussed in the context of interorganizational networks among social service agencies. One method proceeds with a fixed agency list while the other uses a snowball selection procedure. We find that: 1. (i) the agencies included in the two selection procedures differ; 2. (ii) network-based indicators for agencies (such as centrality) differ; 3. (iii) the substantive contents of the included ties differ; and 4. (iv) the structure of the networks differ. The two methods are fundamentally different and cannot be used as surrogates for each other. When data are gathered at multiple points in time, the differences between the methods are more acute. Interagency networks change through time: agencies leave the network while other agencies join. Snowball selection can respond to these changes while fixed list selection does not. © 1992.