METRIC INFERENCE FOR SOCIAL NETWORKS

Using a natural metric on the space of networks, we define a probabili ty measure for network-valued random variables. This measure is indexe d by two parameters, which are interpretable as a location parameter a nd a dispersion parameter. From this structure, one can develop maximu m likelihood estimates, hypothesis tests and confidence regions, all i n the context of independent and identically distributed networks. The value of this perspective is illustrated through application to porti ons of the friendship cognitive social structure data gathered by Krac khardt (1987).