Abstract: This talk will survey my recent work on empirical models of network formation. I will emphasize procedures which incorporate unobserved agent characteristics (i.e., heterogeneity). These characteristics may covary with observed agent attributes or policy variables of interest to the researcher.