This dissertation applies the tools of network analysis to study job mobility. Job mobility is a complex phenomenon, and network theory provides a novel and practical framework for dealing with this complexity in understanding how individuals move from job to job. My first essay measures the effect of job referral networks on search outcomes. The key contribution of this essay is providing evidence of one mechanism by which social interactions affect earnings. An onthe-job search model extended to include social transmission of job information yields an empirical specification in which ones current job offer depends on the average offer of his social contacts. Using block level variation in the quality of jobs held by ones residential neighbors, I find that when changing jobs an individual with better local network contacts will obtain a higher quality job. In addition to the main result, this paper provides new evidence on the spatial structure of the wage distribution within urban areas. In the second essay I apply network algorithms to detect groups of workers and employers with relatively homogeneous patterns of job mobility. Workers with interchangeable skills should have similar patterns of mobility across employers that use those skills in roughly the same way. Grouping workers and jobs solely on the basis of similar mobility patterns reveals labor market sectors with distinct compensation structures. My final essay, joint with John Abowd, uses network models to facilitate identification of employer-specific wage premia in a decomposition of log earnings from matched employer-employee data.