Using the law to define a network for policy analysis and the problem of attrition bias: Three solutions to validate network measures

This presentation will discuss the use of the law to create both a nominally-defined and realist-defined social network and the problem of attrition bias. Specifically, the presentation will address the effect of attrition bias on three network measures: degree, betweeness, and closeness. Using three different sampling approaches: a proportion method, a uniform method and, an opposite method; the presentation will discuss the approaches and their effect on the results of the nominally-defined network when absentee actors are reintroduced to the network based on the three different sampling methods.