Workshop: Handling Missing Network Data

Handling Missing Network Data
Presenter:  Robert Krause, Linköping University

In this workshop I will present the most common methods for handling missing network data, recent developments, and the state of the art. We will discuss both the theoretical foundations as well as how to apply these methods in practice, their assumptions and caveats. 

The workshop will teach participants how to obtain multiple imputations for missing tie variables (and nodal attributes) in cross-sectional and longitudinal network data, and how to proceed from there. The imputations can be used for any form of network analysis (e.g., ERGMs, SAOMs, Blockmodels, estimating descriptives). The imputation procedure makes use of (B)ERGM and SAOM models, thus familiarity with at least one of these model families will be very helpful in following the workshop. Knowledge obtained in the introductory ERGM or SAOM workshops given at the Sunbelt is sufficient. The workshop will be in R and rudimentary knowledge of R is required for the application of the procedure. However, the workshop is explicitly open for all network researchers, who encounter missing tie variables (e.g., because of actor non-response).

Capacity:  95