Prof. KENNETH A. BOLLEN
Departments of Psychology and Sociology
University of North Carolina, Chapel Hill
Abstract:
Instrumental variable (IV) methods provide a powerful but underutilized tool to address many common problems with observational data. Key to their successful use is having IVs that are uncorrelated with an equation’s disturbance and that are sufficiently strongly related to the problematic endogenous covariates. This presentation will briefly define IVs, summarize their origins, and describe their use in multiple regression, simultaneous equation models, factor analysis, latent variable structural equation models, and limited dependent variable models. I will define and contrast three methods of selecting IVs: auxiliary instrumental variable, model implied instrumental variable, and randomized instrumental variable. Also discussed are over identification tests and weak IV diagnostics as methods to evaluate the quality of IVs. I will review the use of IVs in models that assume heterogeneous causal effects. My concluding remarks will suggest ways to improve the use of IVs. The MIIVsem R package will be introduced with examples of input and output.
Anyone interested is invited to participate.
Contact person: Silvia Bianconcini