Partial Effects for Binary Outcome Models with Unobserved Heterogeneity

Unobserved heterogeneity is ubiquitous in empirical research. In this paper, I propose a method for estimating binary outcome models with panel data in the presence of unobserved heterogeneity, called the Penalized Flexible Correlated Random Effects (PF-CRE) estimator. I show that this estimator produces consistent and efficient estimates of the model parameters. PF-CRE also provides consistent estimates of partial effects, which cannot be calculated with existing consistent estimators. Using Monte Carlo simulations, I show that PF-CRE performs well in small samples. To demonstrate that accounting for unobserved heterogeneity has important consequences for empirical analysis, I use PF-CRE to analyze tactical voting during the 2015 British Election. Ignoring the unobserved heterogeneity leads to an overestimation of the effects of being contacted by parties during the campaign. The suggested mechanism is that parties tend to contact voters who are already likely to cast tactical votes. ∗Ph.D. Candidate, Division of the Humanities and Social Sciences, California Institute of Technology. Email: [email protected]