Essays on Interviews and Matching

Abstract

This thesis contains three essays on the topic of quantifying the impact of interviews in a matching market. The first two essays are empirical and use novel preference and matching data from the Canadian Residency Matching Service (CaRMS), and the third essay presents a formal identification result. In the first essay, I measure the impact of interviews on employers' preferences, and in the second essay, I measure the impact of reducing interviews on match outcomes. Both essays require me to quantify employers' pre-interview information about their post-interview preferences, but employers observe information unobservable to the econometrician. To address this econometric challenge, I estimate a joint structural model of interview offers and post-interview ranks in which unobservables may be correlated across the two periods, and thereby I use the information contained in post-interview preferences to correct for employers' additional pre-interview information. The third essay presents a non-parametric identification result that formalizes the possibility of using selection (e.g., interview-offer) data and binary outcome (e.g., job-offer) data jointly to correct for the role of unobservable factors in selection.