Log Odds and Ends: Marginal Effects in Logit Models
Edward Norton, PhD
Seminar date: 6/17/2020
Description: Abstract: The interpretation of coefficients from logit models is challenging because of the importance of the standard deviation of the error term to that interpretation.The presenter will show how odds ratios are computed, how they depend on the standard deviation of the error term, and their sensitivity to different model specifications. Also will be discussed are marginal effects, which, unlike odds ratios, are often robust to alternative model specifications.
Target Audience: Health economists and health services researchers who use logistic regression.