10.7910/DVN/HJCPDKGarrett GlasgowGarrettGlasgowUniversity of California, Santa BarbaraMatt GolderMattGolderPennsylvania State UniversitySona N. GolderSonaN. GolderPennsylvania State UniversityHarvard Dataverse2011Garrett GlasgowGarrettGlasgowUniversity of California, Santa BarbaraPolitical AnalysisIQSS Dataverse Network2011-12-212014-10-023813502727851370177546210790161588943272173235359005text/x-stata-syntax; charset=US-ASCIItext/x-stata-syntax; charset=US-ASCIItext/x-stata-syntax; charset=US-ASCIItext/x-stata-syntax; charset=US-ASCIItext/x-stata-syntax; charset=US-ASCIItext/x-stata-syntax; charset=US-ASCIItext/x-stata-syntax; charset=US-ASCIItext/tab-separated-valuestext/plain; charset=US-ASCIItext/x-stata-syntax; charset=US-ASCIItext/x-stata-syntax; charset=US-ASCIItext/x-stata-syntax; charset=US-ASCII2.0CC0 1.0In recent years a consensus has developed that the conditional logit (CL) model is the most appropriate strategy for modeling government choice. In this paper, we reconsider this approach and make three methodological contributions. First, we employ a mixed logit with random coefficients that allows us to take account of unobserved heterogeneity in the government formation process and relax the independence of irrelevant alternatives (IIA) assumption. Second, we demonstrate that the procedure used in the literature to test the IIA assumption is biased against finding IIA violations. An improved testing procedure reveals clear evidence of IIA violations, indicating that the CL model is inappropriate. Third, we move beyond simply presenting the sign and significance of model coefficients, suggesting various strategies for interpreting the substantive influence of variables in models of government choice.