No CrossRef data available.
Published online by Cambridge University Press: 04 January 2017
In this article, we present a technique and critical test statistic for assessing the fit of a binary-dependent variable model (e.g., a logit or probit). We examine how closely a model's predicted probabilities match the observed frequency of events in the data set, and whether these deviations are systematic or merely noise. Our technique allows researchers to detect problems with a model's specification that obscure substantive understanding of the underlying data-generating process, such as missing interaction terms or unmodeled nonlinearities. We also show that these problems go undetected by the fit statistics most commonly used in political science.
Authors' note: We thank Drew Linzer, Mike Ward, Jacqueline H. R. Demeritt, Jeff Staton, John Freeman, Neal Beck, Patrick Brandt, Phil Schrodt, Teppei Yamamoto, Kevin Clarke, and Will H. Moore for their comments, suggestions, and conversations about previous iterations of the article. Replication materials for all our simulations and data analysis can be found online at the Political Analysis dataverse: http://hdl.handle.net/1902.1/18399. Supplementary materials for the article are available on the Political Analysis Web site.