Logistic Regression: Why We Cannot Do What We Think We Can Do, and What We Can Do About It
Top Cited Papers
- 9 March 2009
- journal article
- research article
- Published by Oxford University Press (OUP) in European Sociological Review
- Vol. 26 (1), 67-82
- https://doi.org/10.1093/esr/jcp006
Abstract
Logistic regression estimates do not behave like linear regression estimates in one important respect: They are affected by omitted variables, even when these variables are unrelated to the independent variables in the model. This fact has important implications that have gone largely unnoticed by sociologists. Importantly, we cannot straightforwardly interpret log-odds ratios or odds ratios as effect measures, because they also reflect the degree of unobserved heterogeneity in the model. In addition, we cannot compare log-odds ratios or odds ratios for similar models across groups, samples, or time points, or across models with different independent variables in a sample. This article discusses these problems and possible ways of overcoming them.Keywords
This publication has 7 references indexed in Scilit:
- Estimation of Limited Dependent Variable Models With Dummy Endogenous RegressorsJournal of Business & Economic Statistics, 2001
- Comparing Logit and Probit Coefficients Across GroupsSociological Methods & Research, 1999
- Testing Collapsibility in Contingency TablesJournal of the Royal Statistical Society Series B: Statistical Methodology, 1986
- A Comment on Presenting Results from Logit and Probit ModelsAmerican Sociological Review, 1985
- Biased estimates of treatment effect in randomized experiments with nonlinear regressions and omitted covariatesBiometrika, 1984
- Regression Models with Ordinal VariablesAmerican Sociological Review, 1984
- Collapsibility of Multidimensional Contingency TablesJournal of the Royal Statistical Society Series B: Statistical Methodology, 1978