Compression and Conditional Effects: A Product Term Is Essential When Using Logistic Regression to Test for Interaction*
Published online by Cambridge University Press: 26 November 2015
Previous research in political methodology argues that researchers do not need to include a product term in a logistic regression model to test for interaction if they suspect interaction due to compression alone. I disagree with this claim and offer analytical arguments and simulation evidence that when researchers incorrectly theorize interaction due to compression, models without a product term bias the researcher, sometimes heavily, toward finding interaction. However, simulation studies also show that models with a product term fit a broad range of non-interactive relationships surprisingly well, enabling analysts to remove most of the bias toward finding interaction by simply including a product term.
- Original Articles
- © The European Political Science Association 2015
Carlisle Rainey is Assistant Professor of Political Science in the Texas A&M University, 2010 Allen Building, College Station, TX 77843 (firstname.lastname@example.org). The author thanks Kenneth Benoit, Bill Berry, Scott Clifford, Justin Esarey, and two anonymous reviewers for helpful comments on earlier versions of this manuscript. The author also thanks John Oneal and Bruce Russet for making their data available, the Center for Computational Research at the University at Buffalo for providing support for the simulations. Code and data necessary to replicate the simulations and empirical analysis is available at http://www.carlislerainey.com/research and at http://thedata.harvard.edu/dvn/dv/PSRM. To view supplementary material for this article, please visit http://10.1017/psrm.2015.59