How Much Should We Trust Estimates from Multiplicative Interaction Models? Simple Tools to Improve Empirical Practice

146 Pages Posted: 29 Feb 2016 Last revised: 13 Feb 2017

Jens Hainmueller

Stanford University - Department of Political Science; Stanford Graduate School of Business; Stanford Immigration Policy Lab

Jonathan Mummolo

Stanford University, Department of Political Science, Students

Yiqing Xu

University of California, San Diego

Date Written: February 13, 2017

Abstract

Multiplicative interaction models are widely used in social science to test whether the relationship between an outcome and an independent variable changes with a moderating variable. Current empirical practice overlooks two important problems. First, these models assume a linear interaction effect that changes at a constant rate with the moderator. Second, reliably estimating the conditional effects of the independent variable at all values of the moderator requires sufficient common support. Replicating 46 interaction effects from 22 recent publications in five top political science journals, we find that these core assumptions fail in a majority of cases, suggesting that a large portion of findings across all subfields based on interaction models are modeling artifacts or are at best highly model dependent. We propose simple diagnostics to assess the validity of these assumptions and offer flexible estimation strategies that allow for nonlinear interaction effects and safeguard against excessive extrapolation.

Keywords: interaction effects, regression models, conditional hypothesis

JEL Classification: C10, C14

Suggested Citation

Hainmueller, Jens and Mummolo, Jonathan and Xu, Yiqing, How Much Should We Trust Estimates from Multiplicative Interaction Models? Simple Tools to Improve Empirical Practice (February 13, 2017). Available at SSRN: https://ssrn.com/abstract=2739221 or http://dx.doi.org/10.2139/ssrn.2739221

Jens Hainmueller (Contact Author)

Stanford University - Department of Political Science ( email )

Stanford, CA 94305
United States

HOME PAGE: http://www.stanford.edu/~jhain/

Stanford Graduate School of Business ( email )

655 Knight Way
Stanford, CA 94305-5015
United States

Stanford Immigration Policy Lab

30 Alta Road
Stanford, CA 94305
United States

Jonathan Mummolo

Stanford University, Department of Political Science, Students ( email )

Stanford, CA 94305
United States

Yiqing Xu

University of California, San Diego ( email )

9500 Gilman Drive
#0521
La Jolla, CA 92093-0521
United States

HOME PAGE: http://yiqingxu.org/

Paper statistics

Downloads
1,460
Rank
9,065
Abstract Views
3,584