A Reinterpretation of Interactions in Regressions

12 Pages Posted: 13 Nov 2007

See all articles by Jenny N. Lye

Jenny N. Lye

University of Melbourne - Department of Economics

Joe Hirschberg

University of Melbourne - Department of Economics

Date Written: November 2007

Abstract

Regression specifications in applied econometrics frequently employ regressors that are defined as the product of two other regressors to form an interaction. Unfortunately, the interpretation of the results of these models is not as straight forward as in the linear case. In this paper, we present a method for drawing inferences for interaction models by defining the partial influence function. We present an example that demonstrates how one may draw new inferences by constructing the confidence intervals for the partial influence functions based on the traditional published findings for regressions with interaction terms.

Keywords: Interaction effects, dummy variables, linear transformation, Fieller method

JEL Classification: I21, C12, C51

Suggested Citation

Lye, Jenny N. and Hirschberg, Joseph G., A Reinterpretation of Interactions in Regressions (November 2007). Available at SSRN: https://ssrn.com/abstract=1029521 or http://dx.doi.org/10.2139/ssrn.1029521

Jenny N. Lye

University of Melbourne - Department of Economics ( email )

Melbourne, 3010
Australia

Joseph G. Hirschberg (Contact Author)

University of Melbourne - Department of Economics ( email )

Victoria 3010, 3010
Australia

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
129
Abstract Views
816
Rank
397,421
PlumX Metrics