Interpreting interaction terms in linear and non-linear models: A cautionary tale

Drichoutis, Andreas (2011): Interpreting interaction terms in linear and non-linear models: A cautionary tale.

Abstract

Interaction terms are often misinterpreted in the empirical economics literature by assuming that the coefficient of interest represents unconditional marginal changes. I present the correct way to estimate conditional marginal changes in a series of non-linear models including (ordered) logit/probit regressions, censored and truncated regressions. The linear regression model is used as the benchmark case.

• Item Type: MPRA Paper
• Language: English
• Keywords: interaction terms; ordered probit; ordered logit; truncated regression; censored regression; nonlinear models
• Subjects: C - Mathematical and Quantitative Methods > C5 - Econometric Modeling > C51 - Model Construction and Estimation; C - Mathematical and Quantitative Methods > C1 - Econometric and Statistical Methods and Methodology: General > C12 - Hypothesis Testing: General; C - Mathematical and Quantitative Methods > C2 - Single Equation Models ; Single Variables > C24 - Truncated and Censored Models ; Switching Regression Models ; Threshold Regression Models; C - Mathematical and Quantitative Methods > C2 - Single Equation Models ; Single Variables > C25 - Discrete Regression and Qualitative Choice Models ; Discrete Regressors ; Proportions ; Probabilities

References

Ai, C, and EC Norton (2003) Interaction terms in logit and probit models. Economics Letters 80 (1), 123-129.

Brambor, T, WR Clark, and M Golder (2006) Understanding interaction models: Improving empirical analyses Political Analysis 14 (1), 63-82.

Drichoutis, AC, and RM Nayga (2011) Marginal changes in random parameters ordered response models with interaction terms. Econometric Reviews 30 (5), 565-576.

Greene, WH. (2003) Econometric Analysis. 5th ed: New Jersey: Prentice Hall

Greene, WH. (2007) Limdep version 9.0, Econometric modeling guide. Vol. 1. Plainview, New York: Econometric Software Inc.

Greene, WH (2010) Testing hypotheses about interaction terms in nonlinear models. Economics Letters 107 (2), 291-296.