Testing hypotheses about interaction terms in nonlinear models

By William H. Greene. 2010

802 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/9cb0eb0d9920b8b5f58a923f3398c806baccf709

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Economics and Econometrics

In Statistics and Probability

Citations in the corpus

Automation and the Welfare State: Technological Change as a Determinant of Redistribution Preferences (2019, background): “Some authors, however, argue that interaction effects in nonlinear models cannot be directly interpreted (Ai & Norton, 2003; Greene, 2010).”
Aid Is Not Oil: Donor Utility, Heterogeneous Aid, and the Aid-Democratization Relationship (2015, background): “See Ai and Norton 2003; and Greene 2010. change varies across periods as predicted.”
R&D Partnerships and Innovation Performance: Can There Be Too Much of a Good Thing? (2016, methodology): “Given the non-linearity of the probit model, the coefficient on the second-order term of collaboration intensity does not provide the change in the partial effect of the intensity variable (Greene 2010).”
Corruption, growth, and governance: Private vs. state-owned firms in Vietnam (2012, positive): “15 Greene (2010) suggests that caution should be applied in the interpretation of standard statistical tests of interaction terms in non-linear model (such as the logit model we estimate for PAYMENTDUM).”
The U.S. left behind? Financial globalization and the rise of IPOs outside the U.S. (2013, background): “interpret interaction terms because the marginal effects differ for each observation (Ai and Norton, 2003) and statistical tests can be uninformative and misleading (Greene, 2010).”
Globalization, Economic Freedom, and Human Rights (2012, methodology): “One way of testing for the statistical significance of the squared GDP per capita term is via a likelihood ratio test (Greene 2010).”
Social structure, reasonable gain, and entrepreneurship in Africa (2016, positive): “Therefore, we follow Greene (2010), who suggests that attention should be paid to fit and statistical issues at the model”
How Anticipated Employee Mobility Affects Acquisition Likelihood: Evidence from a Natural Experiment (2015, methodology): “Greene (2010) recom-”
Shareholder Voting and Corporate Governance Around the World (2015, positive): “However, Ai and Norton (2003) and Greene (2010) show that once an interaction term is included in a nonlinear regression model, one can no longer make inferences using the conventionally estimated interaction coefficient.”
On Interaction Effects: The Case of Heckit and Two-Part Models (2012, result): “In this regard, our discussion is perfectly in line with a recent article by GREENE (2010), who points out that, apart from statistical significance, one should care about economic and policy significance.”