In Defense of Comparative Statics: Specifying Empirical Tests of Models of Strategic Interaction

Posted: 18 Aug 2009

See all articles by Clifford Carrubba

Clifford Carrubba

Emory University - Department of Political Science

Amy Yuen

affiliation not provided to SSRN

Christopher J. Zorn

Pennsylvania State University

Abstract

Beginning in 1999, Curtis Signorino challenged the use of traditional logits and probits analysis for testing discrete-choice, strategic models. Signorino argues that the complex parametric relationships generated by even the simplest strategic models can lead to wildly inaccurate inferences if one applies these traditional approaches. In their stead, Signorino proposes generating stochastic formal models, from which one can directly derive a maximum likelihood estimator. We propose a simpler, alternative methodology for theoretically and empirically accounting for strategic behavior. In particular, we propose carefully and correctly deriving one's comparative statics from one's formal model, whether it is stochastic or deterministic does not particularly matter, and using standard logit or probit estimation techniques to test the predictions. We demonstrate that this approach performs almost identically to Signorino's more complex suggestion.

Suggested Citation

Carrubba, Clifford and Yuen, Amy and Zorn, Christopher J., In Defense of Comparative Statics: Specifying Empirical Tests of Models of Strategic Interaction. Political Analysis, Vol. 15, Issue 4, pp. 465-482, 2007, Available at SSRN: https://ssrn.com/abstract=1448402 or http://dx.doi.org/10.1093/pan/mpm008

Clifford Carrubba (Contact Author)

Emory University - Department of Political Science ( email )

Atlanta, GA 30322
United States
404-727-7915 (Phone)
404-727-4586 (Fax)

Amy Yuen

affiliation not provided to SSRN

No Address Available

Christopher J. Zorn

Pennsylvania State University ( email )

University Park
State College, PA 16802
United States

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
364
PlumX Metrics