Published in:Political Analysis ; 2010 18(1): 78-102
Extent:
1 Online-Ressource (30 p)
Language:
English
DOI:
10.2139/ssrn.926081
Identifier:
Origination:
Footnote:
Nach Informationen von SSRN wurde die ursprüngliche Fassung des Dokuments August 1, 2006 erstellt
Description:
The increased use of models with limited-dependent variables has allowed researchers to test important relationships in political science. Often, however, researchers employing such models fail to acknowledge that the violation of some basic assumptions has in part difference consequences in nonlinear models than in linear ones. In this paper I demonstrate this for probit models in which the dependent variable is systematically mis-coded. Contrary to the linear model, such misclassiffcations affect not only the estimate of the intercept, but also those of the other coe±cients. In a Monte-Carlo simulation I demonstrate that a model proposed by Hausman, Abrevaya and Scott-Morton (1998) allows for correcting these biases. Empirical examples based on re-analyses of models explaining the occurrence of rebellions and civil wars demonstrate the problem that comes from neglecting these misclassiffcations