Binary response models are often applied in empirical marketing research where, for example, individuals or households can be classified as buyers and nonbuyers of certain products or as respondents and nonrespondents to direct mailing. One way to evaluate such models amounts to calculating the fraction of correctly classified cases, that is, the hit rate. This paper brings to the attention a simple test for the statistical significance of this hit rate. An application to a binary brand choice case serves as an illustration.
Get full access to this article
View all access options for this article.
References
1.
BultJ. R. (1993) Semiparametric versus parametric classification models: an application to direct marketing. Journal of Marketing Research, 30, pp. 380–390.
2.
CramerJ. S. (1997) Predictive Performance of the Binary Logit Model in Unbalanced Samples. Unpublished manuscript, Tinbergen InstituteAmsterdam.
3.
HausmanJ. A. (1978) Specification tests in econometrics. Econometrica, 46, pp. 1251–1272.
4.
LongJ. S. (1997) Regression Models for Categorical and Limited Dependent Variables.Thousand Oaks: Sage Publishers.
5.
MaddalaG. S. (1983) Limited Dependent and Qualitative Variables in Econometrics.Cambridge UK: Cambridge University Press.
6.
PesaranM.H., & TimmermannA. (1992) A simple nonparametric test of predictive performance. Journal of Business and Economic Statistics, 10, pp. 461–465.
7.
PregibonD. (1981) Logistic regression diagnostics. Annals of Statistics, 9, pp. 705–724.
8.
VeallM.R., & ZimmermannK.F. (1996) Pseudo R2 measures for some common limited dependent variable models. Journal of Economic Surveys, 10, pp. 241–259.
9.
WindmeijerF.A.G. (1995) Goodnessoffit measures in binary response models. Econometric Reviews, 14, pp. 101–116.
