Bernoulli Regression Models: Revisiting the Specification of Statistical Models with Binary Dependent Variables

TR Number

Date

2010

Journal Title

Journal ISSN

Volume Title

Publisher

Elsevier

Abstract

The latent variable and generalized linear modelling approaches do not provide a systematic approach for modelling discrete choice observational data. Another alternative, the probabilistic reduction (PR) approach, provides a systematic way to specify such models that can yield reliable statistical and substantive inferences. The purpose of this paper is to re-examine the underlying probabilistic foundations of conditional statistical models with binary dependent variables using the PR approach. This leads to the development of the Bernoulli Regression Model, a family of statistical models, which includes the binary logistic regression model. The paper provides an explicit presentation of probabilistic model assumptions, guidance on model specification and estimation, and empirical application.

Description

Keywords

Bernoulli Regression Model, Generalized Linear Models, Latent Variable Models, Logistic Regression, Model Specification, Probabilistic Reduction Approach

Citation