Estimation and testing of binomial and beta-binomial regression models with and without zero inflation
James W. Hardin
Institute for Families in Society
Department of Epidemiology and Biostatistics
University of South Carolina
Columbia, SC
[email protected]
|
Joseph M. Hilbe
School of Social and Family Dynamics
Arizona State University
Tempe, AZ
[email protected]
|
Abstract. We present new Stata commands for carrying out several regression
commands suitable for binomial outcomes. The zib command extends Stata’s
binreg command to allow zero inflation. The betabin command fits binomial
regression models allowing for beta overdispersion, and the zibbin command fits
a beta-binomial regression model with zero inflation. All the new commands allow
the specification of links within the glm command’s collection for both outcome
and zero inflation. The zero-inflated commands optionally calculate a Vuong
test comparing the zero-inflated model with the nonzero-inflated model, and the
zibbin command optionally includes a likelihood-ratio test of the overdispersion
parameter.
View all articles by these authors:
James W. Hardin, Joseph M. Hilbe
View all articles with these keywords:
betabin, zib, zibbin, binomial outcomes, Vuong test, zero inflation, beta overdispersion
Download citation: BibTeX RIS
Download citation and abstract: BibTeX RIS
|
