Chiang Mai Journal of Science

Paper Type 

Title 

Author 

Email 

In statistical studies, generalized linear models (GLMs) are usually preferred for
modeling incidence rates, often with extensions to zero-inflated GLMs when the proportion
of zero counts is large. However Warton has shown that for many ecological studies, simple
linear models fitted to log-transformed counts do surprisingly well. In this study, we used data
comprising a sizable set of pneumonia incidence rates. We compared the negative binomial
GLM with a log transformed linear model, and found further support for this simpler alternative method.

Start & End Page 

Received Date 

Revised Date 

Accepted Date 

Full Text 

Keyword 

Volume 

DOI 

SDGs