ANALYSIS OF CESAREAN-SECTION BIRTHS- AN APPLICATION OF ZERO-TRUNCATED GENERALIZED POISSON (ZTGP) REGRESSION MODEL
Fitting of ZTGP model is done using C-section data. Data set consists of annual total births, hospital type (private and public) and C-sections. The response variable Y denotes the number of C-sections which do not have any zero values. We first regress the response variable, ‘C-sections’ against one explanatory variable ‘number of births’, then we add one more explanatory variable in the form of indicator variable hospital type (public hospital=1, private hospital=0) in the Poisson regression analysis. A measure of goodness of fit of the ZTGP regression model is used on the log-likelihood statistic. The addition of the dispersion parameter α in the ZTGP regression model is justified by testing for the adequacy of the ZTGP model over the ZTP regression model.