Comparison of Generalized and Multilevel Poisson Regression Model with Poisson Model in Fertility Data in Rural of Fars Province (Iran)

Different statistical methods can be used to analyze fertility data. In cases that dependent variable is count, Poisson model is applied. If Poisson model is not applicable in a specific situation, it is better to apply generalized Poisson model and in cases that multilevel variable exists, it is better to use multilevel Poisson model. Our goal in this study is to compare both generalized and multilevel Poisson regression model with Poisson regression model in estimating of coefficient of the effective factors on the number of children.

This is a cross-sectional study. A sample of 1019 women (15-49 years old) from rural area was selected by stratified sampling. The women were categorized into seven groups and in each group the intended samples were selected equally by systematic sampling. Data were analyzed by Poisson regression model, generalized and multilevel Poisson regression model.

The sample mean and sample variance of the number of children were 4.3 and 8.3, respectively. There was a significant relationship between educational status of the spouses, age of marriage, feeding period, economical status and the interval between the children in generalized and multilevel Poisson regression model.

According to the results of this study, generalized and multilevel Poisson regression models were more suitable for data analysis and it can estimate coefficient effective of factors on the number of children exactly.