Comparison of Efficiency in Generalized Poisson Regression Model and the Standard Poisson Regression Model in analyzing Fertility Behavior among Women, Kashan, 2012

Introduction: Different statistical methods can be used to analyze fertility data. When the response variable is discrete, Poisson model is applied. If the condition does not hold for the Poisson model, its generalized model will be applied. The goal of this study was to compare the efficiency of generalized Poisson regression model with the standard Poisson regression model in estimating the coefficient of effective factors onthe current number of children.\n\nMethods: This is a cross-sectional study carried out on a populationof married women within the age range of15-49 years in Kashan, Iran. The cluster sampling method was used for data collection. Clusters consisted ofthe urbanblocksdeterminedby the municipality.Atotal number of10clusters each containing30households was selected according to the health center's framework. The necessary data were then collected through a self-madequestionnaireanddirectinterviewswith women under study. Further, the data analysiswas performed by usingthe standard and generalizedPoisson regression models through theRsoftware.\n\nResults: The average number of children for each woman was 1.45 with a variance of 1.073.A significant relationship was observed between the husband's age, number of unwanted pregnancies, and the average durationof breastfeeding with the present number of children in the two standard and generalized Poisson regression models (p < 0.05).The mean ageof women participating in thisstudy was33.1± 7.57 years (from 25.53 years to 40.67), themean age of marriage was 20.09 ± 3.82 (from16.27 years to23.91), and themean age of their husbands was 37.9 ± 8.4years (from 29.5 years to 46.3). In the current study, the majority of women werein the age range of 30-35years old with the medianof 32years, however, most ofmen were in the age range of 35-40yearswith the median of37years. While 236of women did not have unwanted pregnancies, most participants of the present study had one unwanted pregnancy.\n\nConclusion: According to the achieved results, the generalized Poisson regression model is more suitable for data analysis and can estimate the coefficient of effective factors on the number of children more precisely.