Socio-economic correlates of fertility among Peruvian women, demographic health survey, 1986

This thesis examines the effect of selected socio-economic and demographic variables on fertility for Peruvian women using data from the Demographic Health Survey (DHS, 1986). Linear and Poisson regression were used to predict family size at the time of the survey. Poisson regression best quantifies the effect of the independent variables on the number of children ever born, because the regression directly models the increasing variance, since the variance of a Poisson variable is a function of its means. The log-linear Poisson model used for Poisson distributed data belongs to the family of generalized linear models of Nelder and Wedderbura (1972). In the log-linear regression analysis, the logarithm of the mean of the number of children ever bom was modeled as a linear function of the following independent variables: woman's age, woman's years of schooling, husband's years of schooling, interaction of woman's age with place of residence, and interaction of woman's age with woman's years of schooling. The findings basically agree with previous studies, except that husband's occupation was not found significant.