Factors affecting family size preference: the case of Taiwan, 1970

In recent years, the debate concerning the independent effect of socio-economic development and family planning programs in reducing fertility rates in developing countries has generated many studies in the social science literature. The goal of this thesis is to advance our knowledge about the development and family planning program arguments for developing countries. A multi-level theoretical model and a statistical estimation procedure were specially developed to aid in the analysis of this problem. A data set from a large-scale survey by Wolfgang L. Grichting, The Value System in Taiwan, was secured from the survey Research center at the University of Michigan, Ann Arbor.

ln Analysis of covariance Model (ANCOV) was first applied to the data set. Results from the analysis show that within a residential area, individual differentials, socio-economic development, and family planning efforts demonstrate no significant relationship to actual family size. Furthermore, it is also demonstrated that area socioeconomic development and family planning efforts have a small independent effect on both desired family size and attitude toward family planning. It appears it is the vectors of background and intermediate variables that are exerting a strong independent effect on these two endogenous variables.

Since macro-indicators are implicitly combined in the ANCOV model, a complete specification of the analytic model includes both within individual differentials and area socio-economic development and family planning effort indicators. The variables are re-estimated for the data set with the Ordinary Least Square (OLS) procedure. Results from the regression model and decomposition procedure were consistent with the findings from the ANCOV statistical model. However, with regard to the attitude toward family planning equation, there is a strong independent effect of IOD acceptance rate on the endogenous variable. The present finding is consistent with other independent studies. The findings suggest that knowledge and use of contraceptives are highly associated with a favorable attitude toward family planning programs.

The analysis of covariance model (ANCOV) and multiple regression models are based on the whole sample. A simultaneous effect fertility model is developed from the regression analysis and estimated for male and female respondents in the second stage of analysis. However, it is found that male respondents are characterized with modern attributes and their fertility decisions are different from female respondents. It is postulated that in the early stages of socio-economic development, males benefit more from this transformation. This results from the fact they receive more education and participate in the modern sector of employment. On the other hand, it takes females longer to acquire the modern characteristics associated with education and higher participation in the modern employment sectors and then translate characteristics and value orientations into actual behavior.

Results from the present study tend to agree with family planning advocates who maintain that a family planning program is a facilitating factor for reducing fertility rates in developing countries over short periods of time. The rate of adopting family planning programs by male and female respondents should be estimated to respecify demographic transition models in future research. It is also recommended that population policies designed to intervene and change fertility levels in developing societies must take into consideration the structural residuals of traditional orientations as well as the provision of contraceptives for individuals.