Factors Affecting Fertility and Statistical Inference: A Comment.

Statistical methods used to study factors affecting fertility are often unsatisfactory. First, it is argued that both biological and socioeconomic factors that affect fertility and each other must be incorporated into the statistical model if nonexperimental data are to yield consistent estimates of the postulated causal relationships. Second, the econometric problem known as "simultaneous-equations bias" is viewed as a serious impediment to drawing causal conclusions from existing empirical studies of fertility differentials. Many of the differences between economists and sociologists can be understood in terms of their choices of explanatory variables. The economist attaches importance to predetermined variables, such as wages and prices; whereas the sociologist attaches significance to endogenous variables, such as attitudes and norms. Until both disciplines consider both sets of factors within the appropriate statistical framework, empirical research will have difficulty resolving disagreements over what socioeconomic factors appear to affect fertility. 9 pp.