Theoretical and methodological implications of the influence of sex preferences on the fertility attitude-behavior relationship

Abstract

Data was collected from a sample of 72 volunteers from the University of Colorado introductory psychology subject pool in order to explore the theoretical possibility that measures of preferred family size generate good aggregate but poor individual fertility predictions because they ignore the influence of family sex composition preferences. The average preferred (1st choice) family size of 2.80 for this sample closely matches the average expected value of 2.86 children. Longitudinal studies of fertility intentions sometimes report the proportions of respondents who have exactly the number of children they said they desired at the time of the 1st interview. The expected proportions for this sample again assuming strict adherence to the stopping rules are as follows: 66% of the respondents would have exactly the number of children they desire; 26% would miss their ideal by 1; and 8% would miss their ideal by 2 or more. The maximum correlation between most preferred family size and actual outcomes would be .78. Study results demonstrate that poor attitude-behavior correlations in fertility studies must be partially attributed to the measures of attitude used and to the probabilistic nature of the determinants of fertility behavior as well as to true attitude-behavior inconsistency. The strategy suggested for improving attitude measures is to develop measures which capture the inherent dependence of fertility decisions on probabilistic events.