Preferences in Size of Family and Eventual Fertility Twenty Years After

Assumptions frequently made as to the value for purposes of prediction of stated preferences in size of family and the reliability of the recall of original preferences are called into serious question. Data collected from a small panel of couples married nearly twenty years and living together reveal that, although the estimate from the initial average preference to the eventual aggregate number of children of the whole group is fairly accurate, the correlation between the two variables is quite low. The high level of accuracy of the aggregate prediction appears to be a statistical artifact resulting from the sample proportions of couples who planned all pregnancies and those who did not. Recall of originally stated preferences after passage of time and intervening experience reveals considerable error, associated with both success in planning and acutal size of family.