Abstract
A measure of underlying family size preference obtained for a sample of Detroit married women in 1962 is related to their fertility over a 15-year follow-up period. The data represent completed fertility. The I-scale preference measure used differs from the conventional single-valued statement of number of children wanted; it is a more fine-grained measure reflecting the respondent’s utility for children as evidenced by her entire preference order. The scales are found to be consistently predictive of fertility over the 15-year prospective period, net of a number of other variables usually associated with differential fertility. The results for the just-married sample, in which preferences and expectations are not confounded with the number of children already born, are particularly striking, with underlying preference much better than expected family size as a predictor of fertility over the entire reproductive cycle. The question of prediction for continuous and discontinuous marriages is discussed.