The harmfulness or harmlessness of using an anticipatory regressor: how dangerous is it to use education achieved as of 1990 in the analysis of divorce risks in earlier years?

The case studies presented in this paper show that the chance of making a sensible analysis of the effect of education on divorce risks may be ruined for women who marry as teenagers if the educational variable is measured only at the end of the study period. By contrast, these adverse effects seem to be unimportant once the age at marriage is 20 or more.