The social meaning of the age at second birth for third-birth fertility: a methodological note on the need to sometimes respectify an intermediate variable

In hazard regressions for a number of countries, including Sweden, more highly educated women have been found to have higher third-birth rates than other women. In this paper we show that this positive educational gradient disappears when age at second birth is respecified in order to better catch what age at second birth means to women at the various levels of education. Instead o f a conventional age grouping that is the same for all educational categories, we suggest that the age factor should be defined so as to reflect what is normal and unusual childbearing behavior for each educational level separately. Considerations of a similar nature can be equally important in other contexts.