- Given the earlier termination of NTD-affected pregnancies, at the time of interview, control women were recalling exposures further in the past than case women

- Although the case-control difference in recall time was only about 1 month, the shorter recall time for case women may have caused them to remember exposures somewhat more accurately than control women

- Combined with the possibility that women with affected pregnancies would recall exposures more carefully, this differential exposure misclassification also would have attenuated observed protective folate effects

- If selection bias produced the weak or null folate effect, then control women with higher socioeconomic status (better diets and more vitamin use) must have refused participation more often

- However, we know from vital statistics data that control women in the study were remarkably similar to the entire border population