Abstract
Among methods of studying reproductive effects of occupational exposures, retrospective questionnaire-based surveys of ocoupationally defined populations have an important place. Reservations concerning the validity of such data have proved to be overstated. Recall is remarkably accurate for birth-weight, gestational age, time taken to conceive, and phases of involuntary infertility, and this appears to hold for recall periods up to 2O years or more. Underreporting of miscarriages can be greatly reduced by restricting the analysis to those that have been confirmed both by a test and by a doctor. Interviews are preferable to self-completion format, as they tend to minimize response bias. It is important to consider many methodologic issues that arise in this type of research, such as the selection of confounding factors for inclusion in the questionnaire.

This publication has 10 references indexed in Scilit: