The 30-item General Health Questionnaire (GHQ) and Zung's Self-Rating Depression Scale (SDS) were distributed to 120 pregnant women 4 times--in early and late pregnancy and 5 days and 1 month after the child was born. The validity of the questionnaires was assessed against the subjects' Research Diagnostic Criteria (RDC) diagnoses. Both the GHQ and SDS sufficiently identified cases of minor mental disorder and depressive disorders respectively in early pregnancy; they lost their validity on the subsequent two occasions, but gained it again 1 month after the birth; the optimal cut-off points varied accordingly. This study suggests that the optimal cut-off point for a questionnaire should be validated against an externally determined clinical diagnosis whenever the instrument is used repeatedly on the same population.