ABSTRACT
In reproductive and perinatal epidemiology, as in other fields of epidemiology, extensive use is therefore made of questionnaires — a term that will be used to refer both to self-administered questionnaires and to schedules used in structured interviews. In validating a questionnaire for measuring “physical health in terms of functioning”, for example, construct validity was appraised by seeking the expected associations between the score and other questionnaire measures of functioning, age and income. When practicable, the best way of appraising validity is to find a criterion that is known or believed to be close to the truth, and to compare the results with this criterion. The interviewer, as well as the respondent, may be a source of bias — his or her expectations or preferences may have an influence on the answers that are given or the way the responses are interpreted and recorded.
