Recall and reliability of interview data from older women.

Retrospective fertility histories were collected from a purposive sample of 211 white, ever-married women, residing in four metropolitan areas of the USA and belonging to the birth cohorts of 1901-1910. At the time of interview, these women were aged 66 to 76. A random sample of 50 of the original respondents was reinterviewed by telephone on selected topics from the questionnaire. Responses to questions on dates and numbers of marriages, as well as births and other fertility events were coded to show whether the responses were exact, estimated by the respondent, or calculated by the editors. Respondents provided exact answers to an average of 90% of the questions they were asked. Reliability ratios, which were calculated for the above variables as well as ever-use of contraceptive methods, were comparable with those observed in previous studies of women in the childbearing years.