Collection of survey data on contraception: an evaluation of an experiment in Peru.

The objective of this analysis is to compare two different approaches to the collection of information on contraceptive use. The data for this comparison are derived from the 1986 Demographic and Health Surveys (DHS) in Peru. Approximately 7,500 women were interviewed with the standard DHS questionnaire, whereas about 5,000 women received an "experimental questionnaire." The major difference between the questionnaires is the inclusion of a six-year monthly calendar in the experimental questionnaire that records pregnancies, contraceptive use, and postpartum information, in contrast to the more common tabular format of the standard questionnaire. The analysis demonstrates that although reports of contraceptive knowledge, ever-use, and current use are relatively robust to the variations in questionnaire design, estimates of past use are dependent on the survey instrument. Several different comparisons indicate that reporting of information on contraceptive histories in the experimental questionnaire is superior to that in the standard one.