Design and operation of the 1995 National Survey of Family Growth

In the US, the 1995 National Survey of Family Growth (NSFG) was designed to provide richer data than previous NSFG surveys from 1973, 1976, 1982, and 1988. Planning for the 1995 NSFG took place at a series of meetings beginning in 1990. Pretesting of the expanded questionnaire, the new computer-assisted personal interviewing method, and the audio computer-assisted, self-interviewing method for sensitive topics occurred in 1993 and led to the decision to offer a cash incentive to respondents and to use the new interviewing methods. The revised questionnaire collected information on event histories, pregnancy history and family formation, partner history, sterilization and fecundity, contraception and birth expectations, use of family planning and other medical services, demographic characteristics, abortion history, number of sexual partners, and rape. The sample for the 1995 NSFG included 14,000 civilian, noninstitutionalized women of reproductive age, 13,795 of whom proved eligible. Of these, 79% completed interviews. Quality control measures included careful design and testing of the questionnaire, use of a Life History Calendar, intensive interviewer training, consistency checks, and use of the incentive. Sampling weights for each respondent were used to derive nationally representative statistical estimates. Sampling errors were created to reflect the complexity of the sample. Research based on the results of the 1995 NSFS has only begun to take advantage of the potential offered by these data.