|
Content |
Sample Design The 1998 Motor Vehicle Occupant Safety Survey, like both the baseline Motor Vehicle Occupant Safety Survey in 1994 and its follow-up survey of 1996, was conducted by telephone. Hence, the study procedures called for the construction of a national sampling frame of telephone households from which a random population sample could be derived. For each of the two survey instruments (one focusing on safety belts and the other on car seats, with a common core of questions relating to personal characteristics and driving behaviors), a national probability sample was developed. Each sample was composed of approximately 4,000 persons age 16 and older, including oversamples of persons age 16-39. Since the sampling procedures and data collection methodology for the two samples were identical, procedures described in this report for one sample apply to the other as well. The procedure for developing a population-based sample for this telephone survey involved four stages. The first stage sample involved a population-based sample allocation, distributed in proportion to the geographic distribution of the target population according to the most recent Census estimates. The second stage employed a systematic selection of assigned telephone banks within the geographically stratified first stage sample design. The third stage in the sampling procedure was to conduct a random digit dialing (RDD) sampling of telephone households within the telephone banks selected in the second stage. The fourth stage required the identification and systematic selection of one eligible respondent within each sampled household so that the household sampling frame yielded a population sample of the eligible population. These procedures yielded national estimates of the target population, within specified limits of expected sampling variability, from which valid generalizations can be made to the general public. |