Sampling procedures and survey methodologies for the 1996 survey with comparisons to earlier national roadside surveys

This article describes the multistage sampling system employed in the 1996 national roadside survey and compares it to the sampling methods employed i n the two prior surveys in 1973 and 1986 Also described are the data collec tion procedures at the selected sites, the breath-test devices used to coll ect blood alcohol concentration (BAC) data, and the methods used to impute BAC values where breath-test measures were not obtained Overall, almost twi ce as many (6,298 in 1996 compared to 3,698 in 1973 and 3,043 in 1986) driv ers were interviewed in the most recent national survey as in the previous efforts. The procedures implemented in the three surveys are sufficiently s imilar to permit comparison of these surveys conducted at 10-year intervals .