Abstract
DRIVERS REFUSING TO PROVIDE A BREATH SAMPLE (N=483) IN A ROADSIDE SURVEY WERE COMPARED WITH DRIVERS AGREEING TO PROVIDE A SAMPLE (N=9745) ON A NUMBER OF SURVEY VARIABLES TO DETERMINE WHETHER THE FORMER WERE MORE LIKELY TO BE IMPAIRED BY ALCOHOL THAN THE LATTER. SIGNIFICANT DIFFERENCES IN NONRESPONSE RATE AS A FUNCTION OF TIME OF INTERVIEW, AGE. SEX AND SEAT BELT USE WERE OBSERVED. WITH THE EXCEPTION OF SEX, THESE DIFFERENCES SUGGESTED THAT THE NONRESPONDENTS WERE MORE LIKELY TO BE IMPAIRED THAN THE RESPONDENTS. THE ASSUMPTION OF EQUIVALENT BAC DISTRIBUTIONS FOR RESPONDENTS AND NONRESPONDENTS WAS NOT SUPPORTED BY THE PRESENT RESULTS. METHODS FOR ESTIMATING THE BAC'S OF NONRESPONDENTS ARE DISCUSSED. (Author/publisher).