Citation

Hurst PM, Darwin JH. Accid. Anal. Prev. 1977; 9(2): 119-123.

Copyright

(Copyright © 1977, Elsevier Publishing)

DOI

unavailable

PMID

unavailable

Abstract

This paper addresses the problem of missing data due to refusals in roadside breath surveys. To correct for bias thus introduced, estimation procedures have been developed that incorporate judgments made by interviewers concerning each driver's degree of intoxication. Such judgments are made on initial contact, ideally before there is any knowledge of whether the detainee will be a donor or a refuser. Subsequent knowledge of results from donors is then used to develop weights for application to refusers. Development of these weights depends on choice of assumptions about contingent probability relationships among three variables: Blood alcohol concentration, judged degree of intoxication, and the decision to respond or to refuse. One of the possible alternative assumptions is applied to the data from the U.S. 1973 National Roadside Survey, and the results are contrasted with those of the survey's author, who chose a different assumption.