A survey was taken of the frequency and kind of drugs used at a midwestern university in 1971 by students, at any time and shortly before driving. About 3% of the students were sampled and 823 returned completed questionnaires, representing a return rate of over 90%. The survey provided biographical data of the respondents, their number of accidents, violations and miles driven in the previous 12 months, in addition to the frequencies of drugs used. The drugs most frequently used at least once in the prior years were alcohol (77%), marijuana (41%) and caffeine/nicotine (54%). These basic results at Eastern Michigan University in 1971 are compared with the results of the same survey made in December, 1975 at the University of Illinois, in which about 10% more students reported use of these drugs. Automatic Interaction Detector (AID) analyses were carried out on the EMU data to find those biographical and drug usage variables which were most associated with accident and violation rates. Among the findings were the following:​ persons with greater violation rates were - (1) users of cocaine than non-users;​ (2) juniors, seniors and graduate students than freshmen or sophomores;​ and (3) users of alcohol while driving than nonusers. Persons with greater accident rates were - (1) marijuana users than non-users;​ (2) among marijuana users, females than males;​ and (3) among female marijuana users, those who drank alcohol frequently than infrequently. Among specified high violation and accident rate groups, those who frequently used caffeine or nicotine had lower rates than infrequent users.