And pretty crappy science as well:
But there is another problem with the data. The research appears to assume that alcohol-related motor-vehicle crashes were caused by the presence of alcohol. But since at least 2003, the National Highway Traffic Safety Administration has emphatically warned that the data do not mean that alcohol consumption caused the crashes …
In other words, an accident is considered alcohol-related if there is any measurable alcohol in any participant in the crash — but there is no evidence that the consumption of alcohol led to the accident and subsequent death. Yet about three-quarters of the reported “alcohol-related injury deaths” stem from the motor-vehicle records.
Hingson, in an interview, said “you are raising an appropriate question on the data that I had when I first wrote the paper.” He said research on the issue has been hamstrung by a lack of solid information, though efforts are underway to improve it. He also noted that he took a conservative approach by assuming college students had accidents at the same rate as noncollege students, even though surveys indicate that they are heavier drinkers.
Hingson appeared unaware of the NHTSA warning about the accident data and said he would make a note of that in the latest edition of the report. “I agree that you can’t say that alcohol is the sole cause of these accidents,” he said.