The broader alcohol involvement statistic is so meaningless as to draw the ire of the above slashdot commenters. No doubt that is why it is not used. Your original response implied that was the statistic being used in the document, and was interpreted as such by the responders below you. If you were not talking about the data I presented, why bring up an unrelated and useless statistic? Either you did not recognize the difference to begin with and are now backpedaling on semantic grounds, or you were trying to mislead readers. Either way, "wrong" is an accurate characterization of your comment, especially since the numbers for the definition I gave are the same as those in the original document, so it is obvious what definition is being used. Also, the document I referred to only counts drunk passengers as fatalities if they died. The fact that the passenger was drunk does not make the accident alcohol impaired.