The stats for alcohol "related" deaths are not correct. The way they report it is as follows: when an accident occurs (fatality or not) if anyone involved in the accident has been drinking be it driver of either car at fault or not and the passengers of either vehicle it is considered an alcohol related accident. It is even considered alcohol related if all involved are stone cold sober and alcohol is present in either vehicle (say from picking up a bottle of wine to take home for dinner and accident occurs). Do people die in accidents where drivers are actually drunk? YES however that is something we already knew, what we don't know is the correct statistic. Further, I would like to know the source that this NTSB is using to validate their data on the "drunken reflexes".