There was an ad at the top of my page that said something about not serving alcohol to anyone under 21. With (presumably) most readers of this blog living in Manitoba and Alberta, this makes no sense. Anyways, I followed the link and I found the following statement: "Since laws established 21 as the minimum drinking age, the likelihood that a 15 – to 20-year-old driver will be involved in a fatal crash has dropped by more than half." Really? I found this paper from the American Journal of Preventive Medicine. On page S262, the chart seems to indicate that crash rates for 15-19 year-olds hit a local minimum in 1983 and have stayed relatively flat since then - though they're 10-15% higher in 2004 than they were in 1983. In 1978, 38 states had a drinking age lower than 21; by 1988, only one (Louisiana) was left. The data simply don't support the conclusion that raising the drinking age lowered fatal crash rates.