Bad Science has possibly the best post title I have ever seen: You are 80% less likely to die from a meteor landing on your head if you wear a bicycle helmet all day.
I don't know if the statistic is true, but its a great way to introduce the problems of relative versus absolute risk reduction. The first can make the implications of taking a precaution sound great, the other can make it sound just ok.
So, 80% sounds like a lot, but would you really want to wear a bike helmet all day to increase your chances of surviving something that is very unlikely to happen to begin with?
What if I told you the same statistic about the effect of wearing a helmet if you're hit by a car while riding your bike? It's pretty close to the numbers reported at the Bicycle Helmet Safety Institute.
Personally, I never wear a helmet when riding because the likelihood of needing it is very small. For every 1 million miles Americans ride their bikes, only 15 are ever injured at all.