the claim that helmet wearing actually increases the likelihood of crashing and sustaining head injuries is very interesting to me.

the problem is that it's very tricky to figure out the causal relationship -- correlation itself does not imply causation. it could be, as the article/study seems to claim, that there is something about having a helmet on that makes you crash and hit your head more often. maybe the helmet makes your head heavier and thus more likely to hit pavement when you crash. however, it is also very possible that helmets themselves don't cause more head injuries -- that the positive correlation between helmet use and head injuries is actually due to some additional factor. for example, it could be that helmets are more often worn by inexperienced riders. or, maybe helmets just seem to make riders ride more carelessly due to a false sense of safety. if either of these are true (i'm not saying that they are), then it's some characteristic of the rider that is ultimately responsible for the positive correlation. here, it is the fact that the riders are less skilled at riding and haphazard that is responsible for the crash, not the helmet. it's very possible that there is some characteristic of a rider that both makes him/her more likely to wear a helmet and more likely to crash.

importantly, and obviously, if it's an "external cause" sort of situation, then the argument simply doesn't hold because one isn't in any position to postulate on the efficacy of helmets. inexperienced riders are always going to be more likely to crash, helmet or not. among those riders, it may be the case that head injuries are mitigated by wearing helmets. as for the situation in which a rider becomes less careful when riding with a helmet, the remedy for that is clearly to get the rider to stop being a ****** just because (s)he's rocking a helmet. it wouldn't be to recommend taking off the helmet.

this all said, if it turns out that something about a helmet itself causes more head injuries (the weight /size or some visibility issue) -- i would need to see more studies to be convinced -- then, well, **** a helmet.