Remember kids... always wear a helmet. (The almighty bikeradar helmet thread)

(1) The evidence suggests that bike helmets are not very effective at preventing serious injury assuming a given accident happens.

Not quite. The range of accidents where the difference between the injuries to the unprotected head and those to the protected head is significant is small, so the overall probability of being involved in accident where use of a helmet will have a material effect on the outcome is very small.

(2) The evidence suggests that bike helmets are fairly effective at making what would have been a minor head injury even more minor assuming a given accident happens. But this is not very important.

Not quite, again. Bike helmets are effective at doing what they are designed to do, but what they are designed to do is not very well correlated with what they would need to do to make a significant material difference to the outcome of incidents where the helmet comes into play.

(3) The evidence suggests that wearing a helmet increases cyclists' risk taking and even more importantly the risk taking of drivers around the cyclist. Accidents are therefore more likely when a helmet is worn.

See below, and see general risk compensation research. I'm not sure that adult cyclists ride differently when wearing helmets as long as they are doing he same kind of cycling; that some of them choose helmets for some kinds of cycling and go without for others indicates some kind of risk compensation behaviour, but that doesn't prove that they would ride differently if you added/removed a helmet.

(4) Taking (1) to (3) together we can safely say that there is no evidence that helmet wearing makes any real improvement in safety, helmets should not be compulsory, nor should they be encouraged.

Taking all the evidence together, including evidence of the health benefits of cycling and the fact that helmet compulsion puts off some people from cycling at all, and also taking into account the general libertarian principle that only an enormous public good justifies state impositions on private behaviour, helmet compulsion is a bad thing. The costs and benefits of helmet use should, as with any product, be presented honestly, leaving people to draw their own conclusions.

Okay MDCC, where is this evidence around increased risk taking and increased accident rates when wearing helmets due to sloppy cycling and driving?

(Yes, I do realise you never suggested this!)

Some evidence from practical trials suggests that motorists pass closer to cyclists wearing helmets, I think the CTC probably has the results on their site. Also, looking like a female gets you a wider berth, so go figure.