People do seriously argue that helmets don't reduce injury in the event of a crash. e.g. http://cyclehelmets.org/1039.html
Myself, i'm not sure either way.
Secondly, 'reduce injury in the event of a crash' is not the same thing as 'reduce risk'. It's logically possible that helmets could do the first but not the second, if they somehow increased the probability of crashes. And the implication of the population-level stats is that somehow they do, whether through risk compensation by riders or drivers, reduction in rider awareness, or indirectly though reducing cyclist numbers and the safety-in-numbers effect.
you are getting two things wrong.
People do seriously argue that helmets don't reduce injury in the event of a crash. e.g. http://cyclehelmets.org/1039.html
Myself, i'm not sure either way.
Secondly, 'reduce injury in the event of a crash' is not the same thing as 'reduce risk'. It's logically possible that helmets could do the first but not the second, if they somehow increased the probability of crashes. And the implication of the population-level stats is that somehow they do, whether through risk compensation by riders or drivers, reduction in rider awareness, or indirectly though reducing cyclist numbers and the safety-in-numbers effect.