While the evidence doesn't confirm for definite that they help in certain accidents, I'm pretty sure there are no cases where someone's injuries have been WORSE for wearing a helmet. On that basis, it's probably sensible to put children in them. It's kind of like making kids eat all of their vegetables. There's no proof that an additional carrot or two will make them live longer, be healthier, develop better, but it certainly doesn't do them any harm so you may as well do it anyway. And once that child grows up, it's then their choice whether or not they ever eat carrots again.
Perhaps a weak analogy, but the best I can think of just at the moment.
And as well as that, the reasons most adults give for not wanting to wear them tend to be for more practical reasons that don't really apply so much to children. Or at least, this is the impression that I get.
You have slightly missed the point. If helmet wearing doesn't seem to reduce head injuries overall, which it doesn't if the Austrialian numbers are anything to go by, you can't logically claim they are of benefit - "because they are bound to help in some cases". If that were true, then they must make things worse in other cases else they would help overll - wouldn't they, or have I missed something?
I can think of a number of reasons why they might make some injuries worse, but this would be speculation - but might include:
1 risk compensation - a well document effect whereby people take more risks if they feel safer. It is cerainly plausible that cyclists do this.
2 risk compensation by drivers - some recent research indiciated drivers pass closer if cyclist wearing a helmet. It's certainly plausible, and at first cut, research supported this - or at least suggested there's something in it.
3 makes your head bigger - therefore you will hit your head more oftent - this is simple geometry so can't really be argued against. Turns a near miss into a hit.
4 rotational injuries. The argument goes extra leverage of the hat, and / or extra weight n your head increases the risk of neck injuries and / or brain injuries caused by rotation. Dunno, but not implausible.
5 slight extra weight up high increases stopping distance and over the handlebars risk - not by much, granted, but if helmets not helping much, wont take much to cancel out.
Finally, not a reason for not wearing one, but no one seriously suggests pedestrian helments, but apparently being a pedestrian is similar risk to cycling - so why not pedestrian helmets - fair enough to advocate both, but just cycling - isn't that basically silly, or at least inconsistant.
