The connection is this:
Hypothesis A: ads are useful, people click on them b/c they deliberately want more info.
Hypothesis B: ads are useless, people click on them only when they are somehow tricked into doing so.
It's obviously the case that out of the set of ad-clickers some are type A and some are type B; the interesting question is the composition of the ad-clicking population.
What follows depends on how you interpret adware and spyware clicks: are they clicked mostly out of confusion (b/c the demographic that has adware/spyware-infected computers tends to be the elderly and other "unsavvy" people) or mostly b/c they're useful to the clicker?
Depending on how you do your weightings you might have to revise down the % of type-A clicks that're going on (and revise it down again to take various kinds of clickfraud into account).
I can't actually tell you if those stats are accurate: I've overheard figures like that from people who might know, but figure they're mostly disgruntled employees exaggerating for effect, so I tend to discount the testimony somewhat.