Arguments that affirm the consequent infer conditions based on conclusions. This fallacy is easiest to explain by way of example. Quite simply:
If I am having a stroke, my head would hurt.
My head hurts.
Therefore I am having a stroke.
This type of fallacy, as you can see, is patently obvious to most observers who can apply logic to arguments. However, it is rarely as obvious as in the exaggerated example above. Take, for example, something that has the classic dimensions of denialism: the Teach the Controversy "movement." Step one – fabricate controversy where none exists. Step two – let the fallacies roll in a bald effort to lend legitimacy to the argument.
What denialists such as Discovery Institute (the well-funded folks who bring us Teach the Controversy) or the Institute for Historical Review (Holocaust deniers) do is very basic but sufficient to fool the ignorant:
Legitimate scientific theories are supported by scientists.
(Insert theory) is supported by scientists.
Therefore it is a legitimate scientific theory.
Like all truly pernicious logical fallacies, this contains a kernel of truth. That is why it is so deceptive. Obviously, legitimate scientific theories are supported by scientists. But they are also supported by other things (data, research, and testing). Similarly, a denialist theory is supported by "scientists." But they are supported by only a small number of largely-discredited fringe conspiracists. Finding one or two scientists of questionable qualifications who will support a theory does not mean it is a valid one, especially when nothing else in the form of data/evidence support it. Leaping to conclusions based on half-truths and incomplete arguments is fun!
The goal of denialism isn't to be proven correct; it's simply to muddy the waters and create doubt among those who aren't paying too much attention. Like a cheap infomercial seeks credibility by having men in white lab coats stand around, denialists know that plenty of (not-so-bright) people will see an "expert" and infer validity to the argument.