It effectively tests just one thing: one's prior on how likely climatologists are to lie or misinterpret data.

People don't start out with a high/low claimed prior on lying climatologists and then decide to start arguing about global warming on the internet - it's vice versa, in most cases. The end result tells you about this whole causal history, which includes a fair bit of irrationality along the way.

Of course, where the causal chain terminates is often in stuff like "my parents had political view X," which we don't particularly want to learn about, and thus has to be controlled for if we want to learn about the intermediate irrationality.