Earlier quoted context omitted.
People conspire all the time, so there's bound to be conspiracy theories that are true. I think it's telling that the author uses the moon landing as an example, which to me seems to be among the least likely (but somewhat popular) theories. All credible evidence points to the moon landing actually happening, making it a good candidate for this study. Simply labeling that "endorse conspiracy theories" is a bit though…
The "fake" moon landings is an interesting example; although the number of direct eye-witnesses is tiny, there are thousands of scientists and engineers at NASA that will queue-up to testify they really happened, and nobody who actually knows anything will deny it. But the affirmers are all people in white lab-coats, and many people automatically distrust anyone in a white lab-coat. No "ordinary people" went to the m…
I imagine you could just as easily make a study "people who look at fewer partitions of datasets reach wrong conclusions more often". Just as you can sample more fish from a lake and see if there's a pattern in their color, you can try to look at more perspectives and see if there's a pattern one way or the other.