“STUDIES SHOW….”—-GREEN JELLY BEANS. I have posted several times, including here, about misuse of statistical significance tests. The basic idea is that if you run 100 experiments, you can expect that you can on average expect 5 of the 100 to appear to be statistically significant—to appear to be results that would occur only 5% of the time.
Ethan Siegel in this blog post reprints the xkcd strip number 882 (link here) which dramatizes how you can show that green jelly beans cause acne.