Given that Randall Munroe has already expressed that the Frequentist vs. Bayesian angle of the comic was an afterthought, I think the real point (and the point

Given that Randall Munroe has already expressed that the Frequentist vs. Bayesian angle of the comic was an afterthought, I think the real point (and the point that I actually took away from the comic) is about the blind application by scientists of the 0.05 P-value threshold for significance without regard to the specific circumstances of the experiment, which I can attest is a huge issue in the scientific literature. Another huge issue is using a statistical test on data that is known not to satisfy the assumptions of the test, either out of ignorance or because the test gives a good (i.e. P<0.05) result where the correct test doesn't.