It is obvious that ivmmeta.com is picking up something that is a real signal. The only question is a signal of what. Fraud? Bias by the site authors? One of many complex statistical effects? A working drug?

There were loud, even unreasonable, voices in the debate but nobody was raising an actual problem with the results. Ivermectin has a respectable safety profile. "Maybe it works, no obvious downside" was a reasonable position.

It was always a long shot, the evidence was weak. But loudest voices in the anti-ivermectin crowd are people like, eg, jacquesm in this thread. A lot of bluster, a smidge of bullying and a weak-sauce appeal to authority for why statistical evidence should be ignored. And YouTube et al. believe it to be convincing evidence or they wouldn't bother to censor discussion of it. If that is the opposition then they don't seem to have uncovered a methodological problem yet or they'd raise it.

But Scott raises an interesting theory that would be enough to explain ivmmeta.com. Since it was only ever weak evidence that is enough for me to change my mind.