Bret Weinstein

Okay.

So a natural experiment is something in science where maybe you happen on an archipelago in which you have a bunch of different islands that have different conditions and you can go to each island and measure whatever parameter it is.

Because nature has given you an experiment that you can analyze.

You don't have to build islands, right?

In this case, what Pierre reports is that there were 80 court cases in which...

a family sued a hospital that was refusing to give ivermectin to a desperately sick family member and they wanted the courts to intervene and force the hospitals to administer ivermectin, 80 cases.

In 40 of those cases, the courts granted the family's request and ivermectin was administered.

In 40 cases, they refused to intervene and no ivermectin was given.

In 38 of the cases where ivermectin was given, the patient survived.

In two, the patient died anyway.

In 38 of the cases where no ivermectin was given, the patient died.

And in two, the patient survived.

Wow.

Okay.

Now...

I find this, like this is incredibly ... I cannot vouch for the data itself.

Because it's not published in the scientific paper, I can't go look at the methods, I can't go find the court cases.

But assuming that the data is accurately reported, and I know Pierre well, he didn't make it up.

Assuming that the data is accurate, the level of statistical significance on that accidental study is absolutely astronomical, right?

I had Heather run a chi-squared calculation, and the p-value, I checked it also with two different AIs.