[IfOnlyYouKnew]

Nothing is "banned".

AZ accidentally ran their first trial with only half the intended dose. For some reason, this turned out to work better than the intended regimen (90 % vs 60 % in the other trial).

When they noticed the mistake, they pretended it had been their intention all along.

That's one troubling error, one surprising result that doesn't fit expectations and could, therefore, indicate other, potentially dangerous, problems in thinking or execution, and one deliberate lie.

Personally, I'd tend to agree with you and would probably take their vaccine if offered.

But that's something entirely different than allowing it to be given to millions of people. Getting this wrong would lead to a total breakdown of trust in institutions, and, in due course, a few democracies as well.

To insinuate some conspiracy given these circumstances is evidence for the mechanism of that breakdown, as well as irresponsible and, frankly, amateur populism.

[chromatin]

And it turns out the cohort who received the half dose was on average younger than the total study population, this explaining the apparent improvement in efficacy

[IfOnlyYouKnew]

As far as I can tell, the virus is more prevalent among younger people. Since infection, not death, was the primary endpoint, and also because medicine hasn't used that sort of wishful reasoning for about a hundred years, the age difference doesn't explain anything.

[BillyTheKing]

it certainly does, younger people have stronger immune systems, and therefore generally react better to vaccines than older people. The stronger the immune system reacts to a vaccine the better, since that usually ensures that there will be an immune-response to the actual virus as well.

[SAI_Peregrinus]

> For some reason, this turned out to work better than the intended regimen (90 % vs 60 % in the other trial).

The difference isn't actually statistically significant. It's probably 60-70% effective overall.

[IfOnlyYouKnew]

It is utterly impossible for two trials of that size (thousands) to result in 60% and 90% effectiveness, respectively, by chance. I can't quite do a t-test in my head any more, but I'd guess that difference passes the 5% significance threshold somewhere around n=30.