The error in your line of thinking is that you assume humans are better interpreting statistical evidence than they are broad logical arguments. I don't see any evidence of that.
It's not a error until a) you show that's the case, and b) it's material to my point.
Honestly, I don't think people are "better" at that (if such a fuzzy construction can have a useful meaning). But I do think they are more careful with it because it's harder to get right and more easy to visibly get it wrong. It's fine with me if expecting more concreteness just drastically reduces the total amount of social commentary. Even if it dropped by 99%, we'd still have more than we needed.
