> "I'm not convinced, maybe you just saw black cats but non-black cats exist, after all, other animals, like dogs, horses, cows, ... have different color".

Well the actual argument wasn't nearly as extreme as "only black fur", and a better (but still messy) analogy would be you citing one other animal and we don't have good stats for any other kinds of animal either. That's part of why I'm saying the child example isn't very good at affecting convincedness.

Also you used the word "proves", I feel like if you say X proves Y then it's not weird for me to call that a "claim". I don't think "I'm not the one making any claim" is valid here; you're responding to the original claim with arguments that include your own claims.

> there are different models that explain the situation as accurately

I'd say that so far none of the models here reach "very convincing" for all the data at hand. They're all big maybes.

> I'm saying that the extreme of the distribution shows that we cannot simply assume that the mean is at the same place.

Well even with a completely isolated size factor, the means are different. I don't think anyone was saying that.

And sure the distributions might have different shapes. There's a lot of analysis here that hasn't been done.

> What I don't understand is that on one hand, the claim is that violence is mainly due to a "mechanical fixed parameter" such as the size of the person and that societal messaging has no significant impact.

That's too binary. One thing can be the main effect while another is still quite significant. Like 75/25 just to toss out a number.

The places I've used "prove" are in totally different contexts that the context in which you pretended I was claiming some proof.

For the rest, it feels like you are moving the goalpost. The initial discussion was about the fact that if there is an impact of the societal messaging, it is smart to acknowledge that (and I insisted that it should still be done carefully). If now you are saying 75/25, then you are still saying that the societal messaging has an impact. Therefore I still think it's smart to acknowledge it.

(I understand the objection that it may be too dangerous, but unfortunately you did not go into this direction. I come back on that on my last paragraph.)

Why this impression? Your initial main argument was initially resting on "lesbian couples have the same amount of violence", but this argument does not make sense if it is 75/25: if there is a 25% effect, what does observing the same rate means? If you observe exactly the same rate, does it mean that we have a societal impact that somehow canceled itself, or does it means that in fact the "natural rate of violence from women" is less than men (the exact thing you pretend this data proves impossible), but that the societal impact increases it up to reach the same value? If indeed you believe in the 75/25, then the "same rate amongst lesbians" cannot prove anything. And on top of that, now, you also need to justify why it is 75/25 and not 85/15 or 65/35, and the whole argument seems to be "it's my gut feeling" (not that it is bad in itself, but then you cannot use that as basis to pretend that the lesbian data proves something).

I can be wrong, but because of that, it feels to me that you were not really agreeing to 75/25 at the beginning of the discussion (otherwise you would not have used the "lesbian rate" argument as you had), but now that the discussion advanced, you are making some concessions but still try to find some reasons why the conclusion you prefer is still valid. If my impression is correct, it means that instead of looking at the arguments first and reaching the conclusion based on the arguments, you start with the conclusion that you prefer, and build arguments in order to defend this conclusion.

(and I understand that you instinctively prefer the conclusion where we avoid acknowledging that there is something specific to men, it's not a comfortable prospect)

I've mentioned that we could have discussed around the danger of acknowledging. Unfortunately, if now we start discussing that, it will just reinforce my feeling that you jump to another thread now that one ran its course because you want to defend the conclusion you like.

So, yeah, I guess there is not much more to add here.