Robin HansonVerified account

But don’t neglect base rates. Even if people believe P(Timing|True) < P(Timing|False), we still have to consider the base odds of a true accusation given that she’s making any accusation at all. Which I think are extremely high, and I expect a majority of your pollees would agree

This poll isn't a whole analysis. Yes of course base rates are relevant for a whole analysis.

Yes, but you added the last sentence "We expect this timing for a much larger % of false accusations" and that is not found in your poll (to the extent anything is).

Yes of course, that is another relevant number. I didn't ask for it in a poll, because I see it as more obvious.

Why do you see it as more obvious? That's quite the assumption you've made.

Isn't it obvious that false accusations, done to hurt the accused, are more likely to be done at just the moment when it would hurt them the most?

But it does not appear to be obvious to you that an accusation that is difficult to make is most likely made at a time when the accuser desperately thinks there will be no more options?

That is a plausible effect, but just not as strong an effect as the other. Do you really want me to run another poll on this analogous question, just to be sure?

She reported her accusation to her congressional rep shortly after he was nominated, not “after hearings & just before Congress vote.” Your poll doesn’t reflect the actual accusation.

That wasn't a public accusation.

Yes, but she claims that she didn’t want to go public when she did, either. Her accusation was leaked by a congressional staffer. She was approached by a reporter. Why judge the honesty of the accusation based on the timing if she didn’t choose the timing?

There are many things we don't know about the timing. But we do know when a public accusation was made. At that is the most important timing. So plausible to ask poll about that.

In this case it seems like there was a fair amount of evidence - testimony to friends, to reporters, to her therapist - that predated the Supreme Court nomination.

That may well be true. Doesn't cancel what i said.

alright. I estimate a far higher percentage than 1%. Model I have is "coming forward with an accusation will impose a significant cost on me, and I will only do it if I see a corresponding benefit to society, or if the stakes for the attacker are high enough"

maybe a better framing is "I've done something bad in my past. What are the odds I am accused by the victim if I become a very public figure vs. not"

You're a full blown quack. The results of your poll have more to do with the demographics of your followers than the details of the accusation. But go ahead and slap yourself on the back for the "successful" circle jerk!