If you’re interested in #statistics, #Metascience, #openScience, #pHacking, etc., this is quite a good read.

https://joebakcoleman.com/blog/2023/pcurve/

In any case, it's fascinating to see #statscop twitter fall so publicly and transparently into the prosecutor's fallacy.

@ct_bergstrom they are actually not. They are using Bayesian statistics to evaluate relative evidence. Of course this can be cherry picked. But in this specific field I know 3 you g researchers who left academia because everyone in the field was p-hacking like crazy, they did not see the use of being in science. Given that reality, what should be ones posterior belief this is p-hacking? Extremely high, I think. So statscops, like most often, are just right.

@lakens I've heard the same arguments made about prosecutors, right down to the parallel anecdotes.

@ct_bergstrom They might not specify the Bayesian model (it can not be specified, and you know ut, just like you know this study is unlikely to replicate).

The more interesting thing here is the pushback. No one wants to defend the study. But we are supposed to do what exactly with these p-values? Trust them blindly? No? Criticize them with the only tool we have? Also no?

We do the second. We use the tool. Distrust the study, but not dismiss it, because we could be wrong.

@ct_bergstrom This all falls in a pattern. Just like the recent message where you based John Ioannidis, and I asked you to prove your criticism, and you ignored my post twice, because you knew your criticism was unfounded. I see what you are doing. And I do not like it.

@lakens Sorry, Daniel, but after you glossing over a substantial history here.

As I've said before very publicly , you are a bully in how you present yourself online, you turn academic disagreements personal, and you go from zero to sixty with accusations at the drop of a hat, just as you have here.

I usually ignore you for THAT reason.

And with this post, I'll go back to doing so.

@ct_bergstrom Sure, feel free to call me a bully, and ignore my criticism (instead of responding to it). That is your right. No need to engage. But as I said, I see what you are doing, and the weakness and one-sidedness of your criticism, and I will keep calling it out when I see it.

@ct_bergstrom of course, it is convenient that you called me a bully for criticizing your preprint, and now Stuart als did not like your preprint, and said so in an even more annoyed tone, and now you are using your platform to cririze both of us, even though all factual points we made are very defensible. It is an interesting way to act.

@lakens you ok, bro?