Congratulations.

I can be ok with using utilize but would draw the line at utilizing it.

So good to see!

One way streets here are mostly only one way for cars. Obviously not a blanket solution for every case!

You don't unfollow enough people?

I think without getting independence in there somewhere you suffer a similar problem with ESS (if every sample is effective then effective sample size is just n of that), but, nuance quibbling, the field has it's letters and I guess you've thought more in this domain than me :)

Since you make me think about it, if starting from scratch I would opt for something like ENIS -- estimated number of independent samples.

Nah I kinda love it that every time I'm forced to use a webmail client adding an attachment 'accidentally' uploads to onedrive, where I assume I've signed away the rights to my personhood or whatever already...

Best to have the talk early, you never know what they'll get up to...

Sure, but why is the respect always assumed to only flow in one direction? (To receiver). If my cultural background (bloody convicts) and time needs (needed) make skipping pleasantries seem favourable, shouldn't that carry some weight in a mutual interaction?

Honestly I'd be happy if we'd just name far less stuff and force people to describe what they really think they're doing...

Worth being aware of but also a mathy inevitability.

That's not how it works in psych stats sorry 😆

Does an effect suddenly become interesting at a specific point though? I can see some value to the null hyp when there is a directional hyp that supports an interesting theory, struggle to see value in a SESOI for the kind of case described though.

If they were bigger, would the model / story be compelling? Would alternate causal stories (eg common causes) have been considered? Knowing typical clpm research, I doubt it :/

You really think the problems are *inherent* to the approach? I could see that getting rid of the approach leads to a general improvement (because of low standards for interpretation etc), but also see value in the approach in ideal cases...

I guess if we let 'true' do enough work I agree with you for a while, but I mostly find it pretty difficult to find that level of expertise around...

Why? I'm not convinced you're wrong, but I can also easily imagine current forms of AI displacing GP's with a few more years development. Some level of displacement coupled with some level of collaboration seems inevitable to me.