Literally how I’d pronounce it, i don’t see any problem here

Comes with a bonus point if you define your concept (very) loosely, so that it can be applied anywhere, anytime, for anything.

Yes, I’m thinking about you, attitudes, the “general evaluation of an object”.

I've started a no-phone, no-computer policy this year (except for handicapped students who need it). Best teaching decision I've ever made!

Automating code reviews is one area where LLMs could be highly useful.

I still don't understand why papers are reviewed but code is not.

Merci !

That was my initial thought, after reading your query.

Nice!

Is there a way to get p values from heteroscedasticity robust standard errors?

Or, at least, manually specify these p values ?

Could the question be answered using an IV approach (instrument by treatment)?

After reading this, I actually had to check the news, because anything can happen nowadays…

@noahgreifer.bsky.social I’ll definitely check this out. As someone not familiar with this, why is this plot representing a curve instead of a straight line?

@dingdingpeng.the100.ci this one is for you !

Include their interactions as well, potentially with a squared term if applicable or needed, and then perform a marginal comparison. Much still needs to be repeated…

Is this the new reference, instead of “Efficient estimation of maximum likelihood models with multiple fixed-effects: the R package FENmlm” ?

Sorry, “weren’t allowed to have kids” ?!

Tout le contraire pour ma part, mais c’est une situation également subie : doctorant, puis post-doctorant, j’ai (notamment) besoin de publications pour monter un bon dossier :).

I’d love to see those back-and-forth exchanges turned into a blog post. It would be great to see where you started, how you updated and used it, and where you ended up!

GitHub - MattCowgill/ggannotate: Interactively annotate ggplots Interactively annotate ggplots. Contribute to MattCowgill/ggannotate development by creating an account on GitHub.

Have you played with ggannotate? It opens up a UI that lets you place your ggplot2 annotation, then it gives you the code so you don't have to guess coordinates
github.com/MattCowgill/...

Hi @grunk.comind.network,
Considering where you started, how much do you think you've learned?

I suppose it depends on how you define AI. As a non-native English speaker, I use DeepL Write a lot, as well as many GPT prompts, to check for non-idiomatic sentences, grammatical errors and wordings that are not quite right.

Hello Brian, have you already identified these secondary datasets for replication?

This is 100% the most annoying ggplot() feature

« The hours have to come from somewhere, so when someone ignores the fact that I get the same 24-hour-day as everyone else, it makes me angry — especially so when it’s coming from people with better job security and higher salaries. »

As a postdoc, I could not relate more.

In psych literature, I’m afraid yes for the vast majority of them.

Emmeans() goes brrr 😱

Cross validated ?

The more logistic regression I do, the more I like linear probability models 🙌

Thanks for making them public!

Avec un bonus de charisme pour l’intervalle de confiance.

Je suis un homme simple. La miniature affiche un ggplot(), je suis convaincu.

Are you saying that including the 'mode of collection' factor in the regression isn't enough? 😳😳