Very happy that @dfg_public is funding our (with @Navajoc0d3) proposal "Facing the replication crisis in machine learning modeling" as part of the @meta_rep priority program. We have summarized our planned work program at https://t.co/tZ1PFYfkIi.
If you don't have time to read a 15,000 word paper — not a breezy read, as it's a compressed version of a book's worth of arguments! — we plan to give many talks about it. The next one is at Harvard's Berkman Klein Center on Thursday. Zoom available: https://t.co/gsM4eklYUj
@lakens Ah, no, but that sounds reasonable then. Makes me wonder whether the educational aspect of the mere act of posting the prereg in an osf project etc is worth the potential for confusion.
@lakens I am surprised that even 40% of all preregs were published since it is quite common (at least at my uni) that students preregister their empirical projects or thesis (often because they have to / are supposed to learn good practice) and never indend to publish the results.
4. I watched TED LASSO when I wanted something feel-good and funny with no toxic vibes. It delivered. A (too) friendly American coach has to coach a struggling English soccer team. You would enjoy it even if you don't love sports
Beyond the Hype: A Simulation Study Evaluating the Predictive Performance of Machine Learning Models in Psychology https://t.co/MateD9crdE via @OSFramework@KriJanko
An early Christmas present for you: A cheat sheet on all the different modeling mindsets from Bayesian inference to reinforcement learning.
Basically a 1-page summary of my book Modeling Mindsets: The Many Cultures of Learning from Data.
Logistic regression still the GOAT😤:
When you fix the problems in papers claiming "ML predicts XYZ amazingly well", you end up with ML ~= LR
https://t.co/oOYJ0tCdJN
Exciting news!
Our book Supervised Machine Learning for Science is now published! 🥳🥳🥳
Available in hardcover, paperback, eBook, and PDF. You can even read it online for free!
@JnfrLTackett@LorneJCampbell@ianhussey@DenOlmo This preprint totally confirms my impression - but not only regarding preregistrations but also open code & other materials. In every review process I was a part of this year, I was very obviously the only one that looked at the R code.
📢 The first chapter of the AI snake oil book by me and @sayashk is now available online. https://t.co/BEH9bApr2H
It is 30 pages long and summarizes the book’s main arguments. If you start reading now, you won't have to wait long for the rest of the book — it is available to preorder and will be published in less than two weeks. https://t.co/foQpEhRfhs
We were fortunate to receive positive early reviews by The New Yorker, Publishers' Weekly (featured in the Top 10 science books for Fall 2024), and many other outlets.
Here's an outline of all the chapters: https://t.co/nj2ZRGFn5L
There will be plenty of opportunities to participate in the upcoming projects for all researchers interested in reproducible machine learning modeling in the social sciences.
Very happy that @dfg_public is funding our (with @Navajoc0d3) proposal "Facing the replication crisis in machine learning modeling" as part of the @meta_rep priority program. We have summarized our planned work program at https://t.co/tZ1PFYfkIi.