I wrote this paper arguing that the inferential statistics we report in psychology papers are so disconnected from the hypotheses they're meant to test that they may as well be made up. feedback is appreciated.https://psyarxiv.com/jqw35
-
-
oh, sure. but I'm not arguing for any particular kind of model. I'm saying that researchers need to be clear and honest about the relationship between their models and their verbal claims. that's all.
-
as I say in the last section, the easiest solution to the problem is exactly what you say: use fixed-effects models, and just moderate your claims. if you want to publish papers claiming evidence that there's an effect A > B in these 20 people tested with these stimuli, godspeed!
Kraj razgovora
Novi razgovor -
-
-
minor thought: why don't you square your variance? it's weird to use standard deviation as a parameter of your Gaussian distributions, no?
-
hah, I didn't even think about it, tbh. I think I'm too used to working in Stan and PyMC3, which parameterize distributions in terms of sd rather than variance. but I agree that it's less conventional, and should probably be squared.
Kraj razgovora
Novi razgovor -
-
-
other quick thought (should have been my first): this is an amazing piece, on a very important topic. Reads like a novel (I couldn't put it down), and highlights how a high-level, almost epistemological concern, is woven in every day statistics.
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
Yes, I fully agree with the fact that the failures of science are failures of generalizability, and I have hope in machine learning approaches to address them, because they are robust to mis-specificationhttps://www.sciencedirect.com/science/article/abs/pii/S0959438818301089?via%3Dihub …
-
generalizability in science =! generalizability in ML models
- Još 1 odgovor
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.