Thread by @ChelseaParlett, Here are things that come up a LOT for me as a [...]

Here are things that come up a LOT for me as a stats consultant/collaborator:

use explicit equivalence tests, not p > 0.05 (or use ROPE if ya Bayesian)
https://journals.sagepub.com/doi/full/10.1177/2515245918770963 @lakens @annemscheel @peder_isager

Equivalence Testing for Psychological Research: A Tutorial - Daniël Lakens, Anne M. Scheel, Peder...

Psychologists must be able to test both for the presence of an effect and for the absence of an effect. In addition to testing against zero, researchers can use...

https://journals.sagepub.com/doi/full/10.1177/2515245918770963

use Poisson regression for count data

you can use:

SIMR (mem's: https://cran.r-project.org/web/packages/simr/index.html)

SuperPower ( https://cran.r-project.org/web/packages/Superpower/vignettes/intro_to_superpower.html) for easy power calculations in R @ExPhysStudent @lakens

jPower in JAMOVI ( https://github.com/richarddmorey/jpower) @richarddmorey

richarddmorey/jpower

functions to compute power for various designs. Contribute to richarddmorey/jpower development by creating an account on GitHub.

https://cran.r-project.org/web/packages/simr/index.html

if there's a lot of expertise in your industry/company/field, consider Bayesian stats to take advantage of it (I recommend brms--R--and Stan--R or Python)

https://paul-buerkner.github.io/brms/ @paulbuerkner

brms truly is the gateway bayes.

rules of thumb for deciding if an effect should be random are less useful than just reading up on what partial pooling is. I recommend: https://www.tjmahr.com/plotting-partial-pooling-in-mixed-effects-models/ @tjmahr

Plotting partial pooling in mixed-effects models

There are a lot of bilabial sounds in that title

https://www.tjmahr.com/plotting-partial-pooling-in-mixed-effects-models/

You can follow @ChelseaParlett.

Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: