gravatar

lindeloev

Jonas Kristoffer Lindeløv

Recently Published

Wilcoxon is (almost) the same as t-test on signed ranks
This document presents the close relationship between the p-values of the Wilcoxon signed-rank test and t-test with signed ranks as dependent variable. It is an appendix to the post "[Common statistical tests as linear models](https://rpubs.com/lindeloev/366857)". Since Wilcoxon matched pairs is just the signed-rank on difference scores, the points below apply to that as well.
Common statistical tests as linear models
Demonstrate that many classical tests are special cases of linear models (regression)
How to compute Bayes factors using lm, lmer, BayesFactor, brms, and JAGS/stan/pymc3
Here, I compare different ways of calculating Bayes Factors in R. I start with a TL;DR section showing off the syntax for the simplest of all models: the intercept-only model. Then I go on to demonstrate Bayes Factors for mixed models using the same packages, including a more thorough discussion of pros and cons.