Bayes effect size delta

HannahRap

Hi there,

Two questions about the effect size used in JASP:

1. What is an effect size (delta)? Is it Glass's delta?
2. How do you convert Cohen's d into delta? (this is straightforward if delta is Glass's delta)

Cheers,

Hannah

PS. Thank you Team JASP - the program is fantastic!!

EJ

Hi Hannah,
I assume you are talking about the t-test. Delta is the population version of the sample Cohen's d. So it is just the mean population difference divided by the population sd. You can find a detailed description in http://www.ejwagenmakers.com/2016/LyEtAl2016JMP.pdf (and also in the PBR t-test paper by Rouder et al., which is cited in the Ly paper).
Cheers,
E.J.

HannahRap

Very helpful - thank you! And thanks for your speedy reply! Much appreciated:)

Robert_Voogdgeert

When using the BEST-package in R the effect size I get, has a mode of -1.55 with a 95% CI of -2.36, -0.528. While in JASP the median is -0.685 with a 95% CI -1.211, -0.184. This on the exact same dataframe. As far as I know, the effect size in BEST is δ as well. Can anybody explain te difference?

EJ

This is just a wild guess, but I think that BEST uses a t-likelihood, not a normal likelihood. This *may* explain the difference