jasp-issues
jasp-issues copied to clipboard
jasp-stats
[Feature Request]: Bayesian Mediation
Description
No response
Purpose
No response
Use-case
Likely whenever frequentist mediation would be used, but with the inherent advantages of Bayesian analysis
Is your feature request related to a problem?
Not a problem per se, just the inclusion of a method that the team has already developed
Describe the solution you would like
The inclusion of the Bayesian counterpart to Frequentist mediation perhaps as in: Nuijten, M.B., Wetzels, R., Matzke, D. et al. A default Bayesian hypothesis test for mediation. Behav Res 47, 85–97 (2015). https://doi.org/10.3758/s13428-014-0470-2
Describe alternatives that you have considered
Going straight to using the R package, but it does not seem that this can be easily downloaded.
Additional context
Thank you for implementing frequentist mediation analysis, using the GUI, and for the flexibility of structural equation modelling for more complicated analysis. I recently came across this Bayesian implementation, and it seems the logical counterpart to your frequent test implementation, and would follow nicely with the ethos of making Bayesian analysis accessible. Not least, because it seems like the package can no longer be downloaded from CRAN?
Hi @TarandeepKang, thank you for the request. Can you specify which r package you are referring to in particular?
Yes, would be good. However, the package is no longer on CRAN, and I would want to give this another think (and see how this can be generalized)
@TarandeepKang Is there any update on the state of the R-Package? Currently there are at least 3 that mention Bayesian Mediation:
- bama https://cran.r-project.org/web/packages/bama/readme/README.html
- BayesianMediationA https://cran.r-project.org/web/packages/BayesianMediationA/index.html
- hdbm https://cran.r-project.org/web/packages/hdbm/vignettes/hdbm.html
all 3 download from CRAN. BayesianMediationA however does not finish compilation, complaining about missing dep R2jags.
I can't help you there, I'm afraid, Thomas. EJ is the author on the paper, so I guess it's his baby (well at least partly :-))?