Bayesian statistics for Spearman's

LucieR

In my paper I have computed several Spearman's correlations. I plan to do Bayesian statistics on these correlations but JASP does not have the option for Spearman's. Does anyone know how I can get around this?

JohnnyB

Hi LucieR,

The JASP implementation of Spearman's rho is still a work in progress. What you can consider in the meantime is to use Kendall's tau (another rank-based correlation), or to run the analyses in R, using the code from our paper on Spearman's rho (https://osf.io/gny35/ ; see SpearmanSampler.R and rankBasedCommonFunctions.R for the main functions to get posterior samples and Bayes factor). Please let me know if anything is unclear still!

Kind regards,

Johnny

LucieR

Thanks Johnny!

If I were to use Kendall's tau for my Bayesian statistics, I'm assuming I would also have to rerun my initial correlations using Kendall's tau?

Can you provide a rough idea of when the Spearman's rho implementation is likely to be available on JASP?

Kind regards,

Lucie

JohnnyB

Hi Lucie,

Yes, you would need to redo the previous analyses to make it consistent, so that is probably not ideal.

I am currently a bit busy with teaching and finishing my dissertation, I hope to still do it before the end of the year, but unfortunately I cannot give any promises. If you are unable to work with the R-scripts, and you just need Bayes factors, you can also send me your data and I can quickly run the analyses for you!

My mail address is j <dot> b <dot> vandoorn <at> uva <dot> nl

Cheers

Johnny

LucieR

Hi Johnny,

No problem - thank you for the update!

Thank you for the offer but unfortunately I will probably be needing figures etc too. I am not trained in R and my data is currently on SPSS which is proving difficult to transfer but I will give it a go.

Thanks for your help.

Lucie