comparing Pearson r
Hi everyone!
I have two continous variables (VAR1 and VAR2) and I wish to calculate their correlation in married and not married individuals and compare the two correlation coefficients. I wish to compare the coefficients in the two subsamples. I think that this should be done with a z-test, right? Can this be conducted in JASP? Note that I opt for the Bayesian options.
An alternative that I thought was to run a bayesian linear regression analysis and include the marriage variable as a covariate (and then include the interaction marriage * VAR2 term in the model). However, JASP does not let me put Marriage in the Covariates box (in the classical linear regression this is not a problem).
Finally, I run the correlations separately by group and the results did not update automatically when I was changing the filter. Not sure if this should be happening, but I am letting you know :-)
Best,
Yorgos.
Comments
Hello Yorgos,
I cannot help with the statistical part of your question Im afraid, but as far as the results not updating as you changed the filter... This could be several things, maybe it updated so quickly you didn't notice and it gave the same results or maybe the filter didn't change the data as far as the analysis was concerned and so it wasn't rerun.
To be sure about this it would be useful if you could make an issue with your datafile and a description of how you updated your filter so that we might be able to check it to be sure. This issue can be registered here: https://github.com/jasp-stats/jasp-issues/issues
Best regards,
Joris
Dear Yorgos,
The Bayesian hypothesis test for the equality of two correlations is still under development.
Cheers,
E.J.
Thank you for your answers!