"Condition on" option for correlation in Jasp v0.14
Hi all,
I am looking for information on the "condition on" option in Jasp´s frequentist correlation analysis. It seems it was introduced in the latest version v0.14 as I can not find anything about this in older documentation. However, the documentation for the new version is not available yet. Does anyone know about this new feature? I am wondering if I can apply the "condition on" for my problem:
I have two ratio scaled variables (movement and localization), computed for every subject but for two different conditions (targets, nominal scaled). I neither want to average over the conditions as there is some direction dependency which will be lost then, nor do I want to inflate my results by mashing all samples together without any sort of correction. Applying the "condition on" option with a condition vector that includes my targets (1 or 2) works fine but is it correct? Or is there even another solution to my problem?
Thanks for any advice.
Kind regards,
Matthias
Comments
I believe that "condition on" allows you to compute a partial correlation. So it seems that this is what you want. I will ask our experts.
Cheers,
E.J.
Hello,
As E.J. says, "condition on" is used to compute a partial correlation - whatever is put in this box is used as the "conditioning"/"controlling" variables.
I am not sure whether this is appropriate for your data, but I don't know enough about your research question and the structure of the data. Normally, you would condition on a continuous variable, not a nominal condition. It also seems that you have a within-subject design is that correct (all subjects went under both conditions)? If that is the case, partial correlation does not seem appropriate as the analysis assumes independent measurements. From the description of the data it seems more common analyses would be calculating the correlations for both conditions separately, testing whether or not they are the same, pooling them together, etc.
Best,
Simon
Thanks for the info both!
Yeah, it doesnt seem like the right thing to apply. I went for another approach. Thanks again, Simon.