I'm planning a study with a twice repeated measurement but I'm primarily interested in the correlation between the two measurements. As there are a number of factors potentially influencing the agreement or otherwise between these two measurements, can Cohen's k or Pearson's r be used as the DV in a multiple regression analysis ? If so, are there any conditions or parameters that need to be taken into account ?
Hello Dougal Julian Hare. Do you really want/need to look at correlation coefficients and kappas? Or could you address your questions via regression models with Y = the time 2 measurement and the time 1 measurement as one of the X-variables? This would allow you to add the other variables you want to account for as additional explanatory variables. And presumably, you would want to look at there interactions with the focal variable. (For the situation where you would compute kappa, it would likely be some kind of logistic regression model, the exact type depending on whether your table is 2x2 or larger, and whether the variable is ordinal in the latter case.)
If you really do want to analyze correlation coefficients or kappas, the first thing that comes to mind for me is meta-regression.
HTH.
Surely yes, provided the regressors of the model having correlation coefficient as DV are not the same from where correlation come out. In the paper linked below (Fig.1 panel B)n there is an example having canonical correlation coefficient as dependent variable.
https://www.nature.com/articles/s41397-019-0080-6
Thank you Medhat Elsahookie Alessandro Giuliani Bruce Weaver for these really useful answers - I feel much more confident to proceed with a modified version of this approach.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size