Hi,

I have a dataset including 8 phenotypic traits measured in 10 fish from each of 4 distinct populations (N=40 fish). In addition, I have a second dataset including 4 behavioural traits measured in 12 fish from a second batch of individuals from the same populations than before (N=48 fish). I end up with 8 traits measured at an individual level and 4 traits at a population level (mean+SD).

I would like to perform a multivariate analysis including all these traits to assess their associations and perform post-hoc correlation tests. I was thinking about a PCA that would include the 12 traits for the 40 fish so that each fish ends up with 1) the 8 traits individually measured in this fish + 2) the corresponding population mean for each of the 4 additional traits.

I don't know if I can run a PCA on such a dataset combining individual- and population-level traits and if so, how to add variability in the 4 population traits so that the 10 fish of each population don't show the same value for these traits.

Hope this was clear enough. Thanks for your help!