I sampled three different rivers over one year during the four seasons. In addition, in the autumn and winter seasons I sampled both during an event of high (drought period) and low water (flood period). I have 3 categorical variables (river, season, event) and 11 continuous (environmental parameters, like oxygen, temperature, pH, no3, etc...). The goal is to assess the degree of similarity-diversity of the three rivers, the influence of the seasons and flood-drought periods on the environmental parameters. I was thinking about a set of multivariate techniques among: 1)redundancy analysis 2)canonical correspondence analysis 3)multiple factor analysis 4)factor analysis of mixed data. Which, in your opinion, would better answer the questions stated above? Thanks in advance for the answer!
3
-
$\begingroup$ In your case, I'd go for a Multiple Co-Inertia analysis. See Fig. 3D in this paper esajournals.onlinelibrary.wiley.com/doi/full/10.1890/03-0178 . From your description, it seems this is exactly what you have $\endgroup$– Diogo B ProveteCommented Nov 23, 2023 at 12:35
-
$\begingroup$ Both CCA and RDA require at least two tables/matrices. And it seems you only have one matrix with 11 variables, and then 2 factors, each with 2 levels. River is not a factor, but your sampling unit $\endgroup$– Diogo B ProveteCommented Nov 23, 2023 at 12:37
-
$\begingroup$ Thank you Diogo for the link, I'll study it! Just in this case, I presume, "rivers" is still a factor cause identify three different watersheds, while the sampling stations in each river can be counted as the sampling units. I recently discovered the nMDS maybe it can fit my case? $\endgroup$– Rudy BenettiCommented Nov 27, 2023 at 16:44
Add a comment
|