Need help with this problem (Correlation)

#1
For my research project I'm looking at a database to correlate variables. In particular Im trying to find the effect of phosphorus fertilizer on plant growth with and without mycorrhizae (symbiont soil fungus). I have the following variables which I am trying to correlate in what eventually will become a meta-analysis.

FertP: Variable to indicate whether phosphorus fertilizer was added to both inoculated and control treatments
Trt_mass: Average mass of plants from inoculated treatment. “Value refers to whole plant biomass if given. If not, value is shoot biomass. Units are grams.
Ctrl_mass: Average mass of plants from the non-inoculated control. Value refers to whole plant biomass if given. If not, value is shoot biomass. Units are grams.

Inoculated means mycorrhizae were added and non-inoculated means no mycorrhizae were added.

Now I am wondering what correlation coefficient I could use to determine the relationship between phosphor addition and its effect on growth mediated by the mycorrhizae. Is this possible to be done? I want to transform these values to Pearson's R values or Fisher Z values if possible. I think Ctrl_mass/ Try_mass would be a good response variable. But what does it say if I used FertP as a predictor then?

I'm working in R if you want to suggest any functions
 
Last edited:

katxt

Active Member
#2
It sounds like the simplest approach would be a two way anova. It tests both factors at once to see if either or both make a difference and checks to see if the response from one factor is influenced by the level of the other.
The correlation coefficient is used for different sorts of problem.