Advice on correlation

Question

Dear Cross validated community, I need some guidance with statistics on my sample set. I collected plants at 7 sites following an environmental gradient.

Data consists of
• 7 sites along spatial grid
• 3 replicate samples for soil compound 1
• 3 replicate samples for soil compound 2
• 70 to 90 samples of plants per site, measuring both compound 1 and 2 in them.

I wanted to correlate the concentration of compound 1 (or 2) in the plants with the concentration of the same compound in the soil. But, understand these are not independent samples (so it violates Pearson).

Please advise. Ideally in SPSS terminology.

Answer 1

I can't help with SPSS, but you don't want correlation, you want regression. You can then use a multilevel model or generalized estimating equations to deal with the dependent data.