Hello, regarding point 2: if your sites are similar and located near each other (for example, on the same reef), then they are pseudo-replicated, which is okay if you just want an idea of local effects. But if you want true replication, select at least 3 sites distant from each other and place at least 3 plots in each area. You only need to use non-parametric statistics if the data are not drawn from a given probability distribution (such as normal).

Regarding point 3: as you mentioned, this is the same analysis as in point 2, but done for each species.

If your data are non-parametric (check the residual distribution and also do an autocorrelation test), you can use Chi-squared tests to compare sites and species. If the data are normal in distribution, you can use GLM (ANOVA) with post-hoc tests, using the factors of site and species with the variable being degree of bleaching.

I think the opinion by Andrew Paul McKenzie Pegman is very right

Hi all,

Thank you very much for the input. I will see into the advice and come back to everyone about the data if I have any questions

Thanks again.

threat index may serve

Shannon Weiner Index can also be a good proxy