I am currently working on a project which replaces a COD analyses with a TOC analyses using the COD/TOC-ratio factor and/or linear regression of COD vs TOC.
Which form(s) of statistic tests can I use to determine whether the TOC analyses is a good replacement for the current COD analyses?
I was thinking about the Pearson correlation coefficient, linear regression of course, 2-tailed t-test (or paired?), and maybe a F-test to check the correlation using SSres and SSfactor.
What do you think? Do you have other suggestions? I've added a photo to show you all what I mean.