What do you do if you are just trying to bang out the last gasp of your master's project and inevitably it involves some very noisy environmental data which is leading to not a lot of correlation between the soil amendment and the crop responses.

In short yes it's a dumb project and taking field samples in this way clearly needs more time then I have (maybe 10 years of sampling would eliminate some of the noise) However I don't have that time, and I do need to pass review... I'm talking defensible statistical methods to point to the hypothesis being invalid rather than poor methodology/ sampling practice.

Please link any resoruces that you know of, it's not actually for me just a colleague that I need to free up from lab work so we can launch our own business, ty.

And yes I will ensure he never calls himself a scientist ever again.