It took me a lot of time to find a balanced way to evaluate the results. I felt a bit like Unkar Plutt giving Rey only three portions for her painfully collected material..
I hope I found one: 10 points for participating and doing the work, one point each for correct and meaningful answers to the the initial questions (total 3) and 17 for correct figures, where the relative root-mean-square errors of all given figures are added and scaled from 0 (no answer) to 17 (100% agreement).
One injustice remains and this is that I might not have told you how to evaluate the double injections, so the ones of you who put all the unknown sample results into one bucket to calculate the standard deviations have not 100% of the available points. I apologise for that (and tell me if I told you to do it the wrong way). What the customer wants to see in your bargraph is the biological reproducibility = one result per sample. So the correct way would have been first to calculate averages per sample and then the standard deviations of the groups- see below. The averages come out the same, but your SD goes down. (In fact that means that our method is less accurate than the biological variation, a rare case and a bad sign! One would think about running triplicates and, for sure, improving the method – find out where the variability comes from: Unsuited internal standard, irreproducible recovery, volume errors, interferences etc)
Otherwise the questions were answered mostly ok without systematic problem-