My colleagues above me make some great points. I agree that noticing any deviation earlier is better than after the validation. It would seem we also agree that verification should have as little deviations as possible. But I think srg36 makes a good point in that deviations are common in running new tests. I think this largely takes effect when there is a new member on the team. I started volunteering at a cardiac tissue lab and I was tasked with doing a “simple” DNA Assay. I was given the protocol and followed it meticulously, but undoubtedly I had to make some deviations. When the protocol asked for 0.01 mico liters, I was unsure how accurately I could get that quantity, and ultimately unsure how much that would affect my overall outcome. As it turns out the protocol given to me was quite old and certain changes were executed by those who have done it before, but never changed the document. This asks the question, is the DNA Assay I did valid? Was I right to stick to the original protocol, or should we adopt this new slightly deviated as our new standard? Does that change anything else? Had I known about the deviation earlier it could have saved time and money.
Anyway. I think that if our project that summer where to have been published, I would expect that there would be some mention of a deviation in the verification/validation section, as well as a deviation section to go list out the steps taken and the assumed impact it had.