Within a dataset I am analyzing, I'm examining a line by tester experiment in which only one tester exists and was crossed with 12 other parents. Normal analysis would include error terms associated with Lines x Tester, but given we only have one (2 reps/yr across 3 years) we will not have any variation for this term.
Is there a way to partition data to identify variation on a rep basis without confounding data if other terms are significant? Are there other statistical tools we could employ which will help us in drawing conclusions on the dataset?