Thank you for your input, sld. In regards to treating year as a fixed or random effect: I do realize that two years is not a large sample for estimating the variance between them. When we conduct these studies we repeat in time as a means to determine the stability of our treatment effects (in this case, graft and irrigation). Furthermore, I have conducted a prior analysis with year as a fixed effect - none of the interactions with year (graft x year, irrigation x year, graft x irrigation x year) nor the main effect of year were significant. Thus, instead of keeping it as a fixed effect, we make it random, thus broadening our inferences on both graft and irrigation outside of 2016 and 2017. We do an analysis looking at the total yield at the end of the study. The repeated measures analysis is helpful, especially for a grower, as it can show how plants yield throughout the growing and marketing season. When we did the repeated measures analysis last year, we found that graft X Harvest was significant. From there we modeling this interaction by modeling harvest as a continuous factor. Thanks again for your help.
... View more