I hope my question hasn't been answered specifically on a different forum post. If so, please send me the link since I haven't found the exact problem I'm looking to solve.
My situation is as follows: I have an experiment in which I compare 10 different plant genotypes, with 16 plants per genotype divided in two different treatments (WW and WD). So I have two independent variables (genotype and treatment) with 8 repeats per treatment (so 16 per genotype, or 160 in total). I want to compare these plants in their performance measured as a single dependent variable ("LER") Now, originally I wanted to perform two analyses:
1) T-test within each genotype, comparing the 2 treatments
2) Two-way ANOVA between genotypes including a possible interaction effect between genotype*treatment
BUT here's the problem: I measured the dependent variable on each plant several times over the course of several months. Thus, there is correlation between the subsequent measurements on the same plants (repeated measures). To make this more tricky, there are many NA's for the dependent variable for reasons I will not get in to here, but which means that I do not have the same amount of repeated measures for each plant over the course of the experiment. And finally, the assumption of equal variances is also violated for 4 out of 10 genotypes (normality is fine).
I have read up on repeated measures ANOVA and Paired T-tests, but I don't think these are appropriate here due to the difference in amount of measures between plants and the fact that I want to compare means of genotype vs genotype or treatment vs treatment (and not comparing the means of the same genotype over time)
I believe other suggestions include just performing ANOVA at each measurement point but that seems tedious and unprofessional, or adding the time factor as a random blocking variable but I also have no idea how to perform such analysis. Any help is appreciated.