Hello,
I have 2 groups (carcasses and controls). I sampled beetle species at carcasses repeatedly. Each carcass had a control site that were also sampled at the same time. I'd like to compare both groups over time. I ran a repeated measures ANOVA, but quickly found that my data violates the assumption of sphericity. Control sites had very few beetles across time, with little to no change. However, carcasses had a higher number of beetle species at first and slowly declined over time. As you can imagine, the variance among sampling periods is not consistent. Is there an approach I could use that doesn't assume equal variances across repeated samples? Also, I do not think that the repeated measures ANOVA paired carcasses and controls, and I'm not sure if that is an issue. I don't have experience with mixed models, but I'm reading that they may be an appropriate choice.
If the distributions of your measure (a.o., Shannon's Biodiversity Index to look at the change of beetle evenness and richness over time across?) are quite normal, just run repeated measures ANOVAs; and, indeed, as Adrian suggested, use the GG correction(see: https://en.wikipedia.org/wiki/Mauchly%27s_sphericity_test ). If the distributions are far from normal, you could use nonparametric alternatives; e.g., the F1_LD_F1 procedure from the 'nparLD' R package by Brunner and colleagues (see attachments).
My Dear Josh... firstly I hope all success to you
second I'm have artical in this subject in my researchgate you can read it
With Best Wishes
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA