Hi! I sampled sediment cores from 15 stations in Western Lake Erie at 4 different time points (Nov '14, April '15, Nov '15, and April '16). At each station, I analyzed for a number of metrics (mcyD, chlorophyll a, phycocyanin, PON, POC, etc). I'm interested in using statistics to analyze the data and see if there's a difference between the sites/basin-averages within a single time point and across time points, but I'm having hard time pinning down the correct statistics to use. Given that the sampling was uneven (I wasn't able to make it to all 15 sites the first time I went out) and the data isn't normal (I've tried transformations and haven't been able to transform most things), what should I be using to compare the numbers? I'm assuming it's going to be something in the nonparametric family, but I'm not sure what's the most appropriate. Please and thank you!
Hi Christine,
I am mostly in agreement with Malcolm here. Multivariate analysis with 2 or more factors is indeed what you are looking for. May I suggest PERMANOVA because
"The PERMANOVA framework is readily extended to accommodate random effects, hierarchical models, mixed models, quantitative covariates, repeated measures, unbalanced and/or asymmetrical designs, and, most recently, heterogeneous dispersions among groups." (Anderson, M. J. Permutational Multivariate Analysis of Variance (PERMANOVA). Wiley StatsRef: Statistics Reference Online). This would exactly aid your unbalanced experiment design.
Definitely have a look into R package called Vegan- it is the Bible for any ecologist involved with statistics (there you can find PERMANOVA analysis amongst many others).
Also, as suggested by Malcolm, multidimensional scaling as unconstrained or constrained PCA/PCoA would provide a nice and quick overview of sample dis/similarity and point towards environmental variables influencing these.
Best,
Deni
- Badges
- Science topic
- Similar topics
- Medicine
- Public Health