Hi everybody,
I'm running PERMANOVA analysis for a habitat experiment in which I'm transplanting epiphytes (living and mimic) in some hosts. The factorial design is:
3 densities (0, Low, High) x 2 epiphyte type (Mimic vs Living epiphyte).
Logically my samples should be 6 (with a double sample 0 due to the 0 density) but... statistically talking in the spreadsheet I cannot leave black cells for factors so I'm contrained on considering 3 epiphyte levels adding the 0 level (0, Mimic, Living)... so that I finally have a 3x3 design with 9 samples:
00
0Mimic
0Living
Low0
LowMimic
LowLiving
High0
HighMimic
HighLiving
4 of which (0Mimic, 0Living, Low0, High0) are repeated samples of the first one.
I've doubt on the validity of this analysis (even because the software automatically expects 9 samples) or maybe there's something I don't know on how working with the black cells for factor columns?
Thank you very much
AS
Yep, sorry. I'm working on Habitat Cascades: a host (a seaweed in my case) supports some epiphytes which in turn supports clients (invertebrates).
My experiment will try to reply 2 questions:
- is the density of the epiphyte (low vs high) relevant in driving abundance/richness of clients?
- is there any difference in using living vs mimic of epiphyte?
Abundance and richness are simply count and identification of my invertebrates species and "epiphyte type" and "density" are my factors (the cores ones). I'm testing their effects (mainly if there is one) on clients.
I hope I was clear enough. And thanks.
AS
That would be the right choice but I cannot use it... because the 0 density result in a 0 epiphyte as well, so I must add a third epiphyte level (which is 0).
the good thing to do would be using a 2x2 design (mimic/living and low/high only) and removing the control samples.
Thanks by the way
AS
Thank you for your tenacity.
Well, I've already replied the question "is the epiphyte relevant in driving abundance/richness of clients?" and the answer is yes. Now I'm going deeper and my questions is "are the epiphyte density (low/high - 1st factor) and the epiphyte type (living/mimic - 2nd factor) relevant?"
I collected the host (Cystophora) and some natural epiphytes, I washed them and I manually transplanted them on the host (in 0, low and high densities) and I did the same with some other host at which I attached an epiphyte mimic.
The idea is that the abundance of clients is dependent on epiphyte size and that clients may use the living epiphyte more efficiently than the mimic one. But these are just thoughts.
Replying the other doubts:
- my variables (abundance and richness) are standardised per gDW of the host... in here I'm considering the size of the host
- second, I use PERMANOVA based on resemblance matrix in Euclidean distance.
I remark that I've already run the analysis removing all the controls (with a 2x2 design then... just low/high and living/mimic) and the results don't change. I think the software, specifically created for this kind of analysis, is smart enough to make the right considerations (considering my input rules correct).
Thanks
AS
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Powders