Hello,
I have a set of anatomical variables for around 30 spp (same genus). Along with using these variables in species grouping, I also want to analyze the effect of some other variables (like, habitat type and life cycle pattern) in shaping them i.e. response of the anatomical traits to habitat variation and lifecycle pattern.
I have matrixed all the variables in binary code (0/1) for each species and performed PCA in R, treating all variables as same (same matrix).
So, the main question was that: can we simply depend on the angle range between the "anatomical" vs "ecological" variables in a "variables vs. species PCA biplot" to determine their correlation?
Also, need your suggestions regarding:
1. What kind of analyses can best fit this form of data?
2. Shall I log-transform the data before analyses?
3. Is that a correct form of data-treatment that I have binary-coded the environmental and life cycle variables, or I shall treat them as categorical?
4. Is it necessary to assess the significance of the analysis results statistically (eg., ANOVA...)?
Please note that the ecological variables are of different species, not of different populations of the same species.
I highly appreciate your kind help!
Your's,
Dastan
For categorical data, you would use a Chi Squared test (not anova), which is non-parametric, so you do not need to transform the data. Binary code is the same as categorical data btw:)
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA