Never wish them pain. That's not who you are. If they caused you pain, they must have pain inside. Wish
Idea Transcript
Methods in EEC (BIO 221B) Spring 2013
Dr. Jim Baxter Dept. of Biological Sciences
MULTIVARIATE EXERCISES
We will use the software package “PCOrd” to conduct multivariate analyses on several multivariate datasets. These exercises will provide you with an introduction to several of the multivariate methods I discussed and give you the opportunity to compare and interpret the results of these different methods. The exercises use the data tables from the assigned reading (Chapter 1 in Manly 1986). 1. Conduct a cluster analysis of prehistoric Thai dogs using data Table 1.4. The data file is: ‘dogs.wk1’. Use different distance methods and clustering techniques to compare your results. Which dogs are most closely associated, as defined by mandible size? 2. Using the bird size data in Table 1.1, reduce the number of variables to a single explanatory variable. Data files: ‘bird.wk1’ and ‘bird categ.wk1’. Evaluate different ordination methods in their effectiveness. Generate an ordination diagram showing those that survived verses those that died. Test whether or not there is a difference in “bird size” between birds that survived and those that died. 3. Conduct a PCA of European employment percentages (Table 1.5) using a correlation matrix. Data file: ‘employment.wk1’. How many axes should be used? What proportion of the variance is explained by the first 2 PCA axes? Is there an identifiable pattern to the spread of countries in the PCA diagram? 4. Conduct a CCA of the butterfly data in Table 1.3. Data files: ‘butterfly2.wk1’ and ‘butterfly env.wk1’. Compare the results of the CCA to a regression of the environmental variables on the axes of a straight‐ahead ordination. 5. Conduct an ordination of the Egyptian skull data (Table 1.2). Data files: ‘skulls.wk1’ and ‘skulls categ.wk1’. Evaluate the degree to which skull size differs in the various epochs. Manly, B. F. J. 1986. Multivariate Statistical Methods: A Primer. Chapman and Hall, New York, NY.