PBDB Intensive Summer Course 2007

Foundations Similarity index = a numerical index describing the similarity of two community samples in terms of their species content Similarity matrix = a square, symmetrical matrix with the similarity value of every pair of samples, if Q-mode, or species, if R-mode, in the data matrix The similarity matrix is the basis for all multivariate techniques depicting relationships among community samples or taxa, so the choices made at the initial stage of an analysis will strongly influence the results at the final stage All similarity metrics can be converted to difference metrics by subtracting them from their maximum value (1.0 in all cases presented here); similarity is required for cluster analysis, whereas difference is required for ordination analysis p = number of taxa N = number of samples xAi = abundance of species i in sample A