Comparison between species is hindered by the fact that there are non-independence between observations, causing biased statistics, especially false significance, since closely related species often inflate the sample size when they actually should be treated as technical repeats and combined into just one or a few data points. Phylogeny-Weighted Analysis is a robust phylogenetic comparative method (PCM) that gives a weight between 0 and 1 to each species according to its share of evolutionary time in the phylogeny after correcting for evolutionary inertia (since only inertia traits that hardly change over time cause false statistics). This weight represents reduced effective data point numbers and can be robustly used to correct biased statistics. Conventional statistics, such as correlation coefficients, multiple linear regression beta coefficients (standardized), Student's t test and their corresponding p-values can be conveniently and correctly calculated based on such weight.
Be the first to post a review of Phylogeny-Weighted Analysis!