This function conducts a phylogenetic paired t-test, roughly following Lindenfors et al. (2010; J. Evol. Biol.). This is not a phylogenetic ANOVA, in which we want to compare the means of different sets of species on the tree. Instead, we are interested in the difference between two characters, or two measures of a character within a species, and we want to know if this difference is significantly different from zero controlling for the phylogenetic non-independence of species.
phyl.pairedttest(tree, x1, x2=NULL, se1=NULL, se2=NULL, lambda=1.0, h0=0.0,
fixed=FALSE, ...)a phylogeny as an object of class "phylo".
data vector for first trait, or matrix with two traits in columns.
data vector for second trait (or null if x1 is a matrix).
standard errors for x1.
standard errors for x2.
starting value for Pagel's (or fixed value, if fixed=TRUE).
null hypothesis (to be tested) for the mean difference between x1 and x2.
logical value specifying whether or not to optimize .
optional arguments.
An object of class "phyl.pairedttest" with the following components:
phylogenetic mean difference.
standard error of dbar.
estimated evolutionary variance (of the difference).
fitted (or fixed) value of .
log-likelihood of the fitted model.
t-value ((dbar-h0)/se where se is computed from the Hessian).
P-value.
the degrees of freedom.
the null hypothesis that was tested.
Likelihood optimization is performed using optim with method="L-BFGS-B" with box constraints on (0,1).
Lindenfors, P., L. J. Revell, and C. L. Nunn (2010) Sexual dimorphism in primate aerobic capacity: A phylogenetic test. J. Evol. Biol., 23, 1183-1194.
Revell, L. J. (2012) phytools: An R package for phylogenetic comparative biology (and other things). Methods Ecol. Evol., 3, 217-223.