raref_test: Rarefaction testing with rarefactions within permutations

In bbolker/betararef: Beta Diversity with Rarefaction

Description

Rarefaction testing with rarefactions within permutations

Usage

raref_test(comm, ttt, n_raref = 50, nperm = 200, method = "jaccard",
  binary = TRUE, ptype = c("raref_perm_medianp", "raref_perm_meanF",
  "raref_perm_meanp"), return.all = FALSE, dummy_raref = FALSE,
  seed = NULL)

Arguments

comm	community matrix (patches=rows, species=columns)
ttt	treatment vector (treatment id of each patch)
n_raref	number of rarefactions
nperm	number of permutations
method	beta diversity metric (passed to vegan::betadisper)
binary	reduce data to presence/absence? (passed to vegan::betadisper)
ptype	p-value calculations: "raref_perm_meanF", calculate mean F-statistics within permutations; "raref_perm_meanp", calculate mean p-value across permutations; "raref_perm_medianp" (default), calculate median p-value across permutations.
return.all	(logical) return all information rather than summarized test statistics?
dummy_raref	(logical) [TESTING ONLY] fake rarefaction? (i.e., don't actually do any rarefaction)
seed	random-number seed

Value

A list with class "htest" containing the following components:

method	a string giving the method and number of rarefactions used
estimate	the mean beta diversity (measured as median distance from centroid) across all sites and rarefactions
conf.int	the confidence interval of the mean beta diversities across rarefactions
statistic	the median F-statistic across permutations
p.value	the p-value corresponding to the statistic

Examples

set.seed(101)
sx <- simComm(n.site=20,p.mix=0.5,n.indiv.site=c(10,20),totsp=c(20,40))
plot(sx)
raref_test(sx[,-1],sx[,1])
tmpf <- function(...) {
   sx <- simComm(...)
   tt <- try(raref_test(sx[,-1],sx[,1])$p.value)
   if (is(tt,"try-error")) return(NA) else return(tt)
}
## Not run: 
## compute power for this case
set.seed(101)
system.time(pvec <- replicate(100,tmpf(n.site=20,p.mix=0.5,
                       n.indiv.site=c(10,20),totsp=c(20,40))))
## ~ 10 minutes
hist(pvec,breaks=20,xlim=c(0,1),col="gray")
mean(pvec<0.05)  ## 0; conservative (but very small sample!)

## End(Not run)

bbolker/betararef documentation built on Sept. 17, 2021, 8:49 a.m.