paleoTS (version 0.3-1)

test.var.het: Variance heterogeneity test

Description

Tests for heterogeneity of sample variances in a paleoTS object.

Usage

test.var.het(y, method = "Bartlett")
pool.var(y, nn=NULL, ret.paleoTS=FALSE)

Arguments

y

a paleoTS object

method

test to be used; currently only Bartlett's test is implemented

nn

if y is a vector of variances, nn is a vector of sample sizes

ret.paleoTS

logical, indicating if the function should return a new paleoTS object with variances replaced by pooled variance.

Value

• Function pool.var either returns a paeloTS object, or the pooled variance. Function test.var.het returns a list with the following elements relevant to Bartlett's test
• stat
  test statistic for Bartlett's test
• p.value
  P-vlaue for statistical test
• df
  degrees of freedom, equal to one fewer than the number of samples

Details

Tests for variance heterogeneity among samples. In the absence of evidence for heterogeniety, it may be desirable to replace individual estimates of phenotypic variance with a variance estimate pooled over all samples. Function pool.var is used internally in test.var.het. It returns the value of the pooled variance.

References

Hunt, G. 2006. Fitting and comparing models of phyletic evolution: random walks and beyond. Paleobiology32:578--601. Sokal, R. and F. J. Rohlf 1995. Biometry, 3rd Ed.

Examples

# generate simulated data with no real variance heterogeneity
y <- sim.GRW(20, 0, 1)
bart <- test.var.het(y)   # nonsignificant 95% of the time!
ys<- pool.var(y, ret.paleoTS=TRUE)
cat (ys$vv)		# note all variances now the same