cor.test(x, y, alternative="two.sided", method="pearson")
x and
y must have the same length greater than 2.
Missing values (
NAs) and
+-Infs
are allowed but ignored at calculation.
x and
y.
"two.sided" (non-zero),
"greater" (greater than 0), or
"less" (less than 0).
"pearson",
"kendall", or
"spearman", depending on
what coefficient of correlation should be used in the test statistic. Only
the first character is necessary.
class "htest", containing the following components:
names attribute.
statistic containing
its degrees of freedom when
this is a t-distribution.
x and
y
is zero.
names attribute
which is either
"tau" for
Kendall's statistic,
"cor"
for
Pearson's,
or
"rho" for
Spearman's.
x and
y,
coef=0.
"two.sided",
"greater", or
"less"
.
"Pearson\(aas product-moment correlation",
"Kendall\(aas rank correlation tau",
or
"Spearman\(aas rank correlation".
x and
y vectors.
x
and
y are mutually uncorrelated.
When
method="pearson" the data are assumed to come from a bivariate Normal
distribution.
If this is not the case, the other two methods offer
nonparametric alternatives.
If
method="pearson" the (usual)
Pearson's product moment correlation
coefficient (
r <- cor(x,y)) is computed, and divided by its standard error
to produce a t-statistic with n-2 degrees of freedom,
where n =
length(x) =
length(y).
This statistic is given by
t <- (sqrt(n-2)*r) / sqrt(1-r2).
If
method="kendall" then
Kendall's tau using ranks is computed.
If
method="spearman" the rank correlation coefficient is used.
Conover, W. J. (1980). Practical Nonparametric Statistics. 2nd. ed. New York: Wiley.
murder <- state.x77[,"Murder"] illit <- state.x77[,"Illiteracy"] cor.test(murder,illit,method="k") # Transformations can be used cor.test(log(x),log(y),alt="gr")