miscor (version 0.1-0)

comptest.cor: Comparision of product-moment correlation coefficients

Description

This function statistically compares product-moment correlation coefficients in independent and dependent samples.

Usage

comptest.cor(x = NULL, y = NULL, z = NULL, group = NULL, r.xy = NULL, r.xz = NULL, r.yz = NULL, n = NULL, r.1 = NULL, r.2 = NULL, n.1 = NULL, n.2 = NULL, alternative = c("two.sided", "less", "greater"), conf.level = 0.95, digits = 3, output = TRUE)

Arguments

x

a numeric vector.

y

a numeric vector.

z

a numeric vector.

group

a numeric vector indiating the group membership.

r.xy

alternative specification, product-moment correlation coefficient between x and y.

r.xz

alternative specification, product-moment correlation coefficient between x and z.

r.yz

alternative specification, product-moment correlation coefficient between y and z.

n

alternative specification, number of observations.

r.1

alternative specification, product-moment correlation coefficient in group 1.

r.2

alternative specification, product-moment correlation coefficient in group 2.

n.1

alternative specification, number of observations in group 1.

n.2

alternative specification, number of observations in group 2.

alternative

a character string describing the alternative hypothesis, must be one of "two.sided" (default), "greater" or "less".

conf.level

confidence level of the interval.

digits

integer indicating the number of decimal places to be displayed.

output

logical: if TRUE, output is shown.

Value

Returns an object of class comptest.cor with following entries:

call

function call

dat

data.frame with x, y and z (if available)

spec

specification of function arguments

res

list with results depending on the analysis (independent of dependent samples), i.e., z (test statistic), pval (significance value), r.xy, r.xz, r.yz, r.1, r.2, diff (difference), n, n.1, n.1, lower (lower limit of CI), upper (upper limit of CI)

Details

In dependent samples, the function tests the two-sided null hypothesis H0: $\rho .xy = \rho .xz$ or the one-sided null hypothesis H0: $\rho .xy >= \rho .xz$ or $\rho .xy <= \rho="" .xz$.="" function="" parameters="" are="" specified="" using="" either="" (x, y, z) or (r.xy, r.xz, r.yz, n). In independent samples, the function tests the two-sided null hypothesis H0: $\rho.1 = \rho.2$ or the one-sided null hypothesis H0: $\rho.1 >= \rho.2$ or $\rho.1 <= \rho.2$.="" function="" parameters="" are="" specified="" using="" either="" (x, y, group) or (r.1, r.2, n.1, n.2).

References

Rasch, D., Kubinger, K. D., & Yanagida, T. (2011). Statistics in psychology - Using R and SPSS. New York: John Wiley & Sons.

Zou, G. Y. (2007). Toward using confidence intervals to compare correlation. Psychological Methods, 12, 399-413.

See Also

test.cor, seqtest.cor

Examples

# Dependent samples: Generate random data
x <- c(3, 2, 2, 3, 7, 8, 5, 9)
y <- c(2, 4, 1, 5, 7, 3, 6, 7)
z <- c(1, 4, 3, 3, 1, 4, 2, 5)

#--------------------------------------
# Dependent samples
# H0: rho.xy == rho.xz,  H1: rho.xy != rho.xz

comptest.cor(x, y, z)

#--------------------------------------
# Dependent samples
# H0: rho.xy <= rho.xz,  H1: rho.xy > rho.xz
# r.xy = 0.44, r.xz = 0.21. r.yz = 0.20, n = 120

comptest.cor(r.xy = 0.44, r.xz = 0.21, r.yz = 0.20, n = 120,
             alternative = "greater")

###

# Independent samples: Generate random data
dat <- data.frame(group = rep(1:2, each = 200),
                  rbind(sim.cor(200, rho = 0.3),
                        sim.cor(200, rho = 0.5)))

#--------------------------------------
# Independent samples
# H0: rho.1 == rho.2, H1: rho.1 != rho.2

comptest.cor(x = dat$x, y = dat$y, group = dat$group)

#--------------------------------------
# Independent samples
# H0: rho.1 >= rho.2, H1: rho.1 ! < rho.2
# Group 1: r = 0.32, n = 108
# Group 2: r = 0.56, n = 113

comptest.cor(r.1 = 0.32, n.1 = 108, r.2 = 0.56, n.2 = 113,
             alternative = "less")