weighted_mean_r: Weighted mean correlation

In kim: A Toolkit for Behavioral Scientists

Weighted mean correlation

Description

Calculate the weighted mean correlation coefficient for a given correlations and sample sizes. This function uses the Hedges-Olkin Method with random effects. See Field (2001) \Sexpr[results=rd]{tools:::Rd_expr_doi("10.1037/1082-989X.6.2.161")}

Usage

weighted_mean_r(r = NULL, n = NULL, ci = 0.95, sigfigs = 3, silent = FALSE)

Arguments

r	a (vector of) correlation coefficient(s)
n	a (vector of) sample size(s)
ci	width of the confidence interval. Input can be any value less than 1 and greater than or equal to 0. By default, ci = 0.95. If ci = TRUE, the default value of 0.95 will be used. If ci = FALSE, no confidence interval will be estimated.
sigfigs	number of significant digits to round to (default = 3)
silent	logical. If silent = FALSE, a message regarding the weighted mean correlation and its p-value and CI will be printed. If silent = TRUE, this message will be suppressed. By default, silent = FALSE.

Value

the output will be a list of vector of correlation coefficient(s).

Examples

weighted_mean_r(r = c(0.2, 0.4), n = c(100, 100))
weighted_mean_r(r = c(0.2, 0.4), n = c(100, 20000))
# example consistent with using MedCalc
weighted_mean_r(
r = c(0.51, 0.48, 0.3, 0.21, 0.6, 0.46, 0.22, 0.25),
n = c(131, 129, 155, 121, 111, 119, 112, 145))

kim documentation built on Oct. 9, 2023, 5:08 p.m.