var_error_mult_R: Estimate the error variance of linear regression multiple...

In psychmeta: Psychometric Meta-Analysis Toolkit

Estimate the error variance of linear regression multiple R(-squared)

Description

This function estimates the error variance for linear regression model (squared) multiple correlations (R and R-squared).

Usage

var_error_mult_R(R, n, p)

var_error_mult_Rsq(Rsq, n, p)

var_error_R(R, n, p)

var_error_Rsq(Rsq, n, p)

Arguments

R	Vector of multiple correlation coefficients.
n	Vector of sample sizes.
p	Vector of numbers of predictors in the model.
Rsq	Vector of squared multiple correlation coefficients.

Details

The sampling variance of a multiple correlation is approximately:

var_e = (1 - R^2)^2 \* (n - p - 1)^2 / ((n^2 - 1) \* (n + 3))

The sampling variance of a squared multiple correlation is approximately:

var_e = 4 \* R^2 \* (1 - R^2)^2 \* (n - p - 1)^2 / ((n^2 - 1) \* (n + 3))

Value

A vector of sampling-error variances.

References

Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). Lawrence Erlbaum and Associates. doi: 10.4324/9780203774441. p. 88.

Olkin, I., & Finn, J. D. (1995). Correlations redux. Psychological Bulletin, 118(1), 155–164. doi: 10.1037/0033-2909.118.1.155

Examples

var_error_mult_R(R = .5, n = 30, p = 4)
var_error_mult_R(R = .5, n = 30, p = 4)
var_error_mult_Rsq(Rsq = .25, n = 30, p = 4)
var_error_mult_Rsq(Rsq = .25, n = 30, p = 4)

psychmeta documentation built on Aug. 26, 2022, 5:14 p.m.