var_error_alpha: Analytic estimate of the sampling variance of coefficient...

In psychmeta: Psychometric Meta-Analysis Toolkit

Analytic estimate of the sampling variance of coefficient \alpha

Description

Estimates the error variance of Cronbach's coefficient \alpha.

Usage

var_error_alpha(item_mat = NULL, alpha = NULL, k_items = NULL, n_cases)

Arguments

item_mat	Item correlation/covariance matrix. If item_mat is not supplied, the user must supply both alpha and k_items. If item_mat is NULL, the program will assume that all item intercorrelations are equal.
alpha	Vector of population \alpha values. Must be supplied if item_mat is NULL.
k_items	Vector of numbers of items to be simulated. Must be supplied if item_mat is NULL.
n_cases	Vector of sample sizes to simulate in sampling distribution of alpha.

Value

Vector of sampling variances of the supplied \alpha values.

References

Duhachek, A., & Iacobucci, D. (2004). Alpha’s standard error (ASE): An accurate and precise confidence interval estimate. *Journal of Applied Psychology, 89*(5), 792–808. \Sexpr[results=rd]{tools:::Rd_expr_doi("10.1037/0021-9010.89.5.792")}

Examples

item_mat <- matrix(.3, 5, 5)
diag(item_mat) <- 1
alpha <- mean(item_mat[lower.tri(item_mat)]) / mean(item_mat)
k_items <- nrow(item_mat)

var_error_alpha(item_mat = item_mat, n_cases = 50)
var_error_alpha(alpha = alpha, k_items = k_items, n_cases = 50)
var_error_alpha(alpha = c(alpha, alpha), k_items = c(k_items, k_items), n_cases = 50)

psychmeta documentation built on June 22, 2024, 6:52 p.m.