rstudy: Reliability Study

In talbano/epmr: Tools for Educational and Psychological Measurement

Reliability Study

Description

Functions for estimating internal consistency reliability, adjusting reliability, and estimating the standard error of measurement.

Usage

rstudy(x, n, sigma = FALSE, use = "everything", se = FALSE)

coef_alpha(x, n = NULL, sigma = FALSE, use = "everything", se = FALSE)

coef_omega(x, sigma = FALSE, use = "everything")

sb_r(r, k)

sem(r, sd)

Arguments

x	a matrix of scored item responses, or a variance/covariance matrix for a set of scored item responses when sigma = TRUE.
n	number of participants, defaulting to number of rows in x when sigma = FALSE.
sigma	logical with default FALSE indicating whether or not x is a variance/covariance matrix.
use	string for specifying how to handle incomplete cases
se	logical with default FALSE for finding analytical standard errors and confidence intervals for alpha. Requires that x be a matrix of scored item responses, or n be provided.
r	the reliability.
k	factor by which test length changes, used to predict a corresponding change in reliability.
sd	the standard deviation of total observed scores.

Value

Returns a data frame of estimated reliability coefficients, currently including alpha and omega, with corresponding standard errors of measurement. Reliability coefficients can also be obtained individually with coef_alpha and coef_omega. sb_r returns a new reliability coefficient predicted using the Spearman-Brown formula. sem returns the standard error of measurement.

Examples

ritems <- c("r414q02", "r414q11", "r414q06", "r414q09", "r452q03",
  "r452q04", "r452q06", "r452q07", "r458q01", "r458q07", "r458q04")
rsitems <- paste0(ritems, "s")
rstudy(PISA09[PISA09$cnt == "JPN", rsitems], use = "complete", se = TRUE)

talbano/epmr documentation built on May 1, 2024, 11:10 a.m.