From b56172f9a98dd9ff5c6ddaf61b27fde688951269 Mon Sep 17 00:00:00 2001 From: "Isaac T. Petersen" Date: Thu, 9 Jan 2025 07:18:59 -0600 Subject: [PATCH] 20250109 - composite reliability --- 05-Reliability.Rmd | 1 + 15-Factor-Analysis-PCA.Rmd | 1 + 2 files changed, 2 insertions(+) diff --git a/05-Reliability.Rmd b/05-Reliability.Rmd index abe6a9dd..bb701772 100644 --- a/05-Reliability.Rmd +++ b/05-Reliability.Rmd @@ -1598,6 +1598,7 @@ However, because of the many weaknesses of Cronbach's alpha [@Cortina1993; @Dunn Current recommendations are to use McDonald's coefficient omega ($\omega$) instead of Cronbach's alpha ($\alpha$) for internal consistency reliability.\index{reliability!internal consistency}\index{reliability!internal consistency!Cronbach's alpha}\index{reliability!internal consistency!omega} Specifically, it is recommended to use omega total ($\omega_t$) for continuous items that are unidimensional, omega hierarchical ($\omega_h$) for continuous items that are multidimensional, and omega categorical ($\omega_C$) for categorical items (i.e., items with fewer than five ordinal categories), with confidence intervals calculated from a bias-corrected bootstrap [@Flora2020; @Kelley2016].\index{reliability!internal consistency}\index{reliability!internal consistency!omega} +In addition to estimating coefficient omega ($\omega$) in the ways described below, omega ($\omega$) can also be estimated in latent factor models, and is sometimes called *composite reliability* (see Section \@ref(reliability-cfa) for an example).\index{reliability!internal consistency}\index{reliability!internal consistency!omega} ##### Omega Total ($\omega_t$) diff --git a/15-Factor-Analysis-PCA.Rmd b/15-Factor-Analysis-PCA.Rmd index ab917a60..b48fa59e 100644 --- a/15-Factor-Analysis-PCA.Rmd +++ b/15-Factor-Analysis-PCA.Rmd @@ -2164,6 +2164,7 @@ cor.test( #### Internal Consistency Reliability {#reliability-cfa} [Internal consistency reliability](#internalConsistency-reliability) of items composing the latent factors, as quantified by [omega ($\omega$)](#coefficientOmega) and [average variance extracted](#averageVarianceExtracted) (AVE), was estimated using the `semTools` package [@R-semTools].\index{factor analysis!confirmatory}\index{reliability!internal consistency}\index{reliability!internal consistency!omega}\index{reliability!internal consistency!average variance extracted} +[Coefficient omega ($\omega$)](#coefficientOmega) is an example of *composite reliability*.\index{factor analysis!confirmatory}\index{reliability!internal consistency}\index{reliability!internal consistency!omega} ```{r} compRelSEM(cfaModelFit)