Obtain composite reliability from a fitted model — bmasem_composite_reliability • bayesianmasem

Get composite reliability

Usage

bmasem_composite_reliability(object, interval = 0.9, return_draws = FALSE)

Arguments

object: (bmasem) An object of bmasem-class returned by bmasem.
interval: (real in (0, 1)) Credible interval to return.
return_draws: (LOGICAL) If TRUE, returns the realised distribution of reliability. If FALSE (default), return the distribution summary.

Value

If return_draws = FALSE, returns the summary of posterior distribution of composite reliability for each factor.
If return_draws = TRUE, returns draws from the realised distribution of composite reliability for each factor.

Details

This metric is the ratio of common variance (CV) to total variance (TV) for each factor.

Common Variance: For factor \(f\), the common variance is: $$ \mathrm{CV}_f = (\sum_{i \in I_f} \lambda_{if})^2 \phi_{ff}, $$ where \(I_f\) indexes the indicators of factor \(f\).
Total Variance: For factor \(f\), the total variance is: $$ \mathrm{TV}_f = \sum_{i \in I_f} \sum_{j \in I_f} \omega_{ij}, $$ where \(\Omega = = (\omega_{ij})\) is the model-implied covariance matrix of the indicators.

Examples

if (FALSE) { # \dontrun{
model_syntax <- paste0(
  "distress =~ ", paste0("x", 1:14, collapse = " + "), "\n",
  "anxiety =~ ", paste0("x", seq(1, 14, 2), collapse = " + "), "\n",
  "depression =~ ", paste0("x", seq(2, 14, 2), collapse = " + ")
)
fit <- bmasem(
  model_syntax,
  sample_cov = Norton13$data, sample_nobs = Norton13$n, orthogonal = TRUE
)
bmasem_composite_reliability(fit)
} # }