Classical and Bayesian Uncertainty Intervals for the Reliability of Multidimensional Scales

The reliability of a multidimensional test instrument is commonly estimated using coefficients ωt (total) and ωh (hierarchical). However, point estimates for the are rarely accompanied by uncertainty estimates. In this study, we compare bootstrap normal-theory confidence intervals. addition, develop Bayesian versions sampling from second-order factor model. Results comprehensive simulation study show that studied intervals performed well when sample size was sufficiently large (N≥500). across most conditions. When small low, only bias-corrected accelerated interval approached satisfactory coverage among all This guides on introduces credible easy to use come with benefits parameter estimation.