A Closed-Form Alternative for Estimating <i>ω</i> Reliability under Unidimensionality
Gregory R. Hancock, Ji An
Abstract
As an alternative to Cronbach’s α for estimating scale reliability, McDonald’s ω has attracted increased attention within the methodological community for its less stringent measurement assumptions. Notwithstanding, ω is still seldom used by practitioners, likely due to its unavailability in popular software packages (e.g., SPSS) and the need for confirmatory factor analysis (CFA) parameter estimates for its computation. This study proposes and evaluates an easily-calculated ω estimate that does not rely on factor analysis, which is shown to be highly comparable to CFA-based ω estimates and thus can be used as a more accessible computational alternative for practitioners, especially if programmed into popular software packages.