Dimensionality in confirmatory factor analysis is not in the eye of the beholder: Ancillary bifactor statistical indices illuminate dimensionality and reliability.

Abstract

This tutorial delves into dimensionality assessment within the context of psychological measurement instruments, particularly focusing on bifactor models. It underscores the imperative to move beyond traditional fit indices when evaluating factor structures while highlighting the significance of ancillary bifactor indices such as explained common variance, OmegaH and percentage of uncontaminated correlations in gaining a more comprehensive understanding of the interplay between general and specific group factors. The tutorial offers a step-by-step guide to leveraging the power of R software for confirmatory factor analysis and the acquisition of ancillary bifactor indices. Through practical case studies, it elucidates the potential pitfalls of exclusively relying on fit indices and advocates for a balanced, multifaceted approach to dimensionality assessment. By integrating fit measures and ancillary indices, researchers can draw more informed and nuanced conclusions about measurement instrument dimensionality, ultimately enhancing the precision of psychological assessment.