g-Disattenuation: Using Omega Coefficients to Estimate Effect Sizes for the Underlying General Factors

Resumen

Researchers are increasingly replacing ot supplementing Cronbach’s alpha with model-based coefficients of the omega family, including omega-total (the proportion of observed score variance explained by all the common factors of a test) and omega-general/omega-hierarchical (the proportion of observed score variance solely explained by the general factor). Under the appropriate assumptions, values of omega-general/omega-hierarchical can be used in the disattenuation formula to estimate effect sizes for the underlying general factor(s), net of the contribution of secondary lower-order/group factors, without the need for complex modeling or access to the raw data. In this paper I present the rationale for g-disattenuation and discuss the assumptions that have to be met to calculate correlations, standardized mean differences, and multivariate differences at the level of general factors. I illustrate these procedures with a worked-out empirical example from a large personality dataset.

Áreas y palabras clave

Cómo citar (APA 7)

Del Giudice, M. (2024). g-Disattenuation: Using Omega Coefficients to Estimate Effect Sizes for the Underlying General Factors. https://doi.org/10.31234/osf.io/zutrx