Abstract
The Barkley Deficits in Executive Functioning Scale – Short Form (BDEFS-SF; Barkley, Citation2011) was developed to assess deficits in five facets of executive functioning. Theoretical assumptions surrounding the BDEFS-SF presume that executive dysfunction is an overarching construct that consists of five domain-specific factors (i.e., a hierarchical model; Barkley, Citation2011). Prior research has supported a correlated five-factor model, but the tenability of hierarchical or bifactor models of the BDEFS-SF have not yet been tested. In the present study (N = 1,120 community adults), confirmatory factor analysis was used to compare four theoretically relevant models of the BDEFS-SF (i.e., one-factor, correlated five-factor, hierarchical, and bifactor models). The bifactor model provided the best fit to the data. However, the general factor accounted for the overwhelming majority of variance in BDEFS-SF scores and none of the domain-specific factors exhibited adequate construct replicability or factor determinancy. Further, the general factor accounted for the overhelming majority of variance in criterion variables (i.e., executive attention and health anxiety); the Organization and Emotion factors accounted for a small amount of unique variance in executive attention and the Emotion factor accounted for a small amount of unique variance in health anxiety. Taken together, study findings suggest that the BDEFS-SF has a strong general factor and independent use of the domain-specific factors is contraindicated.
Notes
1 Based on reviewer feedback, the above analyses were conducted a second time with age included in each model as a covariate. The pattern of results, with the age in each model as a covariate, is consistent with what is presented in the main text. More specifically, the bifactor model provided the best fit to the data, and the lower order factors accounted for little variance in the model and demonstrated poor construct replicability and factor determinancy. The general factor accounted for the majority of variance in both structural regressions.