Use the total score of the Barkley Deficits

Use the total score of the Barkley Deficits in Executive Functioning Scale – Short Form, not the subscale scores. Title: Subtitle Testing the Factor Structure of the Barkley Deficits in Executive Functioning Scale – Short Form Goodness of Fit Statistics and Model Comparisons Kate Clauss and Joseph R. Bardeen (kec 0084@auburn. edu) INTRO • The total and subscale scores of the BDEFS-SF are used in research and clinical practice • Use of a total score assumes an overarching general executive functioning (EF) construct and use of subscale scores assumes that they have incremental value beyond the general factor • These assumptions have not been tested • Bifactor modelling can be used to test both of these assumptions simultaneously METHOD • N = 1, 120 Mturk participants • Completed Demographics and BDEFS-SF DATA ANALYTIC PLAN • Confirmatory Factor Analysis (CFA) with weighted least squares estimation (WLSMV) • Compared One-Factor, Correlated Five-Factor, Hierarchical, and Bifactor models MOST of the variance in the model was accounted for by the general factor (ωH =. 94) PUC =. 84 and ECV =. 81, which calls into question a multidimensional conceptualization of the scale EF Deficits Standardized Factor Loadings for the Bifactor Model The subscales didn’t add very much unique variance (ωHS scores from. 08 to. 24) Note. All ps <. 001. Bifactor Evaluation Indices Time Org. Restraint Motiv. Emotion Poor replicability and determinacy scores caution against using subscale scores or modeling lower-order latent variables RESULTS • The bifactor model provided the best fit to the data • Additional indices question the use of a multidimensional solution, as most of the variance was attributable to the total score, the subscales accounted for little additional variance, and had poor factor determinacy and construct replicability DISCUSSION • Use the total score in research and clinical practice • Perhaps scale modification could result in adequate unidimensional fit What do the Bifactor Evaluation Indices Mean? • Omega. H (ωH): proportion of variance in the total score that can be attributed to the general factor • Omega. HS (ωHS): proportion of variance attributable to each domain-specific factor after removing the variance due to the general factor • Explained common variance (ECV): proportion of common variance accounted for by the general factor • Item-level explained common variance (I-ECV): amount of variance for each item attributable to the general factor (>. 80 -. 85 indicates unidimensionality) • Percentage of uncontaminated correlations (PUC): percentage of item correlations contaminated by variance attributed to the factors • PUC and ECV >. 7, indicates unidimensionality • Average relative parameter bias (ARPB): average bias across parameters if items are forced into a unidimensional structure (ARPB <. 10 -. 15 suggests that the multidimensionality isn’t substantial enough to preclude a unidimensional solution) • Factor Determinacy (FD): >. 9 suggests that factors are of practical value for use in measurement models • Construct Replicability (H): degree to which a factor is well defined by its indicators (>. 8 suggests it will replicate) Correspondence concerning this poster should be addressed to: Kate Clauss, M. A. , Auburn University, Dept. of Psychology, 226 Thach Hall, Auburn, AL 36849. Email: kec 0084@auburn. edu