Examining the Predictive Utility of the BRIEF Teacher on Performance-Based Executive Function (EF) Tasks

• In Event: Poster Session 06
  In Poster Session: PS 06 Topic - Attention, Learning, Memory

Abstract

To create an efficient, cost-effective measurement of executive function (EF) in childhood, Gioia et al. (2000) developed a questionnaire, the Behavior Rating Inventory of Executive Function (BRIEF). Unfortunately, previous studies have found that among school-aged children, parent report on the BRIEF is only modestly predictive of EF (Toplak et. al., 2013). School settings provide greater opportunities to observe goal directed behaviors, and teachers have access to a larger and more representative sample of developmentally appropriate behaviors. The current study tested the predictive utility of teacher report on the BRIEF to predict EF among young children. If the BRIEF is an index of EF, then teacher report on the BRIEF would be expected to predict EF performance over and above teacher report of socioemotional functioning.

Children (N=170, 107 males) aged 5-7 (M=5.81, SD=0.34) and their teachers participated in the study. Teachers completed BRIEF, Conners-2, and Social Competence Scale (SCS) questionnaires. Executive functioning was indexed by performance on five tasks of verbal and visuospatial working memory.

The Conners, SCS, and BRIEF scales together accounted for significant variance in EF, F(16, 150) = 4.600, p < .001, R2 = .329. When the Conners/SCS were entered in Block 1, the Conners Cognitive Problems/Inattention scale and the SCS Prosocial Behavior scale both significantly predicted EF (all ps < .02). The BRIEF (entered in Block 2), did not predict EF capacity over and above the Conners/SCS scales, (all ps > .068). In the reverse, when the BRIEF was entered in Block 1, the Working Memory, Plan/Organize, and Organization of Materials scales were predictive of EF (all ps < .030). Both the Conners Cognitive Problems/Inattention scale and the SCS Prosocial Behavior scale continued to predict EF performance over and above the BRIEF.

Because many items were identical across scales, we wished to determine whether items unique to the BRIEF were capable of capturing variance in EF, independently of those shared with other rating forms. After identifying items that were identical vs. unique to the questionnaires, we performed a series of EFAs to create composite factors for the identical vs. unique items. Factors were selected by examining the scree plot of each EFA. Identical items formed three component factors, items that were unique to the BRIEF formed a single factor, and items that were unique to the Conners/SCS formed three factors.

When all components were entered into the model, the only component that was uniquely predictive of EF was the factor comprised of items unique to the Conners/SCS, b = -3.78, t(128) = -3.584, p = .001. These items were related to academic achievement. Items unique to the BRIEF did not predict EF performance over and above items shared with or unique to the Conners/SCS.

Together, our analyses suggest that teacher report on the BRIEF is not predictive of objective indices of EF in early elementary school students. Further, the modest associations between ratings on the BRIEF and EF appear due to the BRIEF’s overlapping content with measures of psychopathology.

Authors

• Christina Olena Hlutkowsky, The Pennsylvania State University
  Presenting Author

• Cynthia Huang-Pollock, Pennsylvania State University
  Non-Presenting Author