Executive Function as a Predictor of School Success: Validity in the Context of Homelessness

• In Event: Promoting Positive Adaptation Among Homeless and Highly Mobile Students

Abstract

This study examined the usability and validity of executive function (EF) measures for predicting early school success among children experiencing homelessness, highlighting the potential of the NIH Toolbox tasks for screening in an urban district with high rates of homelessness and high mobility (HHM). Administrative data has shown the high risk of students identified as HHM for achievement problems: Students in a HHM group had significantly and chronically lower achievement than other low-income children, with large gaps between their average achievement and normative levels of growth (Cutuli et al., 2013). Gaps persisted for reading and widened for math between 3rd and 8th grades. Studies of children residing in emergency shelters have shown that EF can be reliably assessed, with scores forecasting early school success, both academically and socially, even with IQ controlled (Masten et al., 2012; Obradović, 2010). Moreover, evidence suggests that EF skills are malleable in early childhood and may be an important target for boosting school readiness (Zelazo, Blair, Willoughby, 2017). Consequently, interest grows in the potential of EF measures for early childhood screening, particularly in school districts with many children living in poverty and homelessness.
This presentation will focus on a subsample of children attending routine screening whose parents gave permission for participation (91% agreed) and who spoke English (computerized EF tasks were only available in English at the time). Analyses included 461 children: 54% female; 36 to 72 months of age (M = 54.61, SD = 9.53). The study was designed to evaluate the usability and value of EF assessment in the context of routine screening. The sample was racially/ethnically diverse (56% minority). Children were screened at community locations, including emergency housing (10%). EF was assessed using early versions of the NIH Toolbox measures that include developmental extensions (Dext) developed to extend the lower range of valid scores on the Flanker and Dimensional Change Card Sort (DCCS), plus a table-top Peg Tapping task. The three tasks showed high convergent and concurrent validity and predicted achievement in kindergarten (Table 4). Current status as homeless was related to lower scores on all three EF tasks: Peg-Tapping: t(58.16) = 4.29, p < .001; DCCS-Dext: t(428) = 2.95, p = .003; Flanker-Dext: t(62.30) = 4.72, p < .001 (Figure 2). Structural equation modeling (SEM) results indicate that EF predicts achievement at the beginning of kindergarten and also growth over the year (Figure 3), although not outperforming the general screening test used by this district. EF does contribute uniquely to the prediction of teacher-rated kindergarten behavior over and above all three of the screening measures used by this district (Figure 4). Results show promise of EF tasks for use in screening: tasks are brief and appealing to most children, usable in different settings (even shelters), have predictive validity, and may offer an early indication of risk for behavioral problems. Results corroborate growing evidence that homelessness is a risk marker for EF difficulties and school problems, suggesting that targeting EF skills may be an important strategy for assessing and boosting school readiness.

Authors

• Amanda Kalstabakken, University of Minnesota - Twin Cities

• Christopher David Desjardins, University of Minnesota

• Cynthia K. Hillyer, Minneapolis Public Schools

• Maureen J. Seiwert, Minneapolis Public Schools

• Ann S. Masten, Institute of Child Development