Contributions of modern measurement theory to measuring executive function in early childhood: An empirical demonstration

Abstract

This study demonstrates the merits of evaluating a newly developed battery of executive function tasks, designed for use in early childhood, from the perspective of item response theory (IRT). The battery was included in the 48-month assessment of the Family Life Project, a prospective longitudinal study of 1292 children oversampled from low-income and African American families. IRT models were applied to a select set of tasks to demonstrate empirically (a) a principled method for item evaluation, including the utility of item characteristic curves; (b) how to explicitly test whether the measurement properties of executive function tasks are invariant across mutually exclusive subgroups of youths; (c) how the precision of measurement of a given task can vary according to underlying child ability; and (d) the utility of using IRT-based versus percentage correct scores. Results are discussed with respect to the importance of developing psychometrically sound and scalable instruments that facilitate the measurement of interindividual differences in intraindividual change of executive function across the early childhood period.