Measuring executive function in early childhood: A focus on maximal reliability and the derivation of short forms.

Abstract

This study assesses the maximal reliability of a newly developed battery of executive function (EF) tasks for use in early childhood. It also demonstrates how changes in maximal reliability can inform the selection of different short forms of the battery, depending on child age. Participants included children from the Family Life Project (Vernon-Feagans, Cox, & Family Life Project Investigators, in press)--a prospective longitudinal study (N = 1,292) of families that were recruited at the time of the birth of a new child--who were assessed at ages 3, 4, and 5 years. Results indicate that the EF battery had reasonably good maximal reliability (ρ = .73, 95% confidence interval [CI] = .69, .76) in a mixed-age sample of children randomly selected from assessments at ages 3, 4, and 5. In contrast, maximal reliability of the battery ranged from poor to modest for within-age samples (ρs = .47 [95% CI = .37, .52], .62 [.57, .66], and .61 [.55, .66] at ages 3, 4, and 5, respectively). Although the derivation of a 3-task short form of the battery always resulted in statistically significant decrements in maximal reliability, in some cases the relative decrement was quite modest and may be tolerable given the time savings and potential reduction in participant burden. The benefits of using maximal reliability to both evaluate task batteries and derive short forms are discussed, as well as how a focus on maximal reliability informs ongoing questions about the measurement and conceptualization of EF in early childhood.