This study reported test-retest reliability for a newly developed executive function battery designed for use in early childhood. A total of 140 predominantly low-income children (M = 48.1 months; 51% male; 43% African American) completed up to six tasks on two occasions an average of 18 (Mdn = 16) days apart. Pearson correlations between individual task scores indicated moderate retest reliability (mean r = .60; range = .52-.66) similar to that observed in other retest studies of executive function in preschool, school-aged, and adult samples. In contrast, confirmatory factor analyses of performance on the task battery across time indicated high retest reliability (ϕ = .95) that was identical to that observed in a recent study that used an identical method involving a sample of older adults. The short-term test-retest reliability of executive function in early childhood is comparable to that observed in childhood and adult samples. The retest reliability of children's performance on batteries of executive function tasks is appreciably stronger than the retest reliability of their performance on individual tasks. Studies that focus on inter- and intraindividual differences in executive function would be better served by using scores that are derived from task batteries than those derived from individual tasks.