Validation of executive function tasks with young children

Abstract

A multitrait‐multimethod paradigm was used to assess the stability, discriminant validity, and convergent validity of two commonly used tasks of executive functioning in a sample of 96 normal children in preschool and kindergarten. Memory served as the discriminant construct, and each construct was measured verbally and nonverbally. Results of a matrix of cross‐validated, Pearson product‐moment correlation coefficients did not fully support the construct validity of the tasks. Instead, the pattern of derived coefficients suggested that variance due to traits and to measurement methods were roughly equivalent. Research and clinical implications are delineated.