This study tested the longitudinal measurement invariance and developmental changes of a newly developed battery of executive function (EF) tasks for use in early childhood. The battery was administered in the Family Life Project--a prospective longitudinal study (N = 1,292) of families who were oversampled from low-income and African American families at the birth of a new child--at assessments conducted when the child was 3, 4, and 5 years old. All 6 individual EF tasks exhibited strong measurement invariance over time. The EF battery, which was derived from the 6 individual tasks, exhibited partial strong invariance over time. Second-order latent growth curve models revealed individual differences in the levels but not rates of change in latent EF ability. The functional form of change was nonlinear; 60% of the total change in EF ability that was observed between the 3- and 5-year assessments occurred between the Year 3 and Year 4 assessments. Results are discussed with respect to the importance of establishing scalable measures of EF ability prior to investigating experiences that predict or are predicted by changes in EF during early childhood. (Contains 3 tables and 3 figures.)