Using Meta-Analytic Data to Examine Fadeout and Persistence of Intervention Impacts on Constrained and Unconstrained Skills

Recent reviews of the educational intervention literature have noted patterns of intervention impact fadeout on cognitive skills, whereby skill trajectories between children in the intervention and control group converge in the years following the end of the intervention. Some early childhood education (ECE) researchers have suggested that skill type, specifically whether a skill is “constrained” or “unconstrained” may explain variation in fadeout trajectories. The Constrained Skills View proposes that unconstrained skills, which are thought to develop across the life course, may show more persistent impacts than constrained skills, which are eventually mastered by all. For a broad, short-term test of this theory, we used the Meta-Analysis of Educational RCTs with Follow-up (MERF) to examine trajectories of fadeout and persistence by skill type across a variety of educational interventions tested in childhood and adolescence. The majority of impacts in our sample (91%) were on measures of reading and language skills. We modeled patterns of intervention impact persistence and fadeout six to twelve months after the interventions ended. After coding outcomes as “constrained” or “unconstrained,” we found no evidence that impacts on unconstrained skills persisted more than impacts on constrained skills. Rather, in some model specifications, impacts on constrained skills showed slightly more short-term persistence than impacts on unconstrained skills.