Skip to content

Datasets

  • Meta-analysis of Educational RCTs with Follow-up (MERF)

    I created the Meta-analysis of Educational RCTs with Follow-up (MERF) dataset in collaboration with Tyler Watts, Drew Bailey, and a large team of research assistants. We created the dataset to investigate critical questions about intervention fadeout and persistence, namely: Is fadeout a rarity or a regularity? How much do intervention impacts fade? Is fadeout less of an issue for certain outcomes (e.g., social-emotional vs. cognitive skills), participant characteristics (e.g., younger vs. older children), or intervention features (e.g., shorter vs. longer programs)? MERF is comprised of interventions targeting a diverse array of child outcomes across development (e.g., nurse-home visiting, preschool, elementary-level curriculum, adolescent substance use prevention). We screened studies from eight published meta-analyses for inclusion in the sample. To be included, studies had to use randomized control trial design with treatment impacts measured on cognitive and/or social-emotional skills at least six months after intervention end. Details on the dataset are provided in Hart et al. (2024). The data is available for public use on LDbase.

    View on LDbase
  • MERF-EmergeForthcoming

    As a follow-up to MERF, we created the MERF-Emerge dataset to examine the implications of fadeout for the long-run adult impacts of interventions. Specifically, we created the dataset to examine: 1) whether fadeout precludes the emergence of adult effects, 2) whether the magnitude of initial intervention impacts on child skills forecast long-run effects on adult outcomes, and 3) to identify the general pattern of longitudinal impacts observed in the literature. To create MERF-Emerge, we conducted a comprehensive search of the literature to identify all educational interventions that measured initial impacts on child skills and follow-up impacts on adult outcomes. We identified 29 interventions. Details on the dataset are provided in Hart et al. (2025). The dataset will eventually be posted on LDbase for public use. In the meantime, feel free to email me if you are interested in using it and I will be happy to share it with you.