Harvard University
Facilitating Evidence-Based Instructional Coaching With Automated Feedback on Teacher Discourse
Instructional coaching often aims to ground teacher professional learning in classroom evidence, yet it is labor-intensive for coaches to obtain and curate such evidence. This study explores possibilities for applying an automated feedback tool to classroom recordings in order to support… more →
Why aren’t online learning programs working better in schools: Maybe they need to be more integrated with what is actually taught during whole classroom instruction?
Computer-assisted learning (CAL)—educational software providing instruction to help students learn concepts and skills—is ubiquitous in US schools. Despite meta-analyses suggesting benefits, national test scores have declined during rapid CAL adoption. This may reflect insufficient research on… more →
Measures Still Matter: Replicating and Extending a Meta-Analysis of the Effects of Educational Apps on Children's Reading and Math Skills with Six Years of New Data
This study replicates and extends a meta-analysis of 36 studies of educational apps (Kim et al., 2021) with six years of new data. We synthesize 841 effect sizes from 141 studies evaluating the effectiveness of educational apps on reading and math skills of students ages 3-9. Replicating the… more →
Unveiling Racism: A Systematic Review of Survey Measures of Racism in Education
Education policy research aimed at eliminating racism necessitates methodological innovation that fosters both racial equity–centered approaches and robust empirical analysis of the systemic and structural nature of racism. Most quantitative research in educational psychology does not account… more →
Mapping the Mechanisms of Interdisciplinary Learning Transfer from Reading to Math Achievement: Evidence from a Large-Scale Randomized Controlled Trial
Far transfer—the application of learning across distant domains—remains elusive in intervention research, and even when it is found, its mechanisms remain unclear or unexplored. This study analyzes data from the Model of Reading Engagement (MORE), a sustained content literacy intervention… more →
An Applied Researchers’ Guide to Estimating Effects from Blocked Cluster Randomized Trials: Estimands, Estimators, and Estimates
Blocked cluster randomized trials (CRTs) are widely used to evaluate educational interventions. In such trials, researchers face critical choices about how to define and estimate the average treatment effect (ATE). These choices—both the estimand (e.g., person- vs. cluster-weighted) and the… more →
Why Fadeout is (Probably) Worse Than We Think: Adjusting for Correlated Sampling Error in Meta-Analyses of Behavioral Interventions
The extent to which intervention effects persist or fade over time is an important question in the behavioral sciences. In meta-analysis, persistence is often assessed by meta-regressing effect sizes at followup on effect sizes at endline. While common, the standard meta-regression does not… more →
Mechanisms of Effect Size Differences Between Researcher Developed and Independently Developed Outcomes: An Item-Level Meta-Analysis
Differences in effect sizes between researcher developed (RD) and independently developed (ID) outcome measures are widely documented but poorly understood in education research. We conduct a meta-analysis using item-level outcome data to test potential mechanisms that explain differences in… more →
Success Begets Success: The Dynamic Treatment Effects of Financial Aid Tournaments
Financial aid programs in higher education vary widely in design, including how aid is structured and the timing of provision. This paper studies the impact of financial aid provided as a repeated tournament and its dynamic treatment effects. Pooling administrative data that captures 32% of all… more →
Operational Funding and Early Educator Wage Growth: Evidence from Massachusetts
Early educators are among the lowest-paid workers in the United States, in part because most early care and education (ECE) programs operate within constrained business models. In Massachusetts, the Commonwealth Cares for Children (C3) program distributes noncompetitive grants to licensed… more →
Understanding High Schools’ Effects on Longer-Term Outcomes
Improving education and labor market outcomes for low-income students is critical for advancing socioeconomic mobility in the United States. We use longitudinal data on five cohorts of 9th grade students to explore how Massachusetts public high schools affect the longer-term outcomes of students… more →
Residential Segregation and Unequal Access to Local Public Services in India: Evidence from 1.5m Neighborhoods
We study residential segregation and access to public services across 1.5 million urban and rural neighborhoods in India. Muslim and Scheduled Caste segregation in India is high by global standards, and only slightly lower than Black-White segregation in the U.S. Within cities, public facilities… more →
Beyond the One-Teacher Model: Experimental Evidence on Using Embedded Paraprofessionals as Personalized Instructors
Using embedded paraprofessionals to provide personalized instruction is a promising model for differentiating instruction within the classroom. This study examines two randomized controlled trials of paraprofessional-led tutoring in early-grade math and literacy. However, intent-to-treat (ITT)… more →
Labor Market Strength and Declining Community College Enrollment
Declining U.S. college enrollments have triggered questions about the health of the postsecondary sector. Using institution-level data, we make four points. First, such declines are driven not by the four-year sector but by two-year community colleges, which have apparently shrunk by over 30%… more →
Creating Short Forms of Early Childhood Development Measures: A Framework for Quantifying Statistical, Conceptual, and Practical Tradeoffs in Direct Assessment
Direct assessments of early childhood development (ECD) are a cornerstone of research in developmental psychology and are increasingly used to evaluate programs and policies in lower- and middle-income countries. Despite strong psychometric properties, these assessments are too expensive and… more →
ChatGPT vs. Machine Learning: Assessing the Efficacy and Accuracy of Large Language Models for Automated Essay Scoring
Automated Essay Scoring (AES) is a critical tool in education that aims to enhance the efficiency and objectivity of educational assessments. Recent advancements in Large Language Models (LLMs), such as ChatGPT, have sparked interest in their potential for AES. However, comprehensive comparisons… more →
The reliability of classroom observations and student surveys in non-research settings: Evidence from Argentina
There is a growing consensus on the need to measure teaching effectiveness using multiple instruments. Yet, guidance on how to achieve reliable ratings derives largely from formal research in high-income countries. We study the reliability of classroom observations and student surveys conducted… more →
Investing in Human Capital During Wartime: Experimental Evidence from Ukraine
This paper provides insights into human capital investments during wartime by presenting evidence from three experiments of an online tutoring program for Ukrainian students amid Russia’s invasion of Ukraine. Conducted between early 2023 and mid-2024, the experiments reached nearly 10,000… more →
Comparing Machine Learning Methods for Estimating Heterogeneous Treatment Effects in Randomized Trials: A Comprehensive Simulation Study
This study compares 18 machine learning methods for estimating heterogeneous treatment effects in randomized controlled trials, using simulations calibrated to two large-scale educational experiments. We evaluate performance across continuous and binary outcomes with diverse and realistic… more →
The Sensitivity of Value-Added Estimates to Test Scoring Decisions
Value-Added Models (VAMs) are both common and controversial in education policy and accountability research. While the sensitivity of VAMs to model specification and covariate selection is well documented, the extent to which test scoring methods (e.g., mean scores vs. IRT-based scores) may… more →