EdWorkingPapers
Assessing Permanent School Closures: A Conceptual Framework
Amid widespread declining enrollment, the expiration of COVID-19 ESSER funding, and looming uncertainty in federal P-12 education involvement, many school districts may soon consider permanent school closures. While extant permanent school closure literature provides a starting point for future analyses, it often fails to advise the breadth of contexts in which future closures may occur,… more →
Effects of a non-traditional teacher preparation program on non-test outcomes: evidence from relay graduate school of education in New York City
This study examines the effects of a non-traditional teacher preparation program, the Relay Graduate School of Education, on non-test outcomes for New York City public school students in Grades 3–8. By controlling for student and school fixed effects, I use plausibly random variation in Relay teacher assignments within students over time to identify causal Relay program effects. Results… more →
The Influence of Partisanship in Local School Board Elections: Evidence from Exit Polling in Michigan & Rhode Island
Education in the U.S. has long been shaped by local school boards elected in nonpartisan contests, a structure intended to shield schools from broader political forces. Today, many states are considering reforms to make school board elections partisan, yet the impact on voters remains unclear. Using exit poll data from 839 voters in Michigan (nonpartisan elections) and Rhode Island (partisan… more →
Online Tutoring, School Performance, and School-to-Work Transitions: Evidence from a Randomized Controlled Trial
Tutoring programs for low-performing students, delivered in-person or online, effectively enhance school performance, yet their medium- and longer-term impacts on labor market outcomes remain less understood. To address this gap, we conduct a randomized controlled trial with 839 secondary school students in Germany to examine the effects of an online tutoring program for low-performing… more →
Values, Visions, and Variation in American School Districts: A Computational Mixed Methods Analysis of School District Strategic Plans
The decentralization of power is a defining feature of the American education system, allowing schools to reflect community values and needs. Yet, little is known about how values and visions for education hold constant or vary across districts. Through an analysis of 617 district strategic plans, combining qualitative coding and computational topic modeling, we provide insight into how local… more →
Does State-Mandated Third-Grade Reading Retention Policy Improve Achievement? Evidence from a Staggered-Adoption Difference-in-Differences Design
This paper investigates whether the state-mandated third-grade reading retention policy autonomously enhances student achievement or depends on broader literacy reforms. Using district-level data from the Stanford Education Data Archive (2010–2019), I employ a staggered-adoption Difference-in-Differences design, as per Callaway and Sant’Anna (2021), to assess heterogeneous treatment effects… more →
Understanding the decision (not) to become a teacher: evidence from survey experiments with undergraduates in the UK and US
Teacher shortages are widespread, yet the reasons people choose (not) to enter the profession remain poorly understood. We conducted two survey experiments in which thousands of undergraduates chose between pairs of hypothetical jobs. This allowed us to evaluate the effects of differences in pay, working patterns and other job attributes on job choices, as well as explore how personality type… more →
Marginal Returns to Public Universities
This paper studies the returns to enrolling in American public universities by comparing the long-term outcomes of barely admitted versus barely rejected applicants. I use administrative admission records spanning all 35 public universities in Texas, which collectively enroll 10 percent of all American public university students, to systematically identify and employ decentralized cutoffs in… more →
Variations in Pre-Primary Education Infrastructure Within and Across Administrative Sectors in Rwanda
This study examines disparities in structural quality across Rwanda’s pre-primary modalities—centre-based, community-based, and home-based—operating under a single policy framework. Using data from 4,875 settings across 91 administrative sectors in seven districts, we applied multilevel models to separate within-sector differences by modality from between-sector variation, associated with… more →
Ready for What? School and District Responses to State College and Career Readiness Accountability in Tennessee
Tennessee’s K-12 accountability system incorporates three distinct measures of college and career readiness (CCR) for state and federal accountability. Each of these indicators applies its own set of metrics and performance benchmarks, but they all consistently draw upon similar components including participation in Early Postsecondary Opportunities (EPSOs), standardized tests like the ACT and… more →
Selling Student Success: A Critical Analysis of Predictive Analytics Vendors in Higher Education
As predictive analytics become increasingly embedded in higher education, commercial vendors offering these tools play a growing role in shaping institutional decision making, particularly through identifying students deemed “at risk.” In this qualitative study, we analyzed 161 publicly available materials from 15 vendors to examine these companies’ marketing of predictive analytics. Drawing… more →
Fast Track to Success? A Mixed Methods Evaluation of Condensed Course Formats at Tennessee Community Colleges
As colleges face increasing pressure to improve student outcomes, one solution gaining traction is the adoption of condensed courses (i.e., shortened academic terms). We employ quasi-experimental methods to estimate the effect of enrolling in a condensed course on course- and student-level outcomes at all public community colleges in Tennessee. We also leverage interviews with college faculty… more →
The Early Intervention and Early Childhood Special Education Workforce: Descriptive Evidence on Demographics and Turnover from Oregon
Early intervention (EI) and early childhood special education (ECSE) services for children with disabilities have expanded substantially across the U.S. over the past few decades, necessitating efforts to recruit and retain a qualified workforce to meet their needs. Despite widespread reports of staffing challenges in this sector, few contemporary studies provide large-scale evidence on this… more →
Beyond the Classroom: Impact of a High-Dosage Tutoring Program on Student Literacy Achievement
This study examines the impact of a high-dosage tutoring program, characterized by low tutor-to-student-ratio, on the literacy achievement of students in grades two through five in a midsized suburban school district in the southeastern United States. Using a student-level randomized controlled trial, 333 students were randomly assigned to either receive tutoring during the intervention period… more →
Removing Barriers to College Credits: Where and for Whom AP Exam Fee Waivers Work
Do policies that broaden educational access also foster success? We study this question in the context of North Carolina’s universal Advanced Placement (AP) exam fee waiver policy. Using student-course level administrative data, we exploit within-student variation on a sample of students who took multiple AP courses to estimate the policy’s effect on exam participation (access) and pass rates… more →
Learning to Work Towards Goals: A Sequential Evaluation of the Effect of Goal-Setting Course on Academic and Soft Skills
This study sequentially evaluates a soft-skills course implemented in Ugandan and Kenyan primary schools that replaced academic review time with lessons on goal-setting and related skills as students prepared for high-stakes primary school-leaving exams. An exploratory evaluation in Uganda provided evidence of positive impacts on girls' test scores. A confirmatory evaluation in Kenya found… 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) analyses revealed no overall achievement impacts for either program. We then explore two mechanisms… more →
Do Test Scores Misrepresent Test Results? An Item-by-Item Analysis
Much of the data collected in education is effectively thrown away. Students answer individual test questions, but administrators and researchers only see aggregate performance. All the item-level data are lost. Ex ante it is not clear this destroys much useful information, since the aggregate might be a sufficient statistic. Using data from Texas for 5 million students and 1.31 billion… more →
The Architecture of Expected Wage Gaps: Between- and Within-School Sources of Career Education Inequality
This study investigates how school-level variation contributes to social stratification even before labor market entry by examining Career and Technical Education (CTE) as a key mechanism for sorting students into pathways with unequal economic returns. Using Delaware administrative data and Bureau of Labor Statistics occupational wage data, we introduce “expected wage” as a measure to… 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% since the peak of the Great Recession. Second, over one-third of this apparent decline is an artifact… more →
Tutor CoPilot: A Human-AI Approach for Scaling Real-Time Expertise
Generative AI, particularly Large Language Models (LLMs), can expand access to expert guidance in domains like education, where such support is often limited. We introduce Tutor CoPilot, a Human-AI system that models expert thinking to assist tutors in real time. In a randomized controlled trial involving more than 700 tutors and 1,000 students from underserved communities, students with… 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 time consuming for use in large-scale monitoring or national-level surveys. Short forms of direct… more →
The Effects of Immigration Enforcement on Student Outcomes in a New Era of Immigration Policy in the United States
This study presents the first evidence, to our knowledge, of the effects of the surge in interior immigration apprehensions in 2025 in the United States on student academic performance using detailed student-level administrative records from Florida. We find evidence that immigration enforcement reduced test scores for both U.S.-born and foreign-born Spanish-speaking students while also… more →
Do As I Say: What Teachers’ Language Reveals About Classroom Management Practices
Classroom management critically affects students’ academic and behavioral outcomes, yet we lack quantitative methods for observing these practices at scale. This study develops and validates language-based measures of classroom management—such as responding to student behavior and issuing verbal or material sanctions—using natural language processing (NLP) on 1,652 elementary mathematics… more →
How do place-based scholarships affect student borrowing and academic outcomes? Lessons from Atlanta
Previous research shows that Achieve Atlanta’s placed-based scholarship and associated services meaningfully improve college persistence and completion. In this follow up study that uses similar methods but additional and more detailed data, we examine whether scholarship recipients exhibit different student loan portfolios, course-taking patterns, or academic performance. Using regression… 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 of LLM-based methods with traditional machine learning (ML) methods across different assessment… more →
COVID-19-Induced School Closures and Disadvantaged Children’s Post-COVID Academic Growth: A Longitudinal Cohort Study
This study draws on unique, repeated-measures data on a diverse (51% female; 53% Latine, 22% Black, 11% White), low-income cohort of children (N = 680) whose academic skills were assessed before and after COVID-19-induced school closures. Longitudinal models predicted changes in children’s literacy and math trajectories from before school closures (ages 4-6; 2017-2019) to after school… more →
Education Governance and Race: An Analysis of School Board Discourse Using Large Language Models
Despite growing attention to school boards, it is unclear whether they primarily operate as bureaucratic forums, policy-making bodies, or arenas for contentious debate—particularly on issues of race. Recent controversies suggest increasing public engagement and conflict, but little evidence documents how often questions of race arise in board deliberations. This study analyzes over 40,000… more →
Creating Coherence: Does Instructional Alignment Affect the Impact of Tutoring?
This study examines the impact of using instructionally aligned literacy tutoring with students in kindergarten through third grade under a Response to Intervention framework. We conducted a randomized controlled trial to evaluate the impact on literacy assessment scores for 296 students in four schools in a large suburban school district in the southeastern United States. Students in the… more →
Cheaper (and more effective) by the dozen: Evidence from 12 randomized A/B tests optimizing tutoring for scale
Over the course of 12 rapid randomized experiments, we optimize an educational tutoring program. Tutoring is one of the most effective educational approaches yet has remained difficult to scale due to high costs. We adaptively test and improve a technology-enabled tutoring program to enhance cost-effectiveness and scalability. Results show that seven of twelve tests led to efficiency… more →
The Long-Term Effects of Rank in Elementary School
We estimate the long-term consequences of math and reading rank within an elementary school on short and long-term outcomes. We find that higher rank leads to better outcomes. Students ranked at the top in grade 7 perform up to 0.33 standard deviations higher on future school exams, are more likely to graduate high school and university, and earn significantly more at age 28. Math rank is… more →
Schools Never Die: Toward a Dynamic Systems Theory of School Closure
Educational researchers and policymakers typically treat school closures as discrete administrative decisions with clear endpoints. This paper challenges that assumption by applying Dynamic Systems Theory to school closure policy and research. We argue that schools function as adaptive ecosystems embedded within broader networks of relations that span social, cultural, political, and economic… more →
Creating Classes: Elementary school classroom assignments and their implications for student access to high-quality teaching
We investigate the distribution of students across classrooms in North Carolina elementary schools. While tracking is ubiquitous and well-documented in secondary education, limited evidence exists regarding cross-classroom clustering in elementary schools and its consequences. Consistent with qualitative evidence suggesting that educators seek to create demographically balanced classrooms, we… more →
Policy Impacts of Reimbursement Rate Reform: Evidence from the Child Care and Development Fund
The Child Care and Development Fund (CCDF) subsidizes child care costs for families with low-incomes. Reimbursements for cost-subsidized care are paid to child care providers but are extremely low compared with market rates and actual cost of care. We examine how the 2014 congressional reauthorization of CCDF, which recommended states increase subsidy reimbursement rates to the 75th percentile… more →
Influence of Within-Class Age Differences on Adolescents’ Eating Behaviors
This study examines within-class age differences as a novel determinant of adolescents’ dietary behaviors, isolating it from confounders such as absolute age, season of birth, and country-specific school entry rules. Using a multi-country dataset of over 600,000 European students, we find that younger students within a class exhibit poorer dietary habits. Since confounders are controlled for,… more →
Does Expanding Access to High Quality Technical Education Induce Participation and Improve Outcomes?
Over the last 15 years, Career and Technical Education (CTE) has been changing as schools have aimed to better meet workforce needs and diversify pathways into higher education and the workforce. This study provides the first known causal evidence on the impact of CTE program expansion in U.S. comprehensive high schools on student participation and postsecondary outcomes. Using administrative… more →
Higher Education as Regional Development: Labor Market Impacts of Nigeria’s 2011 Federal University Expansion
This paper examines the causal impact of higher education expansion on regional labor markets and human capital development. Exploiting the 2011 establishment of nine federal universities across previously underserved Nigerian states, we implement a difference-in-differences approach to analyze effects on employment, wages, job quality, and sectoral composition. Our results show significant… more →
Sibling Spillovers and Free Schooling
We use administrative data to measure sibling spillovers on academic performance before and after the introduction of Free Secondary Education (FSE) in Tanzania. Prior to FSE, students whose older siblings narrowly passed the secondary school entrance exam were less likely to go to secondary school themselves; with FSE, the effect became positive. A triple-differences analysis, using… 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 by practitioners in a middle-income country. Both instruments can achieve relatively high… more →
Shock Absorption: Did School Turnaround Shelter Schools from the Pandemic’s Effects on Teacher Turnover?
Successful turnaround interventions should build school capacity to promote not just school improvement but also resilience to exogenous shocks that undermine schooling. While a large literature demonstrates that turnaround can improve school outcomes, little is known about whether it can help schools withstand negative shocks. We examine the impact of Michigan’s turnaround policy on teacher… more →
The Impact of Cellphone Bans in Schools on Student Outcomes: Evidence from Florida
Cellphone bans in schools have become a popular policy in recent years in the United States, yet very little is known about their effects on student outcomes. In this study, we try to fill this gap by examining the causal effects of bans on student test scores, suspensions, and absences using detailed student-level data from Florida. Several important findings emerge. First, we show that the… more →
What is the impact of changing schools on the academic outcomes of elementary and middle school students?
We study how different kinds of school changes shape achievement in grades 4–8 using data from six California districts (2016–17 through 2019–20). We estimate the effects of structural (promotional), nonstructural summer, and midyear moves and find that structural and summer moves have near-zero effects on ELA and math, while midyear moves lower ELA by 0.073 SD and math by 0.083 SD. Midyear… more →
A Framework for Building High-Quality Education Data for R&D in the Age of AI: The EDSI Dataset and Expert Insights
The Gates Foundation, the Walton Family Foundation, and the Chan Zuckerberg Initiative have launched a series of collaborative investments in building large-scale datasets that can support and accelerate data infrastructure for AI R&D efforts in education. In partnership with researchers from Harvard University and Stanford University, the Center for Educational Data Science and Innovation… more →
Understanding Disruptions: Causes of and Variation in Lost Instructional Time
Virtual instruction has boomed after the COVID-19 pandemic, including the use of virtual environments within in-person schools. But, research has provided little evidence about student experiences on these virtual platforms, nor how to improve the use of these platforms. Through natural language processing techniques, this study examines over 26,000 virtual tutoring sessions that took place… more →
Causal Mechanisms of Relative Age Effects on Adolescent Risky Behaviours
Age differences between classmates are attracting growing attention in academic research and public policy, yet their underlying mechanisms remain understudied. We examine how relative age affects adolescents’ risky behaviors across Europe. Using Health Behaviour in School-Aged Children (HBSC) survey data and a two-stage least squares (2SLS) strategy, we provide causal estimates that isolate… more →
The Power of Personalized Attention: Comparing Pedagogical Approaches in Small Group and One-on-One Early Literacy Tutoring
Tutoring has played a significant role in pandemic-related learning recovery, supporting student learning and engagement. A recent randomized controlled trial estimated that one-on-one virtual early literacy tutoring was nearly twice as effective as two-on-one tutoring for improving student learning (Robinson et al., 2024). To better understand this gap, we analyze transcripts from 16,629… more →
Behind the Scenes: Faculty-Staff Collaboration in a Student Success Effort
Interventions to improve postsecondary student success often involve supports that are external to the college classroom, although there is growing evidence that faculty involvement in interventions can improve student outcomes. This paper explores the challenges that arise when faculty and staff collaborate to improve student success as well as the organizational changes that support the… more →
Can Peer Group Design Improve Engagement in Online STEM Courses? The Role of Motivation to Lead
Peer interaction is important for student engagement and success in higher education and becomes even more critical in online STEM education, where limited interaction can undermine motivation and belonging—key factors for success in rigorous STEM coursework. Despite the widespread use of peer group activities to foster collaboration, there is limited understanding of how to effectively form… more →
Off to a Great Start: The Potential for Tutoring Paired with the Off2Class Foundational Literacy Curriculum to Boost English Proficiency Gains for Adolescent Newcomer English Learners
Adolescent English learners with low literacy strive to learn a new language with minimal or no reading skills. Their efforts are often complicated by having special learning needs or limited experiences with formal education. Meanwhile, they need English literacy in middle and high school, where teachers expect students to read to learn rather than learn to read. Evidence on effective… more →
Contemporary Child Labor and Declining School Attendance in the U.S.
The United States has experienced a 400% increase in reported child labor violations over the past decade, coinciding with declines in K-12 school attendance and enrollment. We examine the causal relationships between these patterns with microdata from the American Community Survey (ACS) from 2005 to 2023. Using a shift-share instrumental variable approach, our findings show that increased… more →