Stanford University
The Impact of Tutor Gender Match on Girls’ STEM Interest, Engagement, and Performance
Gender disparities in STEM persist despite girls performing as well as boys academically, suggesting girls may benefit from role models who shape their perceptions of STEM. We examine whether female math tutors influence girls’ STEM interest, attendance, and performance. We randomly assigned 422… more →
The Language of Closure: Examining Racial Differences in How A Community Discusses School Closure Metrics
School closures in urban districts disproportionately affect marginalized communities, yet community input often goes unanalyzed or is reduced to simple frequency counts. This study applies BERTopic, a neural topic modeling approach, to analyze 4,159 suggestions from 2,006 community members… more →
IDEA-Aligned Estimates of Racial Disproportionality in Special Education versus Conventional Approaches: A cautionary note on included-variable bias when achievement and socioeconomic status proxy for special education need
Racial disproportionality in special education is a contested policy space. Federal oversight has traditionally focused on minority over-representation through IDEA’s significant disproportionality framework. However, observational studies report that Black students appear under-identified based… more →
Beyond the Shutdown: Tracking Language Growth in Early Head Start Children Before, During, and After COVID-19
Understanding early language outcomes for low-income children in the wake of the COVID-19 pandemic is an important concern for researchers, policymakers, and educators. We examined language environments and language development among infants and toddlers in Early Head Start from pre-COVID,… more →
Capturing Voter Turnout at the School District Level: Validating a Geospatial Strategy
School boards are critical sites of education policymaking, yet scholarship on these institutions is scarce because of severe data limitations. We introduce a geospatial strategy and open-source R package, called “Query, Overlay, Recover” (QOR), that generates high-quality estimates of voter… more →
School-Based Disability Identification Varies by Student Family Income
Currently, 18 percent of K-12 students in the United States receive additional supports through the identification of a disability. Socioeconomic status is viewed as central to understanding who gets identified as having a disability, yet limited large-scale evidence examines how disability… more →
COVID-19, School District Operations, and Student Academic Performance in Virginia
We use longitudinal student-level data and interrupted time series methods to examine the impact of the COVID-19 pandemic on mathematics achievement among 3rd-8th grade students in Virginia, a state that offered particularly low levels of access to in-person learning in the school reopening… more →
The West Texas Measles Outbreak and Student Absences
Declining child-vaccination rates are driving a measles resurgence in the US, yet little evidence documents how these outbreaks may disrupt schooling. Using daily absence data from a school district at the center of the West Texas outbreak, this preregistered analysis finds absences increased 41… 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… 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 →
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… 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… 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… 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… 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… more →
Improving Student-Teacher Relationships Through Feedback: The Development and Evaluation of the Stanford/Leading Educators Wise Feedback Professional Development Learning Series
High-quality academic feedback, especially feedback that highlights errors, mistakes, misunderstandings and shortcomings, is one of the most valuable tools teachers have for promoting student growth and learning. It is how teachers help students go beyond what they could accomplish on their own… 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 →
Item-Level Heterogeneity in Value Added Models: Implications for Reliability, Cross-Study Comparability, and Effect Sizes
Value added models (VAMs) attempt to estimate the causal effects of teachers and schools on student test scores. We apply Generalizability Theory to show how estimated VA effects depend upon the selection of test items. Standard VAMs estimate causal effects on the items that are included on the… more →
Out-of-School Learning: Subtitling vs. Dubbing and the Acquisition of Foreign-Language Skills
The development of English-language skills, a near necessity in today’s global economy, is heavily influenced by historical national decisions about whether to subtitle or dub TV content. While prior studies of language acquisition have focused on schools, we show the overwhelming influence of… more →
From Rural Schools to City Factories: Assessing the Quality of Chinese Rural Schools
The changing pattern of quality in China’s rural schools across time and province is extracted from the differential labor market earnings of rural migrant workers. Variations in rates of return to years of schooling across migrant workers working in the same urban labor market but having… more →