A Randomized Controlled Trial of BookNook Tutoring Program SY24-25

• In Event: The Impact of High-Dosage Tutoring on Student Achievement Outcomes? Evidence from Four Randomized Controlled Trials

Abstract

Tutoring has long been promoted as an effective strategy for increasing student academic achievement (Dobbie & Fryer, 2013; Nickow et al., 2024; Neitzel et al., 2021). More recently, tutoring has received renewed attention as an evidence-based intervention to address the learning loss resulting from the COVID-19 pandemic. Post-pandemic progress in reducing gaps in achievement has been slow; recent reports show that students are performing at a lower level than pre-pandemic in nearly all grade levels in math and reading (Dewey et al., 2025; Lewis & Kuhfeld, 2024; NCES, 2025). As a result, many states, including Arkansas, are encouraging schools and districts to implement tutoring intervention programs to increase student learning and offset learning losses from the pandemic. This study seeks to contribute evidence on the impact and cost-effectiveness of a virtual high-impact tutoring program, BookNook, on elementary students’ proficiency in literacy and English Language Arts through a randomized controlled trial in two school districts. The research questions we seek to answer are:

1. 
   

   What is the impact of BookNook tutoring on student achievement in ELA? 

   
   


2. 
   

   What are the effects of BookNook tutoring on student achievement in ELA with a 1:2 tutor-to-student ratio and a 1:3 tutor-to-student ratio? 

   
   


3. 
   

   What is the cost-effectiveness of tutoring? 

   
   

To assess the intervention’s impact, we randomly assigned 941 Kindergarten through 8th graders identified as eligible for tutoring by their school district. In one district, we randomly assigned students to either the control condition (n = 273), tutoring with a tutor-to-student ratio of 1:2 (n = 272), or tutoring with a tutor-to-student ratio of 1:3 (n = 272). Due to the smaller sample in the second district, we assigned students to only the control condition (n = 62) or to tutoring with a tutor-to-student ratio of 1:2 (n = 61).  Randomizations were blocked on school and grade. The intervention consists of three 30-minute early literacy or ELA tutoring sessions per week over 12 weeks.  

For the impact analysis, we will leverage student-level administrative data on enrollment, demographics, state standardized test scores, and attendance. We will conduct the intent-to-treat analysis using a linear model that regresses standardized state ELA test scores on treatment assignment, controlling for baseline test scores, demographic covariates, and block fixed effects. Through exploratory analyses we will examine the heterogeneity of intervention effects for different subgroups and tutoring’s impact on other outcomes such as ELA proficiency, math achievement, and student attendance. Additionally, we will explore implementation aspects such as participation, dosage, and their variation across grades and student characteristics, as well as the annual per-student cost of the intervention. Overall, this study aims to provide insights into the efficacy and cost-effectiveness of a vendor-based online high-impact tutoring intervention in improving student ELA outcomes.

Author

• Lisa Sanbonmatsu, Harvard University
  Presenting Author