Can technology enable the expansion of small-scale teaching techniques? Improving student learning in the United States involves implementing effective teaching methods on a large scale. Some charter schools have significantly improved test scores through practices such as high-dosage tutoring (HDT). However, because of potential costs and effectiveness questions, it is still being determined whether these successful practices can be replicated widely.
In "Can technology facilitate scale? Evidence from a randomized evaluation of high dosage tutoring" (National Bureau of Economic Research, Working Paper 32510, May 2024), authors Monica P. Bhatt, Jonathan Guryan, Salman A. Khan, Michael LaForest-Tucker, and Bhavya Mishra explore the potential for educational technology, particularly high-quality computer-assisted learning (CAL) platforms, to effectively address the expandability challenge of implementing successful teaching methods.
The authors’ work focuses on incorporating technology into HDT. HDT is a teaching strategy that involves tutoring in small groups during the school day with a structured curriculum. The teaching method, that has been encouraged by the U.S. Secretary of Education to combat pandemic-induced learning loss, has shown substantial improvements in student learning by maximizing task-focused time and addressing academic mismatches. Although HDT is practical and cost-effective at the small scale, scaling it up has been hindered by its cost, particularly in the public sector. Because there are many more students in public schools, hiring enough tutors to keep the group size small enough can be burdensome
The authors examine whether high-dosage tutoring could be combined with a CAL platform so that the same number of tutors could reach more students. Large-scale randomized control trials of HDT indicate substantial gains in math learning for high school students through 2-to-1 tutoring for an hour daily at approximately $3,500 per student per year. However, the authors find that it may be possible to decrease the cost. Following an experiment in which the students alternated time between a human tutor and a CAL platform, the authors find similar learning increases. However, because tutors are not needed for all of the students’ study hours, the learning gains came at only two-thirds of the cost. The authors also find that a fixed number of tutors would be able to tutor 1.5 times as many students with the assistance of a CAL platform.
Saga Education, a nonprofit organization referenced by the authors, suggests that small-group tutoring and classroom teaching are fundamentally different tasks, with classroom management and individualized instruction being the most challenging aspects of teaching. Furthermore, the organization believes that with a significant decrease in class size, people with minimal prior educational training can become effective tutors more quickly, allowing for the adoption of new staffing models. Despite the promise of innovative staffing models, postpandemic cost and labor shortages remain significant barriers to scaling up traditional HDT for all students who could benefit from it.
After the COVID-19 pandemic, the U.S. Secretary of Education encouraged districts to use federal funding for high-dosage tutoring to address pandemic-related learning loss. Bhatt, Guryan, Khan, LaForest-Tucker, and Mishra show that incorporating technology into tutoring has the potential to lower costs while maintaining learning effectiveness, but future research should explore whether additional variations in the structure of high-dosage tutoring could decrease costs while preserving the significant learning advantages. However, the findings of this study suggest the potential of integrating CAL into high-dosage tutoring to help expand this promising intervention to more students.
United States Department of Labor
