The case for virtual tutoring: Translating research into sound policy

To achieve Mastery of grade-level content, the frequency or dosage of tutoring is significant. Tuition that is received at least three times each week, often referred to as “high-dosage” tutoring is more effective than tutoring that is delivered less frequently.5
Tutoring at scale: The relative impact of virtual tutoring
In a widely-cited meta-analysis from 2011, Dr. Kurt VanLehn analyzed the impact of virtual tutoring or “intelligent tutoring systems” as compared with human tutoring.

One initial finding that emerged from the study was that conventional beliefs about the improvements in student learning in a human tutoring setting depended significantly on the presence of highly-trained expert tutors. A second key finding, based on a comprehensive evaluation of peer-reviewed research, was that well-designed virtual tutors performed nearly as effectively as human tutors.

Since this pioneering study a decade ago, additional research emerged with mixed results regarding the efficacy of virtual tutoring. This triggered the need for a renewed scientific analysis of the benefits of virtual tutoring.

In 2015 researchers at the University of Michigan and the Institute for Defense Analysis, which conducts advanced research for the US government, conducted a meta-analysis reviewing 50 controlled studies of the benefits of virtual tutoring.6 The evidence they found was as follows:

Evidence on the Efficacy of Virtual Tutoring
Students receiving virtual tutoring outperformed control group students in 92% of studies.

In more than ¾ of the studies the impact of virtual tutoring were large enough to be considered of substantive importance by the standards of the What Works Clearinghouse maintained by the US Department of Education and the Institute of Education Sciences.

The average effect size equated to moving up a full quartile (50th to 75th percentile) in terms of performance which is a moderate to strong effect size.

Virtual tutoring solutions have become more effective- the average impact on student learning is nearly twice as large as what was found in the earliest versions of computer-aided instruction.

Although 78% of the studies were conducted in the US, the results were highly robust, having been replicated across 9 countries on 4 different continents.

On average, virtual tutoring yielded greater improvements in student learning than the typical effects of human tutoring.

A key variable determining the effectiveness of virtual tutoring was the fidelity of the implementation indicating an important role for training associated with the introduction of virtual tutoring.
Pulling this all together, it is important to note that the research consensus holds that well-designed virtual tutoring yields equivalent learning outcomes to human tutoring, and on-average even outperforms human tutoring because of its reliability, as it is less dependent on sourcing highly-trained expert tutors.

More recent research has shown that virtual tutoring is likely to be particularly effective for the most vulnerable or at-risk students such as students in poverty or students from historically marginalized racial or ethnic groups.

Finally, virtual tutoring can be delivered at 10% of the average cost of human tutoring.

As school systems in the US and across the globe embark on efforts to mitigate the impacts of Covid-19 additional data will emerge that will help inform our perspective on the efficacy of different interventions. Methods for delivering accelerated learning, including both human and virtual tutoring, will continue to represent a very active field for research and evaluation.

We should speak with a clear voice now though- there is a scientific consensus that virtual tutoring works to drive accelerated learning, at a sustainable price, and should be a critical element in the toolkit for education systems worldwide.

1 Bloom, B.S. (1984). “The 2 Sigma problem: The search for methods of group instruction as effective as one-to-one tutoring.” Educational Researcher 13, 4-16.

2 VanLehn, K. (2011, Oct 17). “The relative effectiveness of human tutoring, intelligent tutoring systems, and other tutoring systems.” Educational Psychologist.

3 Elbaum, B., Vaughn, S., Huges, M.T. & Moody, S.W. (2000). “How effective are one-to-one tutoring programs in reading for elementary students at risk for reading failure? A meta-analysis of the intervention research.” Journal of Educational Psychology,92(4), 605-619.

4 Lee, J.Y. (2013). “Private tutoring and its impact on students’ academic achievement, formal schooling, and educational inequality in Korea.” Unpublished doctoral thesis. Columbia University.

5 Reisner, E.R., Petry, C.A. & Armitage, M. (1990). “A review of programs involving college students as tutors or mentors in Grades K-12.” Washington, D.C.: U.S. Department of Education.

6 Kulik, James & Fletcher, J. D.. (2015). Effectiveness of Intelligent Tutoring Systems: A Meta-Analytic Review. Review of Educational Research. 86.