As excitement grows around tutoring as a strategy to combat learning loss, advocates have rightly been encouraged by the growing body of evidence demonstrating the efficacy of tutoring interventions. To date, however, little research has examined the impact of fully virtual tutoring on very young students. Hardly a technicality, this distinction matters because younger children are less likely to have the technical and self-regulation skills upon which virtual learning depends. Now, a new study by researchers from Stanford, Vanderbilt, and UnboundED analyzes the benefits of virtual tutoring specifically for early elementary students.
The authors conducted a randomized controlled trial with 2,085 K–2 students at twelve Texas schools within the same charter network. Those in the sample were randomly assigned to participate either in 1:1 tutoring, 2:1 tutoring, or a control group; the tutoring provider, OnYourMark Education (OYM), is a partner of the unnamed charter network and a science-of-reading-based virtual tutoring program. Students in the 1:1 and 2:1 groups participated in in-school, virtual tutoring for twenty-minute periods, four days a week, from September 2022 until May 2023. For their main measure, the researchers compared students’ beginning-of-the-year performance on a widely used exam, Dynamic Indicators of Basic Literacy Skills (DIBELS), to their end-of-the-year performance on the same assessment. The analysis controlled for demographic factors like gender, race and ethnicity, and economic disadvantage, and the authors also broke down their findings to understand OYM’s effects on students with differing baseline performance and in different grade levels.
Overall, the results show that OYM produced statistically significant reading gains for participants. On average, the students who received the OYM treatment improved their scores by 0.05–0.08 standard deviations. Gains were slightly larger for those in the 1:1 group, a finding in line with other research on 1:1 tutoring.
The effect sizes varied by subgroup. Perhaps most notably, students with the poorest baseline scores saw the largest gains (0.11 standard deviations), and this trend was especially true for the lowest-skill students in the 1:1 group (0.15 standard deviations). Given that this finding was statistically significant, 1:1 virtual tutoring could be a worthwhile intervention for young readers struggling the most. In the results disaggregated by grade level, first graders saw the greatest reading growth, followed by kindergarteners, followed by second graders. This result probably tells us more about the tutoring program and study alignment than it does about student reading skills: OYM focuses on foundational reading skills, whereas by second grade, most assessments include a greater focus on comprehension skills.
Readers will want to interpret all these findings cautiously. First, the student sample was not entirely random (although the assignment of the groups within the sample was random). Prior to the creation of the sample, staff in the twelve schools each selected ten students who most needed tutoring support; these students then participated in OYM that year but were excluded from the study. This limitation suggests that the study findings may actually have been conservative, as the lowest-skill students tended to gain the most from tutoring, and the students hand-selected by staff for guaranteed participation were likely among the lowest-skilled.
A more serious threat to the study’s internal validity surrounds the participation of students with disabilities and multilingual learners. After students were assigned to the three study groups, numerous students from these groups withdrew from the OYM intervention due to a scheduling conflict with their federally mandated services. As a result, there was moderately high attrition, and the patterns of attrition were not random. To account for this issue, the researchers ran additional calculations for a “preferred sample,” which excluded all 731 multilingual learners and students with disabilities—a large proportion of the study’s sample of 2,085. Still, both samples are somewhat problematic, as the “full” sample suffers from disproportionate attrition, and the smaller sample cannot speak to effects on students with disabilities and multilingual learners. This is especially unfortunate, as the study’s implications would be particularly important for these subgroups, which experience substantial and enduring achievement gaps compared to their peers.
Yet the findings remain encouraging, suggesting that many young readers can benefit from virtual tutoring, a more affordable and often more logistically feasible intervention than in-person tutoring.
SOURCE: Carly D. Robinson, Cynthia Pollard, Sarah Novicoff, Sara White, and Susanna Loeb, “The Effects of Virtual Tutoring on Young Readers: Results from a Randomized Controlled Trial,” Annenberg Institute at Brown University (May 2024).
