The good and bad of virtual on-demand tutoring

Tutoring is perhaps the best-studied and consistently effective educational intervention we have. So it should come as no surprise that policymakers and educational leaders across the country are investing money, time, and political capital into tutoring initiatives. The influx of interest in tutoring is, at least in part, to support the struggling and marginalized students whose learning dropped most during the COVID-19 pandemic.

However, educational programs can only benefit students if students access those resources, and tutoring is no different. Many school leaders are grappling with how to get students to participate in tutoring programs—especially for their students who need the most support. This can be particularly challenging for tutoring that happens outside of school hours.

The growth of “on-demand” tutoring programs

On-demand (virtual) tutoring programs have become a popular option among districts and states. Most of these programs are outsourced, with tutors available for one-on-one educational support that is not embedded in the school day. Although the quality of on-demand tutoring programs likely varies greatly, high-quality providers can provide students with well-trained tutors whenever they need academic support.

Perhaps most appealing, education leaders can make these programs available to many more students than would be possible with in-school tutoring. However, the inherent flexibility that on-demand tutoring affords means that educators have less control over whether and when students use the platform. For students to benefit from this on-demand tutoring, someone (either the student themselves or a parent or family member) needs to log on and request help from a tutor.

But, do they? As states and districts make opt-in, on-demand tutoring available to more and more students, we need to understand what types of students use these programs and whether these optional programs reach the students who could benefit the most from them.

New evidence on who uses on-demand tutoring

In a recent study along with Biraj Bisht, we report on the implementation of opt-in, on-demand tutoring in partnership with the Aspire Public Schools (a charter management organization, or CMO) in California. The CMO provided 7,000 middle and high school students with free, unlimited access to one-on-one chat-based tutoring during the spring 2021 semester. Students accessed the program from a mobile device and could request help from an available tutor in any core subject. The topic of each tutoring session was usually driven by student questions and the interaction between tutors and students were chat-based with help from a virtual whiteboard to facilitate joint work.

We found that take-up was low. Only 19% of students ever logged on and received help from a tutor. What’s more, struggling students were only half as likely to ever use the resource than their higher-achieving peers. Figure 1 shows that students who had failed at least one course in the previous fall—and, presumably, could use the most help—were much less likely to log on.

Of course, the question of who uses on-demand tutoring only matters if we think the resource will benefit students. We tested the effectiveness of the program and found positive effects for the students who used it. Accessing the opt-in tutoring increased the likelihood of students passing all of their courses. This is perhaps not surprising, as the tutors were well trained and easily available.

Increasing take-up by communicating with students and parents

So, if students benefit from virtual on-demand tutoring, can we get more of them to use it?

We conducted a randomized experiment testing the impact of sending a range of communications encouraging usage of the tutoring platform to students, to their parents, or to both students and their parents together. Students assigned to the Student Only and Student+Parent conditions received a mailer and a series of email communications that encouraged them to log on at any time to receive help from a tutor. Similarly, parents of students assigned to the Parent Only and Student+Parent conditions received a mailer and a series of text messages that encouraged them to remind their children to log on and ask a tutor for help.

The intervention increased take-up, and by a lot. Engaging both students and their parents simultaneously was far and away the most effective condition. This strategy resulted in students being 46 percent more likely to receive help from a tutor and it more than doubled the number of messages a student exchanged with a tutor. Moreover, the intervention was most beneficial for the students who needed it most. Figure 2 illustrates that struggling students (in dark blue) assigned to the Student+Parent condition were 14 percentage points more likely to log on and ask a tutor for help. This represents a 122 percent increase in usage compared to students who didn’t receive the additional communications.

Even more importantly, when we sent messages to both students and parents, students who had recently failed at least one course became just as likely as their higher-achieving peers to access tutoring.

When opt-in approaches aren’t enough

Yet, there’s a less optimistic way to interpret these results. Even in this most effective condition, with extensive communications to families and students, only about one quarter of students ever logged on to the tutoring platform. As a result, the availability of tutoring rarely translated into the use of tutoring. This open-access program is unlikely to have reduced—and may, in fact, have increased—inequalities in students’ academic experiences and outcomes. While we find that targeted communications to students and their parents dramatically increased the likelihood that struggling students engaged with on-demand tutoring, they only served to eliminate the gap in take-up with their higher-achieving peers. These students deserve more help.

History tells us that how tutoring is delivered matters. Effective, or “high-impact tutoring” programs tend to involve students receiving personalized learning, multiple times each week, from a consistent tutor during the school day. Implementing high-impact tutoring is not easy. Districts that want to implement tutoring often need to make programmatic decisions in the face of logistical challenges and resource constraints. For example, scheduling tutoring during the school day necessitates replacing something else in the schedule. Moreover, the number of students a school can provide with tutoring is limited to the number of tutors they can recruit and afford to hire. Nonetheless, this more intensive and more difficult to implement form of tutoring has demonstrated the most striking effects in helping academically struggling students accelerate their learning and engage in their coursework. The hard work of implementing high-impact tutoring can payoff in increased learning and increased equity.

Stepping back, the lessons of on-demand tutoring speak more broadly to the fair provision of public education. To address systemic inequities and learning recovery, we should not conflate availability of quality educational resources–like tutoring–with access. Results from our study suggest that students who are struggling academically may be less likely to take up optional educational programs, even if those programs could be particularly beneficial for them. As a result, opt-in approaches may be unlikely to address the persistent and increasing disparities in learning between struggling students and their peers. Helping the students who need it most will require a coordinated effort between educators and families to provide students with embedded, personalized learning opportunities.