How well does online tutoring work?
This is an important question. The federal government is pushing schools to spend a big chunk of their $122 billion in federal American Rescue Plan funds on tutoring, but bringing in armies of tutors into school buildings is a logistical nightmare. And now, with the Omicron variant still raging in many states, it’s even more difficult. It’s also hard to find enough physical space to work one-on-one or to rejigger school schedules to make room for tutoring time.
Online tutoring is a tempting solution. It comes in many forms, from text chatting and homework help lines to robo-tutors that use artificial intelligence to deliver prepackaged lessons. Some forms of online tutoring mimic in-person tutoring except sessions take place over Zoom or another video chatting app. All a student needs is a laptop, headphones and a good internet connection to access a one-on-one tutoring session, even in a crowded classroom.
The online tutoring business is indeed booming. Hayley Spira-Bauer, chief academic officer at Jericho, N.Y.-based iTutor, told me they’re in a “hyper growth phase.” Heavyweight investors, including Softbank and IVP, are pouring hundreds of millions of dollars into online tutoring startups, including Vienna-based GoStudent and Montreal-based Paper.
While there’s strong evidence for a particular type of in-person tutoring that takes place every or almost every day called “high dosage” tutoring, it’s still not clear that this success translates to the virtual world.
So far, we have two well-designed studies where students were randomly assigned to receive online tutoring and their academic progress was measured. The first showed promising results for low-income immigrant middle schoolers in Italy in the spring of 2020. When it was repeated during the 2020-21 school year, it again showed that students’ test scores shot up when they received four hours of online tutoring a week from university students. But when students received only two hours a week of online tutoring, the academic gains fell by more than half.
“This suggests that high-dosage tutoring is very effective, but when we decrease the number of hours, the impact gets significantly reduced.” said Eliana La Ferrara, a professor of economics at Bocconi University and a researcher in the study with whom I corresponded by email.
The results were less sanguine in another study that looked at pairing volunteer college students with low-income middle schoolers in the Chicago area. The students who received online tutoring in the spring of 2021 didn’t do much better in reading or math than students who didn’t get the tutoring. Statistically, it was a null result. The study, Online Tutoring by College Volunteers: Experimental Evidence from a Pilot Program, is slated to be published in a forthcoming issue of the American Economic Review, and was recently made public at the end of January 2022.
“We haven’t proven that online tutoring is guaranteed to work,” said Matthew Kraft, an economist at Brown University who led the Chicago study. “But we haven’t gotten evidence to say this is really going to tank. It was a pilot study and it wasn’t particularly large. It would motivate me to want to study this a little bit more, particularly at scale.”
For the Chicago study, CovEd, a non-profit organization run by college volunteers, recruited undergraduate college students from highly selective universities around the country to work with middle school students in Chicago Heights, an industrial suburb 20 miles south of Chicago. Over the course of 12 weeks, more than 250 middle school students – mostly low-income children of color – were supposed to receive 30 minute tutoring sessions twice a week during the school day. In practice, the students received much less, an average of three hours in total. Technical glitches, poor attendance, rolling tutor recruitment and vacation breaks ate away at tutoring time.
It’s unclear whether it was the lower frequency or the awkwardness of remote learning that inhibited the kind of learning gains that are usually seen from tutoring. Kraft noticed that students who started the tutoring earlier because their tutors were recruited in the first waves tended to see larger academic gains. That indicates to him that online tutoring could work if students received more hours.
The unpaid volunteer tutors were encouraged to focus on building personal relationships with their students and then give supplemental help with math and reading.
I talked with Isabella Pedron, a 20-year-old chemical engineering major and pre-med student at Texas A&M University, who was part of the pilot study and continues to serve as a volunteer online tutor to middle school students in Chicago Heights.
“I really love children,” said Pedron, who signed up after seeing a Texas A&M email about volunteer opportunities during the pandemic. “In the future I want to be a pediatrician. So I thought this would be a great way to not only interact with children, but also to give guidance to underserved kids in the United States.”
Pedron told me it was her first time tutoring or mentoring anybody and she appreciated the two-hour online training session, which gave her tips on how to talk to children and what to do when a student doesn’t respond.
Pedron conducted the tutoring sessions over Zoom from her dorm room in College Station, Texas. One of her students never showed up. The other didn’t log on for the first two weeks. Once he did, Pedron said she had a “blast” with her tutee, who had a passion for robotics. Pedron created Kahoot and Quizlet games to play and gave her tutee a virtual tour of the Texas A&M campus and her dorm room. “They were really really attentive whenever I was explaining any topics. And they were always very kind, saying thank you for all the sessions. I always felt like we had a great time during our mentoring sessions,” she said.
But it didn’t surprise Pedron to learn that the students’ test scores didn’t improve much. Often the 30-minute session shrank to 20 minutes once the students logged on to their devices. “You don’t get a lot done in 15 or 20 minutes,” she said.
Sometimes other tutors didn’t show up and once she subbed in for them. Their tutees didn’t have their cameras on and it was hard to establish a rapport.
Pedron said that more frequent sessions throughout the week would have helped the students more, but it would be difficult for many college volunteers to devote more than an hour a week with their school schedules.
She also wished she knew more about what the students were working on at school. “Some tutoring sessions were like, ‘Oh, I finished all my homework. We don’t have any exams. I don’t need to work on anything.’ So we played some other trivia games online, but they weren’t really geared towards what they were actually learning,” Pedron said.
And Pedron admitted that it was sometimes hard to engage students across the computer screen. At times her tutees would look away from the computer for long stretches of time and seem distracted. “I would ask them if they wanted to switch things up and do something different,” said Pedron. “It’s easier to lose focus and attention on Zoom.”
The researchers are hopeful that online tutoring could be more potent if they can increase attendance. The problem they’re now facing is finding tutors.
“We have found it considerably more challenging to recruit college volunteers this year than last year,” said Kraft, explaining that college students are busier now than they were during the first year of the pandemic, when they themselves were attending classes remotely. Kraft is investigating whether high school students could serve as online tutors.
Paying tutors might help too. “In our experience more recently, it seems unlikely that volunteers alone would be a primary source of tutors,” Kraft said.
I asked Kraft how online tutoring companies are claiming such big academic benefits in their marketing materials. iTutor, for example, wrote an impact case study, saying that Alaska students, who were a year and a half behind, had accelerated “growth up to grade level rapidly,” a much larger jump than either the Italy or Chicago studies imply.
“Trust but verify,” said Kraft. “It’s probably true that some kids make huge gains. Whether all kids did is maybe a different question, and not the one that they’re trying to convince you they answered. I think there’s a lot of potential for online tutoring, but this speaks to the need for third-party rigorous empirical evaluation.”
This story about online tutoring was written by Jill Barshay and produced by The Hechinger Report, a nonprofit, independent news organization focused on inequality and innovation in education. Sign up for the Hechinger newsletter.