Teaching assistants may sometimes feel like the Rodney Dangerfields of the classroom. They don’t get as much respect, and certainly not as much pay as the teachers they work with. But recent reviews of effective teaching strategies found that the assistants, often called paraprofessionals, were at least as good as teachers when it comes to one-to-one and small group tutoring. And both sets of paid professionals—aides and teachers—were far more effective than volunteer tutors.
Robert Slavin, director of the Center for Research and Reform in Education at Johns Hopkins University, recently set out to find the most effective programs to teach elementary and middle school students. In the course of analyzing almost 200 studies on particular methods, from “Reading Recovery” to ”Fraction Faceoff!”, he confirmed that students tended to make more progress with tutors than with a teacher in front of an entire classroom.
But he also made an unexpected discovery: teaching assistants were excellent at tutoring.
“In the past, we found that teacher tutors got better outcomes,” Slavin said. “But it was based on very few studies. As we got more studies, we saw that there wasn’t a difference” between teachers and assistants.
Related: Teaching kids not to be scared of math might help them achieve
For schools, this opens up the possibility of offering one-to-one tutoring to more kids who need the extra attention. Teachers are too expensive to deploy as personal tutors. But paras are paid about half as much, making tutoring somewhat more financially practical for the neediest students.
Slavin said he consistently saw the strong results for para tutors across the studies, from math to reading. In some cases, paras seemed to be better tutors than teachers. For example, young struggling readers tutored by paras had 50 percent larger learning gains on reading tests than those tutored by teachers. Meanwhile, paras were nearly three times as effective as volunteer, non-professional tutors with struggling elementary school readers.
The reasons for why paras sometimes outperform teachers are a mystery, Slavin said. But in the studies he reviewed, the paras typically had bachelor’s degrees—considerably more education than the job requires in most school districts. That high level of education might help explain why they performed well in these reviews.
Related: Does a lack of executive function explain why some kids fall way behind in school?
While Slavin’s finding raises all kinds of, as yet unanswered, questions, it’s clear that the paid professionals, unlike most volunteers, are following clear, sometimes rigid methods when tutoring, and they’re frequently tracking students’ progress. “They’re not just saying, ‘Here, go do your best with these kids.’ They’re using structured programs with specific materials that have been proven to work in the past,” explained Slavin.
The volunteer tutors used particular curricula, too, such as those used by Reading Partners and Experience Corps volunteer programs. But Slavin learned that volunteers often didn’t attend all the training sessions and adhere as closely to the prescribed steps. The paid professionals did.
In addition to a standardized curriculum, Slavin noticed that effective tutors were forming positive relationships with students and motivating them. “This may be a very important component of why tutoring works,” said Slavin. “Nobody’s trained in that aspect…. It happens naturally.”
Related: Three lessons from rigorous research on education technology
The great gains from human tutoring are often cited as evidence for so-called “personalized” learning, which sometimes refers to using education software that adapts to each child’s instructional needs. However, Slavin found that technology programs that personalize instruction for struggling readers “made no difference at all” in their test scores compared to traditional instruction. Personalized software did make a small positive difference in math.*
“People think that personalization is personalization whether it’s tutoring or 25 [students]. They’re night and day different,” said Slavin. “Kids who are working on a computer or doing other things with personalized instruction, it’s just not the same thing as a teacher sitting with you and able to see minute by minute how you’re progressing.”
*Clarification: This article was modified from an earlier version to add that personalized software was modestly beneficial in math. A sentence was also deleted that implied that tutors were not personalizing instruction when they were following a standardized curriculum.
Hello Jill Barshay,
I really enjoyed your article, thank you for taking the time to write it.
I wanted to discuss my opinion on why paras may sometimes (I suspect it may be many times) mysteriously outperform teachers, as tutors.
While higher education levels could, in some way, contribute to the better results of the paras, I believe something else could be the reason.
Teachers usually spend most of their time teaching larger groups of students and usually target the average students – let’s say they speak to the middle of the classroom – in an attempt to be as effective as possible and to reach the largest number of students as possible.
In contrast, many paras are usually tasked with helping the outliers – mainly the underperformers that are struggling to keep up with the normal pace of the classroom.
As such, a paraprofessional that has spent years engaging with students that are not in the “middle” may find that they have developed a different skillset than the teacher, whom has spent the same amount of time teaching only the middle.
It may be said that they have become specialists in their ability to tutor underperforming students.
This has been my experience as a professional tutor. I have been tutoring and mentoring students for 18 years now (professionally for the last 7 years).
Prior to hiring me as their tutor, parents have usually tried to get help from the student’s teacher; my results were better every time.
Now, I’m sure many teachers are able to get amazing results when they tutor students. But I obviously only get calls from those parents who have tried and failed using that method.
This has been my observation, I hope you find it interesting.
Thanks again for the great article!