Differentiated Instruction: Not a Solution to Student Struggle (Yet?)

There isn’t a way to differentiate that’s reliably shown to work — but there still is hope.

Gym class didn’t make me healthy and fit. It just showed me how I wasn’t. I couldn’t do a single sit-up, or run a mile fast enough for the fitness cutoff, or hold a flexible yoga pose without collapsing, as hard as I tried. I just felt that putting strain on my body was uncomfortable, and seeing how I compared to my peers was humiliating. These experiences discouraged me from seeing exercise as something I’d want in my lifestyle—so while most of my classmates learned as intended, these classes did the opposite for me.

It was only later that I realized exercise can truly be enjoyable with an appropriate level of challenge. I compare myself to nobody but who I was before, and it’s empowering to see my abilities grow. If only P.E. had been like this all along, I could’ve felt these benefits sooner.

Perhaps this was how academic classes made some students feel, I realized. I worked hard and excelled in math, English, and science courses every year, but I could empathize with those who struggled. They were repeatedly getting assignments they had no idea how to do and scoring poorly on them, and this wasn’t the way to learn. It wasn’t that they didn’t care to work hard; in fact, many poured more effort into those classes than I ever did and felt guilty for still falling short. They just had a lot more to master before things would start to make sense for them, and it wouldn’t be realistic to expect them to get to the same level as their peers in the same time frame.

That’s why education visionaries propose a radical redesign — one of the most common terms describing schools of the future is “tailored,” or “personalized.” In this vision, students do what’s most suitable for them based on their own abilities and goals, getting dedicated assistance when they need it. The description of a teacher’s new role sounds more like a trainer in a gym.

This sounds like it should work; how could holding everyone to the same expectations be an improvement? As of now, confused students struggle with frustration and high performers like me struggle with boredom, which some could point to as evidence that business-as-usual doesn’t “work.” However, we should not make bold claims about which changes to make before consulting past experience.

Troublesome Tries

If the main source of struggle under mainstream teaching methods is that students with great differences are held to the same standards, then in-class differentiation should improve learning outcomes. That is what Carol Ann Tomlinson, a dedicated advocate of differentiation, sought to test in a large-scale trial in schools across three U.S. states, and the results were disappointing — almost no significant effects.

Tomlinson and co-authors attribute this mainly to “implementation” issues. But according to teacher-blogger Greg Ashman, if this is what happened in a thoughtful, rigorous, researcher-directed intervention by those most motivated to cast differentiation in the most persuasively positive light, surely it would not go better if teachers were left to direct it themselves. Whether it’s “hard to train teachers to do it correctly,” “flawed in principle,” or “a bit of both,” argues Ashman, in-class differentiation does not seem to be worth it, compared to whole-class approaches better known to be effective and practical.

The Problem of Choice

Let us say that despite difficulties, instructors could deliver personalized teaching. In a personalized environment, how would students get matched to the right content and methods to help them learn? The common approach, as in Universal Design for Learning and some of Tomlinson’s trial: they get to choose.

If students had a choice, many wouldn’t come to class. If students had a choice, many wouldn’t do their homework. Online learning during the pandemic has shown us clearer than ever. Yet, as this line of thinking goes, if they’re given a choice about what activities to complete, they’ll pick what’s in their best interests, not what’s most comfortable and convenient? It’s hard to believe.

Indeed, research consistently finds that levels of self-reported student enjoyment are negatively correlated with levels of measured learning; students tend to prefer methods that teach them less. When given the choice, lower performers often would rather pick more “permissive” approaches, even though they benefit more from structured instruction.

Ashman points out that preference-based differentiation could especially hinder progress for students with learning difficulties. A student with difficulty writing, for example, could have the option to express herself without writing, but this accepts her issue instead of addressing it, and without the opportunity to improve, “she will fall further behind her peers who are practicing more writing than her.” I know it’s particularly uncomfortable for me to exercise my particularly weak core muscles — but I know that’s exactly why I need to do it. A student who’s required to get an education, as opposed to an exerciser who’s striving to get healthier, is less motivated to want to improve in the right areas.

Algorithms to the Rescue?

So if teacher efforts to differentiate cannot be relied upon to improve student learning, and student choice cannot be relied upon to determine how instruction should be differentiated, new educational technology could fill the gap. Learning platforms could measure students’ performance and match them to reliable tested lessons appropriate for their current skill level, most statistically likely to help them.

This is no mere hypothetical tech utopia. These are real, and they have been undergoing trials and refining for years now. But if a trial fails to find improved learning, is it an issue with the differentiation approach, consistent with in-person trials, or with the tech? If a trial does find positive effects, is it because of an improved learning experience, or because of effective differentiation? There are so many edtech developments with so many differences among them that it’s hard to compare successes and failures and confidently claim what made them such.

I have faith that someday, we will have a more reliable basis to figure out what instruction students need in order to improve and a more reliable way to consistently deliver it. Let us realize, though, that we are not there yet, and trying to implement it anyway, especially by increasing students’ choices, may not provide the solution that struggling students are hoping for.