Starting to think about tracking


I think my school does a nice job handling tracking for math classes.

When kids are in K-2, they stay in their classrooms and get taught math by their classroom teachers. But starting in 3rd Grade, kids split up out of their homerooms for math and get taught by a math specialist.

From 3rd through 8th, we teach six sections of math class for each grade. Three of these classes are “regular” pace class; the other three are “accelerated.” At the end of each year each grade’s math teachers and administrators get together with a huge stack of notecards and make classes for the upcoming year. We think about a lot of things — which kids would do well together, whether a class has a nice mix of personalities, and whether a kid would do better in a regular or accelerated section. And, by high school, students get to pick their pace, and kids absolutely do move between the “tracks.”

For the last few years I’ve been teaching an accelerated 4th Grade class, and it’s definitely not an easy class for me to teach (though consensus here is that accelerated tends to be easier teaching). The spread of interest and abilities is still high. (As you’d expect it to be at a school that has ~50% of students in accelerated sections.) There are two things that I find challenging. First, I don’t have nearly as many curricular resources for the accelerated 4th Grade as I do for the regular pace classes. Maybe you think it’s a social problem, maybe you don’t, but I have way more curricular tools for a struggling class than one that’s ready for more.

And the other thing is that I feel a real responsibility in this class for the kids who come in seeing math as their “thing.” There’s a special responsibility to make sure that these kids are challenged and engaged in my class since there’s nowhere else for them to go.

It’s sad but true: there’s more than one way to fall out of love with math.

Overall, I think it’s very good that my school has half the sections accelerated and half not. This gives us flexibility to make classes that we want, and it avoids some of the ways that tracking can make more problems than it solves.


If we’re looking for a good example of bad tracking, look no further than the first school that I taught at.

Here’s what happened each of the three years I spent there. The school would put students in 9th Grade sections based on what they’d heard from the middle schools. The top two sections were “honors,” and they’d study Geometry. The bottom two sections were taking Algebra 1. Once that placement was made, the rest of their high school enrollments in math were more-or-less locked in.

I used to teach the bottom two 9th Grade algebra classes, 9C and 9D (as they were so lovingly called). At the start of the year the classes would be about the same size, maybe 18 and 18 kids. Slowly, though, the 9D kids would figure out where they’d been placed. They’d tell their parents, the parents would call the school, immediately the kids would be moved to the 9C class, which would typically blossom into a lovely group of 26-28 students, leaving the remaining 8-10 stragglers behind.

I really did love the kids in 9D, but WOW that was a hard class to teach. Thanks to this artificial selection process, all the kids in that class were there because either (a) they didn’t care what section they were in or (b) their parents didn’t but mostly (c) they had diagnosed learning needs that weren’t being met, because it was an under-resourced private school.

I sometimes fantasize about going back to that school and teaching that class again. It wouldn’t be fun, but it really nags at me. Could I do better now, if I tried again? I do know so much more about teaching now, but it’s not a class that sets up a teacher for success. If I’m honest with myself…I don’t know if I’d be any better.

The class was very hard to teach. I don’t want to say unteachable, because there were good kids in that room that needed a good teacher. There were also numerous behavior problems, really all the time, just sometimes punctuated with learning. This says something about me in my first few years teaching; it might say something about me now. I’ll never really know.

But it certainly says something about the school.

I taught other low-track classes at this school — 10D, 11E. The Regents exam was far out of their reach, for the most part. We’d have a couple of passes out of each group, but it wasn’t a realistic goal for most students. The environment in the classroom was often out of control, and the school overall had this reputation for barely contained chaos.

I think I did alright there, but this is just the reality. There are hundreds, thousands of schools like it. And while tracking was clearly not the major problem at this school, there was no question who the losers and winners of this arrangement were.


How common are my experiences? How do they fit into the bigger picture?

I’ve been looking into the research on ability-grouping (within a class) and tracking (making classes by ability), trying to make sense of the state of things.

The story of this field is pretty interesting. It’s a field with a million meta-analyses — even a meta-analysis of the existing meta-analyses! All these reviews exist because there hasn’t been much first-order research since the early 1970s or so. So everyone is bootstrapping their analyses on top of the same old studies. If this is making you think that the evidence base isn’t particularly strong here, you’re getting the picture.

While there isn’t an incredibly strong research base here, there is evidence and even a sort of consensus. Tom Loveless does a nice job reporting on this for the Fordham Institute in a report titled Making Sense of the Tracking and Ability Grouping Debate. Loveless, as others do, frames the research around a debate between two researchers, Robert Slavin and James Kulik.

For those seeking a summary, here’s a condensed version of their debate:

Robert Slavin: booo tracking, you have no evidence

James Kulik: yay tracking, actually we do have evidence

Robert Slavin: no, that’s just evidence from gifted and accelerated programs that are poorly controlled, they no count

James Kulik: no they count

Robert Slavin: yo also I find tracking morally repulsive

James Kulik: really? i like it

That’s sort of it. If you like words, here’s Loveless’ summary:

Kulik finds that tailoring course content to ability level yields a consistently positive effect on the achievement of high ability students. Academic enrichment programs produce significant gains. Accelerated programs, where students are taught the curriculum of later grades, produce the largest gains of all. Accelerated gifted students dramatically outperform similar students in non-accelerated classes. Slavin omits studies of these programs from his analysis. He argues that the gains, though large, may be an artifact of the programs’ selection procedures, that schools admit the best students into these programs and reject the rest, thereby biasing the results.

Loveless is correct to point out that this debate is intractable, though, because Slavin actually finds tracking morally problematic and ugly. The burden of proof for Slavin is on schools that want to track, which explains why he can be so opposed to Kulik, even though they don’t seem to disagree very much at all about what the research shows:

Three things are striking about the Slavin-Kulik debate. First, the disagreement hinges on whether tracking is neutral or beneficial. Neither researcher claims to have evidence that tracking harms achievement, neither of students generally nor of students in any single track. Second, accepting Slavin or Kulik’s position on between-class grouping depends on whether one accepts as legitimate the studies of academically enriched and accelerated programs. Including these studies leads Kulik to the conclusion that tracking promotes achievement. Omitting them leads Slavin to the conclusion that tracking is a non-factor. Third, in terms of policy, Slavin and Kulik are more sharply opposed on the tracking issue than their other points of agreement would imply. Slavin states that he is philosophically opposed to tracking, regarding it as inegalitarian and anti-democratic. Unless schools can demonstrate that tracking helps someone, Slavin reasons, they should quit using it. Kulik’s position is that since tracking benefits high achieving students and harms no one, its abolition would be a mistake.

Loveless seems to be taking a compromise position in all this. “The research on tracking and ability grouping is frequently summarized in one word: inconclusive,” he writes. Since the research is inconclusive, he recommends a live-and-let-live strategy. Schools should have the freedom to choose their tracking structure, he says, but they need to be aware of the ways that each model can fail.

Tracking’s issues are well-known these days. Loveless calls for high-standards for the lowest tracks, and for ending what’s sometimes called the “teacher tracking” of putting the least skilled teachers with the lowest tracks. There need to be clear pathways out of the lowest tracks, a real effort to make sure that there’s room for students that start in one place to end up in another.

Untracked schools have problems of their own, though. “On the political side, anti-tracking advocates need to assuage the fears of parents that detracked schools will sacrifice rigorous academic training and intellectual development for a dubious
social agenda,” he writes, and this seems sensible to me also. The really ambitious students in my accelerated 4th Grade class do have needs, and their parents are legitimately concerned about meeting them.

If this seems wishy-washy and balanced, well, sometimes things just shake out that way.


If you think about it, isn’t it sort of weird that tracking doesn’t have clear and measurable benefits for the top groups in the research? Think about it. How often should ability grouping help strong students? Like, roughly, what percentage of the time should the top-group academically benefit from tracking?

I’d say 100% of the time. Roughly.

Teaching kids more stuff because they’re capable of learning more stuff is the single simplest idea in education I’ve ever heard. There is nothing to it. It’s just teaching more. Add to that the way they’re isolated from some of the toughest-to-teach students in the school, and this seems like it ought to be the clearest slam dunk on the educational menu.

So why should there be anything other than the clearest possible data signalling this? I’m not asking, why hasn’t the educational establishment recognized the evidence? I’m asking why the data isn’t super-clear. Why isn’t there a huge effect? Why isn’t it unambiguous in every single case that top-students benefit from ability grouping?

Two explanations:

  • Ability grouping is not necessarily acceleration. Some teachers don’t use a “top” class as a chance to do anything differently at all. You meet the standards? Great, you’ve met the standards. Let’s chill. (Or, let’s look at cool things on the side that don’t accumulate as knowledge.)
  • The skills students learn in a top group maybe aren’t measured.

Thinking about my own teaching, I think both of these things are probably happening in a lot of “top” track classrooms. I certainly do try to cover topics in my accelerated class that I don’t in my regular paced classes. But sometimes it’s just that we cover the same material but without as much stress, because there isn’t a clear vision I have of what an accelerated class ought to look like. I get little help from the available curricular resources, which are really all about fleshing out support for struggling students over kids who are ready for more.

I’m not complaining about this, mind you, but I think it’s true. There are probably a lot of teachers out there who aren’t making significant curricular changes between their tracks.

The second thing is true also. I try very hard to avoid racing ahead in the “standard” sequence of arithmetic skills with my accelerated class. The easiest way for me to handle an accelerated class would be to just march through the curriculum, teaching 5th, 6th, 7th grade standards to my 4th Graders. But this could create problems for the kids and my colleagues. If I unilaterally decide to teach e.g. fraction division, then I’m stepping on the toes of the 5th or 6th Grade teacher, who now has a handful of bored kids who are skilled at this because I decided to keep marching. 

OK, so the department should make a decision. But once you just set a class off accelerating through the curriculum, you’ve suddenly created a track that is relatively impenetrable to kids who start out of it. Somehow, they’ll have to catch up to join, and that’s going to have to happen outside of class. The only way to get ready for an accelerated class would be to be accelerated already, an unsavory Catch-22.

I think what I try to do is to flesh out the standard, grade-level topics with things that don’t have a strong showing anywhere in the standards. Right now my 4th Grade class is taking a deep dive into probability, a topic that only sort of shows up again in the 7th Grade Common Core standards. Earlier in the year I shared a bit of graph theory. We studied angles at a depth that will only really show up again when they’re studying diagrams in high school geometry.

My dream would be to have a curriculum that had a clear vision for what kids who are ready for more could dive into, beyond the grade-level expectations. There would be to keys to making this work. First, the additional skills would have to actually build and develop throughout the year — we want to equip accelerated students with something useful that builds their mathematical knowledge. But we also want a fresh start each year or so, so that kids can move in between the tracks without requiring some sort of catch-up.

I think something like this would give clarity and purpose to classes that otherwise have no choice but to plow ahead in the standard sequence.


On the margins, should US schools have more or less tracking? I think the answer is probably “better tracking.”

Even Slavin, opponent of tracking, admits that there is evidence for certain kinds of tracking in the elementary years, especially something called the “Joplin plan.” (Named after Joplin, Missouri, the district that gets credit for its invention.)

Joplin-style tracking cuts across grade-levels. A school might have an hour for reading instruction, and each student in the school would go to a classroom they’ve been assessed as ready for. So a 4th, 5th and 6th Grader might be reading similar books together in the same room, working on the same vocabulary. It’s a kind of limited breakdown of the age-grading system, really an artifact of the early 1960s.

Slavin, opponent of tracking, calls the experimental results of studies of the Joplin plan “remarkably consistent” and in support of the program.

Which makes sense, right? This is the simplest possible educational idea: teach kids more when they seem to be ready for more. And, as an extra bonus, since the kids are heterogeneously grouped for most of the day, you don’t run the risk of creating really problematic tracks that lead to wildly varying places. By the nature of the plan, there is curricular guidance for kids who are ready for more. This should work 100% of the time.

And Slavin put his money where his mouth is, co-founding Success for All, a school improvement program that has something very much like the Joplin plan as its cornerstone. (His co-founder is Nancy Madden, another Johns Hopkins education professor. Madden and Slavin are married.)

And, ironically, Slavin’s program has been critiqued for its use of tracking. (Also for its use of scripted lessons, which will never make teachers happy.)

It seems as if the situation is that Slavin’s preferred sort of tracking would be good for students and good for equity and mobility. He’s a noted critic of tracking and ability grouping, and is deeply aware of all the traps. Success for All reassesses students every two months, and students are expected to move between groups. This is the form of tracking with the strongest research pedigree. And yet it comes in for criticism.

What’s confusing about this to me is that we aren’t a country that is shy about grouping students by ability. Loveless notes this: “Ability grouping for reading instruction appears nearly universal, especially in the early grades.” In the elementary years, this is usually within-class grouping, e.g. red group sits at this table and blue is on the rug, etc. But by the time students reach high school, the near universal pattern is separately tracked classes, more like what my old school did.

Here’s the puzzle: why do some forms of grouping and tracking attract more ire than others? Is it just a matter of the devil we know vs the one we don’t? Familiarity breeds begrudging acceptance? I don’t know.

But looking at programs like Success for All and thinking about what happens in the math classrooms that I’ve seen, it seems to me that purely from the standpoint of mathematical learning, there is probably a better way of doing things. Here, as bullet points, are my takeaways from all this, with the most doable items near the top:

  • We could use curricular materials that go beyond the standards for each grade level of math, so that classes who are ready for more can dig in without dashing through the standards.
  • Something like Slavin’s reading plan could be useful in elementary grades. Keep the heterogeneous groups, in general, but assess students every couple of months or so to place them in a class that’s right for them. (If we could clone Success for All minus the scripted lessons it would probably be more popular. Though I’m sure there’s probably something lost when we do that too.)
  • Maybe, even in the upper grades, it would be helpful to split the year into two halves, with an ability grouping move in the middle. Or maybe it wouldn’t help at all. But it might be interesting for a school to try something like that and see how it goes. Maybe?

But all these are speculative recommendations. Overall, I don’t get the sense that there is a huge gap between research and practice because (as Loveless notes) there isn’t a great deal of clarity from the research.

Instead, there are promising ideas with research support (like Slavin’s). This doesn’t exclude the possibility that there are other good ideas out there, and it seems likely to me that if a school or parent body thinks that tracking or untracking is necessary for their students, they’re probably correct.

11 thoughts on “Starting to think about tracking

  1. A couple unrelated thoughts:

    “So why should there be anything other than the clearest possible data signalling this? I’m not asking, why hasn’t the educational establishment recognized the evidence? I’m asking why the data isn’t super-clear. Why isn’t there a huge effect? Why isn’t it unambiguous in every single case that top-students benefit from ability grouping?”

    Your explanations make sense. One other possible explanation: it’s harder for kids who do well on achievement tests to show progress/benefit from an intervention. Not sure how the studies were designed, but I wouldn’t be surprised if at least a few had the same kinds of measurement issues that plagued the early “value-added” models:

    “Here’s the puzzle: why do some forms of grouping and tracking attract more ire than others?”

    One answer: in class grouping tends to be seen as more conducive to movement between groups. Done well, groups are changed frequently–weekly, maybe even daily–based on results of formative assessment. By contrast, once “placements” are formalized, different teachers are involved, kids move to other rooms, etc, the bureaucratic and logistical obstacles to changing groups increase significantly.

    I don’t know that it’s essential that kids change groups every day or week–stability is a good thing too–but I do think that movement between groups happens (not just theoretically, but actually takes place) on a regular basis.

    I totally agree that accelerating (vs. enriching) the regular curriculum makes it virtually impossible for kids to move into higher or more advanced tracks. One model we’ve tried in 6th and 7th: co-teaching larger classes, with 2-3 days a week of whole-class lessons and 2-3 days a week of small-group instruction. Groups were reassigned every unit. The “advanced” small groups did a lot of Beast Academy…

    “Maybe, even in the upper grades, it would be helpful to split the year into two halves, with an ability grouping move in the middle. Or maybe it wouldn’t help at all. But it might be interesting for a school to try something like that and see how it goes. Maybe?”

    I started an advanced track with 7th grade mid-year. Jury is still out on how that will work out long term, but I’m pleased so far.

    I also think motivation is an important aspect in the middle and upper grades–for 12- to- 14- year-olds in particular, in my experience it’s even more than what my best assessments suggest kids are “ready” for. I’ve seen kids get burned out and come to dislike math when placed in classes that are “too hard”…but I’ve also seen kids who have really been on the margin in terms of test scores work tremendously hard to keep up with peers. The seventh grade class I mentioned is using Alcumus at Art of Problem Solving. My highest-rated student on Alcumus has answered 63 of 72 questions correctly to pass the topics we’ve covered so far. One of her friends has answered 234 of 360 questions correctly to pass the same number of topics. She wants to be in the honors class really, really badly, she’s working hard, learning a ton of math, and (so far, at least) enjoying herself. She’s ok with doing five times as many questions…who am I to say she can’t? (I don’t necessarily see that as an argument against tracking, btw, just for lots of humility when it comes to placing kids.)

    Liked by 2 people

  2. Michael, as always, I find your post thoughtful and well written, and looking at an issue from many sides. But it’s one thing to talk about tracking in a private school, I believe, and another thing to talk about tracking in public schools, particularly when students of color may be he majority of the kids who are being tracked into “lower level” classes. There is rarely a way out, once this happens, and it’s difficult (for me), if not impossible, to isolate the racial implications of this system from the academic ones.

    Liked by 1 person

    1. “There is rarely a way out…”

      This sounds a lot like the first school that I taught at. I didn’t like that system, and I don’t like this sort of set up in your school either. And, absolutely, in the current system this usually opens up the door for kids to experience a lot of racism.

      That said, if you told me that you were at a school with many students of color and the tracks weren’t racialized in this way, or the school was doing a good job finding ways to give quality experiences to the “lower level” classes and there was the potential for mobility through the tracks…that sounds good to me.

      And so I’m left with thinking that it really depends on the school. I wouldn’t ever want to be faced with prescribing a practice for a community or school (as advocates of both tracking and untracking frequently do). As it stands, the evidence doesn’t seem to support those kind of universal prescriptions.

      Would love to know more of your thoughts, Wendy!


      1. While I can concede that there may be schools where tracking may be done ‘well’ if such a thing is possible, I would venture to say (without hard facts but only my own experience and anecdotal knowledge to back this up) that students of color being relegated to lower tracks is the wide rule, and not just pattern in particular schools. NCTM has taken an explicit position against tracking for exactly this reason.

        Liked by 2 people

      2. NCTM sort of is against tracking:

        Catalyzing Change draws a distinction between tracking and acceleration, arguing that acceleration of students through shared content may be appropriate if a student has demonstrated deep understanding of grade-level or course-based mathematics standards beyond his or her current level.

        From this.

        Do you feel that public schools shouldn’t have acceleration options? In general, what should a school do with students who are ready for math beyond what’s offered in their courses, in an untracked system?


  3. “particularly when students of color may be he majority of the kids who are being tracked into “lower level” classes. ”

    I agree this is an issue. However, notice that you didn’t assert that kids of color are improperly tracked.

    That is, we automatically jump to politics when we talk about tracking. Particularly in high school. It’s impossible to discuss the subject from an educational standpoint. If all kids were properly tracked, kids of color would remain in the low level classes. And the reason it’s impossible to move out of low level classes once in is because the subject matter is so difficult each year. It’s not about catching up, but rather difficulty level.

    As I said in Twitter, the politics of high school tracking are very different because schools are very likely to be more diverse, and the politics of tracking worse.

    On the other hand, what we mean by tracking changes. High school tracking generally just means moving faster, and generally, in math, that means rushing through the curriculum where all but a few kids don’t understand it anyway. And in order to avoid the racism issue, we use different metrics–parental demand, grades, or state test scores.

    Parents are worried their kids aren’t learning “as much” as they can. But in math, we’re already teaching all but the very strongest kids (say, maybe 20%) far too much. The kids don’t absorb it, even if they get As in the classes.

    I don’t know what the answer is, but if every time someone brings up tracking we still go back to “the best kids are being held back” or “kids of color will be in lowest tracks”, we’ll get nowhere.

    Try: “We aren’t challenging our strongest kids by deeply teaching the math” and “unmotivated kids aren’t interested in jumping through homework hoops”.

    But if we just kept math to the kids who could honestly do the math, most kids–including a hell of a lot of girls with 4.0 transcripts–would be left off.

    So I quit having answers. I do think we should have baseline motivation and ability tests for schools that are politically blocked from any form of tracking.

    Liked by 1 person

  4. One other thought, you mention “My dream would be to have a curriculum that had a clear vision for what kids who are ready for more could dive into, beyond the grade-level expectations. There would be to keys to making this work. First, the additional skills would have to actually build and develop throughout the year — we want to equip accelerated students with something useful that builds their mathematical knowledge. But we also want a fresh start each year or so, so that kids can move in between the tracks without requiring some sort of catch-up.”

    I think this is a solved problem. We have an excellent template in the work AoPS has done for what a deeper curriculum might look like as a coherent whole. I tend to judge alternatives against that standard.

    Liked by 1 person

  5. Might have missed it above, but here’s a recent growth mindset meta-analysis study: “A new study co-authored by researchers at Michigan State University and Case Western Reserve University found that “growth mindset interventions,” or programs that teach students they can improve their intelligence with effort — and therefore improve grades and test scores — don’t work for students in most circumstances.”


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s