The problem with elementary school, some might say, is that the teachers aren’t able to focus on mastering their subject. It’s not their fault! They teach everything. There’s no time to specialize. There’s no time to do anything. Do you have any idea how few bathroom breaks an elementary teacher gets during the day? Zero to one. So of course they can’t master the mathematics curriculum.

And in fact there is research showing that different teachers are more effective at teaching some subjects than others.

Here’s a simple solution: every child should have access to a math or reading specialist.

This is an idea that a lot of people love. For example, the National Council of Teachers of Math and a bunch of other organizations officially support the use of math specialists (though they note that math specialists can be used in many ways besides managing classroom instruction, e.g. as coaches, meeting with teachers or individual students, running professional development for the district, and so forth):

A new paper has added to the surprising(?) research literature that argues that this simple solution does not work. In fact, the teaching goes worse when you’re a specialist. It further suggests that the reason is because you lose out on the benefits of a stronger student-teacher relationship when you specialize. The paper is titled “Spread Too Thin: The Effects of Teacher Specialization on Student Achievement” and if you are inclined to read such things I encourage you to read it.

But, before you do, it might be helpful to list every reason why the researchers might be wrong. Why might researchers have messed this one up? Let’s list the ways:

- Even if teachers are slightly worse at teaching when they’re specialists, schools might be making smart moves about who is best in the classroom and who is best teaching a subject, i.e. Michael isn’t our strongest teacher so we put him in a more targeted role where his knoweldge of math can be put to use
- Every new professional role is hard, so teachers who become specialists will initially struggle but eventually get the hang of it
- Schools that move to having specialists will uniformly improve but this won’t be captured by a change in any individual teacher’s effectiveness

To summarize, we want to know that there’s neither a **school effect **or a **teacher effect**. And we want to make sure that the researchers aren’t just capturing the difficulty of being effective when you’re transitioning to a new role.

Now, on to the paper.

If a policy paper has no experiment then it begins with a dataset. This one comes courtesy of the Indiana Department of Education, which provided researchers with data on students (4th and 5th Graders), their teachers, and whether their teachers were generalists or specialists over a seven year period. This amounts to 591,311 students and 32,996 teachers. (Co-teachers were excluded, because that’s hard to deal with. That was 10% of the teachers in the sample. Can you think of a reason why excluding co-teachers would mess this up? The researchers couldn’t and neither could I.)

Their primary interest is in the **teacher effect**. Since they aren’t running an experiment, they look at *changes *in a techer’s work status as the thing that could cause a change in learning. In their words: “we identify the effect of specialization by comparing the effectiveness of the same teachers in years when they do and do not specialize.” The key assumption being that these changes are **not **correlated with anything else, i.e. you don’t become a specialist once you’re suddenly great (or awful) at teaching, or you don’t become a specialist when your students suddenly become good (or bad) at reading.

Then they run a giant regression that looks to see what contribution the teacher’s role has in explaining student performance, over and above a bunch of other things.

(These “other things”: a student’s prior test scores, gender, race/ethnicity, whether they get free or reduced cost lunch, enrollment in special services, ELL, class size, whether a teacher has a graduate degree, whether a teacher is new to a school, how big the school is, percent of Black and Hispanic students, percent of students in the school who get free or reduced lunch, how effective this teacher was on average no matter the role, how well the grade scored on these tests on average, how well the school performed on these test on average.)

And then the researchers got nervous that they were only capturing the first year of a teacher’s transition, and maybe they got much better after that first year. So they created a model that paid attention to how many years of experience as a specialist the teacher had — one year, two years, three or more years — and took a look at whether experience mattered.

This is all good, but what about the **school effects**? Well, they wanted to look at that as well. “Though assigning a teacher to a specialist role may lower an individual teacher’s average effectiveness, students may still be better off if that teacher is better at a particular subject than the other generalists in the school,” they write. So they tested another model i.e. long equation:

I would like to emphasize that if you are familiar with the idea of linear regression these fancy equations should not intimidate you further. We are interested in the impact of the specialization rate of a given school in a specific grade at a time on the scores of kids in that time in that grade at that school. And beta is the slope, so if that slope is high then voila, the specialization rate explains a lot. If it’s low, it doesn’t. Everything else is in a sense the control — the researchers will measure the explanatory power (i.e. control for) all the stuff that I listed in that parentheses above. What we care about is the sope associated with the specialization rate.

Anyway, the results aren’t good if you like specialization. If you just look at specialization of teachers, there’s a significant reduction of teacher effectiveness, especially for math. If you separate by year, it’s clear that things get better when teachers are more experienced but not that much better.

When you compare schools that specialize more to ones that specialize less, the differences aren’t huge but those numbers all have negative signs in front of them which, if you like specialization, is not what you want to see:

They speculate that this could be explained by the way specialization reduces the strength of the teacher’s relationship with their students, and they have an interesting way to test it: look just at the students who happened to work with the same specialist two years in a row. They find that this reduces the costs of specialization compared to working with a generalist (though not all the way):

Do I believe these results? Yes, I do, for a few reasons. First, because this isn’t the first time researchers have found that specialization backfires in elementary school. There was an experiment in Houston a few years back that was particularly interesting with true random assignment and it didn’t work out. There was also a big study of schools in North Carolina that had negative results for specialization. Here are those studies described by the authors of this paper:

But the other thing is that I am a 3rd Grade math specialist. I work hard, and I think have a good understanding of the math and the kids. Yet I frequently come in and realize that I’m walking in to a situation I only partly understand. The kids main teachers have expectations and routines they work on all day with their students. They have different bathroom policies. Kids are working on things emotionally that I sometimes catch a bit of, things like “trying to deal with frustration productively” or “separating yourself from a tough situation.” And I don’t get to know any of this!

Moreover, these kids worship their teachers. They are the adults who care for these children during the day. Meanwhile I come in and it’s more of a gun-for-hire situation. I’m in, I’m out. It’s a very weird situation compared to the rest of my day.

So, no, I don’t find this surprising, and while maybe someday someone will figure out a way to get the benefits of specialization without the costs, I’ve seen enough for now: specialization in elementary school is not worth it.