Teach Your Teachers Well

Of the (very) many large topics on the Obama administration’s to-do list, one that has slipped off the radar of late is education reform. I assume Arne Duncan et al. are working hard and will retake the spotlight eventually, maybe even in a few weeks when a new school year begins. It will be interesting to see how much attention is paid to one of the most important, albeit touchiest, topics of school reform: teacher skill.

We will touch on the subject in SuperFreakonomics, but here’s a new paper (pdf; abstract) by C. Kirabo Jackson and Elias Bruegmann* that looks at it from an unusual angle: whether a teacher’s success, when measured by students’ test scores, improves when that teacher gets better colleagues in her own school. Their answer: a firm yes.

From the abstract:

Using longitudinal elementary school teacher and student data, we document that students have larger test score gains when their teachers experience improvements in the observable characteristics of their colleagues. Using within-school and within-teacher variation, we further show that a teacher’s students have larger achievement gains in math and reading when she has more effective colleagues (based on estimated value added from an out-of-sample pre-period). Spillovers are strongest for less-experienced teachers and persist over time, and historical peer quality explains away about 20 percent of the own-teacher effect, results that suggest peer learning.

We know that peer effects are strong for students, so perhaps it shouldn’t be surprising that they are strong for teachers as well. But I am guessing this finding is a surprise even to a lot of people within education, and I hope this paper will be read by the people who take school reform seriously.

*For those of you concerned with the alphabetical determinism faced by academic co-authors, consider this interesting footnote in Jackson and Bruegmann’s paper: “Note: In the interest of fairness, the ordering of the authors’ names was determined by a coin flip.”


Was the coin a fair coin? :-)

Interesting that it would be a surprise, because the peer effect seems to be evident in many other things.


I doubt the coin flip was fair.


I would think that as the majority of the teachers in any school improve, it would have a positive effect on student behavior.
Student behavior gets better, and even the bad teacher is more effective than he or she was when kids were disrupting all the time.
That's the "duh" logic I see in it, anyway.

billy bob

The peer effect is real and people make decisions based on this but they don't realize it. There are some educational programs that are truly outstanding and full of bright people (usually the expensive private colleges or a lavishly funded state school). The level of discourse was so much higher when I was surrounded by bright students and even brighter (professors). I don't learn much from people with IQ's one or two standard deviations below mine, but when I was in the supercompetitive programs I was merely average at best, so I was sponging off the even brighter people around me. I learned how they think and how they solve problems. I was lucky to be in a program taught by full professors who continued to do research. I advanced multiple levels over my peers in other less distinguished programs and it became more and more obvious on test scores and career options later. People want to live in neighborhoods that are expensive because these are proxies for high achievement and being in a milieu that values achievement promotes achievement. I suspect the converse is also true. Living around low achieving people and low achieveing teachers probably harms performance.

Peer effects also work in small group settings also. If you have a small group of superbright engineers you can achieve almost anything because of the synergies and energy they give each other.


Ronan L

How does one square this with evidence from the international study using twins that classroom effects (including size and teacher quality) have very little impact on student performance?

Is this something about teachers pick themselves but even still, how does this impact through on student performance, when the (albeit limited so far) evidence is that teachers aren't really that important.


Perhaps it has fallen off of your radar.

But Arne Duncan and his DOE have been very active, and those who care have been quite disturbed by his agree advocacy of/forcing states to implement unproven or even disproven policies.


@ Ronan L - I'm not that familiar with the twin study, but if it is what I think it is, it squares quite nicely. If peers in the classroom or faculty lounge have a strong effect, I would think peers at the dinner table or in the same bedroom would have a monstrous effect. Unless these were twins from separated families, wouldn't using twins be counterproductive? You're just mixing your control every time the kids go home at night and work together on homework, etc.

Justin R

True, Commenter 3. I would add that, even if all students were well-behaved by default (even when being poorly taught), the effect of consistent mental engagement via quality teachers throughout the day makes their synapses more likely to fire in the classroom of any given teacher.


All very interesting. But now how do we improve education? Or even, how do we improve one teacher's performance? Do we have all teachers work at places like Lake Wobegone, where all teachers are above average?

Lily A

Has anyone considered that perhaps students learn good test-taking skills from good teachers, and those skills are applicable across disciplines?

Say I am an inexperienced math teacher, but my students are learning good test-taking and reading comprehension skills from the excellent English teacher down the hall. This will probably improve their math test scores, but it doesn't necessarily mean that my teaching has improved or that the students have learned math any better.


Wouldn't this be true as well for the students? When better students are mixed in with average achievers this improvement benefit should apply as well. We are all affected by the company we keep as wise mothers have always advised.

Former Teacher

It would be silly to deny that teachers, like everyone else with opposing thumbs, are capable of learning from others, so, obviously, the better that knowledge and experience of their colleagues, the better the vicarious learning.

However, the improvement of teaching by the target teachers is just one factor among others including:

1. Better teaching by the new colleagues for those same students in other classes.

2. Better administrators who recruited the better teachers and therefore probably create a better learning environment on the target campus.

3. Possibly or even likely, increased funding devoted to the target campus that enabled the recruitment of better colleagues and provided a range of better resources such as computers, textbooks, and functioning plumbing.

Carol Burnell

I had to laugh at the coin flip. That seems appropriate when the co-authors actually share equal work, expertise, etc. But in academia, the first author is the lead author--the person who contributed the most. (And yes, that is often skewed when a prominent academic is involved--that person is sometimes named the lead author in order to get published.) Authors are not listed alphabetically in academic journals.

antonia r

Excellent point commentator 6. In fact, the Gates Foundation is now giving $250,000 to various states to help them hire consultants to fill out the proper paperwork for the federal bonus money in Obama's "new" plan called "Race to the Top," which is basically NCLB. I believe Freakeconomics authors have written about how performance pay for teachers encourages cheating rather than strengthening teacher/student performance. Moreover, Duncan himself investigated cheating in his own state before becoming Secretary of Education, so it's a cruel irony that he is espousing the very plan that has not been proven to help anyone, least of all the students who, if anything, get turned off by school with all the test prep.


Why do we always avoid the obvious?

Teachers are paid a pittance. Therefore very few people with real potential go into teaching (especially the lower grades, where teachers have enormous impact on young mind development).

Weed out the poor teachers, and compensate the competent ones based on the huge contribution they make to our society!


The US approach is to push teachers to push their students to push their parents into becoming better parents. This is pushing a string.

Only in the US do we blame teachers for the poor parenting their students receive.

Just makes me believe that this is failure by design, making sure that only the well-to-do ever have a real shot at an education.


Okay, I'll accept the research. That said, it does not negate the benefits of paying teachers based on performance. It just adds a component that must be considered.


Can someone please explain this sentence to me: "based on estimated value-added from an out-of-sample pre-period."


Based on my 6 years of teaching thus far, in a very challenging NYC high school, my observation is that the extent of any kid's learning depends upon:

the teacher +
the school administration/leaders +
the parents/community +
the kid him- or herself.

If you get all those stakeholders committed, involved, and working together as a team, you can have success. If you don't, you get reality.

Larry Loganbill

Teachers all need to take a course in Communication Theory and learn to SHOW and TELL and not just tell when they
want to convey new information and concepts to their
students. I don't see any changes in the way teachers teach
from when my great-grandmother taught school in the late
1890s...or even from the Bronze Age. Nothing changes
in education. I guess everyone thinks they're doing just