Study: Great teachers have lifelong impact

Students with an excellent elementary or middle-school teacher don’t just earn higher reading and math scores, concludes a new study that tracked one million students in an urban district over 20 years. A single year with a high value-added teacher leads to higher college attendance, higher adult earnings and even lower teenage-pregnancy rates, according to the authors, economists Raj Chetty and John Friedman of Harvard and Columbia Professor Jonah Rockoff.

All else equal, a student with one excellent teacher for one year between fourth and eighth grade would gain $4,600 in lifetime income, compared to a student of similar demographics who has an average teacher. The student with the excellent teacher would also be 0.5 percent more likely to attend college.

It may be difficult to hire more excellent (top five percent) teachers, but it’s not necessary.

. . . the difference in long-term outcome between students who have average teachers and those with poor-performing ones is as significant as the difference between those who have excellent teachers and those with average ones, the study found.

It adds up: Replacing a low-value-added (bottom five percent) teacher with an average teacher would raise a single classroom’s lifetime earnings by about $266,000, the economists estimate.

“If you leave a low value-added teacher in your school for 10 years, rather than replacing him with an average teacher, you are hypothetically talking about $2.5 million in lost income,” said Professor Friedman, one of the coauthors.

. . . “The message is to fire people sooner rather than later,” Professor Friedman said.

When a high value-added teacher transferred to a new school, student performance went up in the grade or subject area taught by that teacher, matching predicted gains. Scores dropped in the school the high-value teacher had left. Conversely, scores went up significantly when a low-value teacher left and dropped in her new school.

High performing teachers may more than justify much higher pay,” Slate observes.

“Great teachers create great value – perhaps several times their annual salaries,” write the authors. Now a working paper, the study will be submitted to a journal.

About Joanne


  1. Is it the teacher or the kid and his family and environment? How does the study correct for various influences?

    • It is not going to be possible to control for all influences. One issue that occurs to me is that academic magnet schools are apt to include more teachers who are, on the whole, going to be above average or outstanding and also attract kids who are more serious about school and who have parents who take school seriously. If you were to perform this study in Detroit, for example, even if everything else holds true, Cass Tech and Renaissance High School would skew the “long term gain” results. If you’ve been in a bad school you also have some sense of how different things can be in the classroom of an excellent teacher than in the rest of the school and how the impact of excellence may be less a case of “what happens when a teacher is really good” than, “What happens when the average is that inadequate.” And it’s not difficult to see why, in those schools, many excellent teachers find a way to get transferred, seek jobs in other school districts, or choose to leave the profession.

      The dollar figures sound more significant in the aggregate (“rais[ing] a single classroom’s lifetime earnings by about $266,000”) than on an individual level ($4,600 in lifetime income). If that figure represents the present value of the increased future income, assuming 40 years of work at 2,000 hours per year, we’re not even talking six cents per hour. The low number raises the specter of highly motivated students and parents finding the better schools and teachers, and a small number of high earners being responsible for the bulk of the gain; or, given how little money we’re actually talking about, the possibility that as you keep expanding the study to larger populations you’ll see a regression to the mean.

      Joanne’s point about “It may be difficult to hire more excellent (top five percent) teachers” is well-taken, because we don’t live in Lake Wobegon. Only five percent of teachers can be among the top five percent. In fairness, I expect she was suggesting that we try to recruit more excellent candidates to the field of teaching such that we increase the population of teachers with the skills and personality associated with the current top five percent. But you’re unlikely to achieve that without significantly increasing teacher salaries, and in this era of teacher-bashing, cutting of teacher benefits, and complaints that teachers are overpaid and underworked, that’s not going to happen.

      • It’s not necessary to live in Lake Wobegone to replace the lousiest five percent of teachers. But it is necessary to have a reason to do so which currently is not the case.

        Once there is a reason to replace the lousiest five percent of teachers then overpaid and underworked teachers can be given the boot which should free up more then a bit of funding thus allowing good teachers to be paid more in keeping with their value.

        • If you have a decent pool of average-qualty candidates, replacing the bottom five percent, to the extent that you can accurately identify them and fire them, is easy. But I was addressing the question of how to increase the number of exceptional teachers – I would venture that pretty much every “top 5%” teacher has a job already, or would have little trouble finding a job, and you can’t simply make a wish and have more of them enter the candidate pool.

          Are you proposing that a bottom performing teacher, if fired, would not have to be replaced such that their salaries could be redistributed to other teachers in the form of increased pay? That’s not realistic and, even if such an idea were implemented, the resulting raise would be modest (and the workload of other teachers would be increased to cover otherwise empty classrooms).

          • And I was pointing out that teaching skill’s unrewarded by the public education system. If you don’t reward the skill you purport to value then how sincere is the claim that the skill’s valued?

            Teachers don’t get any rewards, other then a sense of satisfaction, for being good teachers. The best teacher in a school isn’t seen as an organizational resource whereby all the other teachers who aren’t the best can be made better. They’re certainly not paid more for being the best teacher. No one in the educational hierarchy has an urgent need to find the best teachers and retain them.

            God knows the ed schools aren’t hunting up the really terrific teachers to try to quantify and distill what it is that makes them terrific.

            So the question is, why would anyone in a position to pursue the sort of policy that’d see the worst 5% of teachers off bother to do so?

  2. This is tremendous news. Now, I can stop working 60-hour weeks, preparing interactive lessons and engaging projects, while providing lengthy narrative feedback that helps my students develop a thirst for lifelong learning. What a relief!

    If all I have to do is get students to pass a standardized test, which requires that less than half of the questions be answered correctly, I can cut my hours in half by simply putting students on a useless computer program and teaching them how to manipulate the test. Thanks Harvard; I can’t wait for my pay to be tripled!

    Let me just say for any readers who don’t understand my sarcasm that as long as institutions like Harvard continue to base years’ of research on standardized testing and average income, we won’t learn a thing about the value of teachers.

    How do we value the art teacher? If he doesn’t churn out dozens of Rembrandts each year, should he be fired? How about history teachers? The bureaucrats don’t consider history important enough to even test it, so do we fire all of the history teachers? What of music? Again, no test. If the 8th grade orchestra teacher doesn’t produce the next Itzhak Perlman, should she, too, be let go? Perhaps we should simply eliminate the arts, history and anything else that isn’t tested.

    Oh, and what of the science teacher who has a class of 35 students? Maybe eight of them come from good homes with parents who foster learning and a respect for education. The rest are poor, hungry and, perhaps, have learning disabilities. I suppose if this teacher doesn’t show “value-added,” she has to be discarded, as well.

    Also, why is salary used as a measuring tool in Harvard’s study? Who is to judge success by income? I know plenty of people, making far less than the median income in my state, who will attribute what they believe to be successful lives to their great teachers.

    Thanks, Joanne, for giving me an outlet to rage against what is quite possibly the most useless study of all time.

    • So, you must teach self-pity then, right?

      I’ll do you the favor of explaining the larger meaning of that Harvard study, not that it’ll have the slightest impact.

      Teaching skill is coming to be valued which is why the results of teaching – learning – are being examined. The self-serving rationale that posits teaching being too aethereal a skill to be measureable by mere mortals is being edged aside by the drearily-mundane assumption that if it’s worth paying for its worth measuring. Sorry if you have a tough time with that trend but other then venting your spleen there’s just not a whole lot you can do about it.

  3. Mark Roulo says:

    “Also, why is salary used as a measuring tool in Harvard’s study?”

    Because it is (a) quantifiable and (b) considered important by a lot of people?

    • Sadly, Mark, you are right. Many people place far too much value on money. I would ask those people, who is more valuable — the pro athlete making $8 million a year, or the teacher making $50 K?

      My answer would be that we can’t value either person on the dollar amounts. It’s their overall contribution to society that matters most.

      • tim-10-ber says:

        Mark — obviously, may not so to some, the teacher is much more valuable than the athlete…I am amazed at the way athletes are treated throughout school and beyond…passed for doing no work, then paid a fortune because they can play a game well…it is painful to watch Rose of the Bull talk, but he can play basketball. Our priorities are so out of whack!

      • Roger Sweeny says:

        who is more valuable — the pro athlete making $8 million a year, or the teacher making $50 K?

        I’m not sure you really want to go there. Who is more valuable — the teacher in America making $50 K, or the teacher in China making 10 K, or the teacher in Senegal making 5 K?

    • Roger Sweeny says:

      “Also, why is salary used as a measuring tool in Harvard’s study?”

      Because that’s how we market education. And when I say “we,” I include the teachers’ unions and ed schools. We tell everyone that if you don’t get a “good education.” you don’t stand much of a chance of getting a “good job.” We say education is an “investment” in future economic performance. We say that the only hope of America continuing to be the richest country in the world is to spend a lot on teaching and research.

      Alas, some of us teachers try to have it both ways: we think we should be valued (both in terms of respect and in terms of pay) but we also think we shouldn’t be held accountable for how much we add to a student’s future pay.

      It would be extremely courageous to go before a legislative committee and say, “I have no idea whether I help to make my students economically successful, but I deserve to be paid well because that’s not what real education is about.” I do not recall ever having seen that.

  4. “Maybe eight of them come from good homes with parents who foster learning and a respect for education. The rest are poor, hungry and, perhaps, have learning disabilities.”

    Mark Barnes,

    There are a couple serious flaws that you make just in the above statement. First of all, when testing to see the impact of a teacher on a group of students, you do not test to see how a batch of students compares to another batch of students in another class. you campare the growth of a group students from one year to the next. In other words, you are eliminating those variables beyond the teacher’s control and you just focus on how the teacher impacts those students in his/her charge.

    Second of all, you imply that students who have less than ideal family/personal conditions do not value education and foster learning as much as those who come from more ideal circumstances. Just because somebody is poor does not mean that they do not value education, Mark Barnes. To makes that assumption is arrogant and insulting.

    • Swede, have you ever seen any of the work of Stephen Krashen, Ron Ferguson or Ruby Payne? These eminent educators and researchers have studied the effects of cultural differences and poverty on education for decades. Not only am I a 19-year veteran classroom teacher and college instructor, I am also an author and researcher. So, my conclusions aren’t based on arrogance, and I’m sorry if you find them insulting.

      My implication, based on nearly two decades of experience, is that most — not all — students who come from poverty have less interest in education. This makes it very difficult for teachers to be effective. The best teachers find ways to build rapport with these students and show them the value of learning — not the value of a bubble test, filled with questions that have no relevance to the real world.

      As far as your statement about the test, you couldn’t be more wrong. Standardized tests are by design meant to test students against one another. This is one of many problems with these tests. As far as the “value-added” part, if a student gets 50 percent of the questions right, he passes this ridiculous test. If he gets 53 percent right the next year, was this a valuable experience? I suppose the Harvard folks might think so.

      What I find insulting is that supposedly learned college professors, most of whom have never been in the K-12 world as teachers, believe that comparing students in this manner is an acceptable practice. The general public is far too quick to place merit in what these people say, simply because they work at Harvard.

      • tim-10-ber says:

        If I remember correctly Ruby Payne has, hopefully, lost credibility. Themoney our district wasted on her…that is a whole other story…

        Yes, proverty can, can be horrible and as a result children (and adults) can and do make horrible choices. However, we are not our environment. Rich kids make horrible choices and ruin their lives, too, as do middle class kids…

        There are too many examples of teachers, other key adults, coaches, etc. truly connecting with kids and helping them see there are options to how to live their lives and break the bonds of poverty or abusive situations. I know to many that have done this to doubt the impact great schools and excellent teachers (as well as bad schools and mediocre or less) teachers have on kids.

        Poverty might be where the child lives but it is not the child…there are ways out and the best way is an excellent education plus caring adults…

      • You’re wasting your time, the fact that you’re an 19 year veteran teacher only makes the “reformers” on this site value your opinion even less.

        Remember, people like Michelle Rhea, who had to tape children’s mouths shut when she was a teacher, and George W. Bush, who had to bribe a journalist to push his NCLB, know more than you do.

        So do political scientists and economists. We teachers know nothing.

      • Standardized tests are by design meant to test students against one another.

        Wrong.  Standardized tests only test students against the standard.  Any comparison to each other is incidental.

        • I wish you were right. Then, maybe my administrators would stop bombarding me with endless paperwork, showing how my students compare to students in surrounding districts and students in the rest of the state, who took the same test.

          • If your students were being compared directly to those in other districts, they’d be tested in something like a spelling bee.

            We both know this isn’t happening.  All of the students, yours and theirs, are being tested against a standard.  The rates at which they meet the standard are being compared, not their direct performance.

          • They take the same test with the same questions. Of course, their performance on the test is compared. The reports don’t name standards. They only show performance in various strands and how each student performed against others.

            If there is no interest in how students perform against one another, then why publish the numbers in the newspaper? Individual reports are mailed to parents. Why are state averages on those reports, if students aren’t being compared?

            Shielding the real intention by saying students are tested against standards is exactly the sort of thing the government does to make the test look acceptable.

          • Roger Sweeny says:

            Students are being tested against a standard. And then people compare scores to see how students compare with each other. Doing the first will pretty inevitably lead to the second. I don’t see why that would be surprising. And I don’t see why that’s bad.

            In my state, each school gets a report showing how their students answered each question, what “standard” each question was supposed to address, and how their answers compare to the rest of the state. Each year’s test is also released so teachers can see exactly what each question asked.

          • They take the same test with the same questions. Of course, their performance on the test is compared. The reports don’t name standards. They only show performance in various strands and how each student performed against others.

            No, they are compared against the standard.  If the standard is easy compared to their abilities, everyone would score 100% and there would be no way to determine who was better or worse.  There is no Superbowl in test scores.

            If there is no interest in how students perform against one another, then why publish the numbers in the newspaper?

            Because who is and is not meeting the standard is a matter of public interest.  If you don’t think the public has any interest in the matter, stop taking the public’s money.

  5. > Replacing a low-value-added (bottom five percent) teacher with an average teacher would raise a single classroom’s lifetime earnings by about $266,000, the economists estimate.

    It’s a pretty good argument for paying teachers $200K then, isn’t it? I’m sure the unions wouldn’t object… in fact, I think the only people who would object are those people trying to fire teachers and replace them with $25K interns straight out of school.

    So how about it? Where’s the campaign for six figures for teachers…?

    • tim-10-ber says:

      Six figures? Sure plus increased accountability, public reporting of value added scores, easier ability to fire (like the real world), 401(k) vs pension and picking up more of the costs of benefits so not to be a drain on the public coffers…

      I am serious…$100K for working a full 12 months is no problem for excellent teachers…

  6. Lightly Seasoned says:

    Well, I don’t put much value on anything education coming out of Harvard. Maybe they produce good lawyers or businessmen or something, but I’ve never seen an ed school grad out of there I thought was worth anything (and I’ve worked with several).

    If anything, this is an argument for better teacher development in mid-career. I guarantee there are a lot of good teachers who could be great teachers if someone invested meaningfully in their professional development — not the pd circuses and checklist evals. There are some good models out there. Unfortunately, they don’t make for good sound bites and aren’t simple enough for the media/politicians.

    • I couldn’t agree more. In my district, they keep asking us what sort of PD we want; then, they provide the opposite. It’s almost laughable.

  7. Peace Corps says:

    I don’t think great teachers are “made”; they just are. However most teachers can improve with the right support and willingness to learn from their fellow teachers and from their own mistakes. So, not-so-good teachers can become good and good teachers can become better. It is the BAD teachers, those at the bottom that have no interest in becoming better, that we need to find ways to remove from the teaching profession.

    • Lightly Seasoned says:

      I think pretty good is just fine. And I think bad teachers are made. The bad teachers who just are get out quick. The rest are made, I think, by the working conditions.

  8. Just noticed, this “study” was not peer-reviewed, so its conclusions are not shown to be anything more than opinion.

    Combine it with the “study” that for profit colleges are a better deal for the time constrained students and I’m beginning to wonder if Joanne has ANY journalistic integrity left.

    • A correct study is no less correct merely because it has not yet been peer reviewed, and an incorrect study that passes peer review remains incorrect. That said, if your argument is that data is more likely to be found wanting, misleading or incorrect if the authors choose not to submit a study for peer review or their study cannot pass the peer review process, you are correct.

      • Its a “working paper’ which means it hasn’t been published anywhere, but HAS been released to the media as gospel.

        A typical “reform” tactic.

  9. Richard Aubrey says:

    5% terrific teachers might be all there are. Paying high dollars to somebody who is good at something else might not be recruiting a TEACHER, who is somebody who can get the material across.
    It ought to be–which is to say changes need to be made–easier to get rid of the worst teacher in the school. Maybe the bottom 2% every third year.
    Thing is, a lousy teacher means the kids either don’t get the material at all or have to be remediated a the cost of learning the new stuff in the next class. It’s a big cost one way or another.
    The continued employment of a teacher generally considered to be lousy makes the whole system look bad, even to parents of kids who don’t have the teacher.