Pay for performance improves performance

In an Arkansas study, students improved scores in low-income, high-minority schools where teachers received bonuses linked to students’ progress. At schools with similar demographics but no merit pay, scores declined slightly.

This gain in achievement after one year’s time is roughly equal to one-sixth of the nationwide average test score gap between black and white students. If the observed benefit of the merit pay program were to compound for six years, it would close the black-white test score gap.

Teachers in the merit-pay schools didn’t report working harder or being more innovative than comparison teachers. The experiment didn’t produce more “counterproductive competition.” In fact, teachers in the pilot schools reported a more positive work environment than the control group and said the program increased collaboration.

This strikes me as very significant:

The ACPP (merit pay) teachers were more likely than comparison teachers to view low-performing students as an opportunity to demonstrate teaching ability rather than as a burden.

Teachers in the merit-pay schools could earn as much as $11,000 extra; school staffers earned a bonus based on schoolwide improvement.

About Joanne


  1. wayne martin says:

    The following study casts a bit of a different light on this:
    Cutting Provisions In Union Contracts Could Free Funds
    By Jay Mathews
    Washington Post Staff Writer
    Monday, January 8, 2007; B02

    U.S. public schools could have as much as $77 billion more a year to improve teaching if they reduced spending on seniority pay increases, teacher’s aides, class size limits and other measures often found in teacher union contracts, a new study contends.

    The report from the Washington-based think tank Education Sector analyzed research on eight common provisions of contracts that require schools to spend substantial sums but have what the report called “a weak or inconsistent relationship with student learning.”

    The provisions include salary increases based on years of experience or educational credentials; professional development days; sick and personal days; class size limits; use of teacher’s aides; and generous health and retirement benefits.

    The study points out that billions are being spent on “higher pay”, but the results for these salary augmentations does not seem to be paying off.

  2. Express, worthwhile recognition of skill results in improved performance?

    Stop the presses!

    I know in five or ten years we’ll all look back and have a good laugh at the notion that rewarding people for doing a good job was a bad idea.

  3. Let’s not pop open the champaign just yet. It is good when any district or state reports success with any initiative. But this appears to be only a first year phenomenon. Let’s see how well this and other districts undertaking experiments in merit pay do over a couple of years. Cautious expansion of the initiative to a few more school districts seems justified. Bloodbaths with skeptical teachers unions to force widespread adoption of merit pay is not.

  4. Just one more brick in the wall Prof210. One more reply to the folks trying to paint the most lurid, frightening picture of the future should the public deviate from the path of righteousness.

    As to spreading to other districts, NCLB may provide some impetus but other then that, this whole procedure consists mostly of common sense. Common sense has been around since before NCLB so I wouldn’t worry about an over-zealous replication of the idea.

  5. Remember – people unwilling to go along with a new policy can kill it. Happens all the time in large organizations. Motivation is an individual thing, and you have to motivate the change.

    Implementation of improved policies and methods depends upon the willingness of individuals to adopt them, and the results will depend upon the type of people who are involved. Remember the results of Head Start – what worked with a small number of dedicated enthusiasts may not work when spread across a burocracy of millions of uncaring parents and mediocre teachers.

    That being said, positive results are always encouraging.

  6. The concern some people have about “merit-based” pay is on objectively defining what achievements earn the extra pay. In one way or another, any measurement scheme that defines performance relies on data and statistics and we all know that data can be fudged and statistics can be manipulated.

    Explain to me what a great teacher is: can all of those attributes me well identified and evaluated? Certainly academic performance is the most important category for measurement but is the only one and if not what else matters and how much?

    It’s nice to use rhetoric that says “better pay for better performance” but true change and progress requires answer tough questions that the people using rhetoric ignore or overlook.

  7. Tom, let me explain how this works out in the real world. Performance is no easier to measure, but most of the time employers manage to do it. Bosses have to evaluate their employees, and promote, give raises, and fire so as to keep a workforce that does the job efficiently. If they fail to do this, their business loses money, and they are looking for a new job.

    I agree that teacher performance evaluations are going to be a joke until non-performing principals can expect to be fired – and considering the state of the school administrations I’ve seen, I could expect a school system trying to mend it’s ways to have to cut most of it’s administrative jobs permanently, and try several people in each of the others before they found enough good ones.