Why incentives don't work in education—or the business world

No readers like this yet.
open source giving


Even as the U.S. economy recovers from a financial meltdown led by a number of white-collar Wall Street swindlers, critics of the public education system push for more “business” in the classroom: specifically the implementation of incentives and the hiring of CEOs for superintendents.

There's just one little problem. According to 40 plus years of academic research, incentives—and disincentives—don't normally work. And when they do, they often don't make people behave quite like their proponents anticipate.

In June the New York Times published an article exposing a number of instances of cheating on standardized, “high stakes” tests in public schools. And the guilty party? Not the students.

From  Houston teachers with a $2,850 bonus on the table to thirteen educators and 2 principals in Georgia who faced the possibility of job loss, cheating is present at alarming levels among today's school staff. Data forensics specialist John Framer told the New York Times, “Every time you increase the stakes associated with any testing program, you get more cheating.”

For educators and administrators, the stakes couldn't be much higher. In addition to financial incentives that reward top performers, school districts and state programs often include provisions for firing or demoting teachers and principals, while their schools may be put on probation or lose crucial accreditation.

Gaming the system

Beyond ordinary cheating lies a gray area that is concerning to most organizations with an incentives plan. Author Daniel Pink offered an example in his recent Telegraph article that should be familiar to anyone who has worked in sales:

“In the early days of the [Red Gate Software] company, [co-founder Neil] Davidson created a fairly straightforward commission scheme. But, of course, salespeople figured out a way to game it – by pushing sales into the time period most advantageous for them, by underselling one month to show a bigger gain the following month, and so on. This wasn't because they were unethical; it was because they were rational humans responding logically to a particular incentive structure.

So Davidson made the system more complex – and salespeople responded by increasing the complexity of their own behavior. On and on it went, until both the management team and the sales force seemed more focused on the compensation system than on making great software and selling it to customers who needed it.”

When the company decided to scrap the commissions plan and simply pay the sales staff a “healthy flat salary,” the sales team generated more revenue.

In a recent FastCompany article, Dan and Chip Heath, authors of the book “Made to Stick,” said that incentives often yield “collateral damage.” The Heaths point to examples like NFL quarterback Ken O'Brien, whose unfortunate habit of throwing interceptions led a team lawyer to add a clause to his contract that penalized him for each interception.

“The incentive worked as intended,” the Heaths said. “His interceptions plummeted. But that's because he stopped throwing the ball.” When the penalties for failure are high, people take fewer risks, which leads to a reduction in innovation and often overall performance.

According to Freakonomics author Stephen Dunbar, you can't simply “divide people into piles of good people or bad people, cheaters or non-cheaters. The point is that people’s behavior is determined by how the incentives of a particular scenario are aligned.”

This is nowhere more true than in the much-lauded business world. Slate places the blame for the Wall Street crisis and BP's oil spill on “the law of incentives.” In each case, journalist Eliot Spitzer said, placing the burden of risk on the public while businesses kept the profits had the effect of removing all sense of responsibility from executives and led to disaster.

If not incentives, then...
So what motivates workers, if not punishments or rewards?

Open source may have some answers there. Every day, programmers around the world generate high quality code for open source projects—often for free.

Yet the Heaths' article mentions a production-increasing initiative by AT&T executives where programmers were paid for each line of (proprietary) code they produced. They weren't more productive; they simply generated a whole lot of extraneous code.

Open source coders and other highly motivated workers share three characteristics, according to Pink: purpose, autonomy, and mastery. The coders contributing to an open source project—some of whom may well be working for AT&T—are intrinsically motivated to do quality work. It seems that offering external motivation (in the form of money) should further increase their interest in a project. But in fact, the opposite occurs.

Social science researcher Alfie Kohn theorizes that the more an executive gets employees to think about what they will earn for doing their jobs well, the less interested they will be in what they are doing.

“Loving what you do,” Kohn said, “is a more powerful motivator than money or any other goody.”

If Kohn and Pink are correct, the key to “motivating” our teachers and our workers may be rather simple. Pay them enough that they don't have to think much about money, then let them do their jobs as they see fit.

User profile image.
Rebecca Fernandez is a Principal Program Manager at Red Hat, leading projects to help the company scale its open culture. She's an Open Organization Ambassador, contributed to The Open Organization book, and maintains the Open Decision Framework. She is interested in the intersection of open source principles and practices, and how they can transform organizations for the better.


In my mind, the two things that every article ignores when it comes to US education (which in my personal opinion is in a woeful state) is that we don't pay teachers enough for what they do, and we don't give them a feeling of job security. There are good and bad apples out there, but no matter which, if people spend their time worrying about their job security they won't be as focused on taking "risks" and getting rewards (providing a positive learning experience). It's less about money, more about security.

One of the most disturbing thing for me about many of the high stakes testing programs is that many great inner-city teachers lose their jobs when test scores won't rise enough. Yet many of these teachers are far superior to their suburban counterparts... they simply have students that need incredible amounts of help, money, and time to succeed. And the same teachers often have their hands tied to specific instruction methods that don't work all that well, and aren't seen in more affluent schools, like drilling and rote learning.

The teachers who most need the freedom to take risks and try new things have the least freedom to do so, under high stakes testing programs. Meanwhile we push for innovation in charter schools, which lack much *needed* oversight and don't even require many certified teachers, at the same time we deny our brightest and most motivated teachers the opportunity to innovate in their own classrooms.

What business-turned-education "thinkers" don't realize is that for just about every teacher, the greatest incentive is students who are engaged in the material and successfully mastering it. If we'd open our schools to more data-driven methods, let our teachers have the freedom to try new things, and provide the budget to attend conferences where other educators present "best practices" from their classrooms and schools, we might just find that we don't need to dangle carrots in front of anyone.


Thanks for a thoughtful and interesting piece. How you would build incentives for education that provide the public with meaningful accountability?

Your comments about charter schools I read as reflecting an ingrained reluctance on the part of the educational establishment to expose schools to competition within school districts. Without meaningful alternatives, the ability to flee from failing schools, or the sense that educators are serious about improving schools, then yes -- parents or voters will turn to imposing external measures on educators, like standardized testing -- with all that entails.

It may be true that teachers work under stultifying conditions in many areas, and no doubt greater engagement would help. But until there is a greater drive to achieve really world-class outcomes, frustration with the U.S. public educational system will only continue to grow among those who are paying for what now is an expensive and marginally effective system.

I believe in innovation, absolutely, but I think we should build a framework to allow teachers to do it. I also think we need to be careful about using at-risk children for the bulk of our experiments, which is one thing that often happens.

I have gone back and forth about the problem of low student achievement (generally constricted to various demographic groups, for example urban or rural low-income children), and I think the problem is that we expect teachers to work miracles, and we blame them when they don't.

We also fail to look at data, which shows that private schools don't perform any better than public when you control for demographics, and that there are twice as many bad charter schools as good ones.

I have to look at my own kids, who come from a middle class family, and see that I could put them into any school, good or bad, and they'd probably be just fine. They'd have above-average scores if they were in an under-performing urban school, average in a typical school, etc.

Yet my own school district (Wake County) has found that even a good socio-economic integration program that avoids creating bad schools isn't enough to significantly raise achievement in most at-risk students. (That doesn't mean the system isn't without its own merits, like stabilizing the housing market and not having any child in a bad school, of course!)

We have studied the handful of urban low-income schools that truly succeed--both public and charter--and no one can seem to apply their methods elsewhere with success. (Again, when you're comparing apples to apples, as far as student population goes. It means nothing to have great test scores if you're pushing your lowest performing students out of your school in various ways.) To me that suggests that it's largely related to the culture of their school and the personalities of their leaders.

We recognize that the world has only a handful of Albert Einsteins and Ben Franklins, but we don't seem to realize that truly gifted teachers who can move mountains are the same class. I don't believe for a second that you can turn a normal, good teacher into one of the few who could close the achievement gap no matter how much you incentivize or share knowledge with them. I may want to be Einstein, but no matter how much you pay me or how much I study under him, it's just not going to happen for me. :) And every math teacher has watched "Stand and Deliver," but we have very few Jaime Escalantes to show for it.

None of that is to say that we should give up; instead, I think we need to fund our schools better, because low-income areas have notoriously poor equipment, programs, etc. We need to realize that charter schools rarely perform better than public schools if you compare student-for-student rather than assuming the children in both are the same. (Even when they come from the same neighborhood.) We also need to realize that when charters take the most motivated and family-backed kids out of the neighborhood school, we can't blame the teachers at the neighborhood school when the grades drop further.

I guess I'd say the first solution is to make sure every child is in a school that they don't need to escape from. Wake County has done a pretty good job of that, although we're now heading down a darker road and heading back toward the creation of high-poverty schools. I think we're going to see student achievement in our at-risk groups decline as a result, but time will tell.

The second solution I think is something that falls outside of traditional school budgets, and that is that we need to address the problems of poverty that are giving rise to these at-risk children. We need to find a way to reach out to parents, especially young single parents, and give them hope for their own future and their children. I don't pretend to know how, but I think we need to at least study the problem and stop assuming that children who come to school late, eat a free lunch, struggle to keep up with the class, and come back the next day with undone homework have parents that are fundamentally flawed and different from ourselves, and people who can't be helped. We also need fewer at-risk children in any one classroom, so that teachers can give them the extra help they need.

And my last solution is to suggest that we need to listen to teachers who work in low-performing urban schools, instead of assuming they are clueless or apathetic. Many are quite committed and passionate. We need to ask them how to help, and try actually listening and making the changes they suggest, instead of assuming they are shifting the blame. We need to realize that if you took the "good" teachers from a typical suburban middle to upper middle class classroom and put them into the schools that struggle, they'd no longer be "good" teachers on paper.

I don't at all disagree with standardized testing, though of course it needs continuous improvement. I do think that we need to stop with the high-stakes variety, and instead use it to direct help and attention to areas and schools that are falling significantly below the curve.

Thanks for alerting me to this conversation, Rebecca! I agree with much of what you said. I fundamentally dislike the idea of standardized tests, though. They will always reflect the biases of the test-maker (cultural, learning & performance styles, etc.) and so will always privilege the kinds of performance produced by people like that test-maker. I think we'd see a lot less "failure" in schools and students if we broadened the means by which they can demonstrate what they can do.

And however we assess people or institutions, I agree that we need to lessen the stakes and use them as opportunities for support instead. After all, people typically don't do their best work when under threat of serious punishment. Unfortunately for the high-stakes testing hawks, Campbell's Law is not the kind that can be repealed!

Thanks for the long and thoughtful answer. Your comments about demographics and about your kids' ability to achieve struck a chord. The role of culture and values is a rightly sensitive one in American society, but it strikes me that there's a lot going on behind the scenes and at home that makes a difference in kids' outcomes -- maybe the decisive difference.

I still think you're letting the educational profession and bureaucracy off too lightly for their own contribution to the mess in many school systems, since educators and educational administrators are as ferociously narrow-minded an interest group -- at least in some places -- as any old-school industrial union. But I agree after reading your response to my post that there are also larger social issues that are being glossed over in the discussion about "fixing" the schools -- and holding teachers accountable for not magically fixing larger social issues is something I'd agree is unfair to teachers. Unfortunately, the standardized testing fetish is about as far as we can get until there's a more nuanced discussion of what's going on, or not going on, in our schools.

Good luck, and thanks for a thought-provoking post!

Low teacher salaries mean that more talented people are drawn to other more lucrative fields early on. All the minor incentive adjustments after people make career choices are unlikely to have effects large enough to be measured. You can't expect to be able to motivate people with money when they've already decided that they will forgo money to teach.

Alfie Kohn hits the nail on the head. I, and many other teachers, didn't choose our work with money in mind. We're driven by the opportunity to make a positive difference in other people's lives. We're driven by meaning. (And many of us are extremely talented, thank you very much; we're just not shallow or greedy! We had the opportunity to choose other paths, but decided that teaching was our calling.)

All studies you've cited are basically observations by mostly empirical observers - outside lookers on. And yet what they all point to is an inner requirement of life - of all organisms - discovered by many biophysicists like Mae Wan Ho pointing to this inner coherence guided by value dynamics/vibrations that we all feel, know by our own inner experience of what we know as "good" and here we find definitions of "good" are as varied as there are individual entities. So why do we find in Open Source , Open Space processes that some form of "good " emerges" and draws people to being their own best "good" they can be if given the space and opportunity to do so? I submit it is because of the inherent desire to create - to be challenged by ones own inner autonomous drive to express self uniquely and in so doing find meaning - quality is part of that meaning - random acts of expression may provide little except distraction for a moment, but leave one feeling rather empty if that's all that occurs. It is not competition, pushing against others to "win" necessarily although there is an element of such as we work collaboratively to get the "best" solutions harmonizing what the greatest number of best efforts can bring into coherence. Complexity theory falls short also in explaining this process because it often relies on belief in "cause and effect" do this and get that steps in bringing about solutions for life and life has quite other ways of functioning. Check out movie In Search of Bobby Fisher to see the effects of competition without the higher values of meaning and diversity.

Not to revive an expired thread, but I was recently pointed towards a site about poverty reduction research, which contained this study summary:


The study is interesting for the parallels to some of the issues addressed (or not addressed) in the discussion over U.S. education: teacher qualifications/conditions of service, class size, test scores and basic academic preparation. It also suggests that even in a very poor context, there are incentives and tools available to improve education in a cost-effective setting.

Unsurprisingly, the bottom line seems to be:

<em>To get the most out of these teachers, implementation details matter. The biggest gains come when local school committees are empowered to effectively monitor these teachers and when extra classes are structured so as to target instruction to students’ initial achievement level.</em>

I may be cynical, but it strikes me that those are some basic principles which are often overlooked in the argument over education here.

But as mentioned in the article, it is important to be diligent and conduct some research before choosing an acceptable <a href="http://www.incentivelogic.com/index.php" title="incentive programs">incentive programs</a> are a great way of keeping employees motivated and accomplished at the same time. More companies need to look into investing into such beneficial programs.

Creative Commons LicenseThis work is licensed under a Creative Commons Attribution-Share Alike 3.0 Unported License.