What Can the Olympics Teach Us About Closing the Achievement Gap?

John List and Uri Gneezy have appeared on our blog many times. This guest post is part of a series adapted from their book The Why Axis: Hidden Motives and the Undiscovered Economics of Everyday Life. List also appeared in our recent podcast “How to Raise Money Without Killing a Kitten.”

If you look at two pictures of two athletes: One is beaming,  the other doesn’t seem too sure what she’s feeling. Which do you think won the silver? Which the bronze? Easy, right? Silver is better than bronze, so the smiling girl on the right must have won the silver.

Not so fast. After all, consider the context: Olympians train for years with nothing but the gold medal on their minds and the silver medalist just barely missed out on gold, whereas the bronze medalist just narrowly made the podium. As it turns out, the ambivalent gymnast won the silver and the happy one won bronze. (And if you think we cherry-picked the pictures here, psychologists have found that this applies to a more general set of Olympic medal winners.)

The idea that context can shape decision-making is a relatively new one for economists, but the cornerstone of this line of thinking is that people will go to much greater lengths to avoid losses than they will to experience the equivalent gain. (Danny Kahneman, author of Thinking, Fast and SlowAmos Tversky, and Dick Thaler, co-author of Nudge, have made seminal contributions to our understanding of this behavioral phenomenon.) For example, suppose you’re at the store and a fancy bottle of wine catches your eye. You notice the price is $50—way out of your price range—and you move on with your errands. Behavioral economists would not be surprised to find that if you later got the same bottle of wine as a gift, it’d take a lot more than $50 to get you to part with it. The idea is just that losing that bottle of wine would be much worse than the price of acquiring it in the first place.

These ideas are incredibly provocative for economists in the abstract, but our work has tried to apply these ideas to get the most out of incentives in environments as diverse as manufacturing plants and inner-city classrooms.

One of the most controversial topics in public education today is teacher incentive pay. We know that good teachers matter, but there’s still only mixed evidence that incentive pay can turn an average teacher into a really good one. The lack of evidence isn’t a coincidence, though. There is a lot of resistance to even testing these ideas, in part out of the fear of the possible findings. For example, when we proposed testing the impact of teacher incentive pay in the Chicago Public School system with bonus payments of up to $8,000, the Teachers Union didn’t hesitate before responding: “Absolutely not.”  The party line is that teachers are trying their hardest already and incentives for performance will not work. 

Of course, not all unions are the same and in Chicago Heights—an economic and demographic microcosm of Chicago, just 30 miles south of the city—we received a more receptive hearing. After extensive negotiations, we were allowed to offer more than 150 Chicago Heights teachers a chance to earn an extra bonus. The experiment was conducted with Roland Fryer, Steven Levitt, and Sally Sadoff.  In one treatment, teachers could earn up to $8,000 (on average, $4,000) if their students’ performance improved beyond a certain threshold. In another, we wrote checks for $4,000 (the average reward) to teachers before the year started with the stipulation that they would have to repay all or some of the money if their student’s performance didn’t improve.

Our results showed that incentives alone weren’t enough to move student performance. But incentives plus context? Well, when teachers were threatened with losing the rewards they had already received, student achievement in math jumped by 6 percentile points and 2 percentile points in reading. 

To put this result in perspective: if the students in Chicago Heights could repeat these gains each year of elementary school, it would be enough to close the entire gap between their achievement and that of wealthier white kids from the suburbs.

If you want to explore our world further, take the Why Axis Challenge: visit www.thewhyaxischallenge.com, post a photo of your copy of The Why Axis, and be entered to win prizes, including a meeting with Uri, John, and Freakonomics author Steven Levitt!

Stay tuned for more posts, which will give a glimpse into more “undiscovered economics.”

Eric H

This is why some version of "at will" employment is so important in schools. At my job, I know that if I don't perform, I can (and should) be fired. I believe that would be the quickest way to improve teacher's performance.


What this dog and pony "research" doesn't address is a core tenant of education: Can a single test measure learning, growth or academic achievement? Further, can this "growth" be sustained? Creating incentives is nothing new in education. Protracted studies within the field, however, have shown - over and over - that the growth, regardless of the incentive, is short-lived. Not everything that counts can be counted; not everything that can be counted counts.

Enter your name...

That isn't an open question. A single test cannot measure everything worth knowing, and no academic-style (written or oral) test can measure everything that a person needs to know and be able to do. There is no test for "keeps his temper under control when provoked" or "is kind to people less fortunate than himself when he believes no one else is looking".

However, these tests can be entirely adequate for measuring whether or not students have learned the things that the state laws require teachers to be teaching them. State law requires teachers to teach their students reading, writing, math, and other assorted academic subjects. It does not require them to teach students to be morally good, interesting conversationalists, emotionally mature, well-mannered while eating, or creative thinkers. Teachers should do their actual, legally required jobs, and not be expected to make up for or tested on whether family, friends, religious institutions, and broader society have done their parts, too.



It's great that we have incentives that are proven to increase scores, but shouldn't we wait to crow about erasing the gap until we've tried offering the same incentives in the white suburban schools? If these incentives work so well and are adopted all over, they might increase everyone's scores by some amount.


My wife works for a large medical group and they do something similar. The doctors are paid a substantial bonus every year, contingent on them accepting certain changes in the practice for the year. These include patient satisfaction metrics, required off-hours meetings, PR commitments on behalf of the practice, extra patient workload, etc.

It is extremely effective. That doesn't mean we like it though... there is considerable social pressure to accept the bonus. In fact, not accepting the bonus might be construed as a lack of commitment to the practice and might affect wages/promotions. And the requirements don't seem to be well defined, or at least not well communicated. There always seems to be a new mandatory commitment we hadn't heard of before. Plus, what if you no longer have the ability to pay back?

How did the teachers respond to this? Did they recommit to the program the next year? Were they able to pay back the bonus when they didn't hit the target?