Most high school science students are introduced to Boyle’s Law, which describes the relationship between volume and pressure of gases. Many people have heard of Moore’s Law, the idea that the complexity of computer chips doubles every two years. And anyone having a bad day is all too familiar with Murphy’s Law, which, if you didn’t know, is apparently attributed to a frustrated engineer working on high-speed rocket sleds for the U.S. Air Force during the 1940s.
But there’s another eponymous axiom the world should probably know a little better, too. Social scientist Donald Campbell coined an adage known as Campbell’s Law that says, interpreted one way, that incentives corrupt. The more we try to use a quantitative measure or test to make decisions—such as whether to pay a bonus or how high of a reward to give out—the more subject it will become to corruption.
I first read about it in The Daily Beast’s coverage of the investigation into standardized test results at Noyes Elementary, one of the schools touted as a model success story of former D.C. schools chancellor Michelle Rhee. Dana Goldstein writes that “the more punishments and rewards—such as merit pay—are associated with the results of any given test, the more likely it is that the test’s results will be rendered meaningless.”
Who knows what really happened at Noyes Elementary, where six of eight classrooms were found to have very high numbers of erasures on their tests, as wrong answers were changed to right ones. A USA Today investigation first turned up the irregularities from 2007-08, which included one classroom that averaged 12.7 wrong-to-right erasures per student. The average erasure for seventh-graders in all D.C. schools on that test? Less than 1. Meanwhile, test scores rose significantly at the school, and teachers each received $8,000 bonuses.
The implication, then, is that students or teachers, motivated by the high bonuses for improved test scores that were part of Rhee’s performance-driven education reforms, may have changed the answers on the tests after the fact. Education testing experts consulted by USA Today said that while the erasures were not proof of cheating, the statistically high numbers suggest the answer sheets may have been tampered with, and should be investigated.
The results have been investigated, and apparently will be again. Under Rhee, the District hired a test security firm to look into the matter, which said it did not find evidence that staff messed with students’ answers. Current chancellor Kaya Henderson (once Rhee’s deputy) defends the security firm’s findings, but has also passed the matter to the D.C. inspector general for further investigation to remove any doubt. Meanwhile, D.C. Council Chairman Kwame R. Brown (D) is calling for hearings on the matter.
Rhee herself has said USA Today’s investigation lacks credibility and centered on one school . She questions why people assume that when academic achievement rates in inner-city districts like D.C. go up, that “it’s because of something like cheating, which in this case is absolutely not the case.” The testing firm also attributed the outsized erasure numbers to better test preparation, which prompt students to go back and review their answers before finishing.
At this point, it’s impossible to tell what really happened with these tests. But it begs the question: Would there be as much attention paid to whether teachers might have been motivated to cheat if there weren’t bonuses dangled before them for higher test scores? The debate over merit pay for teachers is a controversial one, and for good reason. Studies have shown that it can result in teacher morale problems, too much focus on “teaching to the test,” and yes, even cheating by teachers to boost test scores.
Teachers aren’t the only ones whose behavior is changed by bonuses. Duke University behavioral economist Dan Ariely has shown that offering large bonuses to employees doesn’t necessarily improve performance—and in fact, it can negatively affect it. The pressure induced by the possibility of a high bonus can actually cause performance to slip. One need look no further than the financial crisis to see what kind of chaos can be wrought by the magnetic lure of a fat bonus.
I certainly don’t believe merit pay should be entirely done away with, whether in education, on Wall Street or anywhere else. Paying for the results of good quantitative measures—within limits—is a good motivator, as long as it’s properly balanced out by other assessments, such as the quality of someone’s leadership skills, the strength of their character and their ability to produce original ideas. Only time will tell whether any improprieties took place among D.C. teachers under Rhee’s reforms. But I’d guess there would be a lot fewer questions if there had been a lot less emphasis, both financial and otherwise, on test scores.