It was a simpler time when the biggest worry schools had was whether students were peeking at crib notes. More sensational—and more worrisome—is the spate of recent reports about educator cheating on tests.
Atlanta, Philadelphia, and Washington, D.C., are among the districts facing scrutiny about the integrity of their test results. Unfortunately, that doesn’t mean that the problem is confined to schools with the misfortune of winding up in newspaper headlines. And, any superintendents or school officials who think cheating doesn’t exist on their turf has already demonstrated one symptom of the problem: denial.
As an academic with a long-standing interest in the problem of cheating, I’ve had far too much experience with the issue (including assisting in the recent Atlanta investigations by, among other things, actually looking at hundreds of bubble sheets to verify a computer’s judgment that an erasure had occurred).
The most common and persistent question I hear, “Why do educators cheat?” recalls the old joke about why people rob banks: It’s where the money is. Cheating on tests occurs because there are positive consequences for high scores and negative consequences for low ones. In fact, because bonuses and other incentives are often tied to student achievement, cheating is sometimes quite literally “where the money is.”
Incentives matter. As long as important educational decisions are informed in part by test results, cheating will occur. It’s a huge mistake, however, to think that the incentives to cheat in school settings are fundamentally any different from those in any other context. There’s money to be saved by overstating deductions on tax returns; fame and fortune for Major Leaguers who use steroids to improve their sliders or their batting averages; temptation to plagiarize to keep the prestige associated with reporting for The New York Times; and political power to be gained through voter fraud. All such cheating would go away if we did away with baseball, taxes, newspapers, and free elections. Short of that, there will always will be some—hopefully small—percentage of folks who cut corners. Education is no exception.
“Can anything be done to stop cheating?” Sure. There are two potent ways: Remove the incentives, or kill the messenger. The former would mean that we should eliminate bonuses for outstanding teaching. Eliminate evaluations of school performance. Eliminate judgments about which educators are going above and beyond to help kids and those who are actually reducing a kid’s chances of educational success. We could even go back to Lake Wobegon and claim that not only is every student above average—so is his or her teacher.
But why stop there? The easiest way to stop cheating on tests would be to eliminate tests. Although it’s a surefire way to stop cheating, it is also perhaps the dumbest idea of all. As long as important decisions must be made, is there any reasonable argument against demanding that they be made based on high-quality information? For all the yammering about bias, lower-level thinking, bland regurgitation of facts, etc., the fact is that the statewide assessment kids take in April or May is almost certainly the least biased, most rigorous, fair, comprehensive, accurate, and reliable test they’ll see all year.
We’ve also learned that cheating can be addressed, although doing so is going to require modest amounts of funding and large quantities of resolve."
Ironically, the primary reason we are even aware of the extent of cheating is the quality of the information yielded by current tests and the wide range of important educational and policy questions that those data can help answer. The recent findings of unethical behavior by educators are shocking, but beyond the sensational aspects, there are important lessons to be learned from those investigations.
For one, it appears that cheating is limited to a small percentage of educators. For instance, the best data available suggest that it is a serious concern in 4 to 5 percent of elementary school classrooms. That means that the vast majority of educators are acting ethically, playing by the rules to help all students succeed. For another, it’s clear that exclusively blaming teachers is unfair. Counselors, test coordinators, assistant principals, even superintendents are among the co-conspirators.
We’ve also learned that cheating can be addressed, although doing so is going to require modest amounts of funding and large quantities of resolve. Before considering some specific steps for addressing the problem of cheating, however, the first step is to stand back for a moment and contextualize it.
It is important to recognize that accountability systems are still relatively new to K-12 education, and that there are many improvements to come. For example, higher education is marked by fairly strong accountability (think “publish or perish”), with faculty evaluations and raises often based exclusively on three outcomes: quality of teaching, research productivity, and amount of public service.
Although not perfect, that system has been around long enough that it has become fairly well entrenched and, for the most part, accepted. In contrast, nascent K-12 accountability systems rely primarily—or even exclusively—on group test scores and aggregated status of students reaching specified performance levels. Including student achievement seems essential, but improved educational accountability models of the future will almost certainly be based more on individual student progress, and be broadened to capture other valuable educational outcomes.
For the short term, states and districts can take plenty of steps to address the problem of cheating. They include:
• Incorporate less corruptible testing approaches. Performance tasks, essays, and other constructed-response formats can tap knowledge and skills not easily measured by other formats. Although alternative formats are often much more expensive, more time-consuming, and less reliable in the information they provide, they yield clear security advantages: It’s much easier to alter a bubble sheet than to erase a student’s essay and write a better one.
• Invest in test security. It’s clear that having accurate, objective information about student achievement is essential for educators, policymakers, parents, and students themselves. It’s an inconvenient truth, however, that large-scale assessment is actually quite a tiny fraction of education spending. And from my experience, the resources currently allocated to ensuring the quality of assessment data represent a tiny fraction of a tiny fraction. In a June 24 letter to chief state school officers, U.S. Secretary of Education Arne Duncan reminded states that federal funds can be used to promote test-score integrity, and said: “I urge you to make assessment security a high priority by reviewing and, if necessary, strengthening your efforts to protect assessment and accountability data, ensure the quality of those data, and enforce test security.”
• Take advantage of technology. Computer-administered tests are typically much more secure than paper-and-pencil exams. Although computer-based testing presents other challenges, the problem of erasures disappears completely. Moving to computer-based testing carries considerable costs, but the collateral benefits of increased student interest, reduced testing time, and improved psychometric precision go far beyond security advantages.
• Adopt policies on ethical testing. Many states and districts don’t even have the minima in place. Policies that spell out appropriate and inappropriate testing practices—and detail specific consequences for the latter—should be adopted by state and local boards of education and broadly disseminated. Sanctions for violating those policies should be fair, uniform, strict, and meaningful.
• Develop standardized approaches. Currently, a wide range of statistical analyses and procedures are used to identify erasures, unusual score gains, and other indicators of possible cheating. Developing, testing, and applying common procedures across contexts can help us better understand and respond to the problem.
• Support ethical behavior. On college campuses, we have learned that honor codes are effective not because of the penalties they prescribe, but because they contribute to a sense that the broader campus culture values ethical behavior. In K-12 settings, the absence of frequent and explicit communications about ethical testing can foster a culture in which it is assumed that cheating is OK. Silence by school leaders can be interpreted as assent. Real leaders must begin to articulate clear and unequivocal expectations about ethical behavior and fully support those who identify wrongdoing.
Although cheating will never be eradicated, any of these steps will help. Implementing all of them will help a lot. It’s a price that should be paid for obtaining high-quality information about student learning and viable accountability systems. Blaming tests or accountability systems for things we don’t like is a dumb idea. Like banning thermometers for revealing fevers.
Of course, the worst idea of all is to cheat in the first place. It’s a lose-lose-lose situation for all involved. Students and their parents get inaccurate information about their learning. The same bad information feeds into decisions about student placement, instructional priorities, and curriculum decisions. Policymakers are misled about the efficacy of reforms and funding strategies. Even the public—which pays for high-quality tests to measure student achievement—gets ripped off. It’s time that parents, educators, policymakers, and all concerned demand that the high-quality tests they pay for yield the high-quality information that all concerned need. That’s a good idea.
A version of this article appeared in the August 10, 2011 edition of Education Week as Cheating on Tests and Other Dumb Ideas