The Harvard Crimson reports on a new study that concludes that "teachers and school officials cheat in administering standardized tests in a minimum of 4 to 5 percent of elementary school classrooms." Although the headline blares that "High stakes tests lead to cheating," the conclusion is probably not what testing opponents hoped for:
The authors, Kennedy School Assistant Professor of Public Policy Brian A. Jacob and University of Chicago Professor Steven D. Levitt, concluded that local policies attributing more weight to standardized testing made it more likely that teachers would cheat...
Economics Professor Caroline M. Hoxby ’88, who specializes in education, said that it was important to remember that the study’s conclusions were based on inferences. Neither of the researchers actually observed teachers tinkering with tests.
But even with that limitation, Hoxby said the study adds critical information about the current emphasis on high stakes testing. “Before this relatively new era, people just didn’t worry about cheating,” she said.
Jacob said he hoped the study would lead to changes in the current standardized testing system.
“In the future, we hope to try to prevent this kind of behavior by having external monitors, as well as by performing random audits, to discourage these kinds of acts from reoccurring.”
In other words, we notice cheating more now because tests carry more weight, and while the raising of stakes might cause some teachers to cheat, that increased prevalence isn't necessarily an indictment of the tests, or the stakes. This merely suggests that stakes were raised, but the corresponding security controls were not; in any such situation, I'd think you'd see more cheating. Dr. Jacob's conclusion is correct; how do we institute security measures to prevent this from happening in the future?
Here are abstracts for Levitt's recent papers on catching cheating teachers and the prevalence of cheating teachers. Some of you may recognize Levitt's name; that's because he's been mentioned here before, as the Chicago professor who's figured out new ways to detect cheating behavior. Apparently the Chicago public schools provide quite a rich dataset for detecting such behavior.
Posted by kswygert at November 24, 2003 10:48 AM