Cheating on tests: Is it widespread? Is it the wrong issue?

Indiana and other states should be doing more to detect schools that are cheating on standardized tests, according to a nationwide reporting project led by the Atlanta Journal-Constitution newspaper.

But education officials and at least one researcher have questioned the AJC’s methodology and conclusions, suggesting they focus suspicion on schools that probably haven’t done anything wrong.

The newspaper looked at test scores in schools across the country to try to identify anomalies like those that touched off an investigation of cheating in Atlanta schools. That probe found that at least 178 Atlanta educators tampered with tests.

AJC reporters flagged schools with year-to-year test scores changes that seemed too big to result from chance or good or bad teaching. For example, if test scores for a school’s fifth-graders in 2010 were markedly different from those of the same school’s fourth-graders in 2009, suspicions were aroused.

“The analysis doesn’t prove cheating,” the newspaper said. “But it reveals that test scores in hundreds of cities followed a pattern that, in Atlanta, indicated cheating in multiple schools.”

According to the article, a school should expect to have up to 5 percent of its classes experience unusual changes in test scores from one year to the next. Districts that consistently have more than 10 percent of their classes flagged for unusual changes merit further scrutiny, it said.

But Western Michigan University researcher Gary Miron, writing for The Answer Sheet blog at the Washington Post, argues that the AJC’s methodology is flawed, and what may look like cheating, on further examination, probably isn’t.

Miron consulted with USA Today on its prize-winning investigation of possible cheating in Washington, D.C., schools, which looked at year-to-year changes in test scores for individual students as well as rates of erasures on the tests. He said the AJC’s approach – relying on class-level data, not individual scores – probably results in comparing test scores for different groups of students. Continue reading