What with July’s bumper crop of headlines — homies Rep. Michele Bachmann and former Gov. Tim Pawlenty in Iowa, the debut of arena rock at TCF Bank Stadium and the great state of Minnesota coming perilously close to running out of beer, you might have missed news that widespread cheating by educators was found in Atlanta schools.
According to a state investigation, as summarized in Slate, “178 educators in 44 of the district’s 100 schools had facilitated cheating — often with the tacit knowledge and even approval of high-level administrators, including Atlanta’s award-winning former superintendent Beverly Hall, who conveniently parked herself in Hawaii for the investigation’s denouement.”
Maybe it’s Hall’s seemingly “let them eat cake” attitude that’s vaulted this one into the headlines, but Georgia’s is just the most recent of a number of investigations to suggest cheating is a predictable, if deplorable, response to high-stakes tests.
Particularly since we have arrived at the point where a school’s failure to make Adequate Yearly Progress on the tests mandated by the widely reviled No Child Left Behind Act have begun to have serious financial and practical consequences. And, equally particularly, given our seeming inability to agree to throw out the expensive, time-consuming tests, which measure very little of use.
Evidence of cheating elsewhere
In addition to Atlanta, statistically absurd test results and subsequent evidence of cheating by educators has been uncovered in schools and districts in New York City, Los Angeles, Arizona, California, Colorado, Florida, Michigan, Ohio, Texas and, of course, Washington, D.C., where the identity of the villain behind “Erasergate” is still being debated.
(Not on that list: Cleveland, which has posted impressive, apparently solid gains using replicable strategies; more on that in this space as soon as we’re done parsing the compromise K-12 bill.)
The Slate story on the Atlanta scandal, and a March USA Today investigation it cites and links to, offer terrific recaps for those, so I won’t do more than offer the gist:
- U.S. Education Secretary Arne Duncan and other test-defenders aren’t convinced that testing policy, with its penalties, is driving the cheating;
- No Child Left Behind is the redheaded stepchild of the so-called “Texas Miracle,” the supposed revolution spurred by former Bush administration Education Secretary Rod Paige when he was superintendent in Houston. That widespread cheating was found in that city’s schools in 2003 did not dampen W’s resolve to make tests and penalties the law of the land;
- Historically, more schools and districts have claimed their test scores outpace their district/state/national averages than is ever possible, something researchers call the “Lake Wobegone effect.”
“We know that great teachers can and do improve their students’ test scores,” Slate’s Dana Goldstein concluded. “And bad apples certainly do exist, and must be rooted out. But we have to acknowledge that their shenanigans have been incentivized by federal and state education policies, which more and more reward teachers and schools for producing high test scores — not knowledgeable, well-adjusted children. The sad thing is, incentives to cheat will only increase if the Obama administration gets its way: Its education programs, such as Race to the Top, ask states to create new standardized assessments for the full range of grades and subjects, and to tie teacher and principal evaluation and pay to students’ test scores.”
To this might I add: When one school system claims to have achieved eye-popping results, it makes the educators painstakingly pushing scores up a few points at a time seem like indifferent dimwits.
Anybody besides me recall watching Minneapolis’ Bernadeia Johnson (who, by the way, doesn’t make enough to hide in paradise for very long) break down in tears at a school board meeting last fall over her district’s failure to post more than low double digits? She was totally right to wince with pain over the bright futures being squandered, but perhaps overly shamed by the idea that others were performing miracles by sheer dint of will.
Think we’re immune? In 2003, during Cheri Pearson Yecke’s reign, Minnesota experienced a minor testing scandal when the then-brand-new Minnesota Comprehensive Assessments were scored erroneously. The state’s testing director tried to tie the results of the then-new Minnesota Comprehensive Assessments to new state standards.
In the process scores were inflated; when it was pointed out, the functionary in question attempted to brush the error under the rug. A different set of high stakes, certainly, but every bit as political freighted.
Perverse incentives
Lastly, let’s consider the perspective of Jim Angermeyr, director of research and assessment for Bloomington Public Schools, developer of one of the so-called growth model tests policymakers are touting as the solution to the NCLB juggernaut and decidedly ambivalent about tests as we now use them.
“The more high-stakes the test, the higher the pressure on our teachers to cheat,” he told MinnPost.
Cheating’s bad, he continued, but not nearly as widespread and insidious as other effects of the tests’ pressures: Everything from schools and teachers that spend too much time prepping to small-scale thumb-on-scale infractions by teachers who send kids back to their seat to “check again to make sure you’re done.”
And the incentives could easily get more perverse. “As soon as the stakes go way up, teachers think of the other things that might be affected,” Angermeyr said. “This is one reason why pay-for-performance stuff is so frightening. Imagine if your salary is going to start to be based on test scores — man, the stakes go up …. Some places 90 percent of the day is spent on teaching to the test. Arts and music get left to the side.”
Nor, in his opinion, will the clause in last week’s legislative K-12 education compromise that mandates that 35 percent of teacher evaluations be based on tests actually drive progress — not least because the new law allows districts to decide what tests to use to comply. This is no small thing, given that we test math, reading and science but will evaluate teachers of art, social studies, geography, health and so on.
“How fairly will teachers be evaluated when that data’s not evaluated?” he asked. “Now we’re going to be creating tests for kids just to evaluate teachers, and not kids.”
Used this way, he added, tests no longer fulfill the actually useful function of allowing educators to draw inferences about student performance.
