Why The Atlanta Testing Scandal Matters
Once, in a sauna at a Korean spa in Queens, I overheard what sounded like two teachers discussing the cheating practices of a third. "You know how she does it," one said. "She'll lean over a student about to put a wrong answer and whisper, 'Check your work.' "
"Yes, and her finger will just happen to be on the right answer," said the other one.
I thought of that exchange when I read that jury selection is underway in Atlanta for an unusual trial. The defendants are 12 former employees of Atlanta Public Schools. They are accused under the state's racketeering laws of conspiring to falsify their students' results on state standardized tests. Dozens more school employees have faced ethics sanctions in a case that has rocked the city of Atlanta for the past few years.
The trial is unusual. It's likely that millions of dedicated teachers around the country spend their entire careers without engaging in the kind of behavior that happened in Atlanta, or that I heard about in that spa.
But high-stakes state standardized tests of this kind are not unusual. They are mandated in nearly every public school by No Child Left Behind, the 2001 federal education law. These tests are high-stakes because they trigger serious consequences for students (like grade promotion and graduation); for schools (like extra resources, reorganization, or closure); for districts (the loss of federal funds); and for school employees (bonuses, demotion, poor evaluations, or firing).
And so the Atlanta trial should bring two questions: How common is cheating on these tests? And short of cheating, what else might be happening in schools as a result of these tests?
The answer to the first question is: No one really knows. But a Government Accountability Office report last year found confirmed reports of cheating on at least one standardized test in 33 states in the school years 2010-'11 and 2011-'12. Thirty-two of the 33 states, the report said, "canceled, invalidated, or nullified" test scores as a result of suspected cheating.
A 2003 paper by the economists Brian Jacob of the University of Michigan and Steven Levitt of the University of Chicago analyzed test answers from the Chicago Public Schools. They found evidence of serious cheating in 3 to 5 percent of elementary school classrooms. That doesn't sound so bad.
But, the experimenters cautioned, their methods captured only the most egregious kinds of cheating: those that came from teachers or administrators actually altering test papers. It's this type of outlandish behavior that's alleged to have occurred in Atlanta, where some schools allegedly held "erasing parties" to change answer sheets behind closed doors.
Jacob and Levitt note that there are other, more subtle ways teachers can cheat, such as giving students more time, or prepping students beforehand with answers to actual test questions.
And that brings us to a broader point. Research shows high-stakes testing can also produce unintended consequences that fall short of outright cheating. Daniel Koretz, the Henry Lee Shattuck Professor of Education at the Harvard Graduate School of Education and an expert in educational testing, writes in Measuring Up: What Educational Testing Really Tells Us, that there are seven potential teacher responses to high-stakes tests:
1. Working more effectively (example: finding better methods of teaching)
2. Teaching more (example: spending more time overall)
3. Working harder (example: giving more homework or harder assignments)
4. Reallocation (example: shifting resources, including time, to emphasize the subjects and types of questions on the test)
5. Alignment (example: matching the curriculum more closely to the material covered on the test)
6. Coaching students (example: prepping students using old tests or even the current test)
Strategies 1 through 3 pretty much describe what high-stakes testing is supposed to do: raise standards, ignite harder effort from teachers and students, and produce more learning.
Strategies 6 and 7 clearly undermine the effectiveness of tests as a metric of learning, and hurt students in the process. Perhaps 95 percent of educators will never go there.
Strategy 4 (reallocation) and 5 (alignment) are ambiguous. If the test is high quality — if it captures all the most important subjects students need to know — then changing school to prioritize those subjects is, again, exactly what we want to see. In other words, if the test is excellent, then "teaching to the test" can be a very good thing.
On the other hand, if the test captures only a few of the subjects students need to know, or emphasizes, say, memorization over comprehension, then reallocation and alignment can cause students to miss out on other important parts of learning.
There's another kind of reallocation that may go on as well: the reallocation of teaching itself.
Jennifer Booher-Jennings, a sociologist at New York University, describes this as educational "rationing" or "triage." In a pair of research papers published in 2005 and 2006, she draws on field research undertaken at a Texas elementary school after the passage of No Child Left Behind.
She documented teachers diverting resources and time to the group of students who looked like they were closest to passing the test. These kids, who might make the minimum passing score if given intensive help, were described as being "on the bubble."
Correspondingly, the teachers gave less attention to a lower-scoring group that they were encouraged to label as "hopeless cases," as well as the kids who would pass the test no matter what. The upshot was that a relatively small number of students received a disproportionately high amount of the teaching and resources available.
We teach our children not to cheat. The tiny minority of teachers who cheat cross an ethical bright line that harms the entire enterprise of education. But the Atlanta trial should be an opportunity to consider what might be happening in the gray areas as well.