The cheating scandal that has rocked the 48,000-student Atlanta school system was an egregious, but not entirely unexpected, byproduct of accountability pressures, many testing experts say.
The reason: As long as test scores are used in any field to make decisions on rewards or punishments, including for schools or educators, a small percentage of people will be willing to bend the rules—or break them.
But the allegations of systematic test alteration by teachers and principals in Atlanta, along with recent accusations of cheating in Baltimore, the District of Columbia, Philadelphia and other districts, have highlighted a split between those arguing for improved test management and security and those who ask if it’s better to scrap high-stakes testing altogether.
Yong Zhao, the associate dean for global education at the University of Oregon, in Eugene, used the Atlanta situation as a jumping-off point for a five-part series in his blog Zhao Learning. The United States should “ditch testing,” he believes.
In an interview, he said that the country should move to a portfolio-based assessment system that measures students in several areas.
“You can’t fix this by changing internal security,” Mr. Zhao said. “If the stakes are so high that the teachers don’t even believe the measurement itself, they’re going to try to cheat.”
In contrast, Gregory J. Cizek, a professor of educational measurement and evaluation at the University of North Carolina at Chapel Hill, who assisted state investigators in their Atlanta probe, said that there are no suggestions that sports should be eliminated just because some athletes cheat. In his view, tests produce high-quality information that educators need to make good decisions.
States and school districts have grappled with a number of cheating scandals in recent years amid an increased emphasis nationally on high-stakes testing, including in the No Child Left Behind Act, passed in 2001. Here is a sampling.
Montgomery County, Md.: Several teachers and a principal are removed from a middle school after a teacher distributes copies of a state standardized test to her colleagues. The teacher says it was a misunderstanding.
Utah: At least 15 teachers are referred to state investigators for changing answer sheets and otherwise compromising test security on the state’s first year of school accountability.
Chicago: Educators at seven elementary schools are believed to have engaged in testing irregularities on the Iowa Test of Basic Skills. Arne Duncan, the schools chief at the time and now U.S. secretary of education, used a statistical model to find testing anomalies.
East St. Louis, Ill.: Administrators are accused of ordering test sheets and answer booklets for standardized tests directly from the test company, a violation of test-security policy in the district.
California: The Los Angeles Times, using a public records request covering several years, finds that nearly 200 teachers statewide since 1999 have been warned, reprimanded, or fired for providing test answers to students or allowing them more test time.
Camden, N.J.: State investigators blame “adult interference” for suspiciously high test scores at two elementary schools in the district, but do not accuse anyone of cheating. A Philadelphia Inquirer investigation says cheating in the district was extensive and ongoing.
Texas: A state-sponsored analysis flags 114 north Texas schools as having suspicious scores on the Texas Assessment of Knowledge and Skills given the year before. The 3,000-student Wilmer-Hutchins district is dissolved over allegations of cheating there.
Ohio: Teachers around the state are investigated for such activities as making copies of tests, sharing questions with students ahead of time, or giving answers during the test.
California: The principal and founder of the University Preparatory Charter Academy in Oakland, resignes amid allegations of test cheating and security breaches in 2006 and 2007.
Los Angeles: The Los Angeles board of education votes to shut down six charter schools accused of cheating on 2010 standardized tests.
District of Columbia: The U.S. Department of Education and the District of Columbia investigate whether test-score gains in 2009 might have been the result of cheating.
Baltimore: Widespread cheating on state assessments is uncovered at two elementary schools. The tests were taken in 2009 and 2010.
Atlanta: State investigators release a report alleging a long-standing pattern of cheating on state standardized tests in at least 44 schools.
Pennsylvania: The Philadelphia Public School Notebook website brings to light a test-security company’s 2009 report flagging 89 schools for questionable erasures and suspiciously high test gains. The state has launched an investigation.
SOURCE: Education Week.
“Blaming tests or accountability systems for things we don’t like is a dumb idea. Like banning thermometers for revealing fevers,” Mr. Cizek wrote in an opinion essay last month for Education Week’s website. (“Cheating on Tests and Other Dumb Ideas,” July 25, 2011.)
“Tests are a measure of student achievement,” Mr. Cizek added in an interview. “It’s difficult to imagine any measurement system that doesn’t take into account student achievement,” he said.
Nationally, there appears to be no move to lessen the importance of test-based accountability.
Standardized tests remain at the heart of determining whether schools and districts make adequate yearly progress in raising student achievement under the federal No Child Left Behind Act. The tests are also being used to develop “value added” measures that will gauge an individual educator’s effect on a student learning. And two state assessment consortia have received federal funds to design tests for the new common-core academic standards adopted by all but five states.
U.S. Secretary of Education Arne Duncan said that he was “stunned” by the results of the Atlanta investigation, but stressed that other schools and districts across the country are making genuine progress without cheating.
In an interview last month with an Atlanta-area television station, Mr. Duncan said that the situation in that city stood out because of the “clearly systemic” nature of the apparent cheating.
“There’s never been anything like it, and we hope there never is again,” he told the city’s NBC affiliate.
Statistical analysis of the prevalence of teacher- or administrator-led cheating appears to bear out the belief that such misconduct involves a small minority. In 2003, economics professors Brian A. Jacob, of the University of Michigan, in Ann Arbor, and Steven D. Levitt, of the University of Chicago, analyzed the test scores of every Chicago 3rd through 8th grader who took the Iowa Tests of Basic Skills from 1993 through 2000. They detected evidence of “serious” cheating in 4 percent to 5 percent of their sample.
Professors at Arizona State University, in Phoenix, conducted a survey last year of 3,000 teachers, which found that 10 percent reported that they knew of colleagues who had engaged in the most egregious forms of cheating, such as changing answer sheets or somehow preventing low-performing students from taking the tests.
But even though such activity is believed to be relatively rare, it is hardly unknown. And it can sometimes affect the test results of hundreds, or thousands, of students at a time and cast doubt on claims of improvement in achievement.
In 1999, the chief investigator for the New York City schools released a
New York students allegedly were coached to the right answers on the citywide tests, or prepared for the tests using actual questions and answers from the exams. At that time, the cheating scandal was called one of the nation’s largest.
“When cheating occurred, it rendered the use of standardized tests as a diagnostic tool—to evaluate not only student performance, but educator performance as well—meaningless,” the report concluded. The United Federation of Teachers, which represents the city’s teaching force, commissioned its own scathing review, saying that Mr. Stancik’s work had overstated the problem.
In another case, the increased academic performance under then-Gov. George W. Bush that was often called the “Texas miracle” burnished Mr. Bush’s record for his 2000 presidential campaign. He often noted the nearly six straight years of gains in test scores among students, particularly minorities, during his state tenure, from 1995 to 2000.
The Texas testing policy went on to become the foundation of the testing policy created in the No Child Left Behind law and brought Rod Paige, the former Houston superintendent, to Washington as Mr. Bush’s first education secretary.
But those scores in some districts were tainted after newspaper and state investigations revealed gains in some schools that were improbably large. A later state investigation found one Houston school that falsified dropout data.
Scholars often say such cheating incidents are examples of “Campbell’s law” at work. Donald T. Campbell, a social scientist, wrote in the 1970s that “the more any quantitative social indicator is used for social decisionmaking, the more subject it will be to corruption pressures and the more apt it will be to distort and corrupt the social processes it was intended to monitor.”
In other words, the more important test scores become, the more likely it is that test scores will end up corrupted.
Fixing the System
In the debate over testing, however, school officials are focused more on trying to fix the testing system than on upending it.
James A. Wollack, a professor of educational psychology at the University of Wisconsin-Madison and an expert in test security, says that testing systems would benefit from impartial investigators who would independently probe cheating allegations.
“You really do need someone who’s impartial, who doesn’t have friends in the school system, and who doesn’t have friends in the governor’s office,” he said. The state investigation in Atlanta stands among similar probes of cheating incidents because of the involvement of dozens of agents from the Georgia Bureau of Investigation.
Georgia has also been taking steps to combat potential cheating by sending monitors from the state education department to schools around the state that have been flagged in the prior year as having high numbers of test-sheet erasures, said Matt Cardoza, the director of communications for the department. The state has done that for the past two years, and it has seen a drop in scores among some schools that were previously high performers, he said.
Having outside monitors, Mr. Cardoza said, will “slow down that unethical behavior.”
New York state has decided to combat cheating among teachers, administrators, and students by forming a board that will offer recommendations by the beginning of the 2012-13 school year. The board will review all aspects of the state’s testing system, said a spokesman for the New York education department.
But Audrey Amrein-Beardsley, an associate professor of education at Arizona State and the lead author of the study that surveyed teachers on cheating prevalence, recommends focusing on other practices that harm the validity of tests, such as the behavior she says she engaged in during her time as a classroom math teacher.
Ms. Amrein-Beardsley said she never altered a student’s answer sheet, or encouraged low-performing students to stay home on test day. However, she would get copies of the state tests before they were administered, change the details and numbers of word problems, and then use those slightly altered questions to drill her students, she said. She was so successful, she said, that she was asked to be her district’s test coordinator.
Some might call such work savvy test preparation. Ms. Amrein-Beardsley said she looks back on it and believes it was a form of cheating that narrowed classroom instruction to a few items that would be measured on a test. Her survey found those kinds of actions far more prevalent than actively changing test sheets.
“Teaching to the test is still third-degree [cheating] in my mind,” she said.
Library Intern Amy Wickner contributed to this article.
A version of this article appeared in the August 10, 2011 edition of Education Week as Experts Divide on Responses to Cheating