By Bill Turque
Washington Post Staff Writer
Monday, September 7, 2009
Despite two requests from the District's Office of the State Superintendent of Education, D.C. public school officials never provided evidence that they investigated possible cheating at some schools after an analysis showed high rates of erasures on standardized tests in 2008, according to newly released documents.
The analysis was commissioned last summer by then-State Superintendent Deborah A. Gist after more than 20 public and public charter schools showed gains of 20 points or more in reading or math proficiency on D.C. Comprehensive Assessment System (DC-CAS) tests.
The study of student answer sheets, conducted by the test's publisher, CTB McGraw-Hill, eventually "flagged" classrooms in more than 40 schools because the number of wrong answers changed to right answers was significantly above the citywide average.
On Friday, The Washington Post obtained documents related to the investigation through the Freedom of Information Act. But despite repeated requests for a full copy of the study, District officials released only summaries and correspondence describing the analysis.
Two e-mail messages sent Friday to Victor Reinoso, who as deputy mayor for education oversees the superintendent's office, were not answered.
The high-stakes tests are used by the federal government to assess the yearly academic progress of schools under the No Child Left Behind law. The results from the 2007-08 academic year, Schools Chancellor Michelle A. Rhee's first full year in office, were touted by Mayor Adrian M. Fenty (D) as evidence of early progress in her efforts -- closely watched by urban education reform advocates nationwide -- to transform the struggling 45,000-student system.
The percentage of public elementary students demonstrating proficiency rose an average of 11 points in math and eight points in reading; among public secondary students, the percentages increased by nine points in reading and math. Teachers and administrators at public schools received cash awards if proficiency levels increased by 20 points or more in both reading and math.
Gist, who resigned April 1 to become Rhode Island commissioner of elementary and secondary education, was concerned about the validity of the huge increases at some schools.
The superintendent's office does not have authority over school operations, but it oversees administration of the DC-CAS and has the power to invalidate student scores and to fine schools if it determines that test security has been compromised.
However, it relies on the schools themselves to report evidence of cheating. In an unrelated case this past spring, Howard Road Academy, a public charter school in Southeast Washington, fired two teachers and disallowed the 2009 scores of 27 fourth- and sixth-graders after school staff members discovered that copies of the test were distributed before testing day.
A spreadsheet summary of the CTB McGraw-Hill study shows that Bowen Elementary in Southwest Washington was one of schools where test results improved dramatically in 2008. The percentage of children showing proficiency in reading grew by 27 points, from 36.2 to 63.2 percent. The 34 students in one class averaged more than 10 wrong-to-right erasures on the exam. The citywide average for wrong-to-right erasures on the reading test in elementary grades was between 1.4 and 2.3, according District officials.
Some schools with only modest gains in 2008 also had high rates of wrong-to-right erasures. At Bruce-Monroe Elementary School at Parkview in Northwest, where the reading proficiency level edged up from 40.5 to 42.5, the 17 students in one third-grade class had 223 wrong-to-right erasures, an average of 13 per student.
The research scientist who led the McGraw-Hill study, Steve Ferrara, recommended in a March 2009 memo that the superintendent's office "not draw conclusions about cheating" from the data. That recommendation was at least part of the basis for the decision by Gist's successor, Kerri L. Briggs, not to press the matter with Rhee.
Briggs, traveling this weekend, asked that a reporter's questions be e-mailed to her, but she did not respond.
In an interview Thursday, Rhee said: "Given that the people who actually developed the test said that it was inconclusive, we just didn't think it was necessary" to investigate possible cheating.
According to the documents, Gist, who declined to comment Friday, had another view. In a Nov. 20, 2008, memo to Rhee, Gist said the data did not automatically point to cheating. "There are many reasons that a class could have more erasures than other classes," she wrote.
But to guarantee the validity of the scores, she asked Rhee to "please take the appropriate steps to investigate the results enclosed and provide a report" within 60 days.
Gist made the same request of 13 charter schools -- which are publicly financed but independently operated -- that had an elevated number of erasures. Records show that just six responded with accounts of internal inquiries, none of which turned up evidence of cheating.
Officials provided a range of explanations for the high rate of erasures. For instance, students often begin "bubbling," or penciling in the spaces provided for answers, on the wrong line, forcing them to erase multiple answers. E.L. Haynes, a pre-kindergarten-to-sixth grade charter school in Northwest Washington, said some special-needs students had testing accommodations, allowed by law, that included having directions and questions read to them multiple times. That might have led to more erasures, officials said.
Other school officials said that because the DC-CAS is not a timed test, students are encouraged to review their work carefully.
"In math, in particular, students are required to complete the test and then go back and plug in all the possible answers to check their work. This method can result in a high number of [wrong-to-right] erasures," said Susan Schaeffler, executive director of KIPP DC Will Academy, a charter school serving grades 5 through 7 that saw a 39-point jump in math scores in 2008.
On Jan. 7, 2009, Rhee's office asked the superintendent's office for an extension of the 60-day period to Feb. 28. Erin McGoldrick, Rhee's chief of data and accountability, said that "in light of the high volume of classrooms with statistically aberrant erasure rates" she needed the extra time "to provide what we believe will be the most thorough possible response."
But on Feb. 28, McGoldrick wrote again to Gist's office, reporting that none of the schools in question had been investigated. She asked for additional information about the methodology used to derive the erasure rates. She also asked to review student answer sheets.
McGoldrick explained that "given the disruption and alarm an investigation would likely create at schools, [District public schools] must ensure that appropriate due diligence has been performed to maximize the quality of the information provided and minimize the risk of creating unfounded concern at school sites."
The records released Friday do not reflect what, if any, additional information Gist's office provided. The cheating inquiry was also complicated by Gist's departure.
Gist, a former senior policy analyst for the Department of Education in the Clinton administration, was appointed by Mayor Anthony A. Williams (D) in 2004 to run what was then the District's small state education office. She assumed the newly created office of state superintendent when Fenty took control of the public school system in 2007. School officials said that Gist had grown frustrated with her role in the reorganized school bureaucracy, in which she answered to Reinoso.
On April 1, the day Gist announced her resignation, her deputy, Alex Harris, wrote to McGoldrick again asking for a follow-up investigation.
Noting that the 2009 DC-CAS testing period would begin later that month, he provided a slimmed-down list of 11 schools that "would be the ones to prioritize for on-site investigation and follow-through." Those schools are Bruce-Monroe, J.O. Wilson, C.W. Harris, Bowen and Draper elementary schools; Takoma, Langdon, Marshall and Winston education campuses; Shaw Junior High; and Coolidge Senior High.
Closures and consolidations after the 2007-08 school year would have made some investigations more complicated. Bowen was closed and merged with Amidon Elementary; Shaw was closed and merged with Garnet-Patterson Middle School.
According to the records released Friday, the only response to Harris's memo came May 8, when McGoldrick wrote to him saying that her office had "thoroughly examined" the erasure analysis data and "reorganized it into an analytically useful format." She said the District had provided principals with training on the importance of erasure analysis and test security. She also provided a copy of a new test security plan.
Harris resigned soon after Briggs was named superintendent. He could not be reached for comment Friday.