The Louisiana Department of Education (LDOE) did not conduct an erasure analysis of the state’s standardized test scores for the 2009-2010 academic year due to budget cuts, The American Independent has learned through Freedom Of Information Act requests.
While the key monitoring tool was dropped for the 2009-2010 school year, the documents from LDOE also show a host of measures — on-site visits and mandatory internal investigations — were taken to monitor schools during the spring and summer high-stakes testing season for academic years 2007-2008, 2008-2009 and 2009-2010.
An erasure analysis — a once-arcane accountability device that has garnered increased public awareness following high-profile testing impropriety scandals in Pennsylvania,New Jersey and Georgia — is part of a portfolio of precautions undertaken by state education officials to prevent cheating on standardized tests. The process consists of examining electronic bubble sheets on which students record their answers for a high rate of wrong-to-right answer erasures. Typical explanations for why students change test answers include copying answers from other pupils, teacher intervention or as one newsletter released by the Louisiana School Board Association alleged, administrative and state-level meddling to change the answers.
Charles Hatfield, a former superintendent in charge of testing and accountability of what was the unified school district (Orleans Parish School Board) in New Orleans before Hurricane Katrina struck, explained erasure analysis in greater detail. “They [the analysts] scan the bubble sheets following a mathematical algorithm,” he said, “meaning from a statistical point of view, it is highly unlikely that so many wrong-to-right erasures can occur.”
The motivations for test tampering are numerous, but many education experts point to No Child Left Behind (NCLB) as incentivizing testing corruption given the high-stakes nature of the national education policy. States are rated by how many of its schools clear cut-marks; schools that post poor results can lose funding or suffer staff layoffs, while in extreme cases of chronic underperformance, the school is handed to charter operators. In recent weeks, high-profile education officials like U.S. Department of Education Secretary Arne Duncan and firebrand education historian and NCLB critic Diane Ravitch have called for the law to be expunged. The two camps disagree on a replacement.
In seeking an explanation for the one-year suspension of erasure analysis, LDOE spokesperson Rene Greer told TAI the economic downturn burrowed an unexpected $247.9 million budget deficit for 2009-2010 into the state coffers, leading to across-the-board cuts. LDOE absorbed $4.15 million in cuts as part of the department’s effort to “provide the maximum support for students and teachers, while maintaining the integrity of the testing program,” said Greer.
As a consequence, however, the cuts included the elimination of the erasure analysis as well as the development, printing and distribution of interpretive guides and other testing reports. The decision not to have the vendor conduct the erasure analysis saved the department $58,459. According to The Notebook, Pennsylvania — where dozens of schools were flagged recently for wrong-to-right erasures that analysts calculate were very likely the result of testing improprieties– spent nearly double that for its 2009 erasure analysis, though the Keystone State’s population exceeds Louisiana’s, 12.7 million to 4.5 million.
Louisiana’s testing vendor is Data Recognition Corporation (DRC). Through another FOIA request, TAI learned the state entered a contract with the company to the tune of roughly $9 to $13 million a year, spanning a decade of services beginning in 2005. DRC received a no-bid contract extension in 2010 worth nearly $44.5 million. DRC prints, issues, writes, grades and analyzes Louisiana’s No Child Left Behind-compliant state tests. LEAP, iLeap, GEE (the high school exit exam) and several tests geared toward students with limited English proficiency and learning disabilities all fall under the purview of DRC.
“While the erasure analysis is a one tool the agency has used to ensure the integrity of Louisiana’s testing program, as the data indicate, historically this analysis has resulted in an inconsequential percentage of tests being voided statewide,” wrote Greer in an email to TAI. “Thus, the utilization of the erasure analysis alone does not validate nor diminish the reliability of the state’s testing program or accountability system.”
Indeed, for the 2010-2011 school year, 1,533,000 tests were administered statewide, with 214 of those voided for exceeding the number of wrong-to-right answer erasures. In the Recovery School District, the state-run set of campuses that were taken over following Katrina with many then placed under charter operators, 38 tests were voided out of 63,966 administered after an erasure analysis.
The chart below chronicles four academic years of erasure analysis voids, meaning scores for tests flagged are marked with a zero which can impact a school’s overall School Performance Score — the index used by Louisiana in compliance with No Child Left Behind school assessment rules.
LDOE also supplied TAI with testing violations reports beginning in 2007-2008 through 2009-2010.
For ’07-’08, 82 members of LDOE staff visited 238 test sites in 56 school districts during the spring testing season. The report from the Division of Standards, Assessments, and Accountability and Regional Service Centers concluded, “access to test documents was controlled and security procedures were followed.”
However, site visitors recorded springtime violations at 47 sites, 21 of which had “severe testing irregularities.” For heightened violations, the state orders the district to conduct an investigation, identify the cause of the transgressions and form a corrective plan of action to avoid repeat offenses, which is then submitted to the LDOE officials. Violations are represented by a per-incident basis, with the potential for many occurring in one school. At Laurel Elementary in New Orleans, for example, 13 violations were cited over irregularities concerning the answer sheets of six students taking the English Language Arts exam and seven students taking the math portion. Those scores were voided.
Scott Norton, Ph.D., assistant superintendent at LDOE’s office of student and school performance, said districts with significant violations receive follow-up inspections the next year, adding, “sometimes the on-site monitoring is unannounced.”
He elaborated: “The state has a policy in the assessment bulletin that districts have to follow. [These include] a chain of custody for the test documents and a secure locked location for where the testing material is held.”
For 2008-2009, LDOE staff made 257 test site visits in 66 districts, with additional visits to charter special schools. Eighty-three evaluators were sent out, and, according to the report summary, “[a]ny school with a record of prior test security problems was likely to be visited; the remaining visits were scheduled at random.”
The report noted that during the springtime monitoring sessions, 50 sites had minor testing irregularities, while 18 had more severe violations that required a corrective action plan. The Recovery School District, unpopular to teacher and labor groups as well as community activists for its swift school turnaround process and what they call misleading student performance data, had 26 springtime violations noted in ’08-’09. For 2007-2008, RSD was flagged for 27 of the total 137 violations recorded throughout the state. RSD schools are, according to the state, long-struggling schools.
RSD students in New Orleans, a city in which most of the district’s schools are concentrated, struggle with high rates of poverty and higher rates of suspensions, arrests, and draconian school enforcement policies.
Despite the budget cuts and no erasure policy, in 2009-2010, there was a surge in LDOE personnel conducting on-site testing evaluations: 157 LDOE employees were sent out to 378 sites, plus charter and special schools. 54 violations were recorded, of which 12 were significant. 27 occurred at RSD.
Greer told TAI, “Louisiana has the most notable and recognizable accountability processes in the country.”
Relying on 2009 Pennsylvania state data that was never published, The Notebooklearned that in some instances, the likelihood of wrong-to-right erasure marks on student bubble sheets happening in earnest was smaller than the equivalent of dividing one by the product of a trillion multiplied by a trillion. Some 60 Pennsylvania schools were nabbed for possible cheating through the erasure analysis.
New York Daily News reported in early August that the New York State Education Department stopped using erasure analysis in 2001, though is considering reinstating the oversight tool in light of cheating scandals in Atlanta and Philadelphia.Tags: Arne Duncan, atlanta, bubble sheets, cheating, Diane Ravitch, erasure analysis, GEE Georgia, ILEAP, LEAP, louisiana department of education, ncib, new jersey, new orleans, No Child Left Behind, Poverty, Recovery School District, testing, the Notebook