TEA adds 241 schools with suspect scores; Campuses not likely to be part of inquiry into possible TAKS cheating

By Joshua Benton
Staff Writer

Page 1A

Texas officials have released the names of 241 more schools with suspicious patterns in their test scores. But none are likely to be targeted in the upcoming round of state investigations into possible cheating.

The new list, released Friday, brings the total number of schools with suspicious scores to 699. That’s almost one-tenth of all the Texas schools that administered the Texas Assessment of Knowledge and Skills in 2005.

Earlier, the Texas Education Agency had released the names of only 442 schools that had at least one classroom with suspicious scores.

But Caveon – the test-security company the TEA hired to look for cheaters – also looked for schools that had suspicious score patterns schoolwide. Because of differences in the ways Caveon analyzed the scores, some schools were flagged as suspicious schoolwide without raising red flags in any specific classroom.

The TEA had not asked Caveon for the schoolwide list until The Dallas Morning News revealed its existence three weeks ago.

“We wanted to be able to look at all the schools as we think about how to move forward,” spokeswoman Debbie Graves Ratcliffe said.

Later this month, a new task force will decide which of the schools with suspicious scores will be subject to on-site investigations by agency staff members. The TEA is tripling its number of investigators in anticipation of inquiries taking months to complete.

But TEA officials have said they do not consider campuses with suspect schoolwide patterns to be as high a priority as those with classroom-level anomalies.

“We think the original list is the one we need to focus on,” Ms. Ratcliffe said.

Statistically, it is somewhat easier for an entire school to be flagged than for an individual classroom. That’s because the margin of error in Caveon’s analysis grows smaller as the number of students grows larger.

For example, if a 30-student classroom has three suspicious answer sheets, that could just be the result of random chance. But if a 3,000-student school had 300 suspicious answer sheets, that would raise a red flag.

The agency will formulate a recommendation on which schools to investigate in the coming weeks. The task force will make the final decision. Investigations are expected to begin in September.

Even with the new schools, Houston – the state’s largest district – still has the most suspicious schools in the state, 83. Dallas, the second-largest district, is next with 49 schools.

On the previous, shorter list, Dallas had only 39 suspicious schools. Most area suburban districts also saw small increases in the number of schools making the list.

In other large urban districts, El Paso now has 20 schools on the list, Austin has 12, Fort Worth has 11, and San Antonio has six.

The new list also includes 16 schools that had suspicious classroom scores but were not included on the TEA’s initial list.

Caveon looked for a variety of suspicious patterns in a school’s test scores. For example, a classroom might be flagged if a large number of students had suspiciously similar answer choices on the TAKS. Large numbers of erasures on student answer sheets, unexpected gains in student performance, and unusual answer patterns could also earn a school a place on Caveon’s list.

The suspicious schoolwide scores show some schools where test-security problems may be systemic. For example, in Dallas, several high schools had suspicious scores schoolwide in all four subjects tested – reading, math, science and social studies.

Those schools were A. Maceo Smith, Carter, Spruce, Kimball, Lincoln, Molina, Roosevelt, Samuell, South Oak Cliff and Sunset.

Well-off suburban districts continue to be flagged primarily for big jumps in test scores, mostly in their high schools. Many of those districts’ superintendents have said those gains are the result of quality instruction, not cheating.

Urban districts were much more likely to be flagged for students who had identical or nearly identical answer sheets, suggesting that students were copying one another’s answers on the high-stakes graduation TAKS test. Explanations for those patterns have been harder to come by.