Second Look at Tougher Accountability Yields New Results
Two years ago, a pair of Stanford University researchers published a study suggesting that students’ math scores were rising faster in states that had put stronger pressure on schools and students to raise academic achievement than existed in states with weaker school accountability programs.
But last week, the researchers presented new findings suggesting that, over the long run, the achievement picture for those states may be a bit more complicated.
For their new study, researchers Martin Carnoy and Susanna Loeb tracked students in all 50 states over a longer period of time and added data from federal tests in reading as well as math. In most of the states with high-pressure testing systems in place, they found, the mathematics improvements that the researchers had documented in the late 1990s tapered off from 2000 to 2003. And in students’ reading scores, it seemed to make no difference whether a state had attached high stakes to the test scores or not.
The study, one of eight presented at a national conference here last week, hints at some of the surprises—and the potential for new knowledge—emerging now as researchers and states put in place new data systems that allow them to track the progress of schools and students on a large scale, and over a longer period of time.
“A lot of these data are becoming available because of new accountability systems,” said Jane Hannaway, the director of the Center on Education Policy at the Urban Institute.
The Washington-based think tank organized the May 2 conference specifically to highlight findings from those new longitudinal research efforts. “Over the next five to ten years,” she added, “I think we will learn some things that we may not have expected. We do know for certain that we will learn a lot.”
The Stanford researchers cautioned that their findings are still preliminary. They based them on 4th and 8th grade results from National Assessment of Educational Progress tests in reading and math, which were given every two to three years between 1992 and 2003.
No Easy Answers
Mr. Carnoy, the report’s lead author, said he had no easy explanation for the change in the pattern of results that he and Ms. Loeb found. One possibility, he said, may be that the test scores in those high-pressure states simply slowed.
“Or,” he added, “so many other states came on board with stronger accountability mechanisms that we don’t get as much variation among the states.”
For instance, with the adoption of the federal No Child Left Behind Act, which President Bush signed into law in January 2002, states were required to put in place student-testing systems and to report publicly on the results in order to continue to receive funding under the Title I program for disadvantaged students.
The researchers did find that the academic gains in the high-pressure states were significant for black and Latino students. They kept pace with white students during the first part of the study, from 1996 to 2000, and then increased at a faster clip from 2000 to 2003. Though scores improved for students at both ends of the achievement spectrum, the gains were largest over the long run for minority students in the area of basic skills.
But because white students’ scores also improved, the black-white achievement gap decreased only slightly, added Mr. Carnoy, who is a professor of education and economics at Stanford.
David N. Figlio, an economist at the University of Florida in Gainesville, who commented on the study at the Washington conference, said there also may be other explanations for the dropoff in overall test-score gains.
“Maybe the states with the most to gain from school accountability efforts implemented them first,” he suggested.
Another paper presented at last week’s conference previewed results from a study tracking instructional improvements in reading in 107 schools in 17 states.
As part of that study, which was led by researchers from the University of Michigan in Ann Arbor, 1st grade teachers were paid to keep logs detailing their teaching activities. Based on the logs, the researchers categorized classroom reading instruction into four types, which ranged from classes in which teachers emphasized showing students how to decode and analyze words to those that focused heavily on teaching reading comprehension, as well as broader mixes of both those approaches.
“We focused on the comprehension-focused programs,” said lead author Brian Rowan. “We wanted to know: Can it achieve positive effects without sacrificing growth in word-analysis skills?”
It did. Students who started out the school year achieving at the 50th percentile on standardized reading-comprehension tests ended the year scoring, on average, at the 62nd percentile. On word-analysis tests, students moved from the 50th to the 57th percentile, according to Mr. Rowan, an education professor at Michigan.
Ms. Hannaway said the Urban Institute plans to pull together all the papers presented at the conference in a forthcoming book.
Vol. 24, Issue 36, Page 12