A forthcoming study by a pair of Stanford University researchers is further stoking the debate over whether states’ high-stakes testing programs can positively affect academic achievement.
Read “High Stakes Research,” from Education Next. (Requires Adobe’s Acrobat Reader.)
Debates over the value of accountability efforts that determine whether students graduate, which teachers win bonuses, and whether schools are taken over by states grew earlier this year after two Arizona State University researchers published a report arguing that such programs may do more harm than good. (“Researchers Debate Impact of Tests,” Feb. 5, 2003.)
The widely publicized report, coming at a time when the federal “No Child Left Behind” Act of 2001 has effectively made such programs the law of the land, drew a spate of critiques and counter- studies for its contentions that high-stakes efforts had failed to improve achievement and were pushing some students off the high school track.
In their new report, scheduled to be published next month in the magazine Education Next, researchers Margaret E. Raymond and Eric A. Hanushek add some strong wording to that chorus of criticism and also offer some new data of their own.
Different Conditions?
The two researchers contend that the biggest problem with the Arizona study is that the authors, drawing on scores for National Assessment of Educational Progress tests, compare the improvements made by students in states with strong accountability programs with the average national test-score gains made over the same time periods.
The better comparison to draw, they argue, would have been with states that have no accountability programs.
“The number-one precept of good analysis is that you examine the condition and then make an observation without the condition to see if it makes a difference,” said Ms. Raymond, who is the director of CREDO, formerly known as the Center for Research in Education Outcomes. The policy-research group is based at Stanford’s Hoover Institution.
When the data are analyzed that way, the Stanford researchers say, the results are reversed: From both 1996 to 2000 and 1992 to 2000, the average gain made by 4th and 8th graders in mathematics was higher in high-accountability states than it was for states that had not yet not attached any consequences for flat or falling test scores.
That trend held up, according to the authors, even when the data were adjusted to account for any changes in the percentages of students who were being excluded from the tests after the new accountability programs were put in place.
For their own analysis of NAEP mathematics data, the authors focused on states that imposed consequences on schools, rather than on students, for failing to raise test scores. They compared those states’ progress against that for states with no such programs.
The pair of researchers examined what happened to students’ growth in achievement as they moved from the 4th to the 8th grade in those states. While the tests were not given to exactly the same students, the authors say, the tests draw from students who were at least in the same age cohort. The researchers also made adjustments in the data to account for changes in state spending on education and in parents’ educational levels during the time frame they studied.
They found that the average percent test- score gain made by a typical student moving from 4th grade in 1996 to 8th grade in 2000—at 1.6 percent—was more than twice as high for states with consequential accountability programs as it was for those without them.
Audrey L. Amrein, one of the researchers who conducted the Arizona State study, called the Raymond-Hanushek criticisms “old hat.”
“I’ve had a lot of people reanalyze our data,” she added, “and each and every one of them have come up with different results. In no way is this analysis the last word.”
