69传媒

A rigorous study of 38 schools that are using the Success for All improvement program has found that students read better after two years in the program and outpace students in regular classrooms by up to half a school year.

But the long-awaited study, which was posted online last week, is as notable for its research design as it is for its results. Paid for with nearly $7 million in federal and private funds, the study heralds what federal education officials and other experts hope will be a new generation of large-scale experiments that use randomized research designs to give educators and policymakers clearer answers on what works in schools.

鈥淚 applaud them for doing that,鈥� said Grover J. 鈥淩uss鈥� Whitehurst, the director of the U.S. Department of Education鈥檚 Institute of Education Sciences, the federal agency that is spearheading the push for more 鈥渟cientifically based鈥� education studies. 鈥淚t鈥檚 a sophisticated study that uses everything the evaluation field has come to recognize as high-quality.鈥�

is available from the . (Requires Microsoft Word)

Commonly used in medicine but rarely in education, randomized field trials are studies in which subjects are randomly assigned to either experimental or comparison groups. Though proponents hold them up as an ideal research model, some education researchers have been more critical. They say that such studies can cost too much and raise questions of ethics, and that they fail to capture classrooms鈥� inner workings.

Critics Still Skeptical

Used in more than 1,200 schools nationwide, the Success for All program is one of the best-known, most studied, and most debated school improvement models in the country. Though aimed at preventing and resolving reading difficulties in the early grades, the program really takes a schoolwide approach to bolstering learning. It provides schools with materials, training, and an action plan that often requires them to redeploy staff members and make other changes.

Already, 46 studies have weighed in on the program鈥檚 effectiveness; most yielded positive results. None, however, was a pure experiment, leaving the program open to critics who contend the rosy findings were biased.

Whether the new results will quiet critics remains an open question.

Herbert J. Walberg, a professor emeritus of education at the University of Illinois at Chicago, said he remains skeptical because the program鈥檚 developers, Robert E. Slavin and Nancy Madden, were on the research team for the new study.

鈥淢y view is that a contract letting Success for All evaluate Success for All is like asking General Motors to determine if their cars are better than Fords and Toyotas,鈥� Mr. Walberg said in an e-mail interview.

Researchers from the University of Wisconsin-Madison, Baltimore鈥檚 Johns Hopkins University, and the nonprofit Success for All Foundation in Baltimore launched the national study in 2001 with $6 million in funds from the Education Department鈥檚 now-defunct office of educational research and improvement. They retooled the experiment, kicking in $700,000 more in foundation money, when schools balked at participating.

鈥淧rincipals weren鈥檛 excited at the idea of getting their destiny determined by a coin flip in terms of whether they get the program or not,鈥� said Geoffrey D. Borman, the study鈥檚 lead author and an associate professor of educational leadership and policy analysis, educational policy studies, and educational psychology at the University of Wisconsin-Madison.

To persuade more schools to take part, the researchers allowed all the schools to implement the program in at least some grades and at no cost to them. In the end, 20 schools got the program in kindergarten through 2nd grade; 18 provided it only in grades 3-5. That way, researchers could still compare kindergarten and 1st grade students in the program with nonparticipating students in the same grades in other schools.

The 38 schools that eventually signed on were located in predominantly poor communities in 17 Midwestern and Southern states. The sites included cities鈥攕uch as Chicago, Indianapolis, and St. Louis鈥攁nd some small towns and rural communities.

Worth the Investment?

After one year, the researchers found, kindergartners and 1st graders in Success for All were scoring two months ahead of their peers in the comparison group on tests measuring their ability to decode words. The two groups were still evenly matched, though, on other reading-skill measures.

At the end of two years, the program students鈥� edge had grown and spread. In word decoding, the Success for All pupils outpaced their peers by a margin equivalent to 4.7 additional months of schooling. They scored 1.3 months ahead of the control students on tests measuring their ability to understand written passages and 1.7 months ahead when it came to identifying words. Researchers are now analyzing results from the study鈥檚 third and final year.

The researchers said they tried to eliminate potential biases by recruiting independent scholars to oversee the study and hiring an outside research organization to collect the data. It was the independent scholars who made the coin tosses that determined which schools got the study treatment.

Mr. Slavin said critics might legitimately question whether the results are worth schools鈥� investment, but not whether the program worked. 69传媒 pay at least $130,000 over three years to put the program in place鈥攁bout as much as it costs to reduce class sizes in the early grades, according to Mr. Borman.

鈥淭o have research like this that shows some bang is noteworthy,鈥� said Mark Berends, a Vanderbilt University researcher who has studied schoolwide reform programs such as Success for All.