From New York to California, some of the first results of America's bold experiment in school testing are coming in - offering a glimpse into how well a decade of education reform is working.
The early verdict: Class, take out your books again. Many cities and states are reporting declining or below-standard scores in reading, math, and other basic skills. Others are showing mixed results at best.
But experts say what's important is not the early scores themselves. It's what schools do with them. While some districts are using the numbers to change classroom practices, administrators and politicians in other parts of the country are taking stern action:
*Recent test scores in New York City showed a sharp drop-off in student reading and math. Mayor Rudolph Giuliani (R) is calling for placing city schools directly under the control of the mayor's office, while schools Chancellor Rudy Crew says he would use these scores to replace district superintendents, principals, and teachers whose students did not perform well.
*In its first statewide report card, Florida told 78 schools - about 3 percent of the total - that they would have to offer their students vouchers to attend another (public or private) school if they do not improve in the next year. According to the June 24 report, most Florida students attend mediocre or below-standard schools.
*Massachusetts education officials predicted last week that a high percentage of the class of 2003 might not pass the new Massachusetts Comprehensive Assessment System test. One possible result: lawsuits from angry parents.
High-profile tests with teeth are a new development in American education. In the past test results were relegated to district office back files. Parents knew their children's scores, but rarely knew how well local schools compared with others. Congress approved the first nationwide test (the National Assessment of Progress, or NAEP) only on condition that the results could not be used to identify the performance of individual schools or districts.
After a decade of reform, 48 states now have testing systems; 20 have graduation requirements. Increasingly, statewide results are reported down to the individual school level. Newspapers publish them. Real-estate agents post them on Web sites or in ads as a guide to homebuyers.
And consequences are beginning to kick in. For example, 610 high school seniors failed Nevada's exit math and reading exams for the sixth time last week. They have two other chances this summer to pass the test. If they don't, they won't graduate. Some states give teachers or schools bonuses for good results; poor results can drive schools or entire districts into new management.
"It takes a while to sort this out," says Bella Rosenberg of the American Federal of Teachers, the No. 2 teachers' union. "But there are some very irresponsible uses being made out of the new data - in New York City, for example."
Indeed, analysts caution that publishing who's up and who's down doesn't solve the problem of how to improve learning. The systematic use of tests to adjust classroom teaching is a slow process - and is just beginning in some states.
"The good news is that we are beginning to pay attention to evidence of learning. Not so many years ago, all people looked at were graduation rates and per-pupil spending," says Chester Finn, president of the Thomas B. Fordham Foundation in Washington. "But there can still be the problem of having data and nothing changing in the classroom."
Some activists are using test scores to assess the impact of new education reforms - and are quick to seize on results. California's rock-bottom reading scores in the early 1990s fueled a return-to-basics drive that included smaller class sizes in early grades, more phonics instruction, and an end to most bilingual education.
Early results from this year's statewide achievement tests appeared to signal big gains for limited-English students - which would be a vindication for proponents of Proposition 227, the controversial 1998 ballot initiative that ended most bilingual education in the state.
But a closer look showed that test contractors has misclassified some 300,000 students, which likely inflated scores for students incorrectly labeled as limited-English speakers. The result: Perhaps still positive news for proponents of ending bilingual education - but not as dramatic as the early numbers suggested. "I'm cautiously optimistic that the gains will still be greatest in those districts that follow 227," says Ron Unz, a software developer who coauthored the ballot measure.
Overall, California students are now reading better than they were a year ago, but still score well below the national average. Adjusted results for limited-English students will be released by July 15.
While it may take months or years to sort out which reform most contributed to achievement gains statewide, local school officials say they can make immediate use of test scores to improve classroom teaching. Sacramento schools, for instance, tallied some of the biggest gains in reading.
"We did a tremendous amount of staff development," says Sacramento City Superintendent Jim Sweeney, citing financing from the David and Lucile Packard Foundation and a new phonics-based reading program. He's using test scores to determine which classrooms did the best job with comparable kids.
Some Texas school districts are using statewide tests to develop a similar analysis. "Just for Kids," a project based in Austin, provides districts with examples of higher-performing schools that serve students who are equally or more challenging to educate than their own. "This motivates both parents and educators to strive for something better," says Chrys Dougherty, the project director.
Schools in Amarillo, Texas, have been using such data to establish working relationships between their teachers and those in higher-performing schools. "It used to be that each classroom was an icon of solitude where everyone does their own thing.," says Charles Ritchie, director of the Amarillo achievement district's office of assessment. "Now we're looking at a lot of data. And it impacts everyone."