UC Berkeley News
Press Release

UC Berkeley Press Release

NCLB effectiveness tough to gauge, study finds

– Many states provide erratic, exaggerated reports of student achievement trends that make it impossible to determine the effectiveness of the federal No Child Left Behind Act, according to a study released today (Wednesday, July 5) and led by University of California, Berkeley, researchers.

"Parents and policy makers are eager to know whether children are learning more under the federal government's stiff press for accountability," said Bruce Fuller, the education professor who directed the study. "But despite all the new testing, many states are not reliably answering this question."

The 14-month study by Policy Analysis for California Education (PACE), an independent research center based at UC Berkeley, UC Davis and Stanford University, compared state and federal test score trends for math and reading proficiency across 12 states between 1992 and 2005.

The No Child act required for the first time that all states participate in a federal testing program, enabling analysts to compare any state's claims of progress in test results against the uniform federal standard.

Since the act's passage, analysts have worried that states would "dumb down" their tests and set a low bar when defining which children are proficient in basic subjects like reading and math, making it easier for all students to become proficient by 2013 as mandated by the federal act.

The PACE researchers found that governors and state school chiefs often do just that.

This makes it "almost impossible for parents and local educators to determine whether children's achievement is going up or down," said Fuller.

Texas officials, for example, reported in 2002 that 91 percent of their fourth graders were proficient readers, the PACE report says. But that share fell to 76 percent the following year after a new, more demanding testing system was put in place. Meanwhile, the federal National Assessment of Educational Progress revealed that just 29 percent of the same Texas fourth graders were proficient, a statistic unchanged since 1998.

Illinois, since the mid-1990s, has claimed that 35 percent more of its fourth graders are proficient in reading compared, on average, with the federal benchmark. Annual achievement reports by New Jersey officials have found that 42 percent more fourth graders are proficient in reading than the share reported under the federal test.

California officials estimated in 2005 that 49 percent of the state's fourth graders were proficient in math, compared with 29 percent deemed proficient by federal testing officials.

A few states - including Massachusetts - peg their proficiency cut-points somewhat closer to the federal standard, the researchers found. In Massachusetts, the gap in math proficiency between state and federal benchmarks has equaled just 1 percentage point, on average, over the past decade.

Researchers also found that since No Child Left Behind began, many states have continued to report achievement gains which far exceed the rate of progress estimated by federal officials. For example, California has reported a 3.7 percent average annual increase since 2002 in the share of fourth graders proficient in reading, compared with flat test score performance as gauged by federal standards.

Overall, the study reports that federal reading scores have remained flat in the 12 states examined after No Child Left Behind kicked in, reflecting a nationwide trend. Math scores have inched upward on the national assessment, while states continue to report much higher rates of improvement.

"The Washington architects of the No Child reforms promised greater transparency, but we now have wildly varying gauges of achievement trends across the states," Fuller said. "It's a bewildering rendition of transparency."

Congress is beginning a formal review of the No Child Left Behind Act, with reauthorization scheduled next year.

Critics of the national assessment argue that it is not aligned with any state's curriculum and that the federal definition of proficiency is too demanding.

Various steps can be taken to bring more credibility to state testing programs, said the researchers. Those include benchmarking state tests to federal standards, improving tests to better reflect analytical and writing skills as well as higher-ordered thinking, and providing federal support to states to ensure that trends over time can be reliably reported when testing programs change.

The report, available today at http://pace.berkeley.edu/testscoretrends.html, details test score trends for Arkansas, California, Illinois, Iowa, Kentucky, Massachusetts, Nebraska, New Jersey, North Carolina, Oklahoma, Texas and Washington.

Researchers sought out a diverse range of states based on population and urban or rural character, geography and the intensity of the states' test accountability programs in the 1990s. They worked with state education officials, education associations, fellow researchers and news archives.

The report was peer-reviewed by leading experts in the field of student assessment.