(Editorial. New York Times, Sunday, July 2, 2006)

Many of the nations that have left the United States behind in math and science have ministries of education with clear mandates when it comes to educational quality control. The American system, by contrast, celebrates local autonomy for its schools. When Congress passed the No Child Left Behind Act, it tried to address the quality control problem through annual tests, which the states were supposed to administer in exchange for federal dollars. But things have not quite worked out as planned.

A startling new study shows that many states have a longstanding tradition of setting basement-level educational standards and misleading the public about student performance. The patterns were set long before No Child Left Behind, and it will require more than just passing a law to change them.

Policy Analysis for California Education (PACE), a research institute run jointly by Stanford and the University of California, showed that in many states students who performed brilliantly on state tests scored dismally on the federal National Assessment of Educational Progress, which is currently the strongest, most well-respected test in the country.

The study analyzed state-level testing practices from 1992 to 2005. It found that many states were dumbing down their tests or shifting the proficiency targets in math and reading, creating a fraudulent appearance of progress and making it impossible to tell how well students were actually performing.

Not all states have tried to evade the truth. The tests in Massachusetts, for example, yield performance results that are reasonably close to the federal standard. Not so for states like Oklahoma, where the score gap between state and federal tests has averaged 48 points in reading and 60 points in math, according to the PACE report. The states that want to mislead the government - and their own residents - use a variety of dodges, including setting passing scores low, using weak tests and switching tests from year to year to prevent unflattering comparisons over time. These strategies become transparent when the same students who perform so well on state tests do poorly on the more rigorous federal exam. Most alarming of all, the PACE study finds that the gap between student reading performance on the state and federal tests has actually grown wider over time - which suggests that claims of reading progress in many states are in fact phony.

States have always resisted the idea of one national set of tests, citing local autonomy. But if the United States wants to equal its competitors abroad, it must move away from a patchwork system based on weak standards and a frankly fraudulent system of student assessment. Under one promising proposal, the government would finance creation of a rigorous, high-quality test that would be provided free to the states - as long as they agreed to use federal scoring standards. That would finally give the country an accurate and all-encompassing view of student performance.