Ohio public schools need some math miracles, before the state's new graduation test takes effect in the next few years. Alarming math scores during a recent trial run of the new exam prompts the urgency. Only 24 percent of 10th-graders passed the math portion of the test. Almost 95 percent of black 10th-graders failed, as did 90 percent of Hispanics and 73 percent of white students.

The new Ohio Graduation Test (OGT) is much more difficult than the existing proficiency test required for a diploma. It's based on what a 10th-grader should know, according to Ohio's new academic content standards for math. The current so-called ninth-grade proficiency test is based on knowledge students should have by the end of the eighth grade.

The new test raises the bar for learning and graduation in states all over the country. Tests are being created to match the standards to see if kids are learning them. In too many places, the answer is "no."

So what's a state to do? Scrap the exam? Postpone it for years? Write an easier test or lower the passing requirements?

Unfortunately, some states in Ohio's dilemma are retreating from the harder requirements. New York recently voided results from its math exam because appallingly low passage rates would have cost thousands of students their high school diplomas. California recently postponed for two years its exit exam because so many are failing it.

Fortunately, Ohio has some time to plan for success in this area. The new graduation test will be given to sophomores in 2005, the first class (2007) required to pass the exam. The huge challenge now is to make sure teachers know how to teach what needs to be taught.

Ohio Superintendent of Education Susan Tave Zelman and the state board of education seem determined to forge ahead. They've vowed to make the math challenge a top priority.

The trial-run scores confirm the urgent need to focus on math intervention and improvement, both for teachers of math and for students. Our charge to Ohio's education leaders is to dig in and not cave in.