CHEYENNE - Even though Wyoming schools with as few as six students are counted for accountability purposes under the No Child Left Behind Act, the state doesn't separate out the scores of minority groups unless there are at least 30 students - meaning more than half of Wyoming's minority students aren't being counted.

That includes about 50 percent of Hispanics, 40 percent of American Indians, 90 percent of blacks and 100 percent of Asian students, according to an analysis by The Associated Press based on National Center for Education Statistics figures.

In fact, only about 45 schools in Wyoming - 12 percent of the total - are considered to have statistically reliable numbers of minorities in any group, the figures show.

That's significant because lack of progress by any one group brings upon a school the same sanctions as if the entire school weren't progressing. And minorities historically have not fared as well as whites in testing.

The federal government leaves it up to states decide what's a statistically reliable number. The numbers vary widely - from five students in Maryland to 52 in Oklahoma.

In Wyoming, a school needs to test at least 30 members of a minority group for the group's numbers to be separated out from the school's total.

"From a purely statistical view, of course it's arbitrary," said Kay Persichitte, director of teacher education at the University of Wyoming.

No Child Left Behind allows students in schools not progressing sufficiently to switch to schools that are making adequate progress within the same district. Also, teachers in "failing" schools can be fired.

Persichitte said the law's sanctions leave the Wyoming Department of Education little choice but to find ways to help schools avoid them.

"So as long as we continue to use testing and accountability almost as a punishment, all districts will look for ways to minimize the impact of those populations which negatively affect their bottom line scores," she said.

Deputy Superintendent Annette Bohling offered a different interpretation. She said Wyoming can't use a number smaller than 30 because the results wouldn't be reliable and because individual students and their scores could be identified.

"At the state level we are charged with ensuring that the decisions we make for (adequate yearly progress) are valid and reliable. That's our major responsibility. When you're dealing with subgroups that small, you really do get bounced from year to year on what the performance might be. You have to be careful how to build that into the system," she said.

She said 30 is statistically valid.

"Thirty is a number that is textbook - it's sort of a textbook number that's used for validity and reliability. That is a number that is used in statistical methodology," she said.

"We didn't just pull that out of the air."

While many minorities in Wyoming aren't counted separately, they are still counted toward their schools' overall scores. But to ensure that all of Wyoming's small schools are counted, the state brings a smaller number into play, according to the state's rule book for compliance with No Child Left Behind.

If the 30-student minimum were applied to every school as a whole, more than a quarter of Wyoming's schools would be too small to be counted. So for schools with fewer than 30 students taking the Proficiency Assessment for Wyoming Students, or PAWS, test, the state has set six students as the statistical minimum.

"So where did we get that number?" wondered Persichitte.

Again, the U.S. Department of Education allows states to set a lower statistical minimum to ensure that small schools are counted under No Child Left Behind.

Bohling said six wasn't an ideal number. "We would never have gone down that low if we just didn't have to arrive at a number that we could live with. Which is why we put in that margin of error," she said.

Wyoming employs a mathematical formula called a confidence interval to help ensure the reliability and confidentiality of results from schools with fewer than 30 students taking the PAWS test.