The State Office of Education, in a brief presentation to lawmakers last May, further broke down their numbers for ethnicity and other groups of students. The numbers showed every group — even students with disabilities and those still learning to speak English — passed reading at a rate of 96.6 to 96.7 percent. Also, the state reported 3,190 Hispanics passed the reading test and 113 failed it, whereas a second set of numbers also presented to lawmakers indicated nearly 10 times as many Hispanics failed the exam.

In other words, one set of data — the one based on student enrollment — shows there is no achievement gap on the basic skills test. The other — based on students who actually took the test — shows a wide one: Just over 70 percent of Hispanics passed reading vs. 94 percent of whites, for example. Those latter numbers, based on test participation, include students who may have moved out of state, graduated early or dropped out, so they don't show an entirely accurate picture; nor should they add up to enrollment-based numbers.

But mathematically, they should look better than the enrollment-based numbers because they don't include all the students who have never taken the test, who would "fail" by default.

Running the numbers

So where did the numbers come from?

Pass rates, Johnson said.

In all, 34,469 students in the Class of 2006 have passed the basic skills test's reading exam, according to the state's testing contractor.

There were 35,685 students in the Class of 2006 last February.

Divide the numbers, and 97 percent of Utah's incoming senior class is passing. The remainder was plugged in as the fail rate.

But that math doesn't cross check.

The state's basic skills test database shows 3,401 test-takers are known to have failed the reading exam, the State Office of Education said in response to Deseret Morning News questions.

That's nearly triple the state office's public report on the number of students who have not passed the test.

In math, 9,217 students are known to have failed — nearly 50 percent higher than the state's report. In writing, 6,263 students failed — nearly 70 percent higher than the state's report.

Johnson said only pass rates were used in the analysis. She felt they were the only reliable numbers to go on. Fail rates, she says, leave too much room for error, including counting students twice. For example, a student might have failed the math test three times, but the state still has to count it as one student who has yet to pass, and not three.

Testing gurus and state education chiefs say assigning each Utah student a universal ID, scheduled in the fall, will help immensely. When students move, or show academic progress or struggles, the ID would allow the state to track it.