Usually it’s a good thing when a student passes a test.
But when an entire class passes the state exam that might be too good.
“Oh I find them astounding. That’s not predictable.”
That’s Thomas Haladyna. He’s a retired professor at Arizona State University. He’s studied testing in public schools for decades.
Last year, 100 percent of the third graders at Jefferson passed both the reading and the math tests.
“That’s not predictable. What you would predict is a small percentage might pass given their Title I status.”
Title I is special federal money for schools where the majority of students come from low-income families. Those students typically struggle on standardized tests.
Yet the two schools in question outscored the district on most of the exams.
What’s more several grades showed remarkable improvement. We asked Haladyna about that.
“How typical is it for, you know, say 54 percent pass one and then the next year 82 percent pass?”
“You don’t have to be a statistician. All you have to do is look at the state average.”
Looking at the average in Houston elementary schools, the most the district improved was a three-point bump.
“So anything like over three points is statistically improbably. So if you’re getting 12 points or more, not improbable, that’s almost impossible.”
Jefferson and Atherton posted gains in the double digits — as high as 28, even 32 percentage points.
As the probe into cheating continues, Haladyna says investigators will look for outliers — data that’s outside what’s expected.
Linda McSpadden McNeil with Rice University says it’s also important for investigators to look at other things.
“The first question should be who are the children? Have they introduced an honors program or something that is really more robust teaching material? Or on the downside are they doing so much multiple choice drill that the kids can just do multiple choice tests in their sleep?”
McNeil says the high stakes tied to standardized tests is a major problem.
Here in Houston test scores are used for teacher bonuses and evaluations.
“So we’re using the test not only inappropriate ways but it’s actually unethical to use the tests of children’s learning for a purpose it wasn’t designed for and none of these tests are designed and validated for evaluating a grown up.”
In this investigation, 20 teachers have been reassigned from their classrooms.