Twenty-five percent of New Jersey’s third and fourth graders will fail the state tests on math and language arts for 2009. That’s the estimate Education Commissioner Lucille Davy gave yesterday, explaining that the reason for the somewhat disconcerting percentage was because the DOE made the test harder and changed a passing grade from 40% to 50%.

The tests, given this past Spring, haven’t actually been scored yet, but the DOE did a sampling so they could warn districts ahead of time. And it’s the test, not the kids, right? The Star-Ledger quotes Davy:

It is our belief that this doesn’t mean students have done any differently or that students did poorer. It means we have a higher expectation.

At least the DOE is consistent. Last year it changed the definition of “proficiency” (i.e., passing) for 5th –8th graders, also resulting in a a larger percentage of kids who failed the test (oops– kids who were “partially proficient”). But what’s with the timing? Can’t the DOE make these decisions before the kids take the test or, better yet, before subject supervisors and assistant superintendents start poring over spreadsheets to determine how many kids need to be smothered in special tutoring so that the school makes Adequate Yearly Progress?

That’s the problem with all these standardized assessments. Everyone looks at the cut-off and works from there, all eyes on the kids “on the bubble,” those right on the cusp of passing who might be pushed over the edge with extra tutoring. There’s far less attention for kids who are safely proficient or even in the pantheon of “advanced proficient.” And, in many cases, less attention for kids at the bottom of the barrel who won’t pass anyway.

Growth models, anyone? It’s the only way to move from arbitrary definitions of proficiency to really assessing education achievement for each child. Right now it’s just a numbers game.