April 15, 2007

The politics of cut scores

Yesterday, the St. Petersburg Times published Letitia Stein's story describing the twists and turns of setting cut scores in Florida's 10th-grade tests. The deeper story is not the inherent tensions in setting cut scores. There's not doubt that cut scores are arbitrary. The question is whether they are arbitrary in the sense of arbitrary and capricious or arbitrary in the sense of arbitration.

In the end, it's the use that matters. To distribute scarce resources (e.g., interventions), one could justify cut scores. But when careers and diplomas are on the line, the thresholds are far harder to justify.

Update: The original post was written quickly, and one issue I forgot to note was the article's taking the norm-referenced test at face value.  In Florida, students take a test that is supposed to be aligned with state standards, but they also take a few subtests of the Stanford Achievement Tests in math and reading.  Note the word few there; one should be very hesitant to take changes (or stability!) in aggregate norm-referenced test scores at face value in any case, and using a few subscales for this purpose is ... well ... hard to explain.

Listen to this article
Posted in Accountability Frankenstein on April 15, 2007 10:53 AM |