Cut scores: What are they, and why do we care?

“Cut scores”? That certainly sounds like esoteric jargon, doesn’t it? But it turns out to be an important concept, even if the general public doesn’t know the phrase.

Whenever we scale a test to convert raw scores to scaled scores, we determine the cut scores, that is the scores that divide passing from failing, or advanced from proficient, or whatever. How we do this is more of an art than a science. “Use your intelligence guided by experience” is the advice Nero Wolfe gave Archie Goodwin in a wide variety of contexts, and he might as well have been giving the same advice when determining cut scores. They can make the difference between passing and failing, and yet they are not very scientifically determined.

This issue, of course, really comes to the fore when high-stakes standardized tests are involved. I just came across a essay that Diane Ravitch posted in her blog a couple of months ago. She observes that cut scores are

arbitrary and subjective decisions made by fallible human beings. They can raise the passing mark to create large numbers of “failures,” or they can lower the passing mark to create a “success” story, to celebrate their wonderful policies. In some cases, the cut score is set high, so many students “fail.” The next year, or year after, the cut scores are lowered, and HOORAY! Our Wise Leadership Has Create Success!

This has happened with MCAS, and now it’s happening with Common Core. But that’s not the bad news. Ravitch recommends that we “must must must” see the announcement about cut scores in the Smarter Balanced test (the competitor to PARCC among Common Core tests):

You will see that the cut scores predict that most students will “fail” in every grade. Only the top two levels are considered “passing,” that is, proficiency and advanced. In third grade math, 61% are predicted to “fail.” In fifth grade math, 67% are predicted to “fail.” In eighth grade math, 68% are predicted to “fail.”

Who would deliberately create a cut score in which you predict that that 68% of your students will fail? Shouldn’t the test change? Or shouldn’t you wait until the teaching has a chance to change? These are high-stakes test, which can determine whether you get a high-school diploma or not!



Categories: Teaching & Learning