More evidence that all tests are not created equal?

See all posts

While good algebra teachers can raise students' test scores, there's apparently scant evidence that good ninth-grade English teachers--at least those teaching in North Carolina--can do the same, according to a study published this month by economist C. Kirabo Jackson.

Go figure.

While previous studies found that good high school teachers, no matter what the subject, do yield better math and reading outcomes, Jackson's methodology appears more precise, accounting for the influence of multiple teachers in the time frame and distinguishing between different sections of a course, given that one section of English, for example, might contain higher-scoring students than another.

Jackson doesn't have any explanation for a finding that disses on English teachers, as he rules out any number of technical possibilities. We turn to one he didn't consider, the validity of the end-of-course tests used in North Carolina to measure student progress.

The latest findings from the Gates MET project adds fuel to the fire when it comes to our ongoing concerns over the validity of some tests, especially states' tests. For example, the Gates team found that when students took the SAT-9 open-ended reading test, their test scores had a much stronger relationship with teacher quality than when they took state English Language Arts tests.

*Note: CLASS, FFT and PLATO refer to different classroom observation instruments.

Source: "Gathering Feedback for Teaching: Combining High Quality Observations with Student Surveys and Achievement Gains--Policy and Practice Brief", Measures of Effective Teaching Project, Bill & Melinda Gates Foundation

The SAT-9 involves writing--something at the core of many 9th grade English curricula. State reading tests, including those in North Carolina, are multiple-choice and less "cognitively demanding" given the absence of writing exercises.

Jackson's findings come on the heels of the much-publicized study by Raj Chetty and colleagues (see above), showing the positive life outcomes that can be traced back to great teachers, regardless of what some test scores may indicate.

By our count, that makes two studies this month alone showing potential problems with using the current regime of tests to measure teacher performance. With high stakes evaluation instruments right around the corner, fixes are needed sooner, not later.