I read with interest the Manhattan Institute’s report, “Building on the Basics: The Impact of High-Stakes Testing on Student Proficiency in Low-Stakes Subjects.” The authors looked at what happened to science scores in schools that received an “F” grade for their reading and math scores in the previous year. Did focusing on reading and math instruction “crowd out” proficiency in other subjects? The authors conclude that it did not, and that in fact, the F-grade sanction produced small positive gains in science proficiency. If correct and generalizable, this result would have profound policy and curriculum implications across the nation.

After a closer look at the study, I had some questions about the methods.  Since Florida does not test students in science in 4th grade, the authors used the students’ math and reading scores in 4th grade to estimate their performance in science. They then compared this estimated 4th grade performance with 5th grade performance on the Science FCAT. The authors stated that “this procedure assumes that student proficiency in these subjects is highly correlated and that there was no differential relationship in student knowledge among these subjects in the five categories of schools before the F-grade sanction was introduced.” Given that the authors also conclude (tentatively) that math and reading ability enable learning of science, it seems that their method of estimating 4th grade performance assumes the very outcome they purport to discover. A commenter on Eduwonkette raises this and other relevant questions about the methods.

More on “Building on the Basics” tomorrow….