What We Do With Data
Newly installed as the chair of the Senate Committee on Health, Education, Labor, and Pensions,Lamar Alexander is determined to shepherd through a reauthorization of the long-languishing Elementary and Secondary Education Act (ESEA). Seeking to put some daylight between the new ESEA and its widely reviled rebranding as No Child Left Behind (NLCB), Alexander has convened hearings and solicited testimony on several key aspects of the bill.
Among the most debated provisions is the NCLB requirement for annual statewide testing. Currently, the law requires states to administer 17 tests – one each in math and literacy annually in grades 3 through 8 plus at least once in high school, as well as a science test once in elementary, middle, and high school. But these tests are merely the tip of the iceberg. Because of the high stakes attached to the federally-mandated tests – including stakes related to teacher licensure, performance pay, and school closures, among others – a new industry of test design and test preparation has sprung up to ensure students and teachers excel when it matters most. As HGSE associate professor (and former senior advisor to Senator Alexander) Martin West noted in his recent Senate testimony, analyses that students are scheduled to spend a reasonable 1-3 percent of their time taking standardized tests almost certainly understate the reality and that “far too many schools devote excessive time to narrow test-preparation activities in an attempt to avoid federally mandated sanctions.”
Despite this concession, West recommended that the Senate panel preserve annual testing for several reasons: it enables more precise measurements of school performance (thus preventing, say, schools from being falsely designated lower-performing when they serve more disadvantaged students); these more precise measures of school performance can be used to give parents better information when selecting schools; and by testing all students in every school each year, we can collect data about the relative performance of subgroups, for example by race or socioeconomic status.
Rather in spite of myself, I find these arguments convincing. Were I emperor, I might even decree that we continue to require annual testing. After all, the data we collect can and do tell sobering stories about equity, and we should not turn away from that. On the contrary, these stories ought to be catalysts for action. Where I differ with Professor West and others is in just what we – or more precisely, policymakers in a position to exert some influence – ought to do with the data we collect, what actions we should take.
Education policy debates spend a lot of bluster on operationalizing the first two of West’s reasons – better performance measurement and better information sharing – but I think we ought to invest more energy focused on what data tell us about inequality and how to address it. Specifically, I believe we need to start enacting new race- and class-conscious policies. Assuming that we find racial and socioeconomic achievement gaps troubling - which we should and which I believe most people do - we should move away from stubbornly colorblind education policies and embrace policies that take direct aim at racism, poverty, and the steady erosion of civil rights. As UCLA’sGary Orfield has observed, today’s yawning achievement gaps – illuminated with such clarity What We Do With Data | James M. Noonan: