In the past decade, forty-nine of the fifty states have adopted some form of statewide testing, which they are beginning to report in user-friendly "school report cards." Most of the report card programs have no stakes or low stakes, so what purpose do they serve? They create openness and information. They light up schools for the people who are supposed to make informed decisions about schools but who are too often kept in the dark: parents and legislators. The information is imperfect, certainly, but parents and policymakers armed with information can at least start a discussion about school improvement.
How much can one expect from a policy that just informs, with few stakes? I initially thought that one could expect only negligible results. It turns out that I was wrong. The states that started their testing programs at the beginning of the last decade did not break achievement records, but they did improve achievement at a significantly faster rate than states that only began testing in the past couple of years.
One can show this with simple calculations based on the National Assessment of Educational Progress (NAEP)-the only nationally uniform test that is administered to large representative samples of students in each state. State NAEP testing, which began in 1990, is designed to help states track their achievement over time. No one "teaches to" the NAEP test because only a sample of students take it and no individual school's NAEP scores get reported.
I recorded the first year in which each state began its report card program. (I did not count a state if it conducted tests but kept parents in the dark about school and district results.) States with early school report card programs include Texas, Washington, Kentucky, Maryland, Michigan, Connecticut, and Wisconsin. States that got on the report card bandwagon only lately include Iowa, Nebraska, Utah, Virginia, and South Carolina.
I looked to see whether states experienced growth in NAEP scores after they began report card systems (compared to states that had not yet begun such systems). I found that nine-year-olds' reading scores improved 0.26 points faster each year after testing began. Thirteen-year-olds' math scores improved 0.28 points faster each year after testing began. (Full results are available at www.economics.harvard.edu.) These gains are not tremendous, but they are statistically significant and do add up over a decade. At the end of ten years, the nine-year-olds' reading scores would be 2.6 points higher and the thirteen-year-olds' math scores would be 2.8 points higher.
Statewide standardized tests and school report cards may be unpleasant for ineffectual educators, but they should not be controversial with parents or policymakers who want to see higher achievement. Schools conduct themselves better when their constituents are informed.