Abstract

This paper calculates a time series of simple, standard measures of schools' relative performance. These are drawn from a 1997-2004 panel of Chilean schools, using individual-level information on test scores and student characteristics for each year. The results suggest there is a stark tradeoff in the extent to which rankings generated using these measures: i) can be shown to be very similar to rankings based purely on students' socioeconomic status, and ii) are very volatile from year to year. At least in Chile, therefore, producing a meaningful ranking of schools that may inform parents and policymakers may be harder than is commonly assumed. © 2006 Elsevier B.V. All rights reserved.