Standardized Test Scores and Accountability Policy: A Quantile Regression Approach
MetadataShow full item record
Standardized test scores are the primary mechanism that state and federal education policy makers use to hold schools accountable. Student achievement data is used to hold schools accountable for certain thresholds of test scores for all students, and in some cases, subgroups of students. The typical methods, specifically ordinary least squares, employed by policymakers and researchers that study student achievement are generally concerned with averages. The use of average effect models generally results in one-size-fits-all policy interventions meant to raise average tests scores across a large population of students. This approach often does not take into consideration the differential effects of explanatory variables across te entire distribution of standardized test scores. Quantile regression supplements ordinary least squares by generating information across the entire distribution of student achievement, from the lowest to highest performers. As this study shows, federal and subsequent state education policies are concerned with the lowest performers, and not just the average student or school. This study used data from K-12 public schools in Oregon that test students in grades 3-8 and 11 in mathematics and reading. The results show that quantile regression is a useful tool for analyzing school standardized tests scores in an acccounability framework by providing information that ordinary least squares generally misses.