Author: Whitehurst, Grover J., Chingos, Matthew M., & Gallaher, Michael R.

Title: Do School Districts Matter?

University Affiliation: Brookings Institute

Email: gwhitehurst@brookings.edu, mchingos@brookings.edu

Research Question: 1) What is the influence of school districts on student achievement relative to the influence of schools, teachers, and individual differences among students? 2) Are there differences among school districts in their contribution to student achievement that are large enough to be relevant for policy? 3) Can districts be categorized based on patterns of influence on student academic achievement in ways that would inform efforts to improve district performance 4) What are the distinctive features of exceptional districts?

Published: No

Journal Name or Institutional Affiliation: Academic Questions

Journal Entry: N/A

Year: 2013

Findings: School districts only account for 1% to 2% of the variance in student achievement. However, in Florida 10% of districts perform below the expected average given the characteristics of students and 7% perform below. For North Carolina, 10% perform above and 14% perform below. Given that student level characteristics account for 59% of the variance in achievement, this shows that at the extremes, school districts may have an impact. At these extremes, 0.30 standard deviations separate the highest and lowest performing districts in Florida and 0.42 standard deviations of difference exist in North Carolina. This means that the average 4th and 5th grader in the top-performing districts in Florida and North Carolina would be 60% and 80% of a school year ahead of the average 4th and 5th grader in the lowest performing districts, respectively.

Scholarship Type Unpublished Research

Keywords: Academic Achievement, District, Geographic Location, Policy

Regions South

Methodologies: Quantitative

Research Designs: Mathematical models

Method of Analysis: Mathematical Models, Multilevel Models

Sampling Frame: 4th and 5th grade students over the decade spanning 2000-2010 in Florida and North Carolina

Sample Types: Population

Unit of Analysis: School District

Data Types: Quantitative-Panel Data

Data Description:

The dataset is comprised of assessment tests in reading or mathematics taken by 4th and 5th grades students in Florida and North Carolina spanning the decade from 2000-2001 school year to the 2009-2010 school year. Each year there are roughly 500,000 observations resulting in approximately 5 million student data points for the decade.

DV: Reading and math scores

IV: School district

Controls: Student age, race/ethnicity, cognitive disability status, free and reduced lunch program status, limited English proficiency status, and for Florida only whether the parent/student are native English speakers and whether the student was born in the US

Relevance:

Entry Created at: 2013-07-25 02:45:56 UTC
Last Update: 2016-07-21 19:32:22 UTC

Back