Date of Award

4-2019

Document Type

Thesis

First Advisor

Humberto Barreto

Second Advisor

Jamie Stockton

Third Advisor

Steven Bogaerts

Abstract

Federal and state governments want to know that their constituents' tax dollars are being spent well, and that public education is serving the People well. To assess a school's efficacy, they calculate an accountability score. Although state-level scoring systems vary, neither the federal law nor a single state's law require that demographic characteristics of a school's student population be held constant when it is being evaluated. This complicates the assessment of a school's adequacy because factors outside of a school's control influence student performance; when a school is evaluated without taking this into account, it is being unfairly credited with its students' successses and failures. I use OLS regression, holding some population characteristics constant, to create predicted weighted average ISTEP scores for third grades in Indiana to compare to their actual scores. The deviation between actual and predicted scores more accurately reflects how well or poorly the school is educating its students. I propose an alternate, more statistically rigorous grading system which identifies exceptionally performing schools. This approach is by no means foolproof, but it certainly gives a better foundation for assessing a school's influence on student outcomes than a grade based upon a school's average standardized test score. This current approach controls nothing about student population, while my method gets us closer to a fair evaluation of a school's performance.

Comments

Honor Scholar Thesis

Download

Share

COinS