Dissertation/Thesis Abstract

Achievement gap projection for standardized testing through logistic regression within a large Arizona school district
by Kellermeyer, Steven Bruce, Ed.D., Northern Arizona University, 2011, 162; 3490539
Abstract (Summary)

In the last few decades high-stakes testing has become more political than educational. The Districts within Arizona are bound by the mandates of both AZ LEARNS and the No Child Left Behind Act of 2001. At the time of this writing, both legislative mandates relied on the Arizona Instrument for Measuring Standards (AIMS) as State Tests for gauging their effectiveness.

This study utilizes the data from a large Arizona District to measure potential achievement gaps in the AIMS Reading and Math Tests at the tenth grade level. Six years of data were compiled yielding 23,846 valid AIMS Reading Tests and 23,640 AIMS Math Tests. Logistic regression and complex samples logistic regression were employed to determine if District Reading and Math scores would reach the 100% required proficiency rating of "Meets" or above by 2014 with affects the graduating class of 2016. The investigation also satisfied the question of whether high schools which are comparable to the demographics of the aggregated District performed similarly to the District in general.

This researcher has shown that an achievement gap can be defined for a District by the legislated disaggregation categories of student achievement for each subgroup according to gender, ethnicity, disability status (special education), English language proficiency, and low-income status. Furthermore, these subgroups made use of odds ratios for probability of passing and failing for the Reading and Math Tenth Grade AIMS Tests.

Discussions include the ramifications of ignoring achievement gaps in subgroups and revealing evidence that individual schools can make a difference in closing achievement gaps. This form of regression may allow an improved method for judging school improvement based on student performance on high-stakes tests. By ranking schools under a universal paradigm that assesses how a school succeeds with each subgroup compared to its own population could distinguish schools that are genuinely reducing the achievement gap compared to schools that simply have many students who are scoring well.

Indexing (document details)
Advisor: Wiggall, Richard
Commitee: Clement, Nicholas, Dereshiwsky, Mary, Emanuel, Gary
School: Northern Arizona University
Department: Educational Leadership
School Location: United States -- Arizona
Source: DAI-A 73/04, Dissertation Abstracts International
Subjects: Educational tests & measurements, Educational evaluation, Educational leadership
Keywords: Achievement gap, Arizona, Logistic regression, NCLB, No Child Left Behind, Reading and math, Standardized testing
Publication Number: 3490539
ISBN: 978-1-267-11493-8
Copyright © 2020 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy