CEC Theses and Dissertations

Campus Access Only

All rights reserved. This publication is intended for use solely by faculty, students, and staff of Nova Southeastern University. No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, now known or later developed, including but not limited to photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author or the publisher.

Date of Award

2012

Document Type

Dissertation - NSU Access Only

Degree Name

Doctor of Philosophy in Information Systems (DISS)

Department

Graduate School of Computer and Information Sciences

Advisor

Marlyn Littman

Committee Member

William Hafner

Committee Member

Ling Wang

Abstract

The Help America Vote Act of 2002 (HAVA) was a response to the controversial presidential election of 2000. In accordance with HAVA requirements for federal elections, states were mandated to replace punch card voting systems and mechanical lever voting machines with more up-to-date systems that use current technology. As replacements, states selected optical scan (OS) and direct record electronic (DRE) voting systems.

Computer scientists questioned the security of OS and DRE voting systems, and politicians questioned their accuracy. Thus, the goals of this research were to analyze the accuracy of election outcomes generated by electronic voting (e-voting) systems and to document whether e-voting machines were trustworthy (i.e., accurately recorded the voters' intent) and secure (i.e., votes were not altered). To achieve these goals, the author developed an embedded case study and incorporated ethnographic and quantitative techniques. The author observed election officials in two Missouri jurisdictions perform pre-election, Election Day, and post-election tasks.

Specifically, the author observed election officials in Cape Girardeau County perform pre-election tasks, such as logic and accuracy (L&A) testing. In the state of Missouri, pre-election L&A testing involved loading the ballot and was considered finished when the e-voting system was ready for voters. The author identified pre-election adversarial strategies and then used a six-step risk analysis process to identify the most important risks. After following the steps, the author identified 11 e-voting components as high-level security risks.

Additionally, the author observed election officials in St. Louis County, Missouri conduct the 2010 midterm election and post-election activities, which included the manual tabulation of ballots. Election Day culminated with unofficial outcomes generated from the e-voting systems, while the post-election activities yielded official outcomes. To analyze the accuracy of e-voting systems, the author computed confidence intervals for the differences between unofficial and official 2010 midterm election outcomes from statewide races in St. Louis County. Based on these confidence intervals, the author concluded that the e-voting systems used in the state of Missouri were between 99.768% and 99.774% accurate.

To access this thesis/dissertation you must have a valid nova.edu OR mynsu.nova.edu email address and create an account for NSUWorks.

  Contact Author

  Link to NovaCat

Share

COinS