Fraud or Failure? What Incident Reports Reveal about Election Anomalies and Irregularities

Abstract

When things go wrong in elections involving direct-recording electronic (DRE) voting technology, these episodes are viewed by many as proof of the vulnerability, or at least the unreliability, of these systems. Claims by election officials that such problems are “par for the course” in elections or symptomatic of “growing pains” associated with implementing a new technology ring false to many Americans who expect elections to be run without error every time. To date, however, each side in the debate has been able to rely on only limited data and scant research. DRE technology has only recently been introduced on a large scale in the United States, so there is little systematic information regarding the difficulties encountered in its implementation. In this chapter, we examine a novel and potentially very useful source of data concerning the frequency and severity of different types of problems encountered by voters and precinct workers in a DRE environment. The data consist of incident reports collected by poll workers during the May 2, 2006, primary election in Cuyahoga County, Ohio. This election marked the first use of DRE technology in this jurisdiction. Voters cast their ballots on Diebold Accuvote-TSx voting machines -- touch-screen machines equipped with printers to produce the voter-verified paper audit trail (VVPAT) mandated by Ohio election law. Since most voters were unfamiliar with the technology, the Cuyahoga County Board of Elections took the prudent and useful measure of providing poll workers at each precinct with incident report forms to record and to describe difficulties they encountered in conducting the balloting.