Abstract :
The accurate counting of ballots is essential for a functional democracy. In recent elections, particularly the exceedingly narrow presidential one in 2000, widespread concerns surfaced that votes were not being counted accurately. This paper examines major voting technologies, their advantages and disadvantages, and the significance of residual ballots (overvotes and undervotes) across U.S. counties in the presidential elections of 2000 and 2004. In the 2000 election, 1.9 million ballots were voided, as were more than 2.3 million in 2004. The analysis explores three fundamental questions: (1) Do voting technologies tend to favor one political party over its rival? (2) Do voting technologies tend to favor one ethnic group over another? (3) Do voting technologies favor urban areas over rural ones? The empirical results consistently deny the existence of any of these biases at the national level, although the possibility of local bias remains an open question. The conclusion links these issues to the on-going debate about voting technology reform.
