Does Electronic Voting Affect Turnout?
The situation on the voting technology front remains dynamic, both nationally and internationally. As election officials throughout the country contemplate how they'll meet the 2006 for the replacement of punch cards and the implementation of accessible technology, the Election Assistance Commission is contemplating new voting system guidelines according to GCN. The three remaining commissioners are presently considering recommendations of the EAC's Technical Guidelines Advisory Committee. The new recommendations reportedly include new provisions on security and accessibility

After the commissioners' initial review, the proposed guidelines will be publicly available for comment for 90 days, after which they'll go into effect. While states aren't required to follow these guidelines, about two-thirds voluntarily do so. This will likely put state and county officials, as well as voting machine vendors, in a tight spot given that the guidelines won't be final before the end of the summer and new voting technology has to be in place by the first elections of 2006.

The U.S. isn't the only place where electronic voting is the subject of vigorous debate. See this story on the Australian Electoral Commission's plan to issue a white paper this summer on the risks and benefits of electronic voting. Australia is also spending $17 million to upgrade its electoral role and election management computer system.

Meanwhile, as I mentioned Friday, two Berkeley economists have posted this paper on the effect of electronic voting on voting participation. In contrast to some of more sensationalistic reports issued in the immediate aftermath of the November 2004 election (e.g., optical scans were used to steal votes in Florida), the Card/Moretti report seems to be a serious empirical analysis.

The study differs from most analyses of voting technology, which focus on "residual votes" -- those ballots for which no presidential vote is recorded, either because of a undervote or an overvote. That's a good way of determining whether those who went to the polls had their votes counted, but not of assessing whether technology affects turnout. By contrast, the Card/Moretti study looks at the total number of votes cast for the presidential candidates in 2000 and 2004.

The paper's findings are intriguing. On the whole, they find a slight positive correlation between support for Bush and the use of electronic voting. But why? Card and Moretti conclude that it's implausible that this correlation is the result of irregularities by Republican election officials. This is largely because there's a smaller correlation in places with a Republican governor or Secretary of State, and in swing states.

So if tampering with electronic voting doesn't explain this correlation, than what does? It could perhaps be chance, but it also might be that the considerable anxiety that has surrounded electronic voting had a differential impact on voter turnout.

This possibility is suggested by a survey released in August 2004, which I blogged on here . Although it got almost no public attention at the time, this study found confidence in electronic voting to be much higher among Republicans than Democrats -- unsurprisingly, given that most of the anxiety, and in some cases paranoia, regarding electronic voting security was among those on the left side of the political spectrum. More to the point, that study found that many more Democrats (23%) than Republicans (9%) said they wouldn't vote at all due to distrust of electronic voting.

Even the most strident critic of electronic voting would acknowledge that the likelihood of any individual's vote being tampered with is exceedingly small, not enough to warrant one's staying away from the polls entirely -- as someone put it to me, if you don't vote at all, you can be sure your vote won't count. Still, it seems quite plausible that more Democratic than Republican voters might have chosen to stay home because of the widely circulated campaigns against paperless voting that took place in the months leading up to the November 2004 election.

In other words, it may be that the pre-election worries about electronic voting turned out to be a self-fulfilling prophesy (albeit to a very small extent). Electronic voting itself may not have cost Kerry any votes; to the contrary, it's likely that of those who turned out, more had their votes counted and that Kerry voters were helped more than Bush voters. States that moved to electronic voting, like Georgia and some Florida counties, saw steep declines in the percentage of uncounted votes, particularly in heavily minority precincts. But the anxiety surrounding e-voting -- in particular, the widely propogated fear that the computer might "eat my vote" -- might have caused more Democrats than Republicans to stay away from the polls. - posted by Dan Tokaji @ 11:05 PM