Getting the count right

Ian P. Cook
The RAND Corporation

As the votes are finally counted, enormous questions remain about the reliability of voting machines. West Virginia, Colorado, Tennessee and Texas have reported problems with their machines. Paper ballots have been put on standby in Pennsylvania. The District of Columbia is citing voting errors due to “static electricity” found on the voters themselves.

Come Wednesday morning it may be nearly impossible to tell the difference between legitimate votes and those created or lost by the machines.

It’s too late to fix the machine problems. But it is still possible to enhance the integrity of the election. By posting on websites the raw voting data – not just the vote counts ultimately approved by precinct and state election authorities – election bodies can help deter the threat of fraud, raise the likelihood of detecting systematic errors, and elevate the level of public trust in the election results.

The United States adopted electronic-based voting and vote-counting at a rapid rate in the aftermath of the contentious 2000 election, in part supported by the Help America Vote Act. The 2002 legislation called for replacing punch-card voting systems, created the Election Assistance Commission to assist in the administration of federal elections, and established minimum election administration standards.

Confidence in the results of an election must be independent of the machines we use to tally them. This means having a clear and accurate count, and being able to ascertain the provenance of every last vote.

Problems with electronic machines have been noted and dissected by a range of prominent experts in software engineering, computer science, political science, and elsewhere. Meanwhile, the process by which our leaders are chosen becomes more opaque.

Opening the information to the public brings the benefits of transparency, but also allows for mass investigation. Verification and auditing could be placed in the hands of interested, capable reviewers. The function of election commissions, then, is to be available for feedback on various findings reported from official and unofficial sources.

Having data in hand for the entire nation would allow for statistical review to cross precinct, county, and state lines the same way the voting machines have done. The technological glitches of a single machine could affect numerous locations. Yet the standard practices for recounts and audits require only highly localized consideration. To put it another way, aggregated miscounts could affect a national outcome without being noticed at the lower levels.

The open publication of tallies would not necessarily mean all reviews would be done with equal skill, or that they should be given equal weight. Decision-making power would still rests in the hands of local election officials who would be far better informed – as would news sources and thus the public.

Massive vote-counting failure at the polls may drive political will toward a solution to the problem. But a more salient concern is that a series of less cataclysmic mishaps might simply spoil votes from subsets of the population, and remain unaddressed on the national level. Fortunately we can examine statistics, at a level beyond current use, to begin rooting out error or fraud.

We cannot ever really rely on machines being the answer to making sure we have an election we can have faith in. Indeed, no system yet devised has been free of the potential for problems.

So simply pursuing a technological fix to the current situation will never result in perfect security. But we should not compound the threat by allowing the machines to obscure the source and type of erroneous votes.

Delivering on the promise that every vote should count depends fundamentally on our ability to count votes, not just cast them.

Editor's note: Ian P. Cook is a management systems analyst at the RAND Corporation, a nonprofit institution that seeks to improve policy and decision-making through research and analysis.