Working to find actionable solutions to the nation's key challenges.

Paper Trails and Audits Can Boost Confidence in Elections

Wednesday, August 12, 2015

Share on FacebookTweet about this on TwitterShare on LinkedInPin on PinterestShare on TumblrEmail this to someonePrint this page

Fair elections represent a fundamental pillar of our democracy, and universal trust in the electoral process is integral to giving our elected officials legitimacy. This makes it all the more alarming when one considers that voters are not necessarily confident with their own votes. A Bipartisan Policy Center (BPC)/USA Today poll found that one-third of Americans were not “very confident” that their votes were counted correctly in 2012. Other polls have found that over two-thirds of Americans were not “very confident” that votes across the country were accurately counted.

So, why the lack of confidence? We know that partisanship and race play a role in voter confidence. But the more manageable and imminently pressing problem involves voting technology, and concerns that irregularities, even if caught, could change election results.

We do have some ideas on how to combat this risk. The Presidential Commission on Election Administration (PCEA), whose work BPC is continuing, is focused on improving the experience of the American voter. Within their recommendations, the commission advocated for better audits of vote tallies and electronic voting machines to ensure more accurate elections. These proposals stem from our nation’s extensive, and sometimes contentious, history of elections.

The Need for a Paper Trail

Most election results are tabulated accurately. Without a verifiable paper trail, however, we may never know about the ones that are not. In fact, a quarter of Americans went to the polls in 2012 to vote on a machine, and many of those machines had no paper trail.

It is easy to see why this makes voters wary. In 2004, a voting machine in Carteret County, N.C., with no paper trail lost 4,530 ballots because it ran out of memory. Florida’s 13th Congressional District race in 2006 saw a similar controversy, with an unusually high 18,000 blank votes in one county. This under-vote rate of 13 percent was more than double any other county, but since a paperless electronic machine was used, there was no way to dispute the result. Florida’s Palm Beach County actually saw two wrong winners certified in the 2012 elections due to a software glitch, and numerous other races were much closer than originally thought. Unlike the previous two examples, a paper trail was available, and a long legal battle eventually led to the true winners being certified.

These examples, along with a slew of similar stories across the country, are just the ones we know about. The rise of technology in our elections has introduced a new level of risk for simple error or fraud. Thankfully, states such as Ohio, California and Arkansas have recently joined what is now a majority of states that require some form of paper trail for electronic machines.

The Role of Risk-Limiting Audits

With the worst nightmares of election administrators already realized, from lost ballots to wrong winners, the argument turns from whether we need a paper trail, to how to use it effectively.

Many states with electronic voting methods only conduct full recounts randomly or in close races. While this may take care of reported close elections, it does not take into account a malfunction that skews or even reverses the results.

The best solution, as recommended by the PCEA, is for states to implement “risk-limiting” audits. Such an audit would occur for every race after it had been unofficially tallied, but would only take a random sample of ballots determined by the competitiveness and size of the election. For closer elections, a larger sample of ballots would be needed to verify the results, while a less competitive race would only need a small number of hand-counted ballots to ensure accuracy. The process allows elections officials to verify results to a certain level of statistical confidence, a concept universally used by statisticians and researchers. In addition, it can cut down on costly full recounts, which would only need to happen in “too close to call” elections or when audits find an irregularity.

Indeed, some states have already begun to implement this process. Both Colorado and California have moved to use the process statewide.

Going Forward

Three steps, taken at the federal, state, and local levels, can implement the improvements discussed here:

  1. The U.S. Election Assistance Commission, which now has a quorum of commissioners, could share voluntary best practices for risk-limiting audits. This information, gleaned from those jurisdictions already conducting audits, could even be specific to the level of confidence desired. With a simple table, the commission can inform election administrators how many ballots they should randomly sample, depending on competitiveness and number of votes cast.
  2. States could require risk-limiting audits and paper trails for all races.
  3. Localities and states could work together to find sources of funding as their voting technology ages, including costs associated with continued maintenance of existing machines and testing of new machines.

The views expressed here are those of the author, and do not necessarily represent or reflect the views of BPC or PCEA.