Caltech/MIT: What has changed, what hasn’t, & what needs improvement

The Caltech/MIT Voting Technology Project has released a thorough, comprehensive, and insightful new report timed to the 2012 election: VOTING: What has changed, what hasn’t, & what needs improvement <read>

The report itself is 52 pages, followed by 32 pages of opinions of others, including election officials, advocates, and vendors, some of whom disagree with some aspects of the report. Every page is worth reading. The report is not technical. It covers a wide range of issues, background, and recommendations.

We find little to quibble with in the report. We agree with all of its recommendations although we might place different emphasis in particular areas:

As we have studied the areas where progress has been made since 2001, and where progress has stalled, we have developed the following recommendations. All have been discussed earlier in our report, and we summarize them here. They are not in priority order. First, regarding voting technology, we recommend:

  • Legislation mandating effective election auditing, which at a minimum would require post-election auditing of all voting technologies used in an election.
  • Continued strong support for voting systems security research, emphasizing auditing and the verifiability of election outcomes.
  • A movement toward mandating statistically meaningful post-election audits, rather than setting security standards for election equipment, as the primary way to safeguard the integrity of the vote.
  • A new business model led by states and localities, with harmonized standards and requirements.

Second, regarding voter registration, we recommend: » Streamlining the provisional balloting process in many states and the creation of common best practices and voluntary standards across states.

  • The development of voter verification systems in which states bear the cost of stringent voter ID regimes, in those states that desire to increase ID requirements for in-person voting.
  • Continued standardization of voter registration databases, so that they can be polled across states.

Third, with respect to polling places and pollworkers, we recommend:

  • Continued improvement of pollworker training and more reliance on network technologies to facilitate pollworker training.
  • Development of applications deployed on mobile devices that bring more information to pollworkers, and transmit real-time data about Election Day workloads back to the central voting office and the public at large.
  • Increased functionality of electronic pollbooks and their wider adoption.
  • Development of applications that gauge how long voters are waiting in line to vote, so that wait times can be better managed and reported to the public.

Fourth, regarding absentee and early voting our first two recommendations repeat those we issued a decade ago; the third is new:

  • Discourage the continued rise of no-excuse absentee balloting and resist pressures to expand all-mail elections. Similarly, discourage the use of Internet voting until the time when auditability can be ensured and the substantial risks entailed by voting over the Internet can be sufficiently mitigated.
  • Require that states publish election returns in such a way that allows the calculation of the residual vote rate by voting mode.
  • Continue research into new methods to get usable ballots to military and overseas civilian voters securely, accurately, and rapidly and to ensure their secure return in time to be counted.

And, finally, regarding the infrastructure and science of elections: » Continued development of the science of elections.

  • Continued, and expanded, support for the research functions of the Election Assistance Commission.
  • Development of an Electoral Extension Service, headquartered in each state’s land-grant colleges, to disseminate new ideas about managing elections in the United States.

Several items with which we fully endorse were covered in this report which sometimes are missing from the discussion or often underemphasised:

The Risks of Mail-in and No-Excuse Absentee Voting

The report thoroughly covers the disenfranchisement risks of mail voting which are about double polling place voting. Such voting does not increase turnout significantly, except in local elections. We would have liked to seen more coverage of the organized fraud, vote buying, and coercion frequently occurring via such voting. These are  not just theoretical risks. New to us was the surveys showing that the public at some level recognizes the risks and show less confidence in elections with expanded absentee or mail-in voting.

The Emphasis on Election Auditing over Machine Testing and Certification

It is theoretically impossible to develop or test a completely safe voting technology. Extreme testing and slow certification requirements stifle innovation, add costs, delay improvements and are ultimately ineffective. High confidence, efficient statistical audits, paper ballots, combined with a strong chain-of-custody are a necessary solution that eclipse the elusive pursuit of technical perfection.

The Need and Value of Quality Voter Registration Combined with Online Voter Check-in

The report points to the fallacy of votER fraud. Yet there are efficiencies and enhanced enfranchisement available from better, more accurage voter registration databases. There are solutions with online check-in that also provide voter-id without the disenfranchising aspects of the currently proposed voter-id laws.

The Challenges of the Election Technology Industry

My years of experience in the software industry always lead me to the conclusion that the election technology industry is a losing business proposition. While I am not enamored with any of the current voting technology vendors, there is little incentive for them or new players to enter the field. The closest analogy is the defense industry. That industry is not fragmented, has essentially one customer, which designs products and pays for research and development. The voting technology industry is fragmented and has a fragmented customer base, with varying demands, coupled with a very difficult sales environment.

Recognition of One of the Risks of the National Popular Vote Agreement

  • The proposed National Popular Vote (NPV) may have negative security implications, since the opportunity to perform proper post-election audits appears to be considerably diminished.

CTVotersCount readers know that we would go farther and cover the risks of a national popular vote in our current state-by-state fragmented system, not designed to provide an accurate national popular total. Alleged popular totals cannot be audited, cannot be recounted, and electors must be chosen before an official count is available. The National Popular Vote agreement does nothing to address the existing risk issues with the Electoral College and, in fact, adds to the risks.

Facebooktwittergoogle_plusredditpinterestlinkedintumblrmailFacebooktwittergoogle_plusredditpinterestlinkedintumblrmail

Leave a Reply

You must be logged in to post a comment.