 |
Convened by the American Association
for the Advancement of Science
Main | Participants
Before laying out any specific suggestions for research needs relating
to voting, I need to point out that everything I’m writing here is my
own opinion and does not in any way reflect official views of the Congressional
Research Service. CRS does not advocate or take positions on legislative
issues.
The argument can be made that in general, voting in the United States
works pretty well. Nevertheless, there are several areas in which significant
research investment could be very useful. Here are some that come to
mind:
- It seems clear, based on media reports and the activities of various
advocates, that the most important issue to solve with respect to voting
systems is security of the vote. There appears to be widespread consensus
that this is an important issue, but there is substantial disagreement
on what should be done about it. In the absence of significant R&D
activity aimed at providing a level of security that will be widely
regarded as sufficient, many have resorted to demanding that paper ballots
be used. While paper does have advantages, the problems with it tend
to be ignored. The advantages and disadvantages of various kinds of
verification systems and other security approaches — not only from a
security perspective per se but also with respect to usability, cost,
administrative burden, and so forth — need to be much better understood.
In the absence of a major effort to find an optimal solution, it is
possible and perhaps even likely that the nation will continue to drift
into the use of paper ballots in a way that will make potentially better
solutions much more difficult to develop and implement. The potential
of a system that provides true universal verifiability without sacrificing
the secret ballot is especially tantalizing and could do much to improve
the integrity of the voting process. It could be a significant loss
to the future of the democratic process should such possibilities be
denied.
- Accessibility is still awkward for persons with disabilities using
DREs or OS interface systems, and for improved usability for all voters.
This is an area ripe for new ideas and research. For example, the audio
systems used for visually impaired voters are awkward and time-consuming
for both the voter and the pollworker; and with all the attention on
paper ballots, some significant attention to how to make them accessible
would be appropriate.
- Perhaps too much attention has been placed on technological solutions
per se. A significant area for improvement is the reengineering of
the entire process of election administration, from voter registration
through election certification. HAVA and other factors seem to have
made election administration far more complicated, and may actually
have increased the risk of problems by creating more potential points
of failure. While improvements in voting technology appear to be making
things easier for voters — an important goal — they also seem to be
making running elections harder for administrators and pollworkers.
The ability of a small procedural lapse to create major problems for
the smooth running of elections is illustrated by the difficulties in
the September primary in Montgomery County, MD, which stemmed from a
failure to include smartcards in the supply cases. R&D on voting
technology should include research in how to keep procedures as simple,
easy to perform, and error-free as possible. The recent attention that
the healthcare delivery enterprise in the United States has started
to give to process reengineering has already led to enough positive
results to show the potential for such an approach.
- Voting is an area where classic research stovepiping is especially
counterproductive. A lot of the noise about whether paper ballots are
necessary for security has arguably been magnified by differences in
disciplinary perspective, with insufficient attention being paid by
different sides to the overall picture. This is hard to change, but
the development of active, interdisciplinary programs and consortia
to identify key needs and implement needed R&D could be highly productive.
Meetings such as this one, while very useful, cannot by themselves move
the field. There is now a strong core of knowledgeable researchers
and election officials dedicated to doing the necessary work, but there
is still no persistent mechanism for them to work together to determine
what kinds of research and data are needed, what is practical, and how
to make the needed R&D happen.
- Finally, a word must be said about funding and academic priorities.
Until the year 2000, there was no real academic future for researchers
interested in issues in election administration (except certain fairly
narrow questions), despite the intellectual richness of the problems
and the practical potential of solutions. The situation appears to
have improved, but it is arguably still far from where it needs to be.
Part of the reason is that there is little funding for such research
and dim prospects for implementation, caused to significant extent by
the episodic nature of elections, and the fragmented and limited market
for the technology. These are perhaps more policy and resource issues
than research questions, but careful examination of the problems would
nevertheless be likely to help yield good solutions.
|
