Disclosure of computerised voting information from public elections
Dr. Lee Naish
Senior Lecturer
Department of Computer Science and
Software Engineering
University of Melbourne
President, Victorian branch
Proportional Representation Society of Australia
27th October 1999
Full public disclosure of the raw data used for electronic vote counting
in public elections is in the public interest.
Without public disclosure most people have very little information on
voting details. Some people have access to more detailed information,
due to court actions, for example, but only a few have access to
all the data. This creates unfair advantages for those with access,
distorting the democratic process. For example, allowing exclusive
analysis to decide on future campaign strategies or predict who will
be elected if a casual vacancy arises. Furthermore, there are many ways
in which the public can directly benefit from public disclosure. Some of
these are outlined below.
- Validation of election results
- Manual counting of votes from public elections is scrutinised to
detect errors, accidental or otherwise. This is not adequately done
with computerised counting. The source code of the software currently
used for counting votes has not received public scrutiny. Even with
such scrutiny it would be reasonable to be concerned that some bugs may
persist. Public disclosure of data would allow independent
validation of election results using third party software.
- Data security
- Public disclosure of raw data, preferably with secure digital
signatures, enhances security of the data. Because multiple independent
copies will be made, and loss or corruption of the original data can be
detected and corrected. This is particularly important when filling
casual vacancies by "countback" (see below).
- Filling casual vacancies by "countback"
- The better electoral systems fill casual vacancies by examining
additional preferences indicated in the original votes. This makes
access to the original data particularly important. Public disclosure
counters unfairness (exclusive prior knowledge of who will or would be
elected) and data corruption.
- Analysis of voters' concerns
- Some idea of the issues which concern voters can be determined
by just seeing who get elected. More precise analysis can be done using
the details of the preferences each voter gives. Some sketchy data of
this form is often gathered by scrutineers when ballot papers are
examined. Public disclosure allows the concerns of the voters to be
determined more precisely, and (hopefully) acted upon by their elected
representatives.
- Other analyses
- In the longer term, analysis of data from multiple elections may
reveal significant historic/geographic/demographic (et cetera) insights.
Public disclosure will prevent potentially valuable data being lost.
- Improvements to counting rules
- Current rules for counting votes have been developed for counting
votes by hand and are a compromise between easy of counting and
democratic values. With computerised counting, technically superior
rules are feasible and have been implemented.
One obstacle for their introduction is lack of data
from actual elections to demonstrate how they would work and how often
the results may differ from current rules. Public disclosure would
overcome this.
- Education
- The data from actual elections can be useful in civics education.
It can provide better motivation for students compared with artificial
data.
To summarise, full public disclosure of the raw data used for
electronic vote counting in public elections
- safeguards the integrity of and confidence in the electoral process,
and
- provides a source of information which can be used to the benefit
of the public in many ways.
Anything else serves only the interests of those who have access to the
data, and compromises the fairness of the democratic system.