Up: Home page
Previous: Issue 17
Next: Issue 19
Voting matters - for the technical issues of STV
Issue 18, June 2004
(Complete issue in PDF, 120Kb.)
Editorial
I would strongly recommend to all readers that the Interim Report of the Commission on Electronic
Voting, issued by the Irish Government, is studied closely. This Commission, was formed on the 1st March
and required to report by 1st May, on the suitability of the system chosen for use in elections in
Ireland. They recommended that the chosen system should not be used for the local/European elections
to be held on 11th June. The Commission's Report can be downloaded at:
http://www.cev.ie/htm/report/download_report.htm
To avoid any confusion, I need to declare an interest in this report, since I worked with Joe Wadsworth
of Electoral Reform Services in testing the counting engine of the official software. Our work was not
finished until the end of March, which was only 5 weeks before the Commission reported.
Some aspects of their report are of particular interest here:
- The desirability of removing random selection in the counting process;
- Problems associated with full disclosure of the ballot data (discussed further
in this issue);
- Some shortcomings with regard to secrecy;
- The need for a Voter Verifiable Audit Trial.
A Voter Verifiable Audit Trial might work by having a printer attached to the
electronic voting machine which printed out the filled-in paper after it had
been recorded electronically. The voter would then check this, and place the
paper in a conventional ballot box. Hence the ballot box papers can be used
as a (manual) check against the computer count.
Technically, such a scheme has a number of problems. Firstly, printers are less
reliable than a purely electronic device; should the printer jam, the election
officials might inadvertently see a ballot paper. Secondly, the conventional record
would presumably be used for a recount; however, a manual recount is likely to
be less reliable than the initial electronic count. The process whereby the printed
papers are used needs to be very carefully considered.
There is no doubt that the undertaking of a manual count is one that the public
feels gives confidence in the democratic process. What, therefore, needs to be done
to gain the same confidence in an electronic count? The Irish report gives some
insight into this important issue. Is it necessary to have a Voter Verifiable Audit Trial,
it spite of the problems noted above? Since the Irish Government is still planning
to use electronic voting, we will soon be able to see how these issues are being
addressed.
Returning to Voting matters, there are 6 papers in this issue:
- I. D. Hill: What is meant by `monotonic'? What is meant by `AV'?
- M. Schulze: Free riding.
- I. D. Hill: An odd feature in a real STV election.
- I. D. Hill: Full disclosure of data.
- B. A. Wichmann: A note on the use of preferences.
- J. C. O'Neill: Tie-Breaking with the Single Transferable Vote.
David Hill highlights the problem of the meaning of terms and even abbreviations. As Editor,
I am always concerned about this, since the terminology in common use varies substantially,
especially now that papers are authored from outside the UK/Ireland.
Markus Schulze raises the interesting and important question of the extent to which
strategic voting is used in STV elections. Two forms of strategic voting are analysed,
which in one case, can be identified from US ballot data in which voters can
write-in a candidate. Fortunately for STV, the analysis gives no evidence
of strategic voting in the analysable case.
The next three article are all about the use of preferences. David Hill first provides
an example in which a single paper with a large number of preferences has a crucial
effect. His subsequent papers respond to an earlier Voting matters paper on
full disclosure.
In my own article, I consider the actual use made of the preferences specified by the
voter, and how this information could be altered to avoid any undesirable consequences
of full disclosure.
In the final article, Jeff O'Neill analyses the various ways in which ties are broken
which results in a proposal to change the tie-breaking logic in the current Electoral
Reform Society rules.
Readers are reminded that views expressed in Voting matters by
contributors do not necessarily reflect those of the McDougall Trust
or its trustees.
Papers with citations
- I. D. Hill: What is meant by `monotonic'? What is meant by `AV'? (p1, PDF 12Kb)
- M. Schulze: Free riding. (p2-8, PDF 48Kb)
- I. D. Hill: An odd feature in a real STV election. (p9, PDF 12Kb)
[18, 10, PDF].
- I. D. Hill: Full disclosure of data. (p10, PDF 28Kb)
[19, 14-16, PDF].
- B. A. Wichmann: A note on the use of preferences. (p11-13, PDF 28Kb)
[19, 14-16, PDF].
- J. C. O'Neill: Tie-Breaking with the Single Transferable Vote. (p14-17, PDF 12Kb)
[19, 1-5, PDF].
[22, 1-6, PDF].
Main index
Author index