Up: Home page Previous: Issue 23 Next: Issue 25

Voting matters - To advance the understanding of preferential voting systems

Issue 24, October 2007

(Complete issue in PDF, 496Kb.)


The McDougall Trustees recently asked a senior academic is undertake a review of Voting matters, partly due to some comments on the Internet. You may have noticed a difference to the front sheet since the subtitle has been changed from the previous text: for the technical issues of STV. This change reflects the actual content which is not restricted to STV.
An issue which arose from the review was a criticism of the paper by Allard in Issue 5 which gives a low figure for the number of STV elections which are non-monotonic. It is perhaps not obvious that an election can fail to be monotonic in two distinct ways. Given an election for three seats and six candidates in which A, B and C are elected and X, Y and Z fail to be elected, then increased support for C could result in C not being elected; or alternatively, reduced support for X could result in X being elected. It seems clear that a more robust estimate for the occurrence of a non-monotonic election is needed. An academic has agreed to investigate this. It would also be interesting to know if the actual STV counting rule had an impact on this issue.


There are 4 papers in this issue, all of which are comments or reviews of other work:
The Editor must correct a statement made in the last editorial which stated: In electoral terms, Meek has the advantage that the intervention of a no-hope candidate cannot change the choice of the elected candidates - a failing of all the rules used for current hand-counting STV methods. This is not correct in general. For instance, if a Meek count excludes candidates A, B and C (in that order), then rerunning the count without A will get the same result, and also without A and B, or A, B and C. However, a different result may be obtained by excluding B (without A), or by excluding A and C without B, etc. With a conventional count, if the first stage excludes a candidate, then rerunning the count without that candidate does not necessarily obtain the same result.
Readers are no doubt familiar with the problems that were encountered with the Scottish elections this spring. However, the STV elections went off smoothly. The Glasgow election area provided complete details of the voting profiles on the Internet, although it was the only area to do so. This provides a very significant addition to the STV data available for academic study.
The site www.votingmatters.org.uk is now working on the Internet, but bookmarks to the old site should be changed to the new one, since the old site will be removed eventually.

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

  1. David Hill: STV in Northern Ireland. (p1-2, PDF 92Kb)
  2. Jonathan Lundell and I D Hill: Notes on the Droop quota. (p3-6, PDF 200Kb)
  3. H R Droop: On Methods of Electing Representatives. (p7-46, PDF 1.1Mbb)
  4. E Stensholt: Review Elections in split societies. (p47-56, PDF 396Kb)

Main index      Author index