Skip to main contentCambridge University Reporter

No 6299

Wednesday 27 February 2013

Vol cxliii No 22

pp. 379–386

Report of Discussion

Tuesday, 19 February 2013

A Discussion was held in the Senate-House. Pro-Vice-Chancellor Professor John Rallison was presiding, with the Registrary’s Deputy, the Senior Proctor, the Junior Pro-Proctor, and eight other persons present.

The following Report was discussed:

Report of the Council, dated 21 January 2013, on the introduction of electronic voting in ballots of the Regent House (Reporter, 6295, 2012–13, p. 339).

Dr K. B. Pretty (Deputy Vice-Chancellor and Principal of Homerton College):

Deputy Vice-Chancellor, I speak today as the Chair of the Working Group established by the Council during the 2011–12 academical year, to consider whether a system of electronic voting might be introduced for ballots of the Regent House.

Ballots of the Regent House are at the core of the University’s governance arrangements and central to its democratic processes. Ballots are the means by which the Regent House, as the governing body of the University, elects members of the Council and the Board of Scrutiny. They are also the mechanism by which the University determines major strategic decisions (most recently the North West Cambridge Project) and the Regent House exerts its right to question or vary the Council’s decisions. It is vital, therefore, that members of the Regent House should be enabled to exercise their democratic rights as straightforwardly, securely, and transparently as possible. The Council is firmly of the view that an electronic voting system will encourage wider participation and engagement. It also considers that it is a natural and inevitable progression in an institution in which the vast majority of business is transacted securely by electronic means.

The Working Group expended considerable care in reviewing various options for the delivery of electronic voting. The details of these options are set out in the Report. It concluded that an in-house system based on the Basic Online Ballots system offered the necessary standards in terms of security and transparency. Further, such a system can be delivered and maintained at minimal cost. However, it will still be possible for members of the Regent House to receive and return a paper ballot paper if they wish.

The issues of trust, security, and transparency were paramount in the Working Party’s deliberations. We consider that the recommendations which we put to the Council fully meet the Regent House’s expectations in all of these regards. I am pleased that the Council has accepted our Report and is now putting it to the Regent House. I commend the proposals to you.

Mr M. V. Lucas-Smith (Department of Geography):

Deputy Vice-Chancellor, I very much welcome this Report as well as the proposal that this be an in-house system.

Cambridge University Students’ Union (CUSU), for whom I undertake some web development work, has been running online voting now for around five years, with very positive results. I have been involved in running this system and the core voting component which is proposed for the Regent House voting. Any member of the Regent House is very welcome to log in to, and browse around, the existing CUSU system.

The student system is very heavily used for student elections. Some two-thirds of JCRs and MCRs now use it for their elections, as have over half of the Faculties. At the time of writing some 582 ballots have been run, typically involving hundreds of voters per ballot or, in the case of the CUSU and Graduate Union elections, tens of thousands of voters. There has not been a single observed security issue during this period, and the termly feedback that we get back from Returning Officers (ROs) is almost uniformly positive with only small suggestions for changes.

Levels of voting have also increased and graduates in particular have found it easier to vote. Thus this is a well-tested and, I would argue, well-respected system which takes full account of local expertise and knowledge. It is therefore extremely welcome that an in-house system, building on this, is proposed.

I thought it would be useful to give a little background to the system.

The code was originally written by David Eyers at the Computer Laboratory for the Graduate Union elections around seven years ago and was subsequently taken on by CUSU for its own elections. David and I have subsequently collaborated on improving the code so that it is modular and can be deployed either as part of a managed system (as per the CUSU online voting system where JCR, MCR, and other representatives can set up their own ballots directly, which uses a user interface to generate a configuration) or in a standalone system (where people can download the code and configure it manually). The code for the core voting component which is proposed for the Regent House system is open-source and so it can be scrutinized via the link given in the Report1.

The voting workflow can be described as follows.

The Returning Officer sets up a ballot by providing a list of CRSids (i.e. people’s @cam usernames) who can vote. These are stored in a database table. When someone votes, that person’s username has a mark (database entry) put against it indicating that they have voted, thus preventing duplicate votes. Additionally, a voting ‘token’, consisting of a set of four random words (e.g. ‘tour give down this’) is created and this is stored along with the vote that the person has cast in a database table of votes. Additionally, their vote is emailed to the Returning Officer and blind-carbon-copied (so the RO cannot identify them) to the voter by way of a receipt.

Thus there exists a list of votes cast and a list of voters who have voted but no formal link in the database connecting these. At all times, the totals of these are the same, and the number of email receipts also matches. The database consistency is checked on every page hit alongside other sanity checks.

At the end of the ballot, the list of token/vote combinations is revealed, as is the list of usernames of those who have voted, for transparency. This readout, available openly to any authenticated user, can thus be transferred to an STV counting system or spreadsheet. There is also a module for dealing securely with the scenario of electronic voting combined with paper voting.

Anyone wanting more detail on the student system can find this in the ‘election system’ part of the CUSU website.2

The Regent House system proposes to use the same core voting software but with a simplified setup system. The student system includes an additional interface designed to allow JCRs, MCRs, Societies, Faculties, and others to create ballots securely without the need for myself or other administrators to be involved. This clearly is not needed for the Regent House system and the Council’s technical specification outlines what is, in my view, very sensibly proposed.

Over the tens of thousands of votes cast over the last five years, I cannot recall receiving a single email questioning the integrity of the voting process. Thus, as a well-tested, locally-developed system using local knowledge, it is excellent that this should be extended to Regent House voting.

I would commend the Council’s proposals to the Regent House.

Dr S. J. Cowley (University Council, Department of Applied Mathematics and Theoretical Physics, and member of the Working Group on electronic voting):

Deputy Vice-Chancellor, Cambridge is self-governing. Necessary consequences are elections for posts and the occasional ballot on Graces. Legitimacy is enhanced by participation and hence it is essential that it is straightforward to vote in elections and ballots.

In less busy times, all voting was in person in this building. To increase participation, postal voting was introduced and now almost all votes are postal. However, participation still requires access to pigeon-holes or similar, and for many that is not always as straightforward as it used to be (with the necessary travelling and other constraints especially during the Long Vacation/Research Period).

One of the key aims of today’s proposals is to entrench the democratic principles of the University. Those who wish to continue voting by post can do so by opting in. However, those who expect their mail to arrive with a ‘beep’ rather than a ‘rustle’ will have an alternative. That is what the Working Group, and now the Council, recommend.

As a footnote, maybe the Council would now also like to reconsider the type of voting system that it uses. The University was in the forefront of introducing the current version of STV (Single Transferable Vote system) but there are many variants of STV, some of which are more respectable than others. In this light, maybe the Council should revisit the 2005 Report1 and see whether or not it would like to consider using Meek’s method again, because that is what both the London Mathematical Society and the Royal Statistical Society use, and also whether or not when it comes to ballots with a yes/no answer there is a better method than STV.