●
What is the core for research - past, present and future?
E L E C T O R A L
●
What are the boundaries between electoral and political inquiry?
THE CORE AND THE BOUNDARIES
●
What are the benefits and opportunities for cooperation between electoral
administrators and analysts?
●
Where is technology taking the industry and can it be used to further electoral research?
Research Series
South Australian State Electoral Office
June 2000
SA STATE ELECTORAL OFFICE
Editors: Jane Peace and Janet Taylor
R E S E A R C H CONFERENCE PAPERS
Foreword: Elizabeth Ho
Keynote Address: Colin A Hughes
Papers: Nicholas Economou and Brian Costar,
Antony Green, Phillip Green, Andrew Hawkey
and Michelle Davy, Dean Jaensch, Michael
Maley, Alan McRobie, Rod Medew, Gerard
Newman, Jenni Newton, Campbell Sharman,
John Wanna, George Williams
PA P E R S S E R I E S
The papers contained in this volume explore
some of the issues relating to an agenda for
electoral research in Australia from the
perspective of both administrators and
analysts, questions such as:
C O N F E R E N C E
R E S E A R C H
Collaborative ventures, changing technology
and an increasing acknowledgement of the
role played by analysts in their commentary
on the electoral processes has led to renewed
debate on the need for dialogue with the
electoral jurisdictions and investigation of the
basis on which such interaction can occur.
ELECTORAL RESEARCH – THE CORE AND THE BOUNDARIES
Electoral Research
The Core And The Boundaries
South Australian State Electoral Office
Editors: Jane Peace and Janet Taylor
All correspondence
State Electoral Office
134 Fullarton Rd, ROSE PARK SA 5067
GPO Box 646, ADELAIDE SA 5001, AUSTRALIA
Phone: 61 8 8401 4300
Fax: 61 8 8401 4333
Internet: http://www.seo.sa.gov.au
Authors have retained copyright to their material. The opinions expressed are those of the
authors and do not necessarily express those of the South Australian State Electoral Office.
ISSN 1443-3621
Electoral Research – The Core and the Boundaries
Conference Papers
Editors
Jane Peace and Janet Taylor
Research Series
South Australian State Electoral Office
June 2000
ACKNOWLEDGEMENTS
This collection of papers was first presented at a conference of electoral administrators and analysts
in Adelaide on Friday 3, Saturday 4 December 1999. The conference, attended by 61 delegates, was
called ‘Electoral Research – the Core and the Boundaries’. It was organised and hosted by the South
Australian State Electoral Office under the auspices of the Electoral Commissioner Steve Tully and
opened by the Hon. Trevor Griffin, Attorney-General for South Australia.
It was the first open forum in Australia to bring together electoral analysts from the electoral
administrations, academic life, the media and parliamentary libraries in Australia and New Zealand.
The keynote address was given by Colin Hughes, Emeritus Professor, the University of Queensland
and a conference commentary was provided by Elizabeth Ho, Director, The Hawke Centre at the
University of South Australia. Her summary is presented in these proceedings as a Foreword to the
volume.
The New Zealand elections and two referendums precluded attendance of administrators from
there though an independent electoral consultant and a parliamentary analyst were able to be
present. Invitations were extended to the media and academics.
The South Australian Electoral Office thanks all those who participated in the conference - those
who gave up their time to present papers and those delegates who contributed to discussion from
the floor. Professor Dean Jaensch, Flinders University, and Ms Jenni Newton, South Australian
Parliamentary Library, provided valuable support to conference organisation. The office expresses
its appreciation to the Electoral Council of Australia and its Executive Secretary, Janet Taylor, who
provided administrative support for the conference and co-edited the papers for this record of
proceedings. Thanks also to those who offered constructive suggestions for the production of this
volume and to Jade Koopman for her assistance in the final copy.
CONTRIBUTORS
Brian Costar is the Head of the School of Public Policy at Melbourne’s Monash University.
Professor Costar’s areas of academic interest centre on political parties, elections and electoral
systems with a particular interest in the National Party.
●
Michelle Davy joined the Tasmanian Electoral Office in 1995, to assist in the development of new
election material and procedures for the 1996 State election. She has represented the office in a
broad range of electoral areas and issues, including the development of legislation and policy,
dealing with offences and complaints, and the preparation of information booklets and papers.
Michelle came to the electoral world from private legal practice, after completing her BA/LLB in
1989. She returned to the University of Tasmania to complete a Masters degree in Social Science
(Social Research) in 1994.
●
Nicholas Economou teaches Australian Politics in the School of Political and Social Inquiry at
Monash University, Melbourne. Dr Economou has published on Australian and Victorian politics
and elections, and has been a media commentator on, amongst other things, electoral matters for a
variety of media outlets including radio 3AW in Melbourne, the ABC and the BBC. He is co-editor
with Brian Costar of the book The Kennett Revolution: Victorian Politics in the 1990s (UNSW
Press).
●
Antony Green was once employed as a data analyst in the computer industry and has blended this
knowledge with a deep interest in politics to become one of the country’s most respected political
commentators. He appears regularly on ABC election night programs, has been involved in the
coverage of every State, Territory and Federal election since 1989 and also designed the software
used by the ABC on election night. Since becoming the ABC’s election analyst in 1989, Antony has
travelled widely observing elections and television coverage of election counts, including the New
Zealand elections and the poll in Britain that saw Labour’s Tony Blair swept into office.
Antony is a regular commentator on politics for the Sydney Morning Herald and authored the 1996
ABC Election Guide. He has also produced a number of academic publications for the Queensland
Parliament and regular publications on politics for the NSW Parliamentary Library. His academic
background includes a Bachelor of Science (Sydney University) with majors in mathematics and
computing and a Bachelor of Economics (Sydney University) with Honours in Politics.
●
Phillip Green has been the ACT Electoral Commissioner since the establishment of the permanent
ACT Electoral Commission in 1994. Prior to moving to the ACT, Phil worked with the Australian
Electoral Commission, which he joined in 1982. During his 18 years in the electoral industry, Phil
has worked on just about every aspect of the election game. He has written numbers on tally boards
and managed the National Tally Room. He has issued votes in the Northern Territory to soldiers on
manoeuvres and managed election scrutinies in Canberra, Hobart, Launceston, Flinders Island and
Namibia.
Phil is probably the only person in Australia alive today who can claim to have been solely
responsible for writing a major Electoral Act from scratch, and the only person who ever really
understood how the ACT’s ill-famed modified d’Hondt electoral system worked. In 1999, Phil was
the Chairperson of the Electoral Council of Australia, the peak body of Australia’s electoral adminis­
trations. Phil is also lead writer for International IDEA’s forthcoming module of the Administration
and Cost of Election Project on Elections and Technology. Phil has a Bachelor of Arts (Professional
Writing) from the (then) Canberra College of Advanced Education, and a Bachelor of Letters
(Political Science) from the Australian National University.
●
Andrew Hawkey expressed interest in the Australian political system at a very young age in the
post ‘Dismissal’ period. Having a natural inclination towards numbers and in particular statistics,
this interest quickly focussed on elections. In 1989, as a year 12 student, Andrew attempted to
write his own version of ‘EC Count’ in Pascal for his College SRC elections. He completed his
honours degree in Mathematics and Political Science with a thesis entitled ‘Statistical Modelling of
Electoral Systems and the Treatment of the Third Party’. Since joining the Tasmanian Electoral
Office in 1993, Andrew has managed various State Parliamentary elections, designed and managed
the production of both the Hare-Clark Automation System (H-CV Auto) used for ACT and
Tasmanian elections and the Tasmanian Electoral Office 1998 election website. He has also
produced various types of electoral information material and reports.
●
Elizabeth Ho is the Director of The Hawke Centre, a major public education project of the
University of South Australia, devoted to fostering a deeper understanding of the Australian
democratic system and notions of political, environmental and social citizenship through its
programs. Ms Ho is a Councillor of the Institute of Public Administration Australia (SA Division),
an Associate of the Library and Information Association of Australia and a member of the State
Heritage Authority. She previously held senior posts in the State Library of South Australia, and the
South Australian Education Department and holds a BA (Honours) Degree majoring in History
from the University of Adelaide.
●
Colin Hughes, Emeritus Professor, University Of Queensland, came to Australia already a veteran
electoral reformer with experience of campaigning unsuccessfully against the property vote and for
votes for women. He starting writing about Australian elections with an article on the 1957 state
election in Queensland and a chapter on the 1958 federal election.
Since then he has been an inveterate tinkerer, witness before Royal Commissions and parliamentary
committees and member of the EARC that reformed the Queensland system and redrew the most
discussed set of boundaries this century. He is one of two outsiders ever to sit on a Commonwealth
redistribution commission, and for five and a half years was the Commonwealth’s Electoral
Commissioner 1984-89. He currently is a member of Queensland’s Local Government electoral
body and in 1999 finished his term as a member of the Board of Directors of International IDEA.
Dean Jaensch is Professor of Politics at Flinders University. His major field of teaching and
research is Australian politics. He has been analysing elections since 1968 and has written widely
about them, including a series of statistical analyses. He is the co-author of 21 books and numerous
articles and chapters on Australian politics and maintains a major research interest in the politics of
elections.
He is a life member of the Australasian Political Studies Association and a past President and is
widely involved with public education in the field of politics. His publications include text books
and research monographs and he has carried out a number of consultancies concerning elections
and election systems in the federal, state and local areas. He has been a regular political and election
commentator for the media since 1974.
●
Alan McRobie is a political analyst and independent electoral consultant with Masters in both
history and political science. Between 1969 and 1990 he taught courses in New Zealand history,
politics and society at the Christchurch College of Education and from 1990 to 1993 was part-time
lecturer in Political Science at the University of Canterbury. He has been Visiting Teaching Fellow
at the University of Canterbury, Visiting Professor of Political Science at the California State
University. In 1987 he was awarded a Claude McCarthy Fellowship to undertake research into New
Zealand’s electoral system. During 1999 he was a member of an ad hoc commission established by
the Christchurch City Council to review the city’s community areas and ward boundaries and was a
guest presenter at a NZ Society of Local Government Managers seminar considering a redrafting of
the Local Elections and Polls Act.
Alan has written extensively on New Zealand’s politics and history in professional journals both in
New Zealand and overseas and has acted as commentator on elections and electoral politics for the
electronic and print media. Published work includes:
New Zealand Electoral Atlas (1989);
Taking it to the People? The New Zealand Electoral Referendum Debate 1993;
Co-authorship with Keith Jackson of A Historical Dictionary of New Zealand (1996) and New
Zealand Adopts Proportional Representation: Accident? Design? Evolution? (1998). ●
Michael Maley is Director, Research and International Services, Australian Electoral Commission.
He joined the Australian Electoral Office in 1982 after completing a Master’s Degree in Political
Science at the ANU with a thesis on probabilistic modelling of electoral bias in single-member
constituency systems. He has served as an electoral consultant to the United Nations, the
Commonwealth Secretariat, the International Foundation for Election Systems and the International
Institute for Democracy and Electoral Assistance. Michael has occupied senior positions in a
number of United Nations peacekeeping operations which involved elections; and has published
and lectured widely on the subject of elections and peacekeeping. He is the author of a number of
chapters, including that on ‘Administration of elections’, in the forthcoming International
Encyclopedia of Elections.
Rod Medew is Director, IT Applications, Australian Electoral Commission. He holds an Honours
Degree in Geography from Monash University. He joined the Australian Electoral Commission in
1985 and has since been instrumental in implementing many IT innovations in the AEC, including
the use of Geographic Information Systems as part of the redistribution process, the use of matched
polling place data to calculate swings on election night and the development of integrated election
management systems. On secondment to the United Nations, he was the project manager for the
development of the computer system used at the 1993 election in Cambodia, which required the
development of a special keyboard and coding scheme to cope with the Khmer alphabet. He has
also served as an electoral adviser in South Africa, Indonesia, and Eastern Slavonia.
●
Gerard Newman is the Director of the Statistics Group, Information and Research Services,
Department of the Parliamentary Library, Canberra. He has a Bachelor of Economics from
Queensland University and is currently employed as the manager of the group of professionals
providing an information and research service in statistics to members of the Commonwealth
Parliament. Gerard has been employed in the Parliamentary Library for 25 years. His area of subject
speciality is in election results and has published numerous papers analysing Commonwealth and
State elections. He has also published works on electoral redistributions and electoral systems.
●
Jenni Newton has been Research Director at the Parliamentary Library in South Australia since
1985. She was a National Undergraduate scholar at ANU, graduating with an Honors Arts degree
and has since studied demography at ANU and accounting at the University of South Australia.
After graduating she worked in a research capacity with the Commonwealth Departments of Social
Security and, later, Finance. At the library, she is an electoral specialist and has written background
papers and has estimated the effects of State boundary redistributions in 1983, 1991, 1994 and
1998. Her paper Making Sure that Electoral Boundaries are Fair in South Australia was used by all
major parties at the 1998 Boundaries Commission hearings.
●
Campbell Sharman is Associate Professor in the Political Science Department at the University of
Western Australia. He holds degrees from the University of Adelaide, the London School of
Economics and Political Science, and Queen’s University in Kingston, Ontario. He has held
teaching and research positions in several universities including the Australian National University
and the University of British Columbia.
His teaching and research interests focus on Australian government and politics with special
reference to federalism and state politics, the operation of constitutional and electoral rules, the
parliamentary process, and the role of parties. He has published widely on these topics in Australia
and overseas. In association with his colleague Jeremy Moon, he is a principal investigator in a long
term project on Australian politics and government in the Australian states which has had financial
support from the Australian Research Council and, more recently, the National Council for the
Centenary of Federation. From November 1994 for the two years of its existence, he served as a
commissioner on the Western Australian Commission on Government.
John Wanna is Associate Professor and Head of School of Politics and Public Policy, Griffith
University, Brisbane, Queensland and a principal researcher with the nationally-funded Centre for
Australian Public Sector Management (CAPSM) and the Key Centre in Ethics, Law, Justice and
Governance at Griffith University. Dr Wanna has produced around 10 books including two
national text books on policy and public management. He is currently working on a book on
Managing Public Expenditure at the Commonwealth level. His research interests include Australian
and comparative politics, public expenditure and budgeting, and government-business relations.
He also writes on state politics and has been a regular state political commentator for the ABC, The
Courier-Mail, the Australian, other media outlets and commercial TV.
●
George Williams is a Senior Lecturer and Fellow in Law at the Australian National University.
Over the first half of 1999 he was a Visiting Scholar at the Human Rights Institute at Columbia
University Law School in New York. He teaches and writes on topics involving constitutional law,
human rights and electoral law. He is the author of books including:
● A Bill of Rights for Australia (UNSW Press, 1999);
● Human Rights Under the Australian Constitution (Oxford, 1999);
● Australian Constitutional Law and Theory: Commentary and Materials (Federation Press, 2nd ed
1998, with Tony Blackshield).
George also practises as an barrister and has appeared in High Court cases, such as the Hindmarsh
Island Bridge Case and Lange v ABC, raising issues such as freedom of communication as it affects
electoral law, freedom from racial discrimination and the separation of powers.
CONTENTS
Introduction
1
x
Foreword
9
Elizabeth Ho
Keynote article
Electoral Research And Administration – A Brief History
15
The research core – past, present and future
21
1
22
Campbell Sharman
28
36
49
George Williams
Alan McRobie
Dean Jaensch
2
3
4
The Use Of Electoral Data For Research Into The Australian
Political System
A Legal Perspective On Electoral Research
What’s So Special About Specials?
The Key To Progress: A Cooperative Research Centre
Colin A Hughes
The borders between electoral and political inquiry
53
5
54
John Wanna
62
Nicholas M Economou
and Brian J Costar
Michael Maley
6
7
Negotiating The Boundaries Between Electoral And Political
Inquiry
Electoral Inquiry Or Political Debate? Analysis, Commentary And
The Controversy Of Victoria’s Legislative Council Electoral Process
Electoral Research: Spread Over Many Fields
68
The benefits and difficulties of collaboration between
electoral administrators and analysts
73
8
9
74
78
Fostering Cooperation Between Electoral Authorities And Analysts
Developments In The Provision Of Information
Gerard Newman
Andrew Hawkey
and Michelle Davy
Jenni Newton
10 Living In Interesting Times
82
Technology – implications for the industry
87
11 Geographic Information Systems In Political And Electoral Work
12 Why Data Definition Is More Important Than Technology For
Electoral Research
13 Elections And Technology – Implications For The Future
88
Rod Medew
93
97
Antony Green
Phillip Green
TABLES
page
Table 3.1
New Zealand Elections 1972-1996: Fate of special votes cast
46
Table 3.2
New Zealand Elections 1972-1996: Impact of special votes
47
Table 3.3
Disallowed special votes 1972-1996
48
Table 3.4a Fate of special votes disallowed 1996: Reasons for disallowance
48
Table 3.4b Fate of special votes disallowed 1996: Analysis of disallowed votes not enrolled (NOR)
48
Table 6.1
The 1999 Victorian Election Results – Legislative Council
67
Table 6.2a Legislative Council provinces: marginality following the 1996 election
67
Table 6.2b Legislative Council provinces: marginality after the 1999 election
67
INTRODUCTION
The conference at which the papers in this volume were given evolved from a proposal to bring
together a small number of researchers in the electoral industry to discuss the essentials for electoral
research, past and present, the borders between electoral and ‘political’ inquiry, the practicalities of
collaboration between electoral administrators and analysts and the implications of technology for
the industry.
Colin Hughes, keynote speaker, described such a gathering as:
…not completely new. Almost a quarter century ago what was called the ‘First Electoral
Seminar’ was held at the Gold Coast - in June 1976. The first word on the cover of the
transcript of the Seminar’s proceedings is ‘CONFIDENTIAL’, and the first names on the list
of delegates are those of the Under Secretary and the Assistant Under Secretary of
Queensland’s Department of Justice, who were indeed the hosts.
Administrators gave the following papers, addressing topics displaying the problem
addressed in the 1999 conference: ‘Returning Officer’s Appointment, Responsibilities and
Remuneration’ (NSW); ‘Polling Officials’ Fees and Taxation (Cwlth); ‘Postal Charges and
Surcharges’ (WA); ‘Hours of Poll and Certificate Voting’ (Tas); ‘Voting by Persons in Prison’
(Vic); ‘Universal Postal Vote Applications’ (Vic); ‘Availability of Postal Vote Applications’
(NSW); ‘Non-Voter Notices and Erasures by Computer’ (Qld) and ‘Prerogative for
Enforcement of Penalties’ (SA). The Northern Territory and the ACT did not then have
their own electoral authorities.
The two academics present also presented papers: Paul Finn, now Mr Justice Finn gave ‘Electoral
Corruption and Malpractice’, Hughes himself ‘Publication of Election Results’.
Hughes described the papers as portraying in a nutshell the gap which the Adelaide conference had
to try to bridge. ‘On one side of the gap are high particularity and utility; on the other side, bigger
pictures and wider controversies. On the one side immediate responsibilities, on the other side
remoteness from influence’. He noted that it was an all-male gathering and observed that electoral
administration has progressed in both openness and independence since then.1
Research Within The Jurisdictions
A brief survey of Commonwealth, State and Territory electoral administrations prior to the 1999
conference revealed that a statutory provision for carrying out research was first introduced to
Commonwealth legislation in 1984. Since then the Australian Electoral Commission has expanded
the services it offers within the Asian-Pacific region and its research activity has assumed an interna­
tional perspective. By 1999, six Australian electoral jurisdictions had a research obligation. Four
administrations specifically identified research officer positions, seven had contracted out research,
particularly market research. All considered providing database information and most did. All
organisations carried out literature searches with an increasing reliance on the Internet.
Electoral research within the jurisdictions (as distinct from research undertaken by the independent
redistribution commissions) has included: investigation of legislative development; surveys of
electoral knowledge, attitudes and service standards; compilation of database information on
electoral events, candidates, members elected, districts; examination of changes to processes and
procedures; the extension of access to democratic processes and decision making; informality;
compulsory voting; citizen initiated referenda; compliance with how-to-vote cards and the impact
on results of position on the ballot-paper.
1 Campbell Sharman pointed out during proceedings that the Canadians appear to have held such meetings between
officials and researchers in 1970 and 1974. See the web page for the Centre for Election Studies at the University of
Waterloo - http://arts.uwaterloo.ca/PSCI/ces.1htm
1
Introduction
Research output has been used to document and analyse electoral data over time, improve
processes and services and locate areas where education and public information material is needed
to ensure comprehension of voting provisions and procedures.
The result is a fine-tuned sense of the past from the early days of representative democracy.
Changes to boundaries, demography, land-use, modes of communication have altered the electoral
map. In tandem with the physical changes have come developments in electoral thinking, acknowl­
edgment of contemporary issues and the integration of research with mainstream electoral
activities.
Environmental scanning is intrinsic to research. The jurisdictions monitor such matters as evidence
of public awareness of electoral issues, manifestations of a wish for more participative decisionmaking and innovations which may have practical application for electoral operations. Research
within the administrations has not been as ready to explore in depth the political impact and
outcomes of changes to processes and procedures.
The Hon Trevor Griffin in his address at the opening of the conference, reminded those present
that research can be more than critical study, examination and analysis. Conventional practices and
processes can be questioned and the boundaries imposed by traditional and formal structures and
thinking can be explored – the legitimised areas of interest. There are no boundaries to knowledge.
In a world of increasing globalisation and ease of communication, organisations such as the
International Institute for Democracy and Electoral Assistance (IDEA) and the International
Foundation for Election Systems (IFES), and publications such as Electoral Studies provide ready
access to ideas. Likewise contacts outside Australia with similar concerns, with whom issues can be
discussed and explored.
Electoral administrators find themselves increasingly dependent on computerised processes for
operational procedures, facilitating election conduct and disseminating information. This
dependence brings new risks and an impetus to accelerate the pace of procedural change ­
withdrawal from this path is difficult. Technology has brought enormous benefits: a reduction in
efforts expended on proofing and checking materials; improved ways to educate and inform the
public; spin-offs of technological applications in the cross-over from parliamentary to local and
non-government election conduct; better tools to capture and retrieve data, a considerable amount
of which is held by electoral administrations. The allocation of resources to analysis of the data is
often subject to the operational and financial constraints of the electoral cycle.
Software is being used for the management of equipment logistics, processing pre-poll votes,
capturing and transmitting results, as well as mapping electoral information. However, electoral
administrations are increasingly being questioned for still using the pencil and ballot-paper to
record votes of electors – little has changed in that respect for more than a century. The technology
exists for removing the inefficiencies of manual handling of ballot materials and providing
increasingly sophisticated and powerful research tools. The difficulty for the administrations,
separately, is to harness that technology in a cost effective manner. Mechanisms for reducing costs
may be facilitated by increasing interaction within the jurisdictions.
2
Introduction
A major concern before techniques such as telephone or Internet voting can become common-place
is of course ensuring the integrity and confidentiality of the vote.2 The people need to be assured, in
this rapidly changing world, that their voices will be heard, that they can freely and confidently
exercise their democratic rights and their votes will be recorded and counted without violation.
Research into the most effective means for vote protection will run in parallel with systems
development.
The Legal Framework
The Hon Trevor Griffin made reference to the New Zealand and Australian culture of fairness and
compliance with the law. For over 150 years, electoral law in Australia has largely mirrored
procedural and social developments. Reform has included, to some extent, the removal of inequities
and gerrymandering, for example the gradual broadening of the franchise, the development of pre­
poll voting mechanisms to cater for remote and increasingly mobile populations, the adoption of
preferential voting and counting systems to better reflect the voting intentions of the electorate, the
placing of redistribution mechanisms out of the hands of the parliaments.
The law provides an administrative tool and acceptance, security and support for processes and
procedures – often a codification of electoral practice. Sometimes the law has not kept abreast of
change. It took nearly a hundred years for ballot-papers to begin to record candidate party or group
affiliations. In some instances, legislative amendment has preceded change: computerised systems
for vote-counting were only permitted after enabling legislation had been passed. Social acceptance
of the new methodology had to be confirmed.
The challenge for the law is to provide a supportive framework without undue prescription, to be
couched in straightforward language, remove duplication, to clarify the intent of the law-makers
and accommodate the expectations of the electorate. George Williams’ paper dwells on the
difficulties of research into electoral legislation, case law and the structures underpinning the law.
An Administrator’s Viewpoint
Steve Tully, Electoral Commissioner for South Australia, at the commencement of the conference,
provided an administrator’s insight into the difficulties of research activity within the jurisdictions.
He observed that when an organisation is prioritising its services and allocating resources, research
can be seen to be an area from which funding can be diverted. Many electoral positions have a
research component, often project based, which are targeted at operational issues and thus add
value to processes.
Research focused on historical issues adds to an understanding of the whole context in which
electoral administration is carried out. This has to be valued as much as ‘applied’ research.
Contemporary issues also demand investigation: the incidence of informality, ticket voting, the
impact of different systems on voting outcomes and increasingly how to make full use of the
information accumulating on the databases within jurisdictions, particularly with the advent of
computerised counts.
2 Following the conference, publicity was given (March 2000) to the first use of on-line voting in a public election in
Arizona during primaries in the United States of America. Full assessment and more on-line elections before the
November 2000 general election were promised. It was reported that turnout increased by >600% with some
35 000 voters (>50%) ‘clicking’ rather than ‘ticking’ their candidates. Technical glitches were reported to include a
slowed response from the web-site due to traffic volume. Financial Times 13 March 2000.
3
Introduction
Electoral administrations are not only concerned with parliamentary election conduct. An
important part of their operations is advising on and conducting industrial ballots and local
government elections. Another major focus is education – what is needed to encourage partici­
pation in the democratic processes and how to go about it. Resources need careful husbanding to
accommodate a response to all business sectors.
Electoral offices are looking to providers from private and tertiary education organisations to
supplement their own resources. Business imperatives make the identification and selection of
reliable external service providers critical in meeting stated timeframes and quality and cost
parameters.
Tully emphasised that information has to be transferable. Partnerships brokered between analysts
and administrators need to consider customising programs for mutually desirable outputs and
determine the timeframes for service delivery. From practical experience in building data bases,
electoral administrations would need to have early advice of requirements to develop flexible
databases.
Until now administrators have provided what they thought was required and born the costs - for
example importing election statistics from software to more user friendly spreadsheet files for
manipulation by analysts. If more is required the difficult question of ‘who pays?’ needs to be
addressed.
Likewise the question of confidence in the electoral system. Electoral administrations, which serve
the public, are in a ‘catch 22’ situation. They must remain at arms length from government
direction, refrain from political comment while always cognisant that public interest more generally
lies in the outcomes of elections rather than its processes. Agents acting for the office must also
reflect a culture of independence.
The Commissioner saw it as inappropriate for him to comment at any time on a policy matter - for
example voluntary voting, the system of voting, or to publicly surmise an electoral outcome before
a poll is declared. This would threaten the independence and integrity of the electoral adminis­
tration. Rather he saw his role to provide information that can be interpreted, for example by the
parliament, in terms of electoral policy and to satisfy the expectation for more information, not less.
Analysts from academia and the media can provide valuable critical interpretation and public
commentary on results and system intricacies which add to public understanding of the electoral
process, as well as acknowledging the transparency of electoral processes. Parties, analysts and the
media commentators have a freedom to predict likely poll outcomes. They can express opinions on
any policy matter and there will always be a place, Tully believed, for constructive critique of
electoral administration as long as it values the confidence required in the electoral systems.
Other benefits of collaboration with analysts include cross fertilisation of ideas, access to scarce
skills and resources, knowledge pooling and sharing, a facility to analyse events and trends that are
not always a priority when core business is election conduct. The perception and construction
placed on electoral matters can be broadened by their input.
The Commissioner acknowledged that his office had benefited from positive working relationships
established with parliamentary libraries, universities and the media. Processes had been made
transparent and access to staff and resources offered wherever possible. The courtesy had reaped
rewards for the office. Information had been provided on unfolding events which contributed to
4
Introduction
accurate commentary and processes; data had been exchanged for analytical exercises for the office
and the South Australian Electoral Districts Boundaries Commission; useful input had been made
to corporate thinking; there had been understanding of the demands an electoral office is subject to
in the midst of an election.
Tully noted that in general, political trends and data are not explored by the electoral jurisdictions
though the published statistical returns for most administrations over the last ten to fifteen years
showed a marked increase in the range of data with political inferences. Most included such
information as the candidates’ affiliation, swing to lose and safety margin of a seat, election
outcomes in terms of the composition of the government, two party preferred information and,
quite often, analyses of ballot papers as well as a description of legislative changes since the last
elections.
The South Australian Electoral Districts Boundaries Commission does examine two party preferred
voting figures to interpret the voting patterns at the census collector district (CCD) level. This is to
meet the fairness requirement that the group(s) gaining more than 50% of the State-wide vote has a
majority of elected candidates and is able to form government. The 1998 Boundaries Commission
used data from scanned rolls and two party preferred information from the results system
reconfigured with a mapping tool to more readily plot new electorate boundaries.
The Boundaries Commission has also invited comment on redistribution matters from analysts
outside the electoral administration such as the research director from the South Australian
Parliamentary Library, Ms Jenni Newton, and Mr Antony Green from the ABC.
The Commissioner concluded by pointing out that the conference itself was evidence of the
emphasis being placed on increasing liaison between analysts and administrators and freeing up
some aspects of the separation between the public sector, academia and the media. There are
benefits in any shared approach to raising electors’ awareness of electoral issues and understanding
of the systems used.
Conference Outcomes
Conference rapporteur Elizabeth Ho stated that the purpose of the conference was ’wisely
ambitious for such a gathering…there was neither a dearth of good ideas nor any failure to
communicate them well’.
She noted that an additional gift to the conference was the humour of participants:
Conference humour is often ephemeral and generally lofty papers remain as keepsakes and
the intelligent banter is forgotten. Examples of the very lively wit and ease of the gathering
can be seen in the following two (rough) quotes from:
Malcolm Mackerras, veteran electoral analyst:
‘I wear it as a badge of honour that I have been accused of corrupting the electoral process.’
Campbell Sharman, West Australian academic:
‘Australians have an obsession with fairness and it should stop.’
In the light of weighty technological change, it can be pleasantly distracting to hear the
various views on the indomitable political canvasser. Emeritus Professor Colin Hughes
comments on the role of the scrutineer when he talks about Cicero in the Roman scenario –
5
Introduction
‘the freedom of the ordinary person to vote as he wishes shall not be interfered with’. The
ACT administration has recently insisted on a wider berth at polling booths with the
inevitable loss of dramatic effect. However, with the impending spectre of bio-tagged
voters, the truly burning question is: where will the canvasser dwell in the world of on-line
voting?
However, she observed that each of the ephemeral moments connected to more serious elements
with repeating themes providing a framework for debate and comment on the specifics of the
electoral environment. These included a culture of fairness; research as a check and balance in a
democratic society; and, the power of public education.
After the conference, suggestions arising from the closing workshop session were referred to the
Electoral Council of Australia. A number of issues will be addressed by the jurisdictions and
analysts consulted to effect optimum resolutions to some of the problems raised.
The Structure Of This Volume
Ho’s summation at the end of the conference provided a foreword to this volume and the papers are
presented in the order in which they were read at the conference. Referencing is as provided by the
authors.
Jane Peace
South Australian State Electoral Office
Janet Taylor
Electoral Council of Australia
June 2000
6
CONFERENCE PARTICIPANTS
John Alderman
South Australian State Electoral Office
Clive Anson
South Australian State Electoral Office
Andy Becker
Australian Electoral Commission
Doug Beecroft
Victorian Electoral Commission
Brian Beggs
Australian Electoral Commission
Scott Bennett
Department of the Parliamentary Library, Canberra
Judy Birkenhead
Averment Pty Ltd
David Black
Curtin University of Technology
Fiona Colbeck
Western Australian Electoral Commission
Margie Cook
Channel 9
Brian Costar
Monash University
Deane Crabb
The Electoral Reform Society of South Australia
Chris Craven
New South Wales State Electoral Office
Greg Davis
Northern Territory Electoral Office
Nick Economou
Monash University
Tim Evans
Australian Electoral Commission
Rachel Gibson
Australian National University
Tim Glanville
Australian Electoral Commission
Andrew Grant
South Australian State Electoral Office
Antony Green
ABC
Philip Green
ACT Electoral Commission
David Gully
South Australian State Electoral Office
Geoff Halsey
Australian Electoral Commission
Barrie Hamilton
Northern Territory Electoral Office
Andrew Hawkey
Tasmanian Electoral Office
Elizabeth Ho
University of South Australia
Colin Hughes
University of Queensland
Dean Jaensch
Flinders University
Nicole Lugg
Tasmanian Electoral Office
Tan Luu
South Australian State Electoral Office
Gary Lynch
Electoral Commission Queensland
Malcolm Mackerras
Australian Defence Force Academy
7
Conference participants
Michael Maley
Australian Electoral Commission
Haydon Manning
Flinders University
Louis Massicotte
University of Queensland
Dezma Maxwell
Australian Electoral Commission
Ian McAllister
Australian National University
Greg McArthy
University of Adelaide
Clem McIntyre
University of Adelaide
Alan McRobie
Independent electoral analyst
Kay Mousley
Australian Electoral Commission
Gerard Newman
Department of the Parliamentary Library, Canberra
Jenni Newton
Parliamentary Library, Adelaide
Angela O’Neil
Australian Electoral Commission
Graeme Orr
Griffith University
Des O’Shea
Electoral Commission Queensland
Jane Peace
South Australian State Electoral Office
Harry Phillips
Edith Cowan University
Alison Purvis
ACT Electoral Commission
David Quinn
GTV 9
Phil Roberts
Australian Electoral Commission
Campbell Sharman
University of Western Australia
Barbara Stedman
New Zealand Parliamentary Library
Janet Taylor
Electoral Council of Australia
Paul Thornton-Smith
Victorian Electoral Commission
Steve Tully
South Australian State Electoral Office
Di Walker
South Australian State Electoral Office
John Wanna
Griffith University
Lawrie Waters
South Australian State Electoral Office
George Williams
Australian National University
Garry Wiltshire
Electoral Commission Queensland
Michelle Davy was unable to attend the conference and co-author Andrew Hawkey presented their
paper.
Rod Medew was in Papua New Guinea for the Australian Electoral Commission and his paper was
given by Michael Maley.
8
FOREWORD
Overall, the articles in this volume lead to the assessment that the culture of fairness and
compliance with the law is an underpinning theme in Australian electoral systems and the power of
research is to inform that agenda. As a mature democratic nation the focus is on improvement,
protection, and raising community awareness of the nature and importance of the process. It has an
environment where rights and principles of fairness are embedded and if systems are not
necessarily understood by the person in the street, their expectation that those principles will apply
in their electoral system are generally high (see Jaensch).
Australia has a long history of representative government; it has a high degree of literacy; the
compulsory system works against the potential excesses of populism and minority view dominance.
As the democracy is relatively mature, so a maturity of interest is demonstrated among these
specialists in the role, purpose and future of electoral research in support of democratic intent and
practice. Many insights into research, and its power to provide a check and balance, can be gleaned
from the papers that follow.
As electoral consultant Alan McRobie puts it, the electoral researcher will often have as their
objective, to make the electoral system as fair and accessible as possible. Just as there are safeguards
within Australian society to both limit and distribute power, this objective, supported by good
research, can work to dilute the effects of political self interest displayed by the party political fold.
It can also complement the efforts of those administrators who ensure that elections are run
properly and within the appropriate legal parameters.
In this way, the gift of check and balance from the research effort can be achieved. This must be
regarded as a powerful argument for collaboration.
In making the point about the valuable influence of research findings on electoral practice
decisions, a focus is offered on the importance of data preservation (Sharman) and availability
(A. Green) - an affirmation of the importance of information collation, organisation and access in
the maintenance of a democratic society. That libraries and archives struggle to achieve a basic level
of description and resourcing to support research efforts is well highlighted. The cost of access is an
issue.
Turning to public education, the contrast between the deep collective wisdom of electoral analysts
and administrators and the perceived state of public political and civic education is striking. There
are reminders of J. Bronowski in The Ascent of Man and his firm belief in the ‘democracy of the
intellect’ - that unless people with no ambition to control others have knowledge and actively
participate in their society, then we make empty claims to having achieved a civilisation.1
Moving into some of the specifics that are raised in relation to research, the articles range across the
role and purpose of research, and the roles and purposes of those who undertake it.
Respective Responsibilities
Looking first at spheres of responsibility, Emeritus Professor Colin Hughes gives us a compre­
hensive overview of the increasing convergence and strengthening relations between administrators
and academics. This however does not overshadow the message that there is a legitimately
independent role for each, with administrators particularly sensitive to the political potential of
their investigations.
1 J Bronowski, The Ascent of Man, BBC, 1976 pp 435-6.
9
Foreword
In reflection of the power of the region in Australian politics, and the chequered history of electoral
distribution, this 1964 media quote, taken from Gough Whitlam’s The Whitlam Government
1972-52 seems worth repeating:
Daily News 1964 – Murwillumbah
The Minister for the Interior, Mr J D Anthony told a Country Party dinner at
Murwillumbah that in carrying out the proposed redistribution of the Federal seats he
would make sure that fewer people were required to elect representatives in the country
areas than the cities. He said ‘Australia’s electoral system was the finest in the world but
lacked territorial representation. Representation was now based on population not area and
with the result that the power in Parliament was confined to the heavily populated areas.’
Mr Anthony said ‘the task of electoral reform was probably one of his worst jobs as Minister
for the Interior and one on which his political future hinged.’
Australian Electoral Commission Research and International Services Director, Michael Maley
stresses the need for administrators to focus strongly on the scrutiny process, arguing the case for a
strong applied focus in research supporting practical policy programs; a clear delineation of the
rationale for research; an emphasis on an evidence–based approach; and, an assessment of
assumptions at the outset. Maley also refers to the value of undertaking analyses in a national and
international context across a diversity of administrations rather than taking up a single adminis­
tration issue, where the sensitivity may be heightened.
Professor Brian Costar, of Monash University, focused in discussion on the fact that academics,
while perceived to have freedoms not open to administrators, nevertheless are not immune to
political manipulation of their expressions, out of context, with consequences, once they enter the
public affairs arena.
In the context of collaboration, political scientists can play a role in democratic advocacy,
supporting administration concerns in the realm of public comment.
Nick Economou, also from Monash, and Brian Costar use the case study example of the Victorian
Legislative Council to highlight how an area ripe for scrutiny, based on the hard and indisputable
facts of longitudinal analysis, will nevertheless be perceived to have a politically partisan origin if
research is commissioned.
All of these perspectives are useful in focusing attention on the respective roles of administrator and
political scientist.
The Need For Research
Turning secondly to the need for research, the New Zealand examples of Maori and special votes
demonstrate the importance of analysing policies and processes in the protection of fairness. Alan
McRobie’s insightful analysis reveals how even small numbers can affect outcomes. Clearly there
will always be divergent views about what needs to be done, but a powerful force for new questions
and investigations must lie in the increasing sophistication of data availability and the opportunities
to interpret that data in rich ways.
For example, Campbell Sharman, University of Western Australia, demonstrates the effort devoted
by himself and Jeremy Moon to the substantial long run database and at the same time draws
attention to the fact that Australia does happen to have 140 years experience of competitive
2 Gough Whitlam, The Whitlam Government 1972-5 Viking, 1985 Chapter 18, Electoral Laws p664.
10
Foreword
elections and in his words, ‘that it can provide a forum for testing a wide range of propositions
about the nature of representative democracy.’
Opportunities
Turning thirdly to the potential for research, John Wanna, of Griffith University, provides a
sweeping summary of the opportunities for collaboration and partnerships and defines the client
base for such research. His advice on the politics of research coupled with that of administrative
researcher Michael Maley provide some parameter for commissions, which are naturally cautious
about the nature of the research they choose to undertake.
Tasmanian Electoral Office administrators Andrew Hawkey and Michelle Davy point out the
breadth of electoral resources - enrolment, boundary and election data being comprehensive mines
for research projects. Medew (AEC) wishes to encourage this through the use of geographic
information systems. Gerard Newman, Statistics Group Director at the Commonwealth
Parliamentary Library, gives an excellent overview of the Library’s potential contribution to new
research, and the level and areas of cooperation that the Library can support. Jenni Newton,
Research Director at the South Australian Parliamentary Library, refers to the significant contri­
bution that state research data analysis can make to the federal sphere and vice versa. She also notes
that changes to capturing voting and preference distribution figures by data input processes open
up vast research potential.
Impediments
Beyond the opportunities for new research, the conference did not shy away from debating the
problems associated with undertaking research.
Among the impediments, there is a lack of access to longitudinal data although the data itself exists.
There are issues relating to inconsistent deposit in libraries, ISBN identification of material
produced, the effects of Desktop Publishing on legal deposit to libraries and archives, and generally
a dearth of well described and accessible archival and published sources. John Wanna, for example,
talks about the post election information legacy and how this could be utilised if it was more
accessible. From the legal perspective, there is a lack of documentation of electoral cases, even at
the High Court level (see Williams).
For the administrator, research is a secondary activity sitting behind the operational imperative to
organise elections; the resourcing of research may be compromised by these competing priorities.
For the academic, the university funding environment frustrates consistent research activity, and
grants may be the only recourse, meaning a hit and miss approach rather than a steady building of
a body of knowledge and expertise. At the same time, there are good arguments for a strong
industry link, supported through such programs as SPIRT (Strategic Partnership with Industry,
Research and Training) grants.
One time educational leader in South Australia, Garth Boomer, once likened the genetic pool to the
memory pool, encouraging humanity to preserve a diverse ‘meme’ bank for the benefit of all. With
that encouragement in mind, the lack of a consolidated research ‘bank’ or network to advise areas
of current research and share results is another area of concern. Similar statements can probably be
found to be echoing in any number of social science disciplines.
11
Foreword
Fragmentation of the research process and the disparate nature of the product leads Flinders
University Professor Dean Jaensch to deliver a spirited expose on the research boundaries issue. He
celebrates the culture change that he has witnessed in electoral commissions but exhorts that
resources be provided to foster ongoing research into all components of the electoral process. His
prescription for action is to establish a National Election Studies Centre and to use public education
arguments as a lever to secure government resourcing. The effective consolidation of
information/research activity and the international potential of such an exercise are among the
benefits of this vision. In similar vein, Colin Hughes refers to the value of creating a national
research register and elsewhere, others make references to the building of a virtual network of
collaboration.
The Legal Framework
It is worth reflecting on the fact that many Australians assume that our democracy, including
electoral rights, are embedded in the language and intent of the Constitution. Many Australians
would assume that their Constitution as originally framed would include reference to voting
equality. Gough Whitlam’s interpretation is that the Australian Constitution does not entrench the
right to vote for all citizens in both State and national elections, and does not refer to equal voting
rights for all Australians. He notes that Section 24 simply requires that members of the
Commonwealth Parliament are ‘directly chosen by the people’.3
While various challenges may be mounted about the place of the lawyer in the electoral reform
context, ANU legal academic, George Williams opens our eyes to the increasing tendency to refer to
international jurisprudence, where there is either no precedent or, conversely, fuzziness in
Australian law.
While the prevailing ideal is that the law should be to provide a supportive framework without
undue prescription, one is left with the impression that the role of law and especially international
law is a potentially complex arena that could impact on legislative presumptions and thus practices.
This may particularly pertain in a more litigious society. The rebuttal to this position is that
Australian politicians are less likely to mount legal challenges if they feel that the same argument
might come back at them at a later date. However, it appears that international legal determinations
may create new complexities for commissions in the future and this is part and parcel of an
increasingly global environment.
Public Education
The vexed issue of effective public education is ever present in discussions about research, partic­
ularly with the November 1999 Republic Referendum experience firmly in mind. Does it matter
that people were encouraged in this case to vote ‘No’ on the grounds that they did not believe they
had an adequate grasp of the topic? Yes, says George Williams, they should be encouraged to selfinform.
This point stimulates some inquiry. Where is the incentive and encouragement to self-inform in
such situations? What does the experience say either positively or negatively about preparation for
citizenship through our schools, tolerance of political ignorance, and the level of public awareness
of our democratic processes generally? How well developed is our collective national identity and
valuing of the historical foundations of our system? What role can technology play in supporting
more effective public education?
3 Gough Whitlam op. cit. p653.
12
Foreword
It is not possible to adequately explore the very large question of the interface between public
education and research in these proceedings. Reference can be made to government curriculum
investments, and the fact that electoral commissions clearly have a role as ‘parallel education’
agencies. The benefits that electoral research can offer the public education function can be
identified as potentially valuable in stating the case for research resourcing. It is acknowledged that
more effort is required to communicate this position.
Information Technology
With the spectre of the retina-scanned voter before us, the impact of technological change both in
the present and its potential for the future can be considered.
For those who are pre-occupied with the minutiae and the inevitable human frailties of any system,
it is easy to overlook the fact that solutions generally create another set of problems, and it is no
different in the realm of information technology.
For example, there is some preoccupation with the current quality of database design and its
capacity to respond to known and future research needs. For administrators, the resource tension
between IT investments for ‘core business’ and IT investments for research purposes is self-evident.
A further issue is the matter of access and use. There are huge research benefits in computerised
data, but a word of caution can be offered about the potential for manipulation of information
previously inaccessible to external parties.4 Dean Jaensch believes that ethics and probity should be
uppermost in the researcher’s mind when using electoral commission data. These observations
leave little doubt that the ethical context of use of electoral information in a rich IT environment
needs further deliberation.
Turning to some specific points raised by the three articles in the technology section, there is
considerable attention paid to future scenarios.
IT Director at the Australian Electoral Commission, Rod Medew, makes a passionate plea for use of
a Geographic Information System in the electoral context, and claims gross under-exploitation of
this approach. He notes the benefits for electoral commissions, researchers, the media and other
users of data.
Seasoned ABC election analyst, Antony Green, brings attention firmly back to the data and its
definition as the primary issue, not the technology, and the principles involved in secret ballots as
they relate to technological developments.
ACT Electoral Commissioner Phil Green confirms that administrative research efforts encompass IT
applications and other areas of need, not just matters linked to political science. Leading us into the
brave new world, he ranges across the future of electronic scanning in vote counting; voter-direct
input, and smartcard technology applications.
The Future
In summary, if we are to ask ourselves why research is important, then the articles provide an
answer along these lines.
Research is a major building block for transparency of process and effectiveness of policy.
4 Tim Glanville, Deputy Australian Electoral Officer for Victoria.
13
Foreword
Provided that it is well-conceived, generally concerned with advancing principles of fairness, and
offers either immediate or long term benefits, research can be a powerful tool in supporting
confidence in the election process and ‘the legitimacy of the elected body’ (Hawkey and Davy).
How collaboration is to be supported in the future, a practical matter for discussion, is a more
perplexing problem, perversely complicated by the professionalism of administrators and analysts
and a profound wish to ensure that whatever is done in the way of collaboration is well supported.
One can only conclude that the development of a more defined collaborative approach is worth
pursuing, aided by communication power and effective people networking. Indeed, there is consid­
erable scope to ponder the greater uses to be made of the world wide web environment to facilitate
many of the needs raised.
The benefits of the discussion on research in the papers which follow are numerous. The
exploration, investigation, and analysis of research issues cover the roles of academics and adminis­
trators, analysis of data over time, trends, questioning of practices and processes, the changing
environment and the role of technology. The central role of the preservation, access to, and consoli­
dation of research findings (and law cases) are considered. A practical and a profound choice of
topics are offered and an agenda for firm collaborative actions and outcomes is generated.
Being responsible for a public Centre that is devoted to the affirmation of democracy, it is an
honour to provide the forword for this record of proceedings.
Elizabeth Ho
Director, The Hawke Centre
University of South Australia
14
Electoral Research And Administration: A Brief History
Colin A. Hughes
Elections have been around for a very long time and they have always required people to run them.
Other people have also been around for a long time: those who write about political topics,
including elections and sometimes, very rarely, even electoral administration. At the end of the
second millenium of the Christian era, we can still usefully go back to just before the beginning of
the first millenium, and have a look at what Cicero had to say about s.340 of the Commonwealth
Electoral Act (and its State equivalents), the prohibition of canvassing at the entrance to polling
places.
Roman electors, who were grouped according to their notional tribes in stockpens on the Field of
Mars before they were marched off to vote, complained that they were upset by canvassers yelling
at them and intimidating them as they went by. Presumably, though Cicero fails to say so, the more
environmentally-conscious among them were concerned that Rome’s quarries were being exhausted
to produce the how-to-vote tablets thrust into the voters’ hands. A good rule to remember in this
game is that very little occurs for the first time.
Cicero’s starting principle was that ‘the freedom of the ordinary person to vote as he wishes shall
not be interfered with’, and he went on to note that Gaius Marius when tribune in 119 BC had
introduced a law (the Lex tabellaria) to narrow the gangways and thereby keep other people away
from the voters.1 A sound principle and a sensible measure one would have to say. But then Cicero
was not only a political philosopher, he was also a politician and a lawyer. Perhaps his other
interests kept him closer to earth.
Coming back, nearer to the present in the 18th century, Rousseau2, whose ideas about democracy
are still very central to electoral topics, has useful things to say about elections in general and about
details like the allocation of electors among constituencies and the choice between open or secret
voting. Not long after Australia had been colonised, there were Bentham and the Mills, father and
son, writing about electoral reform and in Bentham’s case a great deal of detail about electoral
administration. Item: the title Bentham proposed for the chief administrator of the neutral bureau­
cratic agency he advocated, the Elections Master-General, has a martial swing to it that does tempt
one. Nevertheless, despite such distinguished exceptions, political philosophy for the most part
failed to trickle down.
The end of the 19th century must have been an exciting time for electoral administration. Literacy
had spread sufficiently for secret voting to be practicable. Close to adult male suffrage was
becoming common and the enfranchisement of women was about to start. Mass parties were being
formed. And, on the other side of the gap, philosophising about politics was starting to turn into
the modern discipline of political science.
Notwithstanding such a favourable environment the gap between theory and practice if anything
widens. True, details of electoral law and thus, indirectly, electoral administration can now be
found in useful little textbooks written for lawyers and electoral agents by lawyers. But such details
were seen by few readers.
1 Cicero, On Government, Penguin, Melbourne, 1993, pp.213-34.
2 J.J. Rousseau, The Social Contract and Discourses, Everyman, London, 1913, pp. 89-100.
15
Electoral Research And Administration: A Brief History
There were exceptions like Charles Seymour’s Electoral Reform in England and Wales (1915) and
(with a junior collaborator, D.P Frairy) his How the World Votes (1918)3. Most writers employed a
broad brush and the details were ignored. Not always though. Seymour and Frairy record one
episode that should be more widely known.
In France in the 1820s it was required that the lists of electors be published so that omissions from
the roll could be challenged:
At times, the prefects would not publish the lists until four o’clock on the morning of the
day upon which the election would be held. ... At Nancy, the prefect refused to have the
names listed in alphabetical order, and placed them on placards seven or eight feet in
length in five or six columns. It would take a day to discover whether one’s name were
omitted and a ladder would be necessary to read the names on the tops of the columns.4
It is useful to be reminded what can happen with politically-appointed electoral administrators, and
with drafting which fails to specify timetables and forms with sufficient care. And ingenuity like
that shown by the prefect at Nancy deserves to be remembered. Such nuggets can be found from
time to time in political history and biographies.
Only very late in the day did the, by then numerous, discipline of political science produce a book
that would actually be useful to electoral administrators. Inevitably it was written in the United
States where most of the world’s political scientists then worked, and where the fragmentation and
partial privatisation of electoral administration justified such a book.
Joseph P. Harris got his doctorate at Chicago, then the most ‘practical’ school of political science in
existence - as did Louise Overacker who wrote the first substantial book on Australian political
parties. Whilst a graduate student he became interested in voting machines. Exceptionally, his
interest was then applied. He developed and sold one model (the basis for the Votomatic punchcard system), invested the proceeds wisely, and was able to give a substantial endowment to his last
university.5
Harris began his teaching career at two universities with a focus on social questions and practical
affairs, Wisconsin and Washington. In 1929 the Brookings Institution, one of the earliest and most
productive of the American foundations which link public affairs and academia, published his book
on voter registration. In 1934 Brookings published his more comprehensive study of election
administration in the United States which remains today a mine of information and helpful
suggestions.6
To produce that one exception, Joe Harris and his book, you needed universities with a practical
bent in the social sciences, wealthy foundations which sought to promote good policies, and an
individual who was ready to assemble practical knowledge and then write it down. It is a rare
mixture.
Election Administration’s publication was a false dawn. Harris wrote about Congress for the rest of
his academic career and worked in military government and with United Nations Relief and
Rehabilitation Administration (UNRRA) when he was away from academic life. To the best of my
3 C. Seymour, Electoral Reform in England and Wales: The Development and Operation of the Parliamentary Franchise
1832-1885, Oxford University Press, London, 1915 [reprinted Archon Book, Hampden, CN, 1970]; C. Seymour
and D.P. Frairy, How the World Votes: The Story of Democratic Development in Elections 2v, C.A. Nichols,
Springfield, MA, 1918.
4 Seymour and Frairy, op.cit., v.1, p. 336.
5 ‘Joe Harris’ Electoral Tinkering Funds’, Public Affairs Report (Institute of Governmental Studies, University of
California at Berkeley), July 1995, p.16.
6 J.P. Harris, Election Administration in the United States, Brookings Institution, Washington, 1934.
16
Electoral Research And Administration: A Brief History
knowledge, no one sought to follow in his footsteps and write another book about electoral
administration, even in the United States where circumstances were so much more supportive.
After the Second World War things ought to have got better. Enthusiasm for quantification in order
to make political science more scientific, and for studying actual behaviour (rather than formal
rules) for similar reasons, pushed an increasing number of political scientists into the study of
elections and electoral behaviour. Once a critical mass had been reached, the publishability of that
sort of research added career motivation to disciplinary fashion. The professionalisation of
politicians and those in associated trades like pollsters and campaign managers simultaneously
directed many practitioners towards the same subjects.
Nevertheless a gap remained as evidenced by two journals, each useful in its own way. Election
Administration Reports (begun about 1970), although produced by an academic, Richard Smolka
is, as it says, a newsletter for election officials. In contrast, Electoral Studies (started in 1982) is
wholly academic political science, a specialized journal which deals with elections. Its articles are
about voting and voting change, the course and the outcome of national elections, campaigning and
its legal restraints, the nature of electoral systems and their impact upon parties, and the meaning of
elections in democratic theory.7 By the current (March 2000) number, esoteric statistics have
completely taken over. The two journals began and have remained administrators’ cheese and
academics’ chalk.
Next in this historical survey, the international symposium at Copenhagen arranged by Tokai
University in 1991 can be noted. Its proceedings were subsequently published in the Japanese
Journal of Behavioral and Social Sciences under the promising title, ‘Election Administration in a
Comparative Perspective’. The symposium’s principal organiser, Rei Shiratori, explained that it was
being held because:
…the importance of election administration has not been fully appreciated by academic
scholars in the past. Specialists of election studies have done an enormous amount of
research work on the analysis of election results, voting behavior of the electorate and the
function of party and party systems, but they have neglected the fact that election adminis­
tration was one of the decisive factors in successfully holding democratic elections in all
societies. Like candidate selection processes, election administrations have never been paid
full attention by scholars.8
Five of its twelve papers included the words ‘election administration’ in their titles, and another two
had ‘administering’ and ‘administrative’. The majority of those attending held academic posts, but
quite a few had been closer to the action as electoral advisers and observers, and some even had
experience of administrative responsibilities in the electoral sphere.
Another false dawn. When, a few years later, the International Political Science Review devoted a
special number to what it called ‘the politics of electoral reform’, the editor’s opening sentence read:
Countries commonly introduce minor changes to the legal and procedural regulation of
elections, such as revisions of constituency boundaries or campaign finance rules.9
And that was the last heard of such matters. The rest was about how politicians chose a voting
system.
7 ‘Editorial’, Electoral Studies, vol. 1 no. 1, p.2.
8 R. Shiratori, ‘Preface: Election Administration - New Dimensions in Electoral Studies’, Journal of Behavioral and
Social Sciences, vol. 37, p.i.
9 P. Norris, ‘Introduction: The Politics of Electoral Reform’, International Political Science Review, vol. 16, no. 1, p. 3.
17
Electoral Research And Administration: A Brief History
But by the early ‘90s a new phenomenon had begun to emerge. Bodies predominantly non­
academic and administrator-oriented were edging into the gap between practice and theory from
the practitioners’ side. Admittedly they had a special concentration, on some classes of country
rather than others. But they were doing a lot and doing it quickly.
First into the field was the International Foundation for Electoral Systems (IFES), a private, non­
profit foundation founded in Washington in 1987. Its newsletter, begun in 1990 and called
Elections Today since 1993, is closer to Election Administration Reports than to Electoral Studies. It
is certainly activist in orientation rather than academic. It arranges seminars and colloquia which
are attended primarily by administrators, and promotes regional associations of administrators to
further the exchange of information among them. It publishes manuals and handbooks for the same
public.
IFES was started with financial assistance from the Office of Democratic Initiatives in the US
Agency for International Development, though it also has drawn resources from private
foundations. The driving force behind this new development was the deliberate promotion of
democracy by a number of governments in what used to be called the ‘Free World’ and the identifi­
cation of ‘free and fair elections’ as either the central element in that strategy or else one of two with
‘competitive parties’. More of that part of its history has been traced by Richard Soudriette.10
A significant literature which combines detail about electoral administration with larger questions
about the introduction and maintenance of democratic government has emerged. The title of one
very recent collaborative volume sums it up nicely by combining post-conflict elections, democrati­
zation and international assistance. Another collection interestingly brackets ‘electoral adminis­
tration’ with judicial systems, corruption control and central banks as ‘agencies of restraint’.11 The
latter point is worth thinking about.
Following IFES in time is the International Institute for Democracy and Electoral Assistance
(International IDEA, or often just IDEA), a small international agency within the UN family,
founded in Stockholm in 1995 with originally 14 member states, Australia being one. Sweden’s
government has provided the lion’s share of its resources in keeping with their tradition of overseas
aid.
Some of IDEA’s electoral publications are near to the academic side of the gap. Thus Reynolds and
Reilly’s Handbook of Electoral System Design and Karam’s Women in Parliament sit comfortably on
the shelf (or shelves nowadays) devoted to works written by academics on those topics.12 But other,
shorter publications which contain codes of conduct for various classes of persons and organi­
sations in the electoral administration field are clearly further away in the direction of adminis­
trators.
Third and finally, the involvement of various organs of the United Nations in the field of electoral
assistance should be mentioned, a more complicated story than can be recounted at this point, and
of regional agencies connected with their local sub-global associations eg in Latin America.
In 1991 at Copenhagen it was still correct to lament the limited effort given to the study of electoral
administration. But by 1999 it is possible to point with pride to a truly major achievement of the
three bodies just named, IFES, IDEA and the UN. The ACE project, embodied in CD-ROM and
10 R. Soudriette, ‘IFES: Ten Years of Building Democracy’, Elections Today, vol. 7, no. 3, p.5.
11 K. Kumar, (ed.) Postconflict Elections, Democratization, and International Assistance, Lynne Rienner, Boulder, CO,
1998; A. Schedler, L. Diamond and M.F. Plattner, (eds.) The Self-Restraining State: Power and Accountability in
New Democracies, Lynne Rienner, Boulder, CO, 1999.
12 A. Reynolds and B. Reilly, (eds.) The International IDEA Handbook of Electoral System Design, International IDEA,
Stockholm, 1997; A. Karam, (ed.) Women in Parliament: Beyond Numbers, International IDEA, Stockholm, 1998.
18
Electoral Research And Administration: A Brief History
Website, gives both administrators and academics access to what can only be described as a major
on-going encyclopaedia of electoral administration.
Supporting that spectacular monument are numerous reports, studies, seminar and conference
papers, and networks, by which information, best practices, analyses of failures and successes have
become known to a small but worldwide readership of election administrators though, it is feared,
to only a few academics. The one complaint that can be made is that most of this work has been
directed to conversion of the heathen, to carrying the gospel of free and fair elections to parts of the
world which never, or hardly ever, knew them.
Australia, and quite a few of those here today share the credit for this, made a considerable contri­
bution to this burgeoning body of material, primarily as a donor, very rarely as a beneficiary.
It is now time to ask two questions. What is the purpose of research? What might academics have
to offer administrators and what might administrators have to offer academics? To the first
question, the purpose of research, three separable answers are suggested. To do one’s own job
better by knowing and understanding information that was not previously available. To inform
influentials (MPs and party apparatchiks, journalists, and even non-headquarters electoral officials)
so they understand a lot more about what is going on. And, third, to inform the public, the electors,
so that they understand a few basic things about what is going on.
A quick answer to the second question might be that academics are better placed to ask politically
sensitive questions which may show what influences outcomes and were, though not necessarily
any longer, more likely to look overseas or back in time for comparative evidence. Administrators
are more ready to utilise aggregate data - and better placed to create it - and experience has given
them sounder judgment with which to read often inadequate evidence.
There are undeniably problems when administrators undertake research. Research may end up
attributing causes, but to find that the shifting level of support in a particular electoral district is
caused by demographic change rather than the heroic efforts of the incumbent may be tactless.
Attention given to new phenomena, the rise of Hansonism say, is likely to antagonise those being
examined who are usually unaccustomed to such attention and also the previous principal actors
who believe the problem will go away sooner if nobody says anything. Research on the outcomes of
alternative electoral systems may carry suggestions that one lot have benefited from some
arrangement or other, and that too can be seen as unwelcome and improper intervention in
partisan affairs.
It should be added that the existence of a standing parliamentary committee charged with electoral
matters is a very useful device for reducing those and other risks of research, and such a committee
can also, indirectly, come up with useful ideas for research.
There are now a great many runs on the board. Annual Reports and occasional reports about
technical matters like informal voting now contribute significantly, if discreetly, to the body of
‘research’. Three possible areas where research by administrators ought not to offend too much can
be mentioned. The first might be called gross demography, the enormous volume of roll
transactions. It is unlikely if many people outside the electoral sphere appreciate the extent of
movement, and the potential for change it involves. The second is the extent of relative stability, at
19
Electoral Research And Administration: A Brief History
the polling booth level and even when gross numbers are altering rapidly. And third, again using
polling booth data now there has been some experience of distributing preferences at the polling
booth, seeing how uniformly preferences move within a field.
For the academics, there are different difficulties. It is annoying that over many years, the academic
system has been loaded against applied work. Unless financial arrangements have recently changed,
an article in an overseas journal which is refereed (and usually needs to begin with 20% of the
words going to explain where Australia is) trumps an article in a local journal which isn’t refereed
but can take detail, whilst a submission to a royal commission or a parliamentary committee scores
only as community service if at all. Australia is so expensively distant from the North Atlantic
catchment area, where most of the action is, that our political scientists haven’t been that much
involved in the specialist networks. There have been exceptions though - Ernie Chaples in
campaign finance for one, Marian Sawer and Marian Simms in women and elections for another.
Finally it should be mentioned that academic lawyers have played a very small part in electoral
research, apart from a few articles (sometimes misconceived it might be said). Griffith Law Review
had a special number in 1998 whose editor attributed its ‘Cinderella status’ to the small number of
interested persons - electoral administrators, party apparatchiks and a few practitioners who had
links with parties and were used to advise and conduct the occasional bit of litigation.13
13 G. Orr, ‘Editorial: Special Edition on Electoral Regulation and Representation’, Griffith Law Review, vol. 7, no. 2,
pp. 168-73.
20
THE RESEARCH CORE – PAST, PRESENT AND FUTURE
21
1
The Use Of Electoral Data For Research Into The Australian Political
System
Campbell Sharman1
Running elections has not exactly become a major industry, but the regulation of an increasing
range of activities relating to elections and the role of political parties has led to the growth of
government agencies responsible for electoral matters. This paper, however, is concerned with the
most longstanding concerns of those who run elections, the recording of votes and the determi­
nation of who has won representation. The paper deals with some initial issues about the role of
elections and electoral agencies in the political process, moves to an examination of the accessibility
of electoral information and the way it is presented in terms of its use by the general public and by
researchers, and concludes with a brief survey of a project which uses long run data on elections
and representation to study the evolution of the Australian political system.
The Place Of Elections And The Role Of Electoral Agencies
Our system of representative democracy requires much more than holding free, open and
competitive elections for public office. It rests on a belief in the importance of the rule of law, the
need to limit the powers of government, and a host of other values associated with freedom of
speech, freedom of association and tolerance for those who hold differing opinions. This being said,
elections do play a critical role in the operation of democratic government. Not only do they
provide the component of popular representation in government which is the key to representative
democracy, but they permit that most important of operations, the dismissal of one set of
government officials and their replacement with another. This ability to change regime through an
orderly procedure involving popular choice is the hallmark of democratic government.
What happens as a consequence of regular elections is as important as the results of the elections
themselves. The possibility of losing office creates the apprehension in elected officials necessary for
them to be responsive to our preferences. In addition, it generates the conditions for competitive
politics in which rival groups strive to enjoy the fruits of office and to implement differing views of
the public interest.
All this explains why it is vital that elections are conducted fairly and efficiently. Australia has a
long history of innovation in the administration of elections and of experimentation with the use of
electoral rules to ensure fairness in representation.2 Some would say that the Australian concern
with fairness has reached the point of obsession and that contesting views of what is fair have
become a battleground for partisan advantage.3 Compulsory voting,4 block voting, the single
member district, and state funding of political parties can all be supported or attacked on grounds
of fairness depending on one’s assumptions about what the electoral system should be designed to
achieve.
1 [email protected].
2 Surveys of Australian electoral practices include Bennett (1996), Goot (1985), Jaensch (1995) and Wright (1980).
3 Note the comments by Elkins (1992).
4 There has been a resurgence in interest in this topic; note Mackerras and McAllister (1999), Farrow (1998), Green
(1998) and Johns (1998).
22
The Use Of Electoral Data For Research Into The Australian Political System
The result is that the task of those who administer the electoral process has become highly complex
and remains politically sensitive. A large body of electoral law now regulates the maintenance of the
electoral roll, the registration of candidates and political parties, and the conduct of electoral
campaigns, quite apart from the way in which votes are cast and counted. Organizing a general
election is a major undertaking involving hundreds and perhaps thousands of officials, both
permanent and temporary. There is an increasing demand for speed in the posting of results, and
expectations that the latest technology will be used to display the results, including near instan­
taneous display on the internet. On top of this, a scrupulous observation of the rules is expected
together with a complete absence of any partisan bias on the part of those administering the
electoral process.
This means that government agencies responsible for running elections in Australia have become
large and expensive undertakings with obligations which often run far beyond the administration of
elections. Even so, their activities retain a strongly cyclical pattern. The agencies must contend with
the periodic redistribution of electoral boundaries in addition to the cycle for general elections. As
can be seen in the statistical returns and annual reports produced by electoral agencies, there is a
mountain of detail to be accommodated, both statistical and in terms of fulfilling the requirements
of electoral laws and auditing rules.
The information provided in the publications of electoral agencies is of immediate interest to the
participants in the electoral and governmental processes, but it is also of concern to the wider
public and to those who wish to study electoral data and the conduct of elections. It is now
accepted that electoral agencies have a responsibility to educate the public about the electoral
process and to make relevant information easily available. This has not always been the case and
current electoral publications are vastly superior in most respects to those of thirty or more years
ago. Even so, depending on the information sought, there is still some room for improvement in
both making information available and in the way it is presented.
Availability Of Electoral Publications
On moving to Western Australia in the late 1970s, it was necessary to find where past statistical
returns for Western Australian elections could be bought. The answer was they were not for sale;
they were free. This sounds great, but the result was that they were very hard to find, and only by a
series of phone calls to find the relevant official. Moreover, the statistical returns were not always
regarded as parliamentary papers, and were therefore not bound with such papers in public and
university libraries. This further restricted access, particularly for older returns and for researchers
who lived outside the state. The problem in Western Australia was duly solved but, with the advent
of desktop publishing and corporatisation in all jurisdictions, it has re-emerged and, if anything,
has got worse. There is now a multitude of agencies producing government documents of various
kinds. Some are available from government bookshops, others from the agencies concerned, still
others from commercial outlets. Some are required to be tabled in parliament and some are not.
In order to aid public access to electoral information and to aid research, all electoral agencies
should ensure that a few rules are followed. The first, and perhaps least obvious is that all
publications have an ISBN number and a copy lodged in the state library and the National Library
in Canberra. Without this number, the publication is invisible. With an ISBN number the
publication is available to national and international library databases, and can be tracked down by
an interested citizen or researcher. Most electoral publications in Australia now have such a
number, but some still do not. It should be an auditing requirement that all government reports
and substantial publications have such a number.
23
The Use Of Electoral Data For Research Into The Australian Political System
The second request is that the publication should be tabled in the relevant parliament and treated
as a parliamentary paper. There may be some technical difficulties with this procedure, but the goal
is for each parliament to have an authorised set of statistical returns and to take formal responsi­
bility for their long term availability. The third request is that current statistical returns are made
available at marginal cost in state and commonwealth bookshops or, even better, at commercial
bookshops with appropriate commissions. Again, the goal is to make the information available
easily through regular channels.
Some might say, these requests are unnecessary given the Internet and the presentation of
information on the web pages of electoral agencies. There is some truth in this, particularly in the
provision of background and educational material, but a number of problems continue to dog
electoral data on the Internet. The first is that information is often summarized in a form which
makes sense to web page designers but not necessarily to users. Another problem is that the design
of a web page for the rapid presentation of results during an election count is a different beast from
one which aims to present authoritative results for detailed scrutiny. The Australian Electoral
Commission’s (AEC) web page for the last election is a case in point; it gave excellent coverage of
the progress of the count, but still does not have final figures for most electoral districts.5 The last
difficulty is the question of archiving the material. Who is responsible for keeping authoritative
records once the relevant files are committed to the great server in the sky? The AEC has the correct
approach in producing its final returns in hard copy with its CD and web page publication as
supplementary - and very welcome - material.
The Format Of Electoral Reports
The second issue concerns the way in which information is presented in electoral publications.
Much of the detail published in statistical returns for general elections is of a technical nature or
required to fulfill statutory or auditing obligations. This is of concern to those directly involved in
the political process and a small number of interested citizens, and it is proper that such detail is
publicly available from an authoritative source. University teachers and researchers would like to
make as much use as possible of electoral data for courses which involve the study of elections, and
to have data available for their and their colleagues’ research.
There is no special virtue in having electoral information presented in the identical format but it
would be helpful if summary information was included in all election reports. While the
information can usually be found somewhere in the various state and commonwealth election
publications, it is rarely all in one place and in a form that makes comparisons easy. The
information required is deceptively straightforward; the statewide (or nationwide) first preference
votes gained and seats won (number and percentage) by all parties with more than 2 percent of the
valid vote, or who win a seat, and comparable figures for the previous election. This request may
appear self-serving because it is the kind of information useful for my own research, but there are
good reasons why such information should be presented clearly, and early, in every report. It is also
information which could be posted on the relevant web pages as a summary table.
The people who compile election reports are, understandably, concerned with the detail, and with
their immediate audience who are interested in the detail. But there is a much wider, if diffuse,
audience who want to know only the broad profile and consequences of the election. They see the
election as part of the process of representative democracy and want to know its general outlines.
For them the party share of first preference votes and seats, and the way in which such support is
dispersed between parties, is the best way of summarizing the pattern of partisan support at the
5 The figures for the electoral district of Adelaide, for example, dated 23 October 1998 showed 95.17% of the vote
counted (web page accessed 28 November 1999).
24
The Use Of Electoral Data For Research Into The Australian Political System
election. It also says a lot about the way the electoral system turns votes into seats and the nature of
representative parliamentary government. And for those who want to make comparisons over time
or with other systems, this is the critical information.
A final request would be to separate votes for independent candidates from all others, and not to
lump such votes into an ‘others’ category. This, you might say, is a truly weird request. We live in a
party dominated world, so why should we bother with such anti-system residuals? The answer is
that such candidates predated our current party system, have played an important, and sometimes
critical, role in the evolution of all state and commonwealth parliaments and, notwithstanding the
attempts of political parties to exterminate them, continue to play a potentially important role in all
parliaments. The non-party vote is also an interesting barometer of disillusionment with existing
parties, and the rise and fall of the independent vote tells us a lot about the nature of politics in a
given jurisdiction as well as the operation of its electoral laws. But to chart this support, it would be
helpful if electoral agencies presented the statewide (or nationwide) first preference vote for
independent candidates in summary form even if it falls below a 2 percent threshold. Independents
should be regarded as an endangered species and given special consideration.
The Use Of Long Run Data On Elections, Representation And Government
The last topic involves a return to the general question of the importance of elections in the political
system. This time, however, it is not the issue of the role elections play in the governmental
process, but the way in which electoral data can be used to analyse the nature and evolution of
representative government. Long run electoral data, when coupled with information on changes of
government can reveal a great deal about the effect of electoral systems, the operation of the party
system, and the style of parliamentary government in general. It is often forgotten that the
Australian states with 140 years6 experience of competitive elections are some of the oldest contin­
uously operating systems of broadly based representative government in the world and can provide
a forum for testing a wide range of propositions about the nature of representative democracy. Both
for the study of the Australian political tradition and for comparative purposes, long run data on
elections, representation and governments in Australia are an invaluable resource.
For this reason my colleague in the Political Science Department at the University of Western
Australia, Jeremy Moon, and I have been working on a major project for the last five years which
has involved, as one of its goals, the entering of long run data on elections and government into a
computer database. The aim is not only to collect data for our own research projects and
publications,7 but to make available the information as widely as possible to interested researchers
and institutions. The project was initiated with funding from an Australian Research Council Large
Grant for three years from 19958 and has also been supported by funds from the University of
Western Australia. In a modified form and with the participation of academics around Australia, it
has recently won an award from the National Council for the Centenary of Federation, History and
Education Program, for two years from the end of 1999 to produce several publications and a webbased version of the database for both general information and research use.
This paper is concluded by referring to another of Australia’s invaluable electoral resources, Colin
Hughes. There could be no more appropriate person to give the keynote paper and address the
question of the interface between the practicalities and the study of the electoral process in
Australia. One of his greatest contributions to the study of Australian politics and government has
been the production, with Bruce Graham, of the publication A Handbook of Australian Government
6 That is, all states except Western Australia which did not achieve self-government until 1890.
7 A list of publications which have been derived in whole or in part from the database are listed in a section after the
References at the end of this paper.
8 ARC project ‘Politics and Government in the Australian States’ (No.A79532229).
25
The Use Of Electoral Data For Research Into The Australian Political System
and Politics 1890-1964, and his efforts in producing ten yearly supplements and related
compilations of data on state and commonwealth lower house elections and state upper house
contests.9 This mammoth effort by successive teams of researchers led by Colin Hughes has, by
providing the party labels of candidates and summarizing information on party votes and seats
since 1890, removed the greatest obstacle to long run research on Australian elections. This was the
official invisibility of party affiliations in most jurisdictions until the late 1970s which prevented the
easy calculation of party votes and representation.
The work done by Colin Hughes has earned him the gratitude of current researchers, and future
generations of analysts will be forever be in his debt. His work was certainly the inspiration for the
current project I am working on and represents, among other things, the potential for the fruitful
cooperation between electoral officials and those who study the electoral process. What better note
to end on.
References
1. Bennett, Scott (1996) Winning and Losing: Australian National Elections, Melbourne: Melbourne University Press.
2. Elkins, David (1992) ‘Electoral Reform and Political Culture’ in Malcolm Alexander and Brian Galligan (editors)
Comparative Political Studies: Australia and Canada, Melbourne: Pitman.
3. Farrow, Ian (1998) ‘Compulsory Voting: The Australian Anachronism’ Policy 13(4): 41-43.
4. Goot, Murray (1985) ‘Electoral Systems’ in Don Aitkin (ed.) Surveys of Australian Political Science, Sydney: Allen
and Unwin.
5. Green, Antony (1998) ‘Implications of Abolishing Compulsory Voting’ Australian Quarterly 70(2): 6-7.
6. Hughes, Colin A (1977) A Handbook of Australian Government and Politics 1965-1974, Canberra: Australian
National University Press.
7. Hughes, Colin A (1986) A Handbook of Australian Government and Politics 1975-1984, Canberra: Australian
National University Press.
8. Hughes, Colin A and Don Aitkin (1986) Voting for the Australian State Upper Houses 1890-1984, Canberra:
Department of Political Science, Research School of Social Sciences, Australian National University.
9. Hughes, Colin A and B D Graham (1968) A Handbook of Australian Government and Politics 1890-1964,
Canberra: Australian National University Press.
10. Hughes, Colin A and B D Graham (1974) Voting for the Australian House of Representatives 1901-1964, Canberra:
Australian National University Press.
11. Hughes, Colin A and B D Graham (1975) Voting for the New South Wales Legislative Assembly 1890-1964,
Canberra: Department of Political Science, Research School of Social Sciences, Australian National University.
12. Hughes, Colin A and B D Graham (1975) Voting for the Victorian Legislative Assembly 1890-1964, Canberra:
Department of Political Science, Research School of Social Sciences, Australian National University.
13. Hughes, Colin A and B D Graham (1976) Voting for the South Australian, Western Australian and Tasmanian
Lower Houses 1890-1964, Canberra: Department of Political Science, Research School of Social Sciences,
Australian National University.
14. Jaensch, Dean (1995) Election! How and why Australia votes, Sydney: Allen and Unwin.
15. Johns, Garry (1998) ‘Does Compulsory Voting Distort Electoral Outcomes? Agenda 5(3): 367-372.
16. Mackerras, Malcolm and Ian McAllister (1999) ‘Compulsory Voting, Party Stability and Electoral Advantage in
Australia’ Electoral Studies 18: 217-233.
17. Wright, J F H (1980) Mirror of the Nation’s Mind: Australia’s Electoral Experiments, Sydney: Hale and Iremonger.
9 There is a listing of these publications in the References.
26
The Use Of Electoral Data For Research Into The Australian Political System
List Of Publications From The Australian Politics And Government Project
a) Lusztig, Michael, Patrick James and Jeremy Moon (1997) ‘Falling From Grace: Non-Established Brokerage Parties
and the Weight of Predominance in Canadian Provinces and Australian States’ Publius: The Journal of Federalism
27: 59-82.
b) Moon, Jeremy (1995) ‘Minority Government in the Australian States: From Ersatz Majoritarianism to
Minoritarianism’ Australian Journal of Political Science 31 (1995): special issue on Consensus Policy-making: 142­
163.
c) Moon, Jeremy and Imogen Fountain (1997) ‘Keeping the Gates?: Women as Ministers in Australia, 1970-1996’
Australian Journal of Political Science 32: 455 - 466.
d) Moon, Jeremy and Anthony Sayers (1999) ‘The Dynamics of Governmental Activity: A Long-run Analysis of the
Changing Scope and Profile of Australian Ministerial Portfolios’ Australian Journal of Political Science, 34(2): 149­
167.
e) Anthony Sayers and Campbell Sharman (1997) ‘Electoral and Parliamentary Representation in the Australian States,
1890-1996: Rae’s Index as a Measure of Party System Change’ Proceedings, Australasian Political Studies
Association Conference, Flinders University of South Australia, Adelaide, 29 September to 1 October 1997.
f) Sharman, Campbell (1997) ‘Politics in the States’ in Brian Galligan, Ian McAllister and John Ravenhill (eds) New
Developments in Australian Politics Melbourne: Macmillan, 44-67.
g) Sharman, Campbell (1999) ‘The Representation of Small Parties and Independents in the Senate’ Australian Journal
of Political Science, 34(3): 353-361.
h) Campbell Sharman and Narelle Miragliotta (1999) ‘Dealing with Preferences: Preferential Voting and Minor Party
Representation in the Australian Commonwealth Parliament since 1949’ in John Brookfield (compiler) Proceedings
of the 1999 Conference of the Australasian Political Studies Association Sydney: Department of Government,
University of Sydney, volume III, 719-728.
i) Campbell Sharman and Anthony Sayers (1998) ‘Swings and Roundabouts? Patterns of Voting for the Australian
Labor Party at State and Commonwealth Lower House Elections, 1901-96’ Australian Journal of Political Science
33(3): 339-354.
27
2
A Legal Perspective On Electoral Research
George Williams*
Introduction
This article is written from the perspective of an academic lawyer and barrister. As such, the focus
is not on the outcome of elections, but on how electoral law defines and shapes rights and
obligations, and the way in which the law structures how elections are conducted.
In this paper three topics are dealt with:
1. Research issues;
2. The shifting boundaries of electoral law; and
3. New directions and challenges in electoral law research.
Research Issues
Electoral law is a little developed field and there are very few legal academics who consistently write
on the topic. In part, this is because of the impediments to conducting research on electoral law.
Barriers to entry
The field of electoral law has very high barriers to entry:
1. No Australian text
There is no Australian text on electoral law. This is a major omission, particularly when compared
to works in other nations, such as Patrick Boyer’s two volume work on Election Law in Canada.1
The lack of an equivalent Australian text means that the field has yet to be systemically analysed
and constructed. This can make it difficult to begin or undertake research on electoral law topics.
2. The legislation
Australian electoral legislation is very accessible on the Internet via the AustLII site.2 However, this
merely gives access to the raw materials, and does not filter the information or help by bringing
together the legislation as it deals with common issues or themes. What is needed is a site that
enables comparison between jurisdictions, so as to enable researchers to easily determine how an
issue is dealt with across each jurisdiction and to answer questions such as whether the different
regimes have moved apart or come together. This could be achieved by a comprehensive index,
which would take considerable work to maintain, or perhaps hypertext links between similar
thematic treatment, such as laws dealing with truth in electoral advertising.
Research in this area would also be assisted by the publication of annotated electoral legislation.
3. The cases
Electoral law as developed by courts is a diffuse and hard to locate body of jurisprudence. In other
like fields, specialist series of reports have developed to draw together the cases into one accessible
* Barrister, ACT Bar; Fellow, Law Program, Research School of Social Sciences, Australian National University. Parts of
this paper are derived from Williams, G, Human Rights under the Australian Constitution (Oxford University Press,
1999).
1 Boyer, JP, Election Law in Canada: The Law and Procedure of Federal, Provincial, and Territorial Elections
(Butterworths, 1987), 2 vols.
2 www.austlii.edu.au.
28
A Legal Perspective On Electoral Research
series of reports. This has occurred in fields such as family law, corporate law, taxation law and
industrial law, and in some cases there are multiple series dealing with the same field. There is no
equivalent in electoral law, which can make it time consuming to locate new cases on the topic
(particularly as many go unreported).
An example of a recent research task
A project was recently undertaken for the Commonwealth Parliamentary Library for publication in
a book on the centenary of the Commonwealth Parliament. It involved review and analysis of the
work of the High Court sitting as the Court of Disputed Returns. This proved a difficult task.
Unfortunately, the relevant documents are frequently not accessible, making it hard to determine
what cases have been heard and decided by the Court and how many times it has sat. The answers
to such questions lie buried in decades of Court records in the often unindexed archives of the
High Court.
Neither the High Court nor the Australian Electoral Commission can provide basic information on
the work of the federal Court of Disputed Returns. What can be determined must be gleaned from a
variety of sources: the Commonwealth law reports, the Internet, the Australian Digest, various
archives of federal cases on CDROM, and the list of cases in an appendix to House of
Representatives Practice.3
The difficulty in gaining such information is not only cause for concern to academics, it can also
affect the work of barristers practising in the field. For example, the drafting of the petition in Sue v
Hill4 was made difficult by the lack of accessible materials on appropriate precedents for the precise
wording required.
The Shifting Boundaries Of Electoral Law
Electoral law has never been delineated by clear boundaries. It has always overlapped with other
fields of law.
Over the last decade, one of the major catalysts for legal change has been the influence of interna­
tional and comparative law. We are now more likely to see an increasing use of such materials in
the development of electoral law. This has been a clear trend in other areas, both in the
development of case law and in the interpretation of statutory regimes. For example, Brennan,
Deane and Dawson JJ stated in Chu Kheng Lim v Minister for Immigration:5
We accept the proposition that the courts should, in a case of ambiguity, favour a
construction of a Commonwealth statute which accords with the obligations of Australia
under an international treaty.
There has never been a clear boundary between electoral law and constitutional law. The two have
frequently overlapped. This remains true today.
Some State Constitutions and Territory Self-Government Acts mandate a level of voter equality, or
‘one-vote one-value’. Section 77 of the South Australian Constitution, the Constitution Act 1934
(SA), requires that whenever an electoral redistribution is made for the South Australian Parliament,
the redistribution is to be made by applying the principle that the number of electors in any
electorate should not deviate by more than ten percent from the electoral quota (that is, the figure
3 Barlin, LM (ed), House of Representatives Practice (AGPS, 3rd ed 1997) at 770-775.
4 [1999] HCA 30 (23 June 1999).
5 (1992) 176 CLR 1 at 38.
29
A Legal Perspective On Electoral Research
derived by dividing the number of electors by the number of electoral districts). Section 77 is
entrenched by s 88(2), which requires that any Bill to amend s 77 be approved by the electors of
South Australia at a referendum. The Self-Government Act of the Australian Capital Territory6 and
the Constitution of New South Wales7 also establish a ten percent margin, while the SelfGovernment Act of the Northern Territory8 establishes a twenty percent tolerance.
Of course, much of the federal electoral system is set out by the Australian Constitution. Many
difficult issues remain unanswered in this context, such as the contemporary meaning of the arcane
words of s 44 of the Constitution on the qualifications of members of the Commonwealth
Parliament and whether it is possible to oust the jurisdiction of a court to hear an appeal from a
court of disputed returns. The answer to this latter question is unclear after the High Court’s
decision in Sue v Hill.
Aspects of the electoral process are also being seen as entrenched in the Australian Constitution,
although not as a result of amendment of that Constitution but due to interpretation of the
instrument by the High Court. The Court has held that the Constitution entrenches what the Court
regards as the minimum requirements of representative government.
Sections 7 and 24 of the Constitution require, respectively, that the members of the Senate and the
House of Representatives be ‘directly chosen by the people’. This imperative recognises a system of
government based upon democratic elections.
The words ‘directly chosen by the people’ in ss 7 and 24 harbour many unexplored implications.
The word ‘directly’ indicates that electors are to cast their choice for candidates without any
intervening stage, such as an electoral college.9
The word ‘chosen’ is more significant. The ‘choice’ mandated by ss 7 and 24 would be frustrated by
any law that provided that there could only be one candidate per electorate, or indeed a limited
number of candidates per electorate. It would also be inconsistent with a law that limited eligibility
to stand for office to members of a particular political party, or indeed a law that provided that
members of a certain organisation could not stand for election (hence, the Communist Party
Dissolution Act 1950 (Cth) would now likely be unconstitutional on this basis). A ‘choice’ implies,
if nothing more, a free and ‘genuine choice’,10 perhaps even an informed choice.
The interpretation of ss 7 and 24 by the High Court continues to raise issues that are of importance
to Australian electoral law.
For example, s 303 of the Electoral Act 1992 (ACT) bans canvassing within a 100 metre radius of
polling booths during Australian Capital Territory elections. In particular, it prohibits ‘anything for
the purpose of influencing the vote of an elector as the elector is approaching, or while the elector is
at, the polling place’. There is a strong argument that this prohibition would be unconstitutional.11
This issue also shows the potential relevance of comparative legal materials. In 1992 in Burson v
Freeman12 the United States Supreme Court dealt with the validity of a Tennessee provision that
restricted canvassing within 100 feet of a polling place. A 5:3 majority held that the restriction was
not a violation of the First Amendment right to free speech. This would be highly relevant to any
determination of the similar issue by the Australian High Court.
6 Australian Capital Territory (Self-Government) Act 1988 (Cth), s 67D.
7 Constitution Act 1902 (NSW), s 28.
8 Northern Territory (Self-Government) Act 1978 (Cth), s 13.
9 Australian Capital Television Pty Ltd v Commonwealth (1992) 177 CLR 106 at 227 per McHugh J.
10 Langer v Commonwealth (1996) 186 CLR 302 at 325 per Dawson J.
11 See Williams, G, ‘Freedom of Political Discussion and Australian Electoral Laws (1998) 5 Canberra Law Review
151.
12 504 US 191 (1992).
30
A Legal Perspective On Electoral Research
Other electoral law issues that are also of constitutional significance are examined below:
1. The implied freedom of political communication
In Australian Capital Television Pty Ltd v Commonwealth,13 the High Court forged the text of ss 7
and 24 into a constitutionally guaranteed freedom of political discussion that impacts upon the
laws of both Federal, State and Territory Parliaments. Six members of the High Court found that
the system of representative government created by the Constitution necessarily requires that
Australians be free to discuss matters relating to Australian government. The Court struck down
sections of the Political Broadcasts and Political Disclosures Act 1991 (Cth) which imposed a
blanket prohibition on political advertisements on radio or television during election periods.
Later cases, most recently Lange v Australian Broadcasting Corporation, 14 have recast and
confirmed the protection offered. In Lange, the implied freedom of political communication was
unanimously affirmed as an implication from the text of the Constitution:
Freedom of communication on matters of government and politics is an indispensable
incident of that system of representative government which the Constitution creates by
directing that the members of the House of Representatives and the Senate shall be ‘directly
chosen by the people’ of the Commonwealth and the States, respectively.15
Although the implied freedom extends beyond speech at election time, the focus on the freedom (as
derived from ss 7 and 24) is nevertheless on speech during that period. Hence, it is possible that a
higher degree of scrutiny will apply where electoral laws overstep the line. This means that
particular care must be taken in drafting electoral offences that target speech or association, such as
laws that regulate truth in electoral advertising (and hence the content of such advertising) and
emerging campaign practices such as push polling.
2. Compulsory voting
Section 245(1) of the Commonwealth Electoral Act 1918 (Cth) states: ‘It shall be the duty of every
elector to vote at each election.’ The section then goes on to establish a penalty for electors who fail
to vote in federal elections. Might it be argued that the ‘choice’, or ‘genuine choice’, required by ss 7
and 24 includes the right not to make a choice? In Adelaide Company of Jehovah’s Witnesses Inc v
Commonwealth,16 Latham CJ argued: ‘The prohibition in s 116 operates not only to protect the
freedom of religion, but also to protect the right of a man to have no religion.’ It may be that s 245
does not force a person to make a choice, and that while it requires attendance at the polling booth
and the depositing of the voting paper in the ballot box, the voter need not mark the ballot paper.17
Nevertheless, it might be argued that the system of compulsory voting established by s 245 is
unconstitutional.18
The result in Langer v Commonwealth19 suggests that compulsory voting is not inconsistent with the
‘choice’ required by ss 7 and 24. McHugh J concluded in that case that ‘this Court has held that
compulsory voting in federal elections is within the power of the Parliament’.20
13 (1992) 177 CLR 106.
14 (1997) 189 CLR 520. The author appeared in this case as counsel for the Media, Entertainment and Arts Alliance.
15 Ibid at 559.
16 (1943) 67 CLR 116 at 123.
17 Twomey, A, ‘Free to Choose or Compelled to Lie? – The Rights of Voters After Langer v The Commonwealth’
(1996) 24 Federal Law Review 201 at 208-210.
18 Cf Judd v McKeon (1926) 38 CLR 380, where the High Court upheld compulsory voting.
19 (1996) 186 CLR 302.
20 Ibid at 340, referring to Judd v McKeon (1926) 38 CLR 380.
31
A Legal Perspective On Electoral Research
Section 240 of the Commonwealth Electoral Act stated that: ‘In a House of Representatives election
a person shall mark his or her vote’ by numbering every square ‘1, 2, 3, 4 ...’. Under s 270 a ballot
paper ‘shall not be informal’ if it includes a sequence of consecutive numbers beginning with ‘1’,
even if numbers are duplicated and even if one square is blank. Thus a paper numbered ‘1, 2, 3, 3
...’ will be counted as indicating a preference for candidates ‘1’ and ‘2’. Under s 329A it was an
offence, attracting up to 6 months’ imprisonment, to ‘print, publish or distribute ... any matter or
thing with the intention of encouraging persons ... to fill in a ballot paper otherwise than in
accordance with’ s 240.
At the 1993 federal election Albert Langer urged electors to vote ‘1, 2, 3, 3 ...’, with the major
parties being placed equal last. The High Court unanimously held that s 240 was valid. The Court,
with Dawson J dissenting, held that s 329A was also valid.21 Underpinning these conclusions was
the view that the Constitution, in ss 31 and 51(xxxvi), gives the Commonwealth Parliament a broad
role in selecting, and protecting, the means by which the members of the federal Parliament are
elected. Section 240 was valid as it could ‘reasonably be regarded as prescribing a method of freely
choosing members of the House of Representatives’, while s 329A was valid as ‘a law which is
appropriate and adapted to prevent the subversion of that method’.22 Section 270 was characterised
not as establishing an alternate method of voting, but as being a mere saving provision designed to
enable the counting of the vote of persons who had not filled in a ballot paper in accordance with
s 240. This meant that s 329A did not breach any implied freedom of political communication.
The High Court applied the same reasoning in Muldowney v South Australia23 in holding that ss 76
and 126 of the Electoral Act 1985 (SA) were valid. These cases show that compulsory voting, or at
least requiring a person to attend the polling booth and deposit a voting paper in the ballot box, is
consistent with ss 7 and 24 of the Constitution. The High Court has given the Parliament consid­
erable leeway to shape and determine the electoral process. While ss 7 and 24 require that a ‘choice’
be made, it is up to the Parliament to determine how that ‘choice’ is to be made, including whether
such a ‘choice’ is to be compulsory. Moreover, the Parliament may protect the system it has created
by limiting the information provided to electors about how electors may cast an effective vote.
3. A right to vote
Do the words ‘directly chosen by the people’ in ss 7 and 24 confer a right to vote? Even though the
High Court held in R v Pearson; Ex parte Sipka24 that such a right is not conferred by s 41 of the
Constitution, this does not preclude ss 7 and 24 supporting an implied right to vote. After all, these
provisions require a ‘choice’ by the ‘people’. McHugh J stated in Australian Capital Television Pty
Ltd v Commonwealth:25 ‘[T]he proper conclusion to be drawn from the terms of ss 7 and 24 of the
Constitution is that the people of Australia have constitutional rights of freedom of participation,
association and communication in relation to federal elections.’ If these are guaranteed, how could
the right to vote not be?
Decisions on ss 7 and 24 have not addressed the question whether each Australian is vested with a
constitutionally guaranteed right to vote. The issue has never arisen for decision in this way. It
might arise if a disenfranchised person, say a person excluded under s. 93(8)(b) of the
21 In the aftermath of this decision, in its Report of the Inquiry into all Aspects of the Conduct of the 1996 Federal
Election and Matters Related Thereto (AGPS, June 1997, p. 32), the Joint Standing Committee on Electoral Matters
recommended that s. 329A be repealed and that s. 240 be amended to include the words ‘consecutive numbers,
without the repetition of any number’. These recommendations are implemented by the Electoral and Referendum
Amendment Act 1998 (Cth).
22 (1996) 186 CLR 302 at 317 per Brennan CJ.
23 (1996) 186 CLR 352. The author appeared in this case as counsel for the plaintiff.
24 (1983) 152 CLR 254.
25 (1992) 177 CLR 106 at 227.
32
A Legal Perspective On Electoral Research
Commonwealth Electoral Act 1918 (Cth) because he or she is serving a sentence of five years or
longer for an offence against the law of the Commonwealth or of a state or territory,26 were to seek a
declaration as to his or her entitlement to cast a vote.
Obiter dicta in several cases has approached the question from the converse, but equivalent,
perspective of whether ss 7 and 24 limit the Commonwealth’s power to restrict the federal franchise
as provided for by s 93 of the Commonwealth Electoral Act. The universal adult franchise
recognised by several members of the High Court in obiter as entrenched by ss 7 and 24 may make
the question of a separate implied right to vote obsolete. Whether a personal right to vote, or at
least an immunity from legislative and executive interference with that right, can be implied from
the Constitution may be irrelevant when the Commonwealth lacks the power to legislate other than
for universal adult suffrage.
In Attorney-General (Cth); Ex rel McKinlay v Commonwealth27 McTiernan and Jacobs JJ stated:
the long established universal adult suffrage may now be recognized as a fact and as a result
it is doubtful whether, subject to the particular provision in s 30, anything less than this
could now be described as a choice by the people.28
In McGinty v Western Australia29 Toohey J argued that ‘according to today’s standards, a system
which denied universal adult franchise would fall short of a basic requirement of representative
democracy’. Gaudron30 and Gummow JJ31 also supported the notion that universal adult suffrage is
now entrenched in the Australian Constitution. Only Dawson J rejected this.32 In Langer, McHugh J
supported entrenchment of the franchise by stating that: ‘[I]t would not now be possible to find
that the members of the House of Representatives were “chosen by the people” if women were
excluded from voting or if electors had to have property qualifications before they could vote’. 33
According to these judges, the right to vote of, say, Australian women or indigenous peoples could
not now be abrogated. This would be inconsistent with the requirement that the federal Parliament
is to be ‘directly chosen by the people’ (emphasis added). This conclusion depends upon a view of
the Constitution as a evolving document, one that embraces a very different notion of ‘the people’ at
the end of the twentieth century than at the beginning. After all, the uniform federal franchise, as
enacted by the Commonwealth Franchise Act 1902 (Cth), extended the vote to women, but, in s 4,
denied it to any ‘aboriginal native of Australia’.
4. Equality of voting power
Even though ss 7 and 24 might constitutionally entrench universal adult suffrage, they have not
been interpreted to entrench equality in voting power, or the idea of ‘one-vote, one-value’. This
issue arose in McKinlay, where it was argued that the election of 18 May 1974 for the House of
Representatives had not been in accordance with s 24 because the electorates had not been of equal
population size. Provisions of the Commonwealth Electoral Act and the Representation Act 1905
(Cth) were accordingly said to be invalid. The High Court, with Murphy J dissenting, held that s 24
of the Commonwealth Constitution does not imply a constitutional requirement of as near as
practicable equal numbers of people per electoral division for the House of Representatives.
26 See Orr, G, ‘Ballotless and Behind Bars: The Denial of the Franchise to Prisoners’ (1998) 2 Federal Law Review 55.
27 (1975) 135 CLR 1.
28 Ibid at 36.
29 (1996) 186 CLR 140 at 201.
30 Ibid at 221-222.
31 Ibid at 287.
32 Ibid at 183.
33 (1996) 186 CLR 302 at 342. Cf McGinty v Western Australia (1996) 186 CLR 140 at 243 per McHugh J.
33
A Legal Perspective On Electoral Research
In dissent, Murphy J found that s 24 requires as a ‘standard of equality the alternatives of equal
numbers of people and equal numbers of electors’.34 He based this conclusion upon an analysis of
the United States authorities, particularly Wesberry v Sanders.35 On the other hand, the majority in
McKinlay held that a requirement of equality of voting power could not be implied from ss 7 and
24. Stephen J acknowledged that representative democracy ‘has finite limits and in a particular
instance there may be absent some quality which is regarded as so essential to representative
democracy as to place that instance outside those limits altogether.’36 Thus, an electoral system
might lack some quality such that any representatives elected under the system would not have
been ‘directly chosen by the people’ pursuant to ss 7 and 24 of the Constitution. However, one
vote, one value, was not such a ‘quality’. In finding this, the majority rejected any analogy with the
words ‘directly chosen by the people’ in article I of the United States Constitution, or with the
interpretation of those words by the United States Supreme Court.
The majority did not totally reject the notion that s 24 requires some form of equality. Obiter dicta
suggested that, in some situations, there might be such a degree of malapportionment between
electoral divisions as to bring into question whether the Parliament had been ‘directly chosen by the
people’. For example, Mason J stated:
It is perhaps conceivable that variations in the numbers of electors or people in single
member electorates could become so grossly disproportionate as to raise a question
whether an election held on boundaries so drawn would produce a House of
Representatives composed of members directly chosen by the people of the
Commonwealth.37
McTiernan and Jacobs JJ, in a joint judgment,38 and Stephen J, in a separate judgment,39 voiced
similar views.
Since McKinlay, electoral malapportionment has continued to be an important political issue at the
State level. In Western Australia, the number of voters per electoral district differs markedly
between districts. At the 1993 Western Australian election, the most populous electorate in the
Legislative Assembly was Wanneroo, which had 26 580 enrolled voters, while the least populous
electorate was Ashburton, which had only 9135 enrolled voters. There were even greater
differences (up to 376%) between the quotients for members of the Legislative Council.
The plaintiffs in McGinty challenged the legislation giving rise to these differences in voting power.
It was argued that a system of representative government was created by the Commonwealth
Constitution, which was breached by the electoral boundaries in Western Australia. Dawson,
McHugh and Gummow JJ reaffirmed the decision in McKinlay and found, in the words of Dawson
J, that ‘electorates of equal numerical size are not a necessary characteristic of representative
government’.40 The six-judge Court was not decisive on whether the Commonwealth Constitution
contains a guarantee of voter equality as Brennan CJ expressed no final opinion and Toohey and
Gaudron JJ dissented. However, all six judges held that, even if there were such a requirement at
the Commonwealth level, it could not extend to State electoral systems. It was found that in
significant ways the Commonwealth Constitution is inconsistent with any notion of equality in
voting in State elections. For example, under s 128 of the Constitution the aggregate of votes of
persons living in one of the less populous States is given the same weight as the aggregate of votes
of persons living in one of the more populous States for the purposes of achieving a majority of
votes in a majority of States.
34 (1975) 135 CLR 1 at 70.
35 376 US 1 (1964).
36 (1975) 135 CLR 1 at 57.
37 (1975) 135 CLR 1 at 61.
38 Ibid at 36-37.
39 Ibid at 57.
40 (1996) 186 CLR 140 at 185.
34
A Legal Perspective On Electoral Research
In 1978, s 73(2) was inserted into the Western Australian Constitution. Section 73(2)(c) entrenches
Western Australian laws, including the Western Australian Constitution itself, against a Bill that
‘expressly or impliedly provides that the Legislative Council or the Legislative Assembly shall be
composed of members other than members chosen directly by the people’. In McGinty, it was also
argued that s 743(2)(c) supports a requirement of equality of voting power in Western Australian
State elections. A majority, with Toohey and Gaudron JJ dissenting, held that there was no such
implication. Brennan CJ pointed out that to discover such an implication ‘would be to find a
legislative intention destructive of the means by which the enacting Parliament was elected’.41
The majority gave little attention to the further question of whether the Western Australian electoral
districts might be unconstitutional due to malapportionment. Although such a question could not
arise in regard to the Commonwealth Constitution, which had nothing to say about equality of
voting power in State elections, it could arise in regard to the Western Australian Constitution if
McKinlay were to operate by parity of reasoning. This issue was not addressed explicitly by
Brennan CJ, McHugh or Gummow JJ. Dawson J briefly stated that the ‘extreme situations’
considered in McKinlay were ‘markedly different from that which exists under the relevant Western
Australian legislation’.42 It was on this point that Toohey and Gaudron JJ dissented. They found that
the level of malapportionment meant that ‘persons elected under a system involving significant
disparity in voting value, could not … now be described as “chosen by the people” ’.43
New Directions And Challenges In Electoral Law Research
What are the possible future focal points of electoral law research? What do recent developments
suggest as appropriate topics for electoral law research?
Recent events suggest the need for new directions in electoral law research. The 1999 republic
referendum exposed the underbelly of Australian democracy. It has revealed deep, entrenched,
problems in our system of government, and the need for significant structural reform. Such reform
would go beyond a change to our national symbols and head of state. It would seek to rebuild the
connection between Australians and their government. Electoral law has a large role to play in any
such reform process.
Issues for consideration include whether the mechanism for the amendment of the federal
Constitution in s 128 ought to be altered. Bills for citizens’ initiated referenda have also been
introduced across the Australian parliaments. The legal aspects of any such changes must be
carefully examined and scrutinised. If they are to succeed, they must be developed within our
constitutional structures.44
41 Ibid at 178.
42 Ibid at 189.
43 Ibid at 222 per Gaudron J.
44 See Williams, G and Chin, G, ‘Australian Experiments with Community Initiated Referendum: CIR for the ACT?’
(1998) 7 Griffith Law Review 274; Williams, G and Chin, G, ‘The Failure of CIR Proposals in Australia: New
Directions for Popular Participation?’ (2000) 35 Australian Journal of Political Science 27.
35
3
What’s So Special About Specials?
Alan McRobie
Inevitably, and predictably, politicians, administrators and academics have a different perspective of
the issues involved in establishing and maintaining fair, free, and competitive elections; from time
to time, friction is likely to occur. The interface between these three elements is explored in this
paper through two New Zealand case studies relating to the casting and counting of special votes.
The role played by politicians is a key one because they are the ones who control the final shape
and content of a country’s or state’s electoral legislation. The 1992 and 1993 referendums which
resulted in New Zealand abandoning its long-standing first-past-the-post (FPTP) electoral system in
favour of the mixed member proportional (MMP) system recommended by a royal commission in
1986 is a rare example of politicians losing control of an electoral system. Since MMP was endorsed
by a majority (53.9 per cent to 46.1 per cent) of voters in 1993, the country’s politicians have been
endeavouring to regain the initiative in this area. The most recent example is the attempt by an
opposition MP to enact an amendment to the Electoral Act to declare any MP who resigns from the
party under whose banner they were elected to parliament, to have also resigned from the
parliament itself.1
Electoral administrators on the other hand have, as their primary goal, the conduct of a ‘clean’ and
smooth running election, one that takes place within the framework of the law as established by the
parliament. Unusually, New Zealand’s electoral administration is divided between three different
authorities - or four if the Representation Commission, the quasi-judicial statutory body charged
with carrying out electoral redistributions after each quinquennial census, is included. The Chief
Electoral Office (a division of the Department of Justice) is responsible for the detailed organization
and administration of the actual elections; it has a permanent staff of 12 but employs some 15 000
temporary staff to man polling paces on election day. The Electoral Enrolment Centre (a division of
New Zealand Post whose Chief Executive is, by designation under s.21 of the Electoral Act 1993,
also Chief Registrar of Electors) is responsible for maintaining, updating and revising the electoral
rolls over a three year cycle, while the statutorily independent Electoral Commission has responsi­
bility for registering political parties and their logos, supervising parties’ compliance with the
financial disclosure and return of election expenses provisions of the Electoral Act; promoting
public awareness of electoral matters, and providing advice on matters referred to it by the Minister
or House of Representatives. Of necessity, these three bodies work closely together.
Lying beyond the practicing politicians and administrators are the academics involved in electoral
research. Their primary objective is to trawl a fertile ocean in search of interesting and, at times,
intriguing aspects of electoral systems that may warrant further study, to identify shortcomings and
propose possible solutions for public debate. Their overriding objective may be summarized as
making the electoral system as fair as possible for both political aspirants and electors.
Special Voting In New Zealand
In order to cast a valid vote in a New Zealand election, those who are qualified to vote must first
register as electors. It is compulsory for all persons aged 18 years and over, who are New Zealand
citizens or permanent residents, and who have at some stage lived in the country continuously for
1 Following the 1999 election the Labour-Alliance coalition government has introduced a bill to achieve this objective.
36
What’s So Special About Specials?
at least a year and in an electoral district for at least one month to do so.2 Registration is not,
however, permanent as electors are required to re-register prior to each general election.3 But,
unlike Australia, although registration as an elector is compulsory, voting is not.
New Zealand has fairly liberal laws governing the casting of votes by those who, for a variety of
reasons, are unable to attend a polling place on election day to cast their vote. Such votes are
categorised as ‘Special Votes’ (the equivalent of declaration voting in Australia or the absent voter’s
ballot in the USA) and are available to any qualified person whose name does not appear on the
printed electoral roll, who is away from his or her electorate on polling day, is prevented from
visiting a polling place ‘by reason of illness, infirmity, pregnancy or recent childbirth,’ or whose
attendance would involve ‘hardship or serious inconvenience.’4 The practical effect of the special
voting provisions is to allow all registered electors to cast a vote at any polling place in New
Zealand, or at a designated polling place (usually an Embassy, High Commission office, or Consular
office) anywhere in the world.
These ‘Special Votes’ should not be confused with two other forms of special votes, hospital votes
and special votes cast in a district before polling day, both of which are also covered by the electoral
regulations. Where a hospital is situated within an electoral district, or where it is located in an
urban area represented by several electorates, hospital votes are, in effect, ordinary votes cast by
patients who are confined to bed. Special votes cast in a district before polling day, which may,
more aptly, be described as ‘early votes’, are available to electors who know that they will be absent
from their electoral district on election day. Although early voters are required to complete a
statutory declaration5 stating the reason for their application, their entitlement to vote is checked
against the printed electoral roll for the district. These categories of votes are treated as ordinary
votes and included in the election night preliminary count because their validity has already been
determined.
This paper focuses on the special votes whose validity must be confirmed after polling day before
they can be included in the official count.6 In the nine general elections between 1972 and 1996
special votes have, on average, accounted for just under 12 per cent of all votes cast (see Table 3.1).
Yet, when the outcome of the 950 individual electoral contests in these elections is examined, only
68 (7.2%) recorded election night majorities of fewer than 500 votes, and in only five of these was
the election night result reversed as a result of the inclusion of special votes. Further, only in one
case where the preliminary result was reversed (Waitaki, 1993), did the final result decide which
party would be able to form a government (see Table 3.2). At best, the impact of special votes is
minimal although, as the 1993 reversed result shows, on very rare occasions they may determine
the eventual formation of a government.
Interest in special votes was first aroused in 1969 when 2622 (24.1%) of the 10 866 votes cast in
the Northern Maori electorate were ruled out as informal by the returning officer for a variety of
reasons. In stark contrast, the average number of informal votes cast in the other 83 electorates was
109.7 As a direct consequence of this experience detailed analyses of the decisions taken in respect
of special votes have been included in all official election results since 1972. These data set out the
number of special votes allowed in each electorate, along with the number disallowed and reasons
2 Stat NZ (Statutes of New Zealand), 1993, no 87, The Electoral Act, 1993, ss.74 & 82.
3 Ibid., s.83(3).
4 Ibid., s. 61; Electoral Regulations, 1996, Parts II–IV & First Schedule, Form 18.
5 Ibid., Regulation 24 & First Schedule, Form 16.
6 Ten days are allowed from the close of the poll on election day for special votes to be received by Returning Officers.
Thus, the official count cannot be completed until after 7.00pm on the second Tuesday after the election.
7 Informal votes in the 80 General (called ‘European’ prior to 1975) electorates averaged 101. In the three remaining
Maori electorates the average was 324.
37
What’s So Special About Specials?
for their rejection.8 Since then the fate of special votes cast in each election, and the grounds on
which returning officers exclude perhaps as many as one-quarter of those cast, has been much more
transparent.
Two Case Studies
(1) Special voting in Maori electorates: the tangata whenua vote. As Table 3.1b shows, the
proportion of special votes cast in the Maori electorates since 1972 is approximately four times
greater than for General electorates. For the most part this is because the number of polling places
that can accept ordinary votes from electors on a Maori roll (that is, those who attend a polling
place in their electorate and are able to have their right to vote verified by reference to the roll for
that district) is far fewer than for General electorates. A 1977 analysis of the frequency and distri­
bution of polling places where Maori could cast an ordinary vote highlighted the fact that the ‘cost
of voting’ was significantly greater for Maori voters than for non-Maori. Because ordinary Maori
electorate polling places were far more thinly spread than General polling places - despite servicing
the same geographic area - Maori electors were often required to travel far greater distances than
non-Maori electors in order to cast an ordinary vote.9 The alternative was to visit one of the General
electorate polling places and cast a special vote - a time consuming exercise which involved both an
accurate knowledge of the electorate where the elector was registered and the accurate completion
of a special vote declaration. For individuals with a lower level of educational attainment (and many
Maori belonged to the category) this was a somewhat forbidding exercise.
Over the past two decades the number of ordinary Maori electorate polling places has greatly
expanded. During its post-election inquiry following the 1996 general election, parliament’s
electoral officials told the Electoral Law Committee that the number of polling places with provision
for Maori to cast an ordinary vote had been increased by more than 50 per cent (from 534 to 829)
between 1993 and 1996 and that a further 298 were likely to be provided for the 1999 election.
The tangata whenua vote, first introduced in 1987, has contributed significantly to this
improvement because it greatly simplified the voting procedure for New Zealand’s indigenous
people.10 All polling places in General electorates now have a copy of the Maori Reference Roll for
the Maori electoral district where they are located, and this is supposed to be checked whenever an
elector applies for a voting paper for a Maori electorate. Where the applicant’s name appears on this
roll, a simplified declaration is completed. In each of the last four elections (1987–1996), the
tangata whenua vote has accounted for between one-fifth and one-quarter of all votes, and over
one-half of the special votes, cast in the Maori electorates. Although the tangata whenua vote
declarations are subject to later verification, the facility which allows Maori voting entitlements to
be verified at a polling place has greatly reduced the number of Maori special votes which are
disallowed. In the five elections before the tangata whenua vote was introduced, disallowed special
votes in Maori electorates averaged 41.8 per cent, but in the four elections following its
introduction this figure has dropped to an average of 19.3 per cent.
Notwithstanding the undoubted gains made over recent years, parliament’s Electoral Law
Committee noted in its review of the 1996 election that the tangata whenua voting process is still
unsatisfactory:
8 The principal reasons for rejection are that the voter is not registered on the roll of the district where he or she has
voted (or has not registered at all), that the special vote declaration has not been signed or witnessed (or properly
witnessed), that the special vote reached the returning officer after the statutory deadline for receipt, or that no
ground for seeking a special vote was stated.
9 Alan McRobie, ‘Ethnic Representation: the New Zealand Experience’, in Stephen Levine (ed), Politics in New
Zealand: A Reader, Sydney, George Allen & Unwin, 1978, pp 278–79.
10 See Electoral Regulations , 1996, regulation 22.
38
What’s So Special About Specials?
Voters must still queue in the special voting queue, they must deposit their voting papers in
a special votes box and the issuing officer is required to enter the names of tangata whenua
voters on a list of special voters. Casting a tangata whenua vote takes longer than casting an
ordinary vote. This is unfair to those tangata whenua voters who are on the electoral roll
and who are voting in their electorate.11
Although Maori electorates encompass a much larger geographic area than General electorates, a
fact that undoubtedly creates additional problems for returning officers, it should not be beyond
the bounds of administrators and politicians to devise an electoral structure that will preserve the
secrecy of each vote cast, while at the same time ensuring that the ease with which a vote is applied
for and cast is the same for all electors. Ideally, all electors on the Maori roll should be able to
attend any polling place within the boundaries of their electoral district and cast an ordinary vote.
Ballot secrecy could be preserved by requiring all ballot boxes containing Maori electorate votes
which have been cast in polling places where there is no dedicated Maori electorate booth, to be
taken to a central point and brought together before being counted. According to the select
committee, the manual procedures used during the scrutiny and counting phase of elections are
slow and labour intensive. Since this greater difficulty many Maori have in casting their vote was
identified two decades ago, the case for the an urgent and major overhaul of the administrative
process is compelling.
(2) Votes that may not count. One of the more curious aspects of each New Zealand election is
the number of votes cast which are subsequently disallowed. Since data was first published, the
number of special votes which have been disallowed in each election has averaged just over 49 000
(Table 3.1c). In a small number of instances, votes have been rejected because the special vote
declaration has not been signed or witnessed, or the vote has not reached the electorate’s returning
officer within the time-frame provided for in the Electoral Act. Overwhelmingly, however, the
reason special votes are disallowed is that the electors in question are not registered on the electoral
roll of the district where they reside (Table 3.3). It is, indeed, curious to say the least that so many
people fail to register as electors - despite repeated and intensive campaigns mounted by the
Electoral Enrolment Centre before each election12 - but still go to a polling place on election day
and cast a vote. One cannot help but wonder if the same people have their votes disallowed election
after election, because electors whose votes have been disallowed are never advised of this.
Like a number of other jurisdictions New Zealand uses finite geographic units, called electoral
districts or constituencies, as the basic building blocks of parliamentary representation. Eligible
electors are required to enrol in the electoral district in which they normally reside. Registration has
been compulsory since 1924 (1956 for Maori); the penalty for failing to do so is a fine on summary
conviction of up to $100 for the first offence and up to $200 for each subsequent offence. There is,
however, no record of anyone ever having been prosecuted for failing to register. Statistics on the
level of registration have only been kept since 1987 and these show that approximately 92 per cent
of the estimated number of eligible electors actually register.13
Undoubtedly, electors who fail to register have only themselves to blame if the votes they cast are
subsequently disallowed. One can only guess at the reasons why people fail to register but still go to
a polling place on election day where they have to cast a special vote. One possible reason is that
they have registered (or believe that they have registered) at some point in the past and - despite
11 AJHR, 1996–99, I.17A, p 79.
12 New electoral rolls are compiled before each election. Anyone whose name appears on a roll used in the previous
election but who does not re-register is transferred to a dormant file which can be used to verify the validity of any
special vote application provided the applicant is still residing at the address recorded on the previous roll.
13 The final registration figures when enrolments closed on the day before election day were: 1987: 92.4%; 1990,
91.7%; 1993, 92.6%; 1996, 91.6% and 1999, 91.1%.
39
What’s So Special About Specials?
personally addressed roll update and revision requests sent out at regular intervals14 - believe that
no further action is necessary. There is, however, some anecdotal evidence extending over a
number of years which suggests that a significant number of registered electors apply for and cast
special votes for an electorate other than the one in which they are registered.
Between 1853 and 1993 New Zealand’s simple plurality, FPTP, electoral system required each
electoral district to elect one person to represent it in the parliament. An elector could only cast a
valid vote for a candidate standing in the electorate where he or she lived and was registered. The
votes of electors who were not enrolled, or who cast their votes for a candidate in an electorate
other than where they were registered, were, quite properly, disallowed.
With the adoption of the MMP electoral system, electors are now able to cast two votes - one for an
electorate MP and one for his or her preferred political party. Because the party vote is aggregated
across the whole country and determines the final composition of the new parliament, many
electors see the party vote as very important. Political parties also acknowledge the primary
importance of the party vote and now campaign actively to increase their share of the party vote.15
Where an elector votes for a candidate standing for election in an electorate other than the one
where he or she is registered, both votes — constituency and party — are disallowed even if the
elector is correctly and legally registered in another electoral district. Thus, the adoption of MMP,
and the wording of the Electoral Act 1993, has resulted in some party votes cast by people who are
properly registered being disallowed. This situation applies only to electors who need to cast special
votes. It does not apply to electors who vote at a polling place within their electoral district because
their eligibility is able to be checked against the district’s printed roll. As long as their name appears
on the printed roll they can cast an ordinary vote sure in the knowledge that it will count.
Because MMP is, unashamedly, an electoral system that places political parties at the centre of the
political process, party votes cast by electors who are properly registered should be counted, even if
the constituency vote is disallowed because it has been cast for a candidate standing in an electorate
other than the one in which the elector is registered. The party vote is, in effect, cast for a 68th
(multi-member) electorate covering the whole of New Zealand.16 Provided an elector who casts a
special vote is correctly enrolled, the party vote should not be disallowed because he or she makes a
mistake which results in the constituency vote being cast for the wrong electorate.
Electors who are away from their home electorate on election day - or whose name, for some reason
has been omitted from the printed roll17 - must cast a special vote. If an elector requesting a special
vote gives the poll clerk an incorrect electorate name and is, as a consequence, given a ballot paper
for the wrong electorate, both votes - constituency and party - are disallowed because he or she is
not registered on the electoral roll for which the votes are cast. Given the number of votes rejected
because electors are ostensibly ‘not enrolled’, this suggests that disallowed party votes could have a
significant bearing on the final outcome of an MMP election.
14 National roll updates are carried out prior to local authority elections (normally the year before each general
election) and full roll revisions take place in each election year. The roll update programme does not require
registered electors to take any action if the information shown on the card is still accurate but the roll revision
programme requires all registered electors to return their registration cards, duly signed and, where necessary,
corrected. It should be noted, however, that only those who are already registered receive the update and revision
cards. Given New Zealanders’ high population mobility, coupled with the fact that New Zealand Post now only
redirects mail (on request) for three months as opposed to its former practice of six months, many currently
registered electors may not receive the update or revision cards. Failure to respond to a revision request results in
an individual’s name being transferred to the dormant file.
15 The major parties (Labour and National) were reluctant to recognize this in the lead-up to the first MMP election in
1996.
16 Increased from 66 in 1999 as a consequence of the creation of two additional electoral districts in the 1998
electoral redistribution.
17 They may, for example, have registered after writ day when the roll is closed off for printing.
40
What’s So Special About Specials?
When this issue was first raised in the public arena towards the end of 1995, the response of
politicians was interesting to say the least! In December 1995 (a year before the first MMP election)
Tony Ryall, then chair of the Electoral Law Committee, defended the special voting arrangements
contained in the 1993 Electoral Act. He pointed out that the Act’s provisions were the same as
those applying to previous elections under FPTP:
This is not a change to what we’ve ever had before to our open and honest voting and
registration system. … It’s not a question of being on the wrong roll and when (a special
vote is opened the Returning Officer relies) on the declaration. If the declaration is wrong
then our system has always been that the vote is not counted.
When challenged as to whether this was not a spurious argument ‘if it affects seats’, he replied that
it did not affect seats because the votes did not count.18
A few days later the Minister of Justice, Hon Doug Graham, was asked, in parliament, whether the
government intended that registered electors who cast a special vote in the wrong electorate would
have both their constituency and party votes disallowed. He replied:
Under the Electoral Act 1993 voters must vote in the correct electoral district in order to
cast a valid party vote as well as a valid electorate vote. The Government does not favour
amending the law to allow party votes cast by registered electors in the wrong electoral
district to be counted. Such a change would cause unacceptable delays in the declaration of
the final result because of the time taken to check special votes to see whether the voter is
registered in another electoral district.
In response to a supplementary question, Graham, replied:
… In other jurisdictions such as the United Kingdom, one is told which church one can
vote at, and if one does not vote there one misses out entirely. We do at least have special
votes, and 43 000 people at the last general election cast invalid, disallowed special votes. It
is not because of some clerical error; it is because they have got themselves totally muddled
up.
And, to another supplementary question:
… I hasten to advise the House that the 65 registrars in the constituency seats would take,
on average, nine working-days each to go through the 43 000 expected disallowed special
votes to determine which electorate, wherever it is in the country, those people might be in.
I do not think that is a reasonable thing to do.19
These dismissive comments suggested that the politicians had little genuine interest in the fairness
of the electoral system. If Ryall was correct in his claim that most of the disallowed votes were
rejected because the voters concerned were not on any roll, the endemic nature of the problem begs
the question of why successive governments have not made greater attempts to eliminate it. It is
strange, to say the least, that politicians appear to accept that, at each election, more than 40 000
people who have not enrolled as electors still take the trouble to go to a polling place and cast a
vote.
The nature and extent of the problem was amply illustrated by data analysed by the Electoral
Enrolment Centre after the 1996 election. The results of its analysis, the first occasion that such an
exercise had ever been carried out, are set out in Table 3.4 which shows that nearly 40 per cent of
those whose votes were ruled out on the grounds of non-registration were, in fact, correctly
18 Radio New Zealand Morning Report, 1 December 1995.
19 NZPD (New Zealand Parliamentary Debates), 1995, vol 552, p 10432.
41
What’s So Special About Specials?
registered but, on the surface, appear to have given a poll clerk incorrect or insufficient
information. Since the votes of nearly 20 000 electors were disallowed because they were deemed
to be ‘not enrolled’ (even though they were legally enrolled in another electorate district), nearly
20 000 party votes were not able to contribute in any way to the final composition of New
Zealand’s 45th parliament. In the 1996 election just under 2.1 million electors cast valid votes.
Under MMP rules all parties reaching the 5 percent threshold effectively gained one seat for every
17 338 party votes won. Thus, in this election, the nearly 20 000 disallowed votes could have
affected the allocation of the 120th parliamentary seat. For a party whose party vote falls just short
of the threshold, party votes disallowed under this rule could be critical to its survival.
Even if only a small percentage of the rejected votes are cast by people who are otherwise correctly
registered, this should be cause for concern. Both Ryall’s and Graham’s 1995 responses fail to
acknowledge that the new MMP electoral system has shifted political parties from the periphery to
the centre of the electoral process. This is why the party vote has assumed such importance. The
basic point is that where an elector is correctly and properly registered, the party vote should count
even if the constituency vote is disallowed solely because the voter has inadvertently cast it for an
electorate where he or she is not registered or entitled to be registered.
Just how serious is this problem? In 1996, the upper quintile of electorates recording the highest
percentage of votes disallowed because electors were deemed to be not enrolled comprised all five
Maori electorates and eight of the 14 electorates in urban Auckland. In the last FPTP election in
1993 the same pattern was evident. An analysis of the ethnic and socio-economic characteristics of
these electorates indicate that factors such as ethnicity, education, and employment are significant
factors in determining which groups are most likely to have their votes disallowed.20 Each of these
electorates contain large numbers of Pacific Islanders; all have above average numbers of Asian and
other minority groups; most have Maori population numbers that place them within one of the top
three quintiles. Other socio-economic measures, educational attainment, occupational status and
unemployment characteristics reveal that most have low levels of education, an over-abundance of
people with limited occupational skills and high proportions on unemployment benefits. Although
data on population mobility is not available, the socio-economic characteristics of these electorates
make it highly probable that most have high numbers of transients.21
The current Elections New Zealand handbook, A guide for polling day staff, instructs Deputy
Returning Officers22 that they must first check in the Index of Places and Streets23 to establish the
correct electoral district, and then ascertain whether the applicant is on the General roll or the
Maori roll before issuing a special vote. Ideally, if deputy returning officers and poll clerks adhered
scrupulously to this instruction, and faithfully checked the address and roll of all special vote
applicants to make sure that they were given ballot papers for the correct electorate, the problem of
electors casting a vote for an incorrect electorate would not arise and we would know that all votes
rejected as ‘not enrolled’ were just that. Unfortunately, while this has been a requirement since at
least the 1995 referendum, 24 the pressures on poll clerks on election day appear not to be
universally fulfilled.25 Whether such a failure on the part of an electoral official would save a party
20 See ‘When votes may go to waste’, National Business Review, (Wellington, Fourth Estate), 8 Mar 1996, pp 24–26.
21 Many people may not realise that if they do not complete New Zealand Post’s postal redirection order form when
they move residence, the Electoral Enrolment Centre will not be notified that they have shifted. Others may be
aware of this but do not lodge redirection orders for other reasons.
22 The person designated as being in charge of a polling place.
23 A manual which lists all streets in New Zealand and the electorate for each.
24 A referendum held under the provisions of the Citizens Initiated Referenda Act 1993.
25 Another worrying feature is the competency of some poll clerks employed on election day. In 1998 the Electoral
Law Committee reported that ‘…some clerks were not able to correctly follow the alphabetical order of the roll.’
See AJHR, 1998, I.17A, p 50.
42
What’s So Special About Specials?
vote cast by a voter who was properly registered but who had cast his or her constituency vote for
the wrong electorate has yet to be tested in court.26
Following the 1996 election, parliament’s Electoral Law Committee considered this matter.
Electoral officials told MPs that the current manual procedures used for checking the electoral
entitlement of special voters were very time consuming. Further, if all rolls had to be checked to
determine whether a voter was registered on the roll of the electorate where his or her constituency
vote was cast, unacceptable delays in announcing the final election result were bound to occur. MPs
concluded that enhanced technology was likely to provide a solution, and the select committee
recommended that electoral officials should continue to investigate developing a solution that was
administratively feasible so that it might be applied to future general elections.27 Despite this, no
progress appears to have been made to date. The government’s response to the select committee’s
recommendation was that, while the investigations sought were under way, ‘…the government
considers it unlikely that this matter can be resolved and necessary legislative amendments made in
time for the next General election.’28
Conclusion
Under the MMP electoral system the nationwide party vote is all-important in determining the final
composition of the parliament, and this was clearly recognized by all parties in 1999 as they
campaigned assiduously to increase their share of the party vote. Thus a party vote cast by an
elector who is registered but who, for whatever reason, applies for and casts a special vote in an
electorate other than the one he or she is registered in, may be highly significant in determining the
outcome of an election.
An analysis of the aggregate votes cast for parties in the 1993 and 1996 elections reveals that the
maximum change in their respective shares of the total vote was only 0.2 per cent, insufficient to
affect the final allocation of parliamentary seats. But while the final result of an election conducted
under FPTP rules has depended on the distribution of special votes only once in the past 30 years,
this may not necessarily be the case under MMP. A party which garners 5 per cent of the aggregate
party vote wins at least six parliamentary seats but one which gains only 4.99 per cent of the party
vote gets no reward in terms of parliamentary seats unless, of course, one or more of its candidates
wins a constituency seat, in which case the 5 per cent threshold is waived. Special votes may,
therefore, play a more significant role in the future, especially where a party fails to reach the
threshold by a few hundred votes.29
Despite their relatively small numbers and minor impact, special votes are important. If modern
elections are a key mechanism linking governors and governed, all voters need to be assured that
their individual vote will contribute to the final result. Whenever votes are disallowed because an
26 The Electoral Act 1993, s.178(a)(ii) contains a proviso that ‘…no ballot paper or part of a ballot paper shall be
rejected as informal by reason only of some error or omission on the part of an official, if the Returning Officer is
satisfied that the voter was qualified to vote at the election:’ (emphasis added).
27 AJHR, op.cit. p 59–60.
28 Quoted in correspondence, Phil Whelan, Chief Electoral Officer, to author, 19 Nov 1999.
29 In the election night count in 1999 the Green Party had 4.87 per cent of the party vote. The subsequent inclusion
of valid special votes lifted its share of the party vote to 5.16 per cent, sufficient for it to win parliamentary
representation. The ultimate success of the Green Party had the effect of transforming what appeared to be a
majority coalition government (Labour-Alliance) on election night into a minority coalition government. Had the
Green Party not won a constituency seat, and had its party vote increased but not sufficiently for it to surmount the
threshold, disallowed special votes where voters were otherwise correctly registered might have been critical to the
outcome. (The inclusion of special votes in the Coromandel electorate count resulted in a 114-vote majority for the
National Party’s candidate being eliminated and the Green Party’s candidate took the seat with a majority of 250
votes.)
43
What’s So Special About Specials?
elector has cast a constituency vote for the wrong electorate, even though he or she may be
correctly enrolled in another electoral district, the integrity of the election is compromised to some
degree.
A number of reasons for the apparent reluctance to address this problem may be deduced from the
public statements of politicians and administrators: it is not the responsibility of the State to ensure
that electors cast their votes in the correct electoral district because they should know which roll
they are registered on; the time taken to search other electorate rolls, the dormant file, and the
master roll used at the previous general election is too time consuming; and that insufficient
resources are allocated to the Chief Electoral Office to develop more sophisticated verification
strategies. Given modern computer technology, it is inconceivable that a strategy or strategies
cannot be devised that will, where required, allow the prompt, swift, and accurate checking of the
electoral roll to ascertain whether a person who cast a special voting right is properly registered as
an elector.
Failure to ensure that such votes are counted is a serious anomaly, but it is one which could be
rectified relatively easily by amending the Electoral Act 1993 to ensure that, even where electors are
properly registered but cast their constituency vote for an incorrect electorate, their party vote will
still be counted. The information provided by those casting special votes - full name, current
address, occupation, telephone number, and date of birth (optional) - should provide registrars of
electors or returning officers with sufficient information to determine quickly whether or not an
applicant is properly registered.30 At present, however, returning officers do not have access to the
computerised master roll of electors (although candidates and political parties are able to obtain
computer tapes containing elector details31), so all queries must be directed to registrars of electors.
Providing returning officers with computer access to the master electoral rolls would seem to be a
logical step towards validating the claims of special voters. If returning officers were able to access
the master rolls directly, the declarations and votes of persons who, prima facie, were properly
registered could be forwarded directly to the returning officer of the electorate concerned for final
verification and, where validated, inclusion in the count.
The case studies presented here illustrate the convergence of a number of different interests. The
reluctance of New Zealand’s politicians to grasp the nettle and deal with the issues involved as a
matter of high priority, appears to be rooted in self-interest. They appear to assume that all citizens
are as interested in politics as they are. They seem to believe that because they know the boundaries
of their electoral district and the rules governing voting procedure, everyone else does. This,
however, is far from the case. Politicians must take on board the fact that not everyone is as well
informed about politics and electoral matters as they are. Many citizens believe that they have much
more important things to do!
The failure of many of our politicians to recognize that the new MMP electoral system is very
different from the old FPTP electoral system lies at the heart of this issue. Approximately 85 percent
of the clauses in the Electoral Act 1993 have been transferred directly, or with minor modification,
from its FPTP predecessor, the Electoral Act 1956. The new MMP electoral system has, in effect,
been grafted on to legislation describing an electoral system that electors effectively discarded as a
result of the 1993 electoral referendum. Primary responsibility for this state of affairs rests with the
politicians.
30 The author tested the time it took to recover a name from Telecom New Zealand’s White Pages Internet site. It took
7.2 seconds for the Internet to search for ‘Smith New Zealand’ and list 13 179 Smiths along with their addresses
and telephone numbers. The information provided on a special vote declaration provides a much greater number
of filters.
31 See Stat NZ, 1993, no 87, s.114.
44
What’s So Special About Specials?
Electoral administrators have the task of ensuring that the mechanics of a country’s electoral system
operate smoothly, and that the organization of elections generates public confidence in the result.
The organization and administration of general elections is not easy since they occur relatively
infrequently. The Chief Electoral Office is also required to provide advice and recommendations on
electoral matters to the Minister of Justice. Successive Electoral Law Committee reports to
parliament suggest that fundamental issues are often a lower order of priority than administrative
fine-tuning undertaken to ensure the smooth conduct of the next election.
In his introduction to the current Guide for polling day staff the Chief Electoral Officer, Phil
Whelan, observes that, ‘Voting in elections and referenda is a fundamental democratic right enjoyed
by all eligible voters…’, and the mission of his office is:
…to maintain the integrity of the electoral system in New Zealand and our vision is to
achieve excellence in electoral administration. Our services are provided so that all eligible
people have the opportunity to cast a vote.32
With more than 3000 polling places (one for around every 850 electors) the opportunity to cast a
vote is clearly present. What is surprising, however, is that the emphasis is not firmly placed on
casting a valid vote, one that counts towards the final result. To be fair, if the instructions relating
to special voting contained in the handbook are followed to the letter, no vote - constituency or
party - cast by a person who is properly registered should be rejected as ‘not enrolled’, because each
special vote declaration and ballot paper will be for the correct electorate. If this was universally
achieved, we could be confident that all the votes rejected because the persons casting them had
not enrolled, were properly disallowed. Once electors whose votes have been properly disallowed
have been identified (overwhelmingly persons who are not registered), registrars of electors should
be charged with attempting to contact each one, at the address given on the special vote
declaration, and advise them, with reasons, as to why their vote has been disallowed and inviting
them, once again, to enrol. Following this, continued failure to register could then be dealt with by
applying the penalty provisions for non-registration included in the Act.
The politicians, who largely determine the details of the electoral law, are primarily responsible for
providing an environment where all qualified and registered electors can be confident that the votes
they cast contribute to the eventual result. For them to shrug their shoulders and claim that this is
the way special votes have always been handled in the past - as they have frequently done - is
simply avoiding the issue. Nor, for that matter, is it appropriate for public servants to justify the
maintenance of the previous system on grounds of cost when it has clearly been shown to be
deficient. The electoral system belongs to the people of New Zealand; it is not the preserve of the
politicians and bureaucrats.
Taken overall, the administration of New Zealand elections is thoroughly professional and generally
competent. This is not to say that the electoral system is without its weaknesses which require
attention. Those responsible for the election night count in 1999 were subject to considerable
criticism, mainly from senior politicians, at the slow release of those preliminary results.
Unfortunately, political pressure is likely to mean that a much greater effort will be made to rectify
these perceived difficulties than to seek a just and fair solution to the problem of including valid
party votes which are currently disallowed, in the final count.33 What is required is an urgent
commitment from both politicians and administrators to give priority to matters such as the ones
described here so that electors’ confidence in the electoral system is enhanced.
32 Chief Electoral Office, A guide for polling day staff, Wellington, 1999, p 1.
33 Shortly after the 1999 election the Secretary for Justice established an Electoral Review Team, to carry out an
urgent review of the factors that resulted in the slow release of election results. The committee reported in
February 2000.
45
46
138 612
182 696
194 626
1 367 546
1 568 181
1 672 599
1972
1975
1978
207 048
197 192
231 848
1 877 040
1 787 850
1 785 277
1 869 992
1 977 658
1 739 134
1984
1987
1990
1993
1996
Averages
Source: AJHR, E.9, 1972–1996.
190 691
193 750
183 887
1 746 062
1981
186 560
Total
Specials
Election
Year
Total
Valid
Votes
11.0
11.7
10.5
11.6
10.8
9.8
10.7
11.6
11.7
10.1
42 243
47 347
38 962
37 252
35 586
36 490
43 625
44 305
64 312
32 308
22.2
20.4
19.8
18.0
18.4
19.8
23.4
22.8
35.2
23.3
Specials
Specials
Specials
as % of Disallowed Disallowed
(%)
total votes
3.1a General Electorates
24 726
48 995
58 011
102 884
23 821
41 274
28 715
24 626
55 111
64 168
21 090
15 162
48 844
60 320
19 586
43 839
22 256
16 953
42 694
55 241
Total
Specials
Total
Valid
Votes
41.1
40.1
44.7
50.5
44.7
35.0
40.3
31.0
44.7
39.7
6 949
7 286
4 970
5 591
4 835
5 542
6 638
5 870
11 940
9 871
29.2
17.7
17.3
22.6
19.6
26.3
29.8
38.7
61.0
58.2
Specials
Specials
Specials
as % of Disallowed Disallowed
(%)
total votes
3.1b Maori Electorates
202 282
209 788
1 612 020
1 721 443
1 797 145
2 080 542
1 934 160
1 834 272
1 842 961
1 937 360
214 512
273 122
225 907
231 774
218 376
204 977
208 816
155 565
1 410 240
1 801 303
Total
Specials
11.9
13.1
11.7
12.6
11.8
10.6
11.6
12.2
12.5
11.0
49 192
54 633
43 932
42 843
40 421
42 032
50 263
50 175
76 252
42 179
22.9
20.0
19.4
18.5
18.5
20.5
24.1
23.9
37.7
27.1
Specials
Specials
Specials
as % of Disallowed Disallowed
(%)
total votes
3.1c All Electorates
Total
Valid
Votes
Table 3.1 New Zealand Elections 1972–1996: Fate of Special Votes Cast
47
8 (8.2)
97
99
65
950
1990
1993
1996*
Totals
* First MMP election.
Source: AJHR, H.33, 1969; E.9, 1972–1996.
5 (0.5)
–
1
1
?
–
–
1
1
1
–
Result reversed as
result of inclusion of
Special Votes
No. (%)
‡ Includes final result for Hunua following electoral petition.
68 (7.2)
4 (6.2)
10 (10.1)
8 (8.2)
5 (5.4)
92
1981
97
8 (8.7)
92
1978‡
1987
7 (8.0)
87
1975
7 (7.4)
5 (5.7)
87
1972
95
6 (7.1)
84
1969
1984
Election night
majorities
under 500
No. (%)
Electoral Districts
No.
Election
National retained seat and thus able to form a
government
Wellington Central – National election night majority
of 169 became final Labour majority of 246.
Waitaki – Labour election night majority of 127
became final National majority of 53
Labour retained seat
Successful Wairarapa election petition makes it
impossible to determine whether a reversal occurred.
National retained seat
Labour-held seat lost to National
Western Hutt – Labour election night majority of 28
became final National majority of 109
Kapiti – Labour election night majority of 78 became
final National majority of 23
National retained seat
Outcome
Wellington Central – Labour election night majority of
98 became final National majority of 27
Details of reversed results
Table 3.2 New Zealand Elections 1972–1996: Impact of Special Votes
Table 3.3 Disallowed Special Votes 1972–1996
3.3a General Electorates
Election
Year
Total of Disallowed
Disallowed
Votes
Votes
NOR*
%
3.3b Maori Electorates
Total of Disallowed
Disallowed
Votes
Votes
NOR*
%
3.3c All Electorates
Total of Disallowed
Disallowed
Votes
Votes
NOR*
%
1972
32 308
28 196
87.3
9 871
8 106
82.1
42 179
36 302
86.1
19751
64 312
34 904
54.3
11 940
10 033
84.0
76 252
44 937
58.9
1978
44 305
39 895
90.0
5 870
5 108
87.0
49 413
45 003
91.1
1981
43 625
37 723
86.5
6 638
5 883
88.6
50 263
43 606
86.8
1984
36 490
32 336
88.6
5 542
5 229
94.4
42 032
37 565
89.4
1987
35 586
32 244
90.6
4 835
4 590
94.9
40 421
36 834
91.1
1990
37 252
35 213
94.5
5 591
5 287
94.6
42 843
40 500
94.5
1993
38 962
37 176
95.4
4 970
4 901
98.6
43 932
42 077
95.8
1996
47 347
43 281
91.4
7 286
7 010
96.2
54 633
50 291
92.1
* NOR – Not On Roll.
1 It is probable that the 1975 Disallowed Votes NOR are understated. In some electorates Disallowed Votes NOR were classified
under the heading ‘Other reasons’.
Source: AJHR, E.9, 1972–1996.
Table 3.4 Fate of Special Votes Disallowed 1996
3.4a Reasons for Disallowance
No. of Special Votes Disallowed
Reasons for disallowance – Not enrolled (NOR)
Other
No.
% of total votes cast
54 633
50 291
4 342
2.6
2.4
0.2
3.4b Analysis of Disallowed Votes Not Enrolled (NOR)
No.
For those whose votes were disallowed NOR —
% of total votes cast
50 291
Enrolled but voted on wrong roll type
Enrolled but voted in wrong electorate —
Voted in neighbouring electorate
Voted in another electorate
On dormant file but voted in another electorate
Not on any roll
48
3 843
7.6
15 729
31.3
4 080
4 438
7 211
8.1
8.8
14.3
30 719
61.1
4
The Key To Progress: A Cooperative Research Centre
Dean Jaensch
The focus of this article is ‘the research core’. The first question (and the bulk of the article is
concerned with asking questions): how does that relate to the (official) ‘core business’ of the
electoral commissions/departments, that is, the ‘conduct of elections’? Some commissions have
certainly elevated ‘research’, in either a specific or a broad sense, to a key function, either in their
mission statements or in the legislation. The common wording of the specific reference is along the
lines ‘… to conduct and promote research into matters relating to elections’. The commitment in a
broader sense is in such phrases as ‘… shall promote public awareness of electoral matters …’, and
‘… publish material …’. It should be noted, however, that the explicit commitment to such
‘missions’ varies across the commissions.
What Is ‘The Research Core’?
The broad answer is: everything that concerns elections. That is, there should be on-going research
into all components of the election process. It is suggested that this has two separate aspects,
defined by the foci of the data, but the two are linked by being integral parts of the electoral
process.
One aspect includes election laws and systems, election results, from aggregate to micro, contem­
porary, historical and comparative. These data, mostly numbers, are in the commissions, and
should be freely available for research – by the commissions, by academics, and by joint
endeavours.
The second focus may be perceived by some as a more radical approach for a ‘research core’ for the
commissions – the voters. To date, most research on voting behaviour, apathy, party identification,
and the myriad of other topics, has been almost exclusively the area for academics. But it is
suggested that commissions have a legitimate and important role in this area. If the mission
statements include ‘public education’, then, before educating, commissions need to establish just
where the education is needed and education on what topics. This is not to suggest that the
commissions should enter the broad field of electoral behaviour research, should compete with the
Australian Electoral Study, for example, but there are many areas of ‘the voters’ which are central to
political education responsibilities.
What is the state of the past and present ‘research core’? Not as good as it could and should be. In
order to justify that criticism, excuses for the shortcomings should be examined.
Academic researchers need to recognise and welcome the culture change in the electoral
commissions. When this author started research in this field in the early 1960s, when most present
day colleagues were still in nappies, there were few welcome mats in the, then, electoral
departments. Today, there has been a transformation from the formal structure of ‘departments’ ­
embedded in the culture of government, dominated by secrecy, almost paranoia, and limited public
access - to independence (not yet universal), a commitment to public education (admittedly still
varied), a commitment to research (to varying degrees), and a different approach to researchers
outside the organisations. Relationships between commissions and researchers are now generally
warm and welcoming, if occasionally fuzzy. This author has absolutely no complaints about the
‘service’ received.
49
The Key To Progress: A Cooperative Research Centre
This has resulted in the potential for an avalanche of research on electoral matters. But it is a
potential which has not yet fulfilled its full opportunities. Why not? One reason is simply resources
– that is, money. For commissions, themselves, to conduct the research which needs to be done,
they need resources and funding, either for their own staff, or for joint work or for consultancies.
Offering the research potential to ‘outside’ researchers is a relatively cost-free option. But in this
case, what are the potential problems? That is more in the realm of ‘borders’ between electoral and
political inquiry, dealt with elsewhere in this volume.
There has been considerable advance on what used to be secretive bodies protecting their data
behind locked doors. There has been, especially in the last two or three decades, a relative
explosion in the access freely offered and freely available, and certainly there has been an explosion
in the research product. Commissions, parliamentary libraries, academics, parties, groups and
individuals have become more research oriented. This has been a major advance, but it is only a
start in what is possible in the areas of research, publication and, it is suggested more important
than anything, public education. The Australian public is notoriously uninterested in matters
political, and only a little more interested in matters electoral. If there is a core research task related
to public education, and there is in fact, and recognised in the mission statements, then a
commitment to research is one step to fulfilling the promise.
Towards A Research Agenda
There needs to be a firm commitment to history. The basic material is generally available, at least in
raw and aggregate data. There has been some work on the first stage – establishing a data-bank
which can be the prime source for further research. Here the work by Campbell Sharman et al and
David Black, and others working in the field can be publicised. Their efforts need to be replicated
across Australia to provide a complete, published data-bank for all colonial, state, territory and
federal elections. Then there needs to be a detailed secondary analysis of these data to provide a
full, published account of all components of our electoral history. This project needs to be on­
going. Whose responsibility? All of us, and the commissions should play a major role in the task.
After all, they have the data, and most have a formal commitment to research and public education.
Since the 1960s, this task has proceeded a bit. But there is plenty of scope for mining the rich
source of material in historical data – not only in terms of results, but in relation to election laws,
systems, procedures and processes.
Towards The Future
The best way to ensure that the necessary research is done, and done well, is to provide a means to
organise and to facilitate the enterprise. There are many people and groups vitally interested in
electoral research, but the result, to date, is a fragmented process and a disparate product. Most of
us only find out what the other interested parties are doing when the results cross our desks, or are
published. Why not make the process more efficient and effective – even more productive – by
introducing a co-operative enterprise, and formalising it.
The establishment of a National Election Studies Centre is proposed. This should be a jointly
funded enterprise of federal and state governments, federal and state parliaments, federal and state
commissions, with a research institute structure and a charter which establishes it not only as a
repository of data, but as a generator of, and a clearing-house for, electoral research. The
involvement of governments and parliaments is a crucial one; they should recognise that they are
the products of the electoral process, and hence have a responsibility in the research process. Why
not include local government as well?
50
The Key To Progress: A Cooperative Research Centre
However, in the meantime (because establishing any such project in a federal system will be
difficult), each commission should establish its own research section which is either willing and
able to do the research necessary, and is funded to do so, or which can be involved in joint efforts,
or, at the least, facilitate research by others.
If the conduct and promotion of research is accepted as an explicit core function, then there is a
range of questions which needs to be addressed.
(i)
How widely should ‘research’ be defined? The answer is that any focus should be
targeted and any data should be available, unless such access and publication offends the
probity of the commissions or the probity of the election processes. Application of an
‘ethics’ regime is crucial. Outside of such constraints, there is no reason why any data,
any research, and any public publication should be denied.
(ii)
Research by whom? There should be a partnership between the commissions and private
(‘academic’) researchers. This partnership should be based on a commitment to the
benefits of research, and should be encouraged by the commissions.
(iii) Research on what? On the broadest possible definition of data. Only where data is a
matter of probity and necessary secrecy should it be denied to legitimate researchers.
(iv)
What is ‘legitimate’? This question raises the issue of the role of parties and groups with
‘partisan’ purposes. Is this a reason for denial of access? Again, that question relates to
‘borders’.
(v)
Research for what purpose? The best answer is ‘for public education’ – defined broadly.
This may be a ‘buzz-word’, but what is wrong with furthering understanding, public
education, the field of knowledge, and a contribution to the quality of the polity?
(vi)
Research under what conditions? The only limitations appear to me to be the probity of
the electoral process.
The constraint on such an agenda? The most obvious is money. Establishing a national election
studies centre will be an expensive programme. Involving commissions in research will involve
staff, resources, and facilities. Where can the money come from? Governments must be convinced
to recognise the value of such a project, and provide the funds.
A second constraint is a commitment by the commissions. This may require a further culture
change for some, merely a renewed emphasis by others. But the programme suggested cannot come
to fruition without it.
But the realpolitik is defined by money. Hence the conclusion that the ‘research core’ has a prior,
necessary condition – the lobbying of governments and parliaments to recognise their public
education responsibilities, and to supply the funds that are necessary. That is the real ‘research core’
issue. This will require a combined effort of both commissions and academic researchers, which
some may perceive to be a hard task, but this author has been convinced by recent developments
that this is already in process.
51
52
THE BORDERS BETWEEN ELECTORAL AND POLITICAL INQUIRY
53
5
Negotiating the Boundaries Between Electoral and Political Inquiry
John Wanna*
General Introductory Points
Negotiating the boundaries between electoral and political inquiry is a difficult task. The task is not
made easier because the boundaries are inherently ambivalent, ever-changing and dependent on
context. There may be some territorial markers [such as Freedom of Information (FOI) or privacy
legislation], but the frontiers around what are core areas of electoral inquiry and what are more
peripheral or more ‘political’ will never be precise nor universally accepted. Rather the boundaries
will inevitably be informed by what is conventionally considered ‘political’ at any time as well as
how various beholders eye the distinction between electoral and political.
This distinction involves an appreciation of what types of information should be presented on our
behalf by professional electoral managers. But such information has many dimensions, one of
which is its inherently political character. Information is conceived, defined, collated and presented
in ways that have political consequences, and may be used in a diversity of political ways. These
ways may be sanctioned and intended or accidental and unintended; they may be normative and
non-controversial, or alternatively may be used in ways which are not approved.
Hence, at the most basic level all electoral information is by its very nature political and potentially
has political aspects that may not be easily controlled. In a consensual polity or where there is a
certain bipartisan consent over electoral rules and conventions, there may be broad agreement that
some electoral information can be neutral, relatively apolitical, and relatively non-biased in partisan
terms. But, on closer examination even the definition or inclusion of certain types of ‘facts’ can be
political. Even some basic description and categorisation can be treated as controversial and
perhaps unintentionally political. Two examples illustrate this point and underline how much we
become accustomed to the conventional. First, the identification of political parties in electoral
material and information kits (say over votes received or seats held/contested or even the party
alignment of candidates) in many ways advantages political parties as a group over independents
and newer pressure groups – it is not a level playing field. Second, in determining the ‘political
status’ of seats, the labelling of specific electorates as ‘safe seats’ in official publications may cause
voters to feel they can ‘kick the sitting member in the pants’ and is, according to some politicians,
an invitation to protest especially in volatile times. Arguably, a margin of 10% is definitely not safe
given individual swings in some state and federal seats in recent years.
If knowledge is power, then almost all information provided by electoral commissions will be open
to criticism of political bias. For instance, the publication of factual data by electorate such as
household income, levels of education, employment and unemployment can be of enormous
benefit and significance to selected players. Some of the factual material may be generated by other
agencies and available from a variety of readily available sources (such as the Australian Bureau of
Statistics (ABS), parliamentary libraries); the question, therefore, is should Electoral Management
Bodies (EMBs) be in the business of producing or reproducing such material. It may advantage one
side or candidate over another, or even discourage potential candidates from contesting or
campaigning. The EMB’s factual material may arguably assist the major parties by providing profiles
of electorates that only they are in a position to exploit with their resources and political
wherewithal. The way information is made available (or how it is promoted) is also very political;
54
Negotiating the Boundaries Between Electoral and Political Inquiry
electronic and web-based information privileges the information rich, the younger electors and
those largely from educated backgrounds. Less advantaged voters may still have difficulty accessing
information presented through traditional methods of delivering information.
Whereas disseminating information is political, analysis is even more so - whose analysis, what
questions are asked, or not asked, what categories are used, what conclusions are explicit or are left
implicit. Electoral research represents a further muddying of the boundaries between the statutory
provision of essential information (eg the seat boundaries, polling booth locations and candidates’
names) and the pursuit of particular research agendas. As one politician from one of the major
parties stated when interviewed: ‘the Electoral Commission should stick to the facts - and nothing
more - anything else becomes political’. Research agendas are generally related to the EMBs
educative role and the promotion of information. But again there are real dangers in the active
prosecution of an educative role. Even if research aims to simply better inform the citizenry, this
steps into political territory as a more electorally educated population will potentially advantage or
disadvantage certain political players.
However, if the EMBs adopt an ostrich-like attitude and prefer to eschew electoral analysis, they
may equally be accused of failing in or neglecting their duties. For most EMBs in Australia, electoral
research is a statutory responsibility but the exact nature of such research not defined and is subject
to the discretion of commissioners. The community and the polity have legitimate expectations that
EMBs will present an appropriate amount of basic information to assist them in making informed
political choices. Electoral research is at one level the EMBs community service obligation (CSO) –
and as with such obligations in other public authorities the definition of a CSO is highly
problematic. Ascertaining what is expected and who should perform the research will go someway
toward negotiating the boundaries between electoral and political analysis.
Criticisms Of Existing Electoral Research
In general, EMBs tend to see electoral research as a lower order priority compared to their principal
responsibilities – administering elections. Some take the view that research is the responsibility of
other institutions and is not their core business. There is evidence that some EMBs perceive
electoral research merely as an additional task and dispensable if budgets and resources become
tight. Such attitudes are unfortunate for three reasons: the EMBs are the repositories of a vast
amount of primary data which is at present untapped; EMBs can themselves benefit from research
in terms of electoral administration and management; and third, we do not yet have in place
enduring research relations, partnerships or protocols of access which will encourage others to
work on the materials. EMBs are in a strategic place to facilitate research, if not directly undertake
such exercises.
Accordingly, the current state of electoral research throughout Australia displays many gaps, with
even the most basic factual information often overlooked. For instance, not all EMBs have a
comprehensive data-base on by-elections and analysis of swings. In Queensland, the Electoral
Commission Queensland (ECQ) now produces a detailed statistical return for by-elections. Like
most of the states and territories, it has not produced a historical series of all by-elections giving
information such as average number of candidates/swings recorded/proportion won by the
government [as the Australian Electoral Commission (AEC) has with a series dating from 1901].
There is also a wealth of basic electoral material of wide interest that could be produced after an
election, such as whether the candidate lives in the electorate, age profiles of candidates,
occupations etc. Collating such information generally takes individual researchers a fair bit of
digging – and in cases where candidates provide only PO boxes or don’t answer questions then this
is often impossible. But the point remains, EMBs often possess interesting sets of data and detailed
55
Negotiating the Boundaries Between Electoral and Political Inquiry
electoral information yet, for a variety of reasons, do not themselves compile the data or undertake
research projects to find ways to present the data in relatively neutral guises. Often electoral officials
do not recognise the importance of this information regarding it as routine data collected for
recording and compliance purposes.
On the other hand, EMBs will generally help facilitate the research projects of others to the extent
they are able to cooperate. They are usually very generous in answering queries and sending out
published and documentary information. Most EMBs will usually allow researchers to look at files
and records, but would they allow independent researchers to conduct their own ballot surveys? If
an honours student was investigating the extent of deliberate versus accidental informal voting
would the EMBs permit access to ballots to assist this research? Generally, such questions are
governed by either or both the statute (dictating say that the commission would have to be involved
in the research) and the commissioner whose discretion could be sought.
Other weaknesses with the existing stock of electoral research include:
●
EMBs have traditionally been reactive in research perhaps working on specific requests or an ad
hoc basis; but not pro-active or systematic in establishing research agendas.
● Source
data is often incomplete with many gaps in the historical data kept in files. Despite the
archive acts, materials may not be archived, or be deliberately destroyed or dumped. Politically
sensitive information can be preserved and perhaps released at a later date. EMBs could
individually or collectively implement a ‘five year’ or ‘ten year’ rule over sensitive information
similar to the thirty year rule for cabinet papers.
● Existing
studies of particular electoral phenomena are often episodic, ‘one-off’ cases. Even when
topics are investigated more than once there are often differences in the ways data are reported
and collated – making comparison difficult. This means that research generally cannot provide
consistent historical sequences. Moreover, research today is still hampered in some states by the
‘hands on’ political interference of the past.
●
Often research is solely jurisdictional specific, and there is a dearth of comparative electoral
statistical analyses.
●
There is not much research into what the various users of EMB material actually expect (for
example, needs-demand surveys). This client-driven approach to research could help specify the
bounds of appropriate electoral research undertaken by or with EMBs.
In terms of the boundaries of political and electoral analysis – the EMBs to date have generally
responded by being conservative, cautious, and deliberately avoiding contentious issues. Often
EMBs have resorted simply to presenting research and information other agencies have collated (eg
ABS and census data) – but then packaged to make it electorally specific. In other words, in
‘electoral atlas’ productions [eg Parliamentary Research Information Service (PRIS) and AEC] some
correlations or rankings are suggested but the presentation stops short of presenting any analysis –
the reader/commentator is left to draw their own conclusions. Perhaps this approach is justified in
that such research already exists – and the EMBs are merely packaging multiple sources of
information for the convenience of the reader.
56
Negotiating the Boundaries Between Electoral and Political Inquiry
Who Are The Clients/Users/Purchasers/Audiences Of Electoral Research
This is an important question because the type of research undertaken will be guided or influenced
in different ways depending on the answer. For instance, if political parties are the main users of
detailed electoral-based research this may indicate that the EMBs need to consult more fully and
regularly with the parties to establish a longer term research strategy. Standing and sessional parlia­
mentary committees could facilitate this process but EMBs could also venture to establish research
advisory boards and suggest major users nominate suitable members.
One of the problems in determining what type of research is expected, is that the number of
potential users are many and varied. EMBs may view the end users of their research as clients to
whom they deliver services, users of merit goods, purchasers of public goods, or audiences they
attempt to interest. The many potential clients/users/purchasers/ audiences include:
●
the electorate or the population at large – the need to inform the citizenry even if, or especially if,
they are not particularly interested in politics and the mechanics of electoral systems (the EMB’s
community service responsibility to the polity at large);
●
parliaments and legislators – especially parliamentary committees with responsibility for electoral
matters;
●
taxpayers and the budget (the EMB’s fiduciary responsibility to spend money wisely and provide
value for money in its activities);
●
EMBs themselves have an interest in research to assist in operational management and to inform
decision-making (the EMBs managerial and statutory responsibilities);
●
redistribution commissioners – faced with the tasks of making judgements of value about
redefining electoral boundaries; this is a statutory responsibility for most commissions;
●
political parties and party machines – the main partisan players who are likely to have both a
professional interest in the quality of EMB research and a partisan interest in what research is
produced and the findings presented (the EMB’s responsibility to principal clients and
stakeholders);
●
political activists and potential candidates – providing an information base which shows/suggests
an open political process that is both accountable and fair to potential contestants – including
material presented so that intending candidates have a checklist of formal requirements to stand
for parliament such as citizenship, residency etc (the EMB’s responsibility to the liberaldemocratic norms of representation);
●
other government departments and agencies are users of EMB data/research;
●
interest groups, advertising agencies and marketing firms, including commercial consultants and
corporations (how far does the EMB’s responsibility extend to provide basic ‘political’ knowledge
for political activists not intending to stand for election but to influence votes – eg through direct
mail-outs on particular topics);
57
Negotiating the Boundaries Between Electoral and Political Inquiry
● journalists,
commentators and the media - many of whom will be non-technical and unfamiliar
with the nuances of electoral processes and rules (the EMB’s responsibility to deliver information
on candidates and on the current electoral laws and provisions to disseminators who ought to be
provided with simple but accurate information packages);
●
academics and psephologists (a professional responsibility to assist with more scholarly studies of
electoral statistics, trends and voter behaviour);
●
international audiences interested in comparative analysis – interest in how Australians vote,
results, major party shares etc (the EMB’s international community service obligations).
While these potential clients/users/purchasers/audiences are varied, between themselves they would
often be at odds over the types of information and research preferred. In EMB publications it is
often unclear not only to whom the material is aimed, but also what is the intent behind the
information provided. For example, is material presented intended to inform with relevant detail or
stimulate interest in aspects of the political process. Generic and ‘one-size-fits-all’ publications may
do neither well. This issue is crucial in determining the depth of information provided to potential
users and meeting their expectations.
It is also worth noting that some EMBs (eg the AEC, South Australian Electoral Office and ECQ)
undertake research tasks, but this does not happen often largely because of the question of respon­
sibility and the likely costs involved. The government or the relevant parliamentary committee
could request particular research agendas but one suspects that this is not widely appreciated by
legislators.
Pure Research Versus Operational/Instrumental Research
Some might find it useful here to distinguish between types of electoral research: pure research and
operational/instrumental research. EMBs may become interested in pure research for its intrinsic
merits and undertake projects aimed at research development (the conduct of original primary
research as distinct from research collation exercises). Such research will involve both a qualitative
and quantitative dimension – and the qualitative aspects may be more sensitive to political
concerns. So, EMBs could investigate qualitatively: how voters make choices; how well do voters
feel informed to make a choice; how do voters feel about the task of voting (duty, obligation, right,
chore, irrelevant); what do they feel about compulsory enrolment and compulsory voting or would
they prefer more voluntary provisions; what are their attitudes to different voting methods (full
preferential, optional preferential, first-past-the-post and different allocative systems); and what are
their attitudes to the retention of the position of the local member elected from a constituency. Pure
research may also be conducted into the nature of the electorate – how do we identify notions such
as ‘community of interest’. Alternatively pure research may usefully investigate how electoral
research is used (by the players, media, etc) and how much is it used (citation, commentary,
referencing). The boundaries between electoral and political inquiry will generally be more
problematic in pure research – and this may entail active collaboration with research partners or
parliamentary committees.
Alternatively, in more applied or operational/instrumental research the EMBs will be investigating
issues associated with how well they are performing their functions. Operational/instrumental
research will be/should be driven by the relevant electoral acts. For instance, EMBs bear statutory
responsibility for ensuring the electoral rolls are up-to-date and as accurate as possible –
operational research and investigation (whether in a published form or not) can be commissioned
to ascertain the quality of this information. Other topics may include: where would people in a
58
Negotiating the Boundaries Between Electoral and Political Inquiry
locality prefer booths to be located, what time of the day would they choose to vote, how would
they choose to vote if given an option (electronic, extraordinary voting)? Operational/instrumental
research may also investigate EMB staff competencies: such as, to what extent can we be satisfied
that electoral workers around election times are adequately trained and perform their tasks
competently? With respect to this type of operational research two comments are worth making;
first, this operational function has not always been conducted as assiduously as it ought, and
secondly, the results of operational research are rarely released publicly. However,
operational/instrumental research may be of high interest to parliament and the media, yet at the
same time be less sensitive to criticisms of ‘political’ inquiry. Accordingly, the bounds between
electoral and political research may be easier to negotiate in the area of operational/instrumental
research.
There are many other agendas under the heading operational research in which the EMBs could be
more pro-active – time and budgets permitting. For instance, how much responsibility does an
EMB bear for ensuring that the information it is supplied with by parties, candidates and other
public officials is accurate and reliable? Should EMBs merely take the information provided on trust
or, if not, to what extent should they check the information? At present there is a fairly loose set of
procedures at the state level for the registration of political parties and limited checking of the
required information by EMBs. Moreover, some investigations will involve contested definitions and
terminology – what precisely is a political party, who determines what constitutes a party member,
should the EMB verify that the candidate is a party member and has met the party requirements to
be nominated, how do we verify candidate occupations, does it matter if a candidate declares
unemployed, investor or consultant or provides a false address or convenient PO box number?
Such research would better satisfy the EMBs statutory responsibilities but risk upsetting one of their
major clients – the established political parties.
Potential Research Topics
A series of potential topics that directly relate to electoral research and do not involve straying too
far into the realms of overt ‘political inquiry’ is included below. Such research agendas over time
would greatly assist many end-users of EMB material and add to the quality of other research being
undertaken.
First, EMBs have an important role in gathering data on candidates, parties, voting patterns,
variations in seat boundaries over time, and so on which would be of invaluable assistance to
researchers. For example, the information collected on candidates is rarely collated and presented
in digestible form. EMBs could also collect and record ‘how to vote’ cards especially as such
material is often important in helping explain particular results or in documenting particular
relations between political players. Parties for instance may have multiple how to vote cards issued
for various purposes, and produce them under different guises. The EMBs are uniquely placed to
collect and assemble this information and make it available for study – there is no particular onus
on EMBs to actually publish or analyse this material themselves.
Second, applied research is needed into levels of electoral knowledge and the degree of
understanding in the electorate about the basic mechanics of elections, basic voting procedures, the
political system etc. This is needed to inform the types of public interest information the EMBs
produce around election time. In operational terms, such research will indicate how successfully
the EMB has met its corporate goals in ‘promoting informed participation and confidence in the
electoral system’. Occasionally changes to the voting system occur by law and the EMB then has a
statutory role to inform the public as to changes in voting procedures. In Queensland, after the
adoption of optional preferential voting, the ECQ arguably made little effort to inform voters of the
59
Negotiating the Boundaries Between Electoral and Political Inquiry
flexibility contained in the new system in the 1992 or 1995 elections, despite high numbers of
voters choosing to exercise the option of plumping (validly voting only for one candidate under the
optional preferential voting system). From some perspectives, the ECQ’s decision not to seek to
educate the electorate about the new voting system was highly political – favouring the established
major parties at the time, and subsequently leaving it to a political party (One Nation) to point out
explicitly to voters their right to an effectively ‘first-past-the-post’ vote (resulting in increased rates
of ‘exhaustion’).
Third, there is merit in conducting regular and sequential ballot surveys after elections – especially
at the state level. Moreover, the information is more valuable to observers if it is timely and
produced relatively quickly after the election. To illustrate this point, a short ballot survey was
conducted in 1992 and a longer and more detailed study produced in 1995. The ECQ commis­
sioned a further ballot survey covering the extraordinary and fascinating election of 1998, but this
survey had not appeared by early 2000. Such detailed and immediate information on voter
behaviour is a significant issue – especially as NSW evidence has indicated that the incidence of
voting only a ‘single preference’ (plumping) tends to increase steadily, whereas the early evidence
from Queensland suggested that this had not occurred in the first two elections under optional
preference voting. The ballot paper survey based on the 1998 Queensland election would possibly
confirm the NSW experience and perhaps fit with Malcolm Mackerras’ notion that it takes the
electorate at least three elections to familiarise themselves with such changes.
Fourth, with electronic scanning of the rolls at a booth level it is now possible to have much more
detailed information on voting habits and attendance patterns. This implies that there is consid­
erable scope for better polling booth catchment area analysis. In operational terms, it is remarkable
that so little is done given the importance of this empirical data for the optimum location and
relocation of polling booths in the public interest. It is also of interest to the political parties to have
accurate information on where particular voters tend to cast their votes, in order to target their
campaigns and literature better and to more effect.
Fifth, it would be beneficial to researchers and media commentators to have better profile
information on electorates. Already there is some good work here (eg, electoral atlases) but it is
patchy and lacks longitudinal analysis. Besides the usual suspects (EMBs, ABS, parliamentary
libraries) other government departments could assist with this research. Also a wider cross-section
of information would be of interest to observers, for example: how much public money was spent
per electorate, what projects have been announced by electorate, levels of government assistance or
financial grants per electorate, the number of public agencies and service centres in the electorate
(or closures) etc. Much of this information is already produced by government departments and is
either made available or can be acquired through FOI.
Sixth, by-election results are often a neglected item of research. With minority governments, and
governments relying on slender majorities, the issue of voting behaviour at by-elections has
becomes more significant. For instance, is the typical swing against the government affected by the
size of the government’s margin and if so by how much. How often do governments win byelections (and when does this occur – at the beginning of their term, in their first term, etc). When
do oppositions/independents gain the largest swings. Consolidated historical lists and crosstabulated statistics are increasingly essential electoral ‘facts’ which often we cannot find with ease.
60
Negotiating the Boundaries Between Electoral and Political Inquiry
Conclusions
The distinction between electoral and political inquiry is problematic and very blurred. From one
interpretation, the two are inextricable. From another, it is clear that conventional distinctions are
made and will continue to be made between what is legitimate for EMBs to publish and what is ‘too
political’. If such a distinction is difficult to eradicate, then it may be better faced explicitly with
EMBs finding acceptable ways of managing the problems.
Clearly, electoral research is a collective responsibility of many agencies and participants, including
academics, the media, parties, interest groups and interested citizens. But, while electoral research
involves bodies other than the EMBs, it is also true that electoral officials have a lead-agency role to
fulfil which actively involves negotiating the bounds of legitimate electoral inquiry. At present it is
not apparent that the EMBs around Australia are proactive in pursuing this lead-agency role.
Moreover, the establishment of a research advisory committee on which major users and
stakeholders could serve and discuss future agendas is a long overdue innovation.
Partnerships with other key research institutions may assist the EMBs in their research endeavours
without causing them major problems in straying too far into overtly political areas of inquiry.
These other bodies can not only contribute their expertise and resources but shoulder much of the
accountability for the analysis and presentation of the research. We can only urge the various EMBs
in Australia to develop and nurture enduring research relations with such bodies for the benefit of
all their end users.
* I wish to acknowledge the assistance of Tracey Arklay and Glenn Rhodes in preparing this paper.
Both are researchers attached to the Centre for Australian Public Sector Management at Griffith
University.
61
6
Electoral Inquiry Or Political Debate?
Analysis, Commentary And The Controversy Of Victoria’s
Legislative Council Electoral Process
Nicholas M. Economou and Brian J. Costar
On 18 September 1999, Victorian electors went to the polls and voted the Liberal-National
coalition government headed by Jeff Kennett out of office. The fall of the government did not
happen immediately the votes were cast and counted on polling day. Two factors contributed to
delay the transition. First, the result in the general election in terms of actual seats won was very
close. Neither the ALP nor the coalition parties won an absolute majority of seats in their own right.
The transition in Victorian government relied on the say-so of three independents who following
the election held the balance of power in the Victorian lower house. The process of transition was
also delayed by the need for a supplementary election in the lower house electorate of Frankston
East caused by the death of one of the candidates on polling day. Once the supplementary election
was held, and the independents gave their approval to the formation of a new Labor government in
the form of an agreement about a ‘Charter of Good Government’, the transition was able to occur.
Contrary to general expectation held during the campaign, the ALP’s Mr Steve Bracks duly became
Victorian premier.
The Charter of Good Government was clearly an important part of the transition process. The
Charter1 is a document that was drafted by the three independent MPs in the interim period
between the general election and the Frankston East supplementary poll. Ostensibly a document
designed to highlight issues such as the answerability and accountability of the executive, the
Charter was to emerge as something of a bargaining chip between the major parties and
independents to determine who would be able to govern in minority. Included in the Charter is the
issue of parliamentary reform, and central to this is the matter of reforming Victoria’s upper house,
the Legislative Council. Implicit in the Charter (including all the component parts making reference
to the need for institutional reform) was a critique of the way the Parliament had operated under
the previous Coalition government. Because the Kennett government enjoyed massive majorities in
both houses, the Victorian Parliament had become something of an irrelevant institution – a
situation reflected in the decline in parliamentary sitting days, for example (Harkness 1999). As the
mechanism by which the Coalition majority was obtained in the review chamber, the Legislative
Council’s electoral system became the focus of the Victorian manifestation of the ‘decline of
parliament’ thesis.
The sense of the parliament being an institution in decline in the face of such rampant executive
power has been a particularly salient issue in discussions about the role of Victoria’s Legislative
Council. Like all the other Australian upper houses, the Victorian Legislative Council is viewed as
the review-oriented chamber in a bicameral parliamentary system whose primary function is to
hold the executive answerable for its actions, and which has substantial blocking powers to
reinforce this accountability function (Sharman 1990:3; Maddox 1996:230). However, unlike all
the other Australian upper houses, the Victorian Legislative Council employs essentially the same
preferential voting system used in the Victorian Legislative Assembly. This differs with the Senate
and most of the other state upper houses in which proportional representation is used. Preferential
voting is used in Tasmanian Legislative Council elections, but elections for the lower house in that
state utilise the Hare-Clark proportional representation system. Victoria is the only state where a
1 A version of the Charter can be found at the website of independent MP Susan Davies at
http://home.vicnet.net.au/~susandavies/Charterfinal.htm
62
Electoral Inquiry Or Political Debate?
system that delivers exaggerated majorities based on majoritarian outcomes in single-member
electorates is used for both the upper and lower houses, and this matter of fact about the electoral
system has important consequences for the normative political debate in that state.
In the aftermath of the 1999 general election and the post-election negotiations that went on over
the Charter of Good Government, the political debate about the Legislative Council has tended to
be influenced by popular perceptions about the proper role of upper houses. During a period prior
to the election when the Kennett government appeared to be completely in the ascendancy, the
Melbourne Age newspaper (26/4/97) undertook a minor campaign to highlight what it saw as a
core failing of the upper house.2 Asking the rhetorical question ‘who is in the upper house, and just
what is it that they do?’, The Age argued that the upper house becomes a mere cipher of executive
power. This was very true, although the deeper point about that, which is structurally problematic
about the Council, is not simply that it has been a rubber stamp in recent times. Rather, the real
problem with the Council is to be found in its electoral system.
In the 1999 general election for the Legislative Council (the broad features of which are contained
in Table 6.1), the ALP won 42.2 percent of the primary vote, and 50.1 percent of the two party vote
across the state.3 This returned the ALP 8 or 36 percent of the available 22 seats being contested.4
With 39.7 percent of the primary vote the Liberal party was able to win 11 or 50 percent of the
seats. The National Party, however, managed to secure 13 percent of the seats with 7.2 percent of
the primary vote. If this outcome of seats to vote is expressed as a ratio, the sense of the National
Party being the big winners out of the system is reinforced. The Labor ratio of vote won to
percentage of representation obtained in the 1999 Legislative Council stood at 1:0.8. The vote-to­
seat ration for the Liberal party was 1:1.2, and for the National party, 1:1.8. However one looks at
these figures, the impression that there is a very great imbalance in the amount of representation
won by the non-Labor side relative to its vote is very hard to avoid, however non-partisan one
would wish to be in this debate.
What has made things even worse for Labor has been that, because Legislative Councillors enjoy
eight-year terms and only half the Council is elected with each lower house electoral cycle, ALP
representation continues to be influenced by the 1996 ‘landslide’ loss to the Coalition. But even
here the sense of a disproportionate outcome in terms of seats to votes is apparent. In 1996, Labor
won 40 percent of the primary and 49 percent of the two party vote for a mere 5 (or 22 percent) of
the available seats. The Liberals, on the other hand, secured 63 percent of the seats with 44 percent
of the primary vote and with 54 percent of the two party vote, the Coalition was able to win 78
percent of the available seats. As it currently stands, the means of electing representatives to the
upper house advantages the Liberal and National parties to the point where they will control the
powerful upper house even if Labor is preferred by an absolute majority of voters. In the meantime,
a host of non-major party participants in the Victorian political debate has been frozen out of the
representative process. The Australian Democrats, for example, nearly always perform well enough
to win a seat in Victorian Senate contests, but are virtually absent from the state electoral contest.
Little wonder that demands to reform the Legislative Council in the Charter of Good Government
have focused on reforming the house’s electoral system.
2 This campaign followed on from a controversy that emerged after revelations that a Liberal MLC had been
‘moonlighting’ as a local government consultant whilst holding his upper house seat.
3 A complicating factor arose during the 1996 election when the ALP decided not to contest the provinces of
Higinbotham and Templestowe. In these seats the ALP issued how-to-vote cards directing primary votes to the
Greens and Democrats.
4 Another complicating factor in this election was the conduct of three by-elections for provinces whose sitting
members were half way through their tenure. Of these three by-elections, one for the province of Ballarat was of
some significance because the ALP was able to obtain a gain at the expense of the Liberal Party. The two party vote
for the Ballarat by-elections was ALP 52.2 percent, Liberal 47.8 percent.
63
Electoral Inquiry Or Political Debate?
It is in the context of this debate that political analysts who have an interest in electoral matters
confront the conundrum of being seen to be highly political when commenting on something as
overtly structural as an electoral system. The situation is exacerbated by the suspicion that general
understanding of the system in question is, at best, rudimentary when the structural problems of
the offending electoral system are quite complicated. Victoria’s Legislative Council electoral system
is a case in point. At a cursory glance, the Legislative Council system looks fair enough. It is based
on electoral boundaries that incorporate four lower house electoral divisions into each Council
province. As such, the provinces are reasonably equal in voter population. They are also coherent in
the sense that they aim to bring similar regions and communities together, thereby being faithful to
a strong theme in Australian electoral redistributions of recognising and maintaining local
community identity when drawing electorate boundaries. This guiding principle militates against
the undesirable practice of gerrymandering, yet it is precisely at this point that the inherent fault in
the Victorian upper house electoral system occurs.
We have argued elsewhere (Economou and Costar 1997:112-115; Costar and Economou 1992)
that the Victorian electorate comprises a set of distinct regional clusters. Metropolitan Melbourne
dominates Victorian elections in the sense that regional and rural voters make up only 30 percent of
the total electorate. Yet Melbourne itself comprises distinct demographic regions in which there are
clearly identifiable areas of blue-collar or ‘working class’ or ‘traditional Labor’ voters, and areas of
much greater affluence which also tend to vote Liberal by very healthy margins. Herein lies the key
point: Labor’s core constituency reside in a contained and concentrated part of Victoria – specif­
ically, the western and northern suburbs of Melbourne, and another concentrated area in the
southern suburbs around Dandenong and Springvale. In these regions, the Labor vote averages well
into the 60 percent range. Beyond these clusters, Labor seeks to maximise its support usually in
marginal seats covering the outer fringes of the Melbourne suburbs (although, somewhat unusually,
Labor’s very good 1999 result was due almost entirely to its ability to win seats in provincial centres
around Bendigo, Ballarat, Geelong, Seymour and Maryborough).
While there is a sense that there are marginal lower house seats that can be won by Labor, there has
not been a concomitant notion of marginality in the Legislative Council. Indeed, if a Mackerras­
pendulum chart were applied to the Council on its 1996 results (a variation of which is presented
in Table 6.2), only one province – Eumemmerring – could be considered as very marginal. Two
other provinces would have changed hands with swings of less than 4 percent, and another two
with swings of 5 percent. Nearly everything else on the pendulum – some 11 provinces in all – was
very safe for the Coalition while Labor, of course, had its core of 5 ultra-safe provinces. In order to
win a majority of provinces, Labor would have had to hope for a uniform swing of 4.8 percent on
its 1996 result, which would have meant Labor winning 52.1 percent of the two party vote in the
1999 contest.
The reason for this is geographic. There are five upper house provinces that cover western and
northern Melbourne that Labor wins with very substantial majorities. However, there are nine other
provinces that are either very safe for the non-Labor parties, or are very difficult for Labor to win.
This latter situation arises particularly in the south-eastern suburbs where strong pro-Labor
divisions such as Clayton, Dandenong and Springvale are distributed amongst three provinces
(Waverley, Chelsea and Eumemmering) which also include very strong Liberal voting adjacent
divisions. Put simply, Labor has far too many of its committed voters concentrated in too few areas.
By the same token, the Nationals are assisted by the concentration of their support in particular
regions that allows the party to win provinces despite their low state-wide vote.
64
Electoral Inquiry Or Political Debate?
Commentators whose normative understanding of democracy embraces the idea that representation
should roughly reflect a sense of the party winning a majority of the two party vote getting
somewhere close to winning a majority of available seats, would surely advocate a review of the
Victorian Legislative Council electoral system. This is precisely the rationale behind the call for
electoral reform contained in the Charter of Good Government, but herein lies the dividing line
between electoral analysis and politicisation of the debate. The Charter’s call for proportional
representation as the necessary reform now resonates as a partisan demand rather than as the only
realistic panacea to the peculiarities of Victoria’s electoral geography. The reason for this is quite
simple: at the moment, one side of Victorian politics – the Liberal-National side – is advantaged by
the current system. There is absolutely no incentive for either of these parties to alter a system that
is so difficult for their opponents to succeed in. In the meantime, the call for upper house reform
resonates as a party partisan demand because it is really only Labor and the Australian Democrats
who champion this cause. The normative democratic debate is thus obscured by an apparently
party-partisan approach. The apparently apolitical commentator who espouses the virtues of a
democratic system that bestows legitimate victory upon the party that wins the most votes runs the
risk of being cast as the intellectual face of an ALP-Democrat grab for parliamentary power.
There is an important second theme accompanying this apparent party politicisation of the debate.
Much has been made recently of the urban-rural divide in Australian society – a discussion that was
given even greater impetus in the light of contribution made by non-metropolitan voters in Victoria
to the fall of the Kennett government. Yet the contemporary re-discovery of this divide tends to
gloss over the fact that the urban-rural cleavage has been one of the long-standing characteristics of
Australian political society (see Jaensch 1983:56-60). It is certainly at work in Victoria, where the
National Party has historically been an important participant in the political debate by virtue of its
ability to win representation to both parliamentary chambers. It is at this point that an interesting
symbiosis between the electoral fortunes of the Nationals and the political aspirations of regional
and rural Victoria starts to emerge. Under this approach, that which is good for the Nationals is
soon recast as being in the interests of non-metropolitan Victoria. The current upper house electoral
system is a case in point. Normative democratic theorists might question the legitimacy of an
electoral system that allows a party such as the Nationals with 7 percent of the vote to win 13
percent of the seats. Defenders of the status quo however can quickly justify such outcomes on the
grounds that regional and rural interests need to be protected in the face of the tyranny of the
metropolitan majority.
Thus the coalition parties defend the status quo on the grounds that it protects the political
interests of non-metropolitan Victorians. This is a very powerful argument in a liberal democratic
political culture in a community in which suspicions about the tyranny of the majority does
resonate. It allows for defenders of the status quo to demonise advocates of electoral reform for the
upper house not only as Labor sympathisers, but also for being ‘anti regional’ and/or ‘anti-rural’.
Ironically, the 1999 general election outcome has simply served to enhance the notion of a gulf in
aspirations between metropolitan and non-metropolitan Victoria. Even though the defeat of the
Kennett government depended upon regional and rural voter realignment, this has simply served to
reinforce non-metropolitan appreciation of the power of the ballot box as a mechanism for keeping
metropolitan-based governments answerable and accountable for their actions. Advocacy of upper
house electoral reform based on proportional representation has been quickly criticised by its
detractors for the way in which it will result in a decline in regional and rural representation. It is
politically unlikely that non-metropolitan voters will accept a proposal that would diminish their
electoral influence right at the moment in which they are basking in the power they exercised over
the Kennett government via the existing electoral system.
65
Electoral Inquiry Or Political Debate?
These political realities mean that, for the electoral analyst, it is unlikely that any participation in
the normative debate about what should be done with Victoria’s upper house electoral process can
occur in an apolitical or non-partisan environment. The proposal to alter the upper house electoral
system is, in fact, a proposal to significantly alter the balance of power in Victoria. The adoption of
proportional representation based on electorates of equal voter number will invariably result in the
diminution of the power of non-metropolitan voters, and in a reduction in the power of the parties
of traditional non-metropolitan preference, the Liberal and National parties. No political discourse
that involves the contest for power can be apolitical. In this environment, it is not possible to be
value-neutral and to avoid being seen to be partisan. One either supports the status quo (and sides
with the Coalition and its traditional rural voter base), or one advocates a major change to the role
of the upper house based on changing its electoral system (and, as a result, is seen to be on the side
of coalition’s opponents).
This is the reality of politics, of which the electoral contest is a subset of the ongoing competitive
struggle for power. So, too, are normative debates over which should be the core values upon
which a political system is based. The electoral analyst may try to avoid being seen to be political by
retreating to pure empiricism, but, in the case of Victoria’s upper house, this is just not possible. A
cursory review of the facts simply serves to reinforce the notion that there is something not quite
right with a system that can deliver such a large majority to parties with a minority of support.
Electoral analysts are political analysts: they not only describe the results of an electoral system,
they also seek to explain the politics behind these systems. No electoral system operates in a
vacuum of norms and ideology, and no electoral system is not implicated in the central task of
seeking ways in which political power may be distributed (Taylor and Johnston 1979:37).
Consequently, analysis of this critically important component of the liberal democratic state cannot
operate in a political vacuum either.
References
1. Costar, B. and Economou, N. (1992) ‘Elections and electoral change’ in Considine, M. and Costar, B. (eds) Trials in
Power: Cain, Kirner and Victoria 1982-1992, Melbourne University Press, Melbourne.
2. Economou, N. and Costar, B. (1997) ‘Victoria: against the slide’ in Bean, C., Simms, M., Bennett, S. and Warhurst,
J. (eds) The Politics of Retribution: the 1996 Federal Election, Allen and Unwin, Sydney.
3. Harkness, A. (1999) ‘The Victorian parliament: an institution in decline?’ in Costar, B. and Economou, N. (eds) The
Kennett Revolution: Victorian politics in the 1990s, UNSW Press, Sydney.
4. Jaensch, D. (1983) Power Politics: Australia’s Party System, Allen and Unwin, Sydney.
5. Maddox, G. (1996) Australian Democracy in Theory and Practice, Third Edition, Longman Cheshire, Melbourne.
6. Sharman, C. (1990) ‘Australia as a compound republic’ Politics 25 (1), pp 1-5.
7. Taylor, P. and Johnston, R. (1979) Geography of Elections, Croon Helm, London.
8. Victorian Electoral Commission (1996) Report to Parliament on the Administration of the 1996 Victorian State
Election, Government Printer, Melbourne.
66
Table 6.1
The 1999 Victorian Election Results – Legislative Council
Party
Seats
% of seats
Primary Vote %
Australian Labor Party
Liberal Party
National Party
Australian Democrats
Other
8
11
3
0
0
36
50
14
0
0
42.2
39.7
7.2
6.7
4.2
8
14
36
64
50.1 (+2.8)
49.9 (-2.8)
% Swing to change
+1.7
-4.2
+0.6
+1.0
Two party preferred vote
Australian Labor party
Liberal/National
Source: Victorian Electoral Commission (1999).
Table 6.2a
Legislative Council provinces: marginality following the 1996 election
Labor provinces
Doutta Galla
Melbourne West
Melbourne North
Jika Jika
Melbourne
% swing to change
-11.2
-11.2
-11.7
-12.8
-13.8
Liberal/National provinces
Eumemmerring
Geelong
Chelsea
Ballarat
Monash
Waverley
Central Highlands
Templestowe
Koonung
South Eastern
Silvan
Higinbotham
Gippsland
North Western
East Yarra
Western
North Eastern
% swing to change
-2.8
-3.3
-3.3
-4.4
-4.7
-4.8
-7.6
-9.2
-9.6
-10.0
-10.4
-10.8
-11.0
-11.6
-12.2
-14.7
-19.3
Source: Victorian Electoral Commission (1996).
Table 6.2b
Legislative Council provinces: marginality after the 1999 election
Labor provinces
Chelsea
Ballarat
Geelong
Melbourne
Melbourne West
Melbourne North
Doutta Galla
Jika Jika
% swing to change
-1.2
-1.3*
-2.3
-13.9
-14.5
-14.6
-15.6
-15.8
Liberal-National provinces
Eumemmerring
Gippsland
Waverley
Central Highlands
Monash
North Western
Templestowe
Higinbotham
Koonung
South Eastern
Silvan
North Eastern
East Yarra
Western
% swing to change
-0.3
-0.3
-0.7
-3.2
-5.3
-5.7
-6.4
-6.4
-6.7
-6.8
-6.8
-10.2
-10.3
-11.9
*Note that a second Ballarat province seat was also won by Labor at a by-election. The swing required to
change this seat is -2.2 percent.
Source: Victorian Electoral Commission (1999).
67
7
Electoral Research: Spread Over Many Fields
Michael Maley
Introduction
Three issues are addressed in this paper. It is first argued that the question of the border between
‘electoral’ and ‘political’ research is really only significant for electoral administrators, and need not
cause much concern to the broader community of social researchers. Secondly, the rules of thumb
which have structured the approach of the Australian Electoral Commission (AEC) to the conduct
of electoral research are identified. Finally, it is pointed out that there are other fields of research
besides political science, and some ways in which learning from those areas has been, or has the
potential to be, usefully applied in the electoral field are identified.
Why Do Borders Matter?
For the political scientist, the question of a boundary between ‘political’ and ‘electoral’ analysis is
not a matter of concern. Wearing the cloak of academic freedom, he or she may address whatever
issues appear to be of importance. On this basis, scholars have been studying elections, as an
integral aspect of political science, throughout the 20th century.
Electoral administrators, on the other hand, have to balance a number of imperatives in deciding
how to approach research. Their primary task is to conduct elections in a manner that is ‘non­
partisan and neutral’, and is accepted as such by the communities they serve.1 Research which has
partisan implications, or which is capable of being so characterised, is a potential minefield.
Since 1984, the AEC has been empowered by paragraph 7(1)(e) of the Commonwealth Electoral Act
1918 ‘to conduct and promote research into electoral matters and other matters that relate to its
functions’. Read in conjunction with the Shorter Oxford Dictionary’s definition of ‘research’ as ‘an
investigation directed to the discovery of some fact by careful study of a subject; a course of critical
or scientific inquiry’, this is a very broad power, which would permit the AEC, if it so chose, to
engage in projects covering a wide scope.
It is, however, one thing to have a power; it is another thing to use it wisely. Although the AEC has
never adopted a formal policy on the areas which it regards as appropriate or inappropriate for
research, a number of quite clearly defined practices have developed since 1984.
●
The AEC’s research has had a strong applied focus: it has sought to address issues which are
relevant to contemporary policy on the administration of federal elections, rather than devoting
resources to more abstract or esoteric issues which might be seen as only of ‘academic’ signif­
icance. Most of the AEC’s research output has been published in the form of submissions to the
Federal Parliament’s Joint Standing Committee on Electoral Matters, which has become the
primary forum for dialogue on federal electoral policy; though AEC officers have also published
in academic journals.
●
A clear distinction has been drawn between the study of electoral behaviour and the study of
electoral systems. The AEC is not interested in the question of why people vote for one or
another candidate or party at an election, or for a particular option at a referendum: those are
1 On this, see the International Institute for Democracy and Electoral Assistance’s Code of Conduct for the Ethical and
Professional Administration of Elections, Stockholm, 1997.
68
Electoral Research: Spread Over Many Fields
matters which are best left to political scientists. It is interested, legitimately, in examining
electoral procedures and mechanisms, in looking at how votes of electors, taken as exogenous
variables, are reflected in election outcomes. These enduring questions regarding the effects of
systems can be examined outside the framework of a particular party structure, and indeed of a
particular country.
●
The AEC has sought to maintain a high level of independence and rigour in its work. The results
of its analyses, by virtue of being published as submissions to the Joint Standing Committee on
Electoral Matters, are available for all to see, and are open to criticism by those who might wish
to dissent from the conclusions reached. The AEC has often expressed views at variance with
those of the Government of the day, but has had to be prepared to do so on the basis of logical
argument and clear evidence which could stand up to close public scrutiny.
These approaches have worked well in practice. The AEC has regularly been thanked by the Joint
Standing Committee on Electoral Matters for the quality of the input it has provided. It has seldom,
if ever, been suggested that the research output of the AEC has been politically tendentious. And
most importantly, the AEC has not had to shy away from dealing with potentially controversial
subjects.
Learning Which Can Be Imported From Other Fields Of Inquiry
Because elections are inevitably political and embedded in a broader political process, there is a
tendency to give priority to considering the connections and areas of overlap between electoral
studies and political science. From the point of view of electoral administrators, however, an
election can also be considered as a purely practical and logistical exercise, and indeed, an election,
involving (as it does) putting almost the whole adult population of a country through a prescribed
process in a limited period of time, is the most extensive (if not necessarily the most complex)
logistical operation which a country ever undertakes, at any rate in peacetime. In this sense, an
election can be likened to an industrial process, and it is to fields such as operations research,
statistical quality control and applied mathematics that we must look for inspiration. A few
problems can be mentioned which arise in the electoral administration field, which have been, or
could be, susceptible to illumination from disciplines other than political science.
Coming to terms with random effects
While there is a tendency to think of elections as deterministic processes, some electoral systems
have inherently random features. One prominent example of this was the Senate scrutiny system in
use prior to 1983, which required a random selection of ballot papers to be transferred as surplus
votes. The potential for this system to produce different results if a scrutiny had to be repeated, and
the probability that it would do so, were analysed in detail by Alastair Fischer in his 1978
monograph on the subject.2 His work significantly motivated the shift to the fractional transfer
system at the 1984 election.
Much more recently, attention has been focussed on the failure of the Robson Rotation system, as
implemented at the 1998 ACT election, to randomise the ‘donkey vote’ fully among all candidates
of a party. Various approaches to the solution of this ‘glitch’ have been identified, and at least one
sophisticated analysis of the problem, based on Latin Square experimental design models, has been
undertaken.3
2 A J Fischer, The Random Sampling of Votes in the Australian Senate, Economics Department Paper No. 7, The
University of Adelaide, January 1978.
3 Ken Brewer, Fixing the Glitch in Robson Rotation - An Initiative of the Statistical Society of Australia, Canberra
Branch, Inc., mimeographed.
69
Electoral Research: Spread Over Many Fields
Probabilistic modelling of the seats-votes relationship
Modelling of the seats-votes relationship is a field in which statisticians rather than political
scientists have tended to take the lead, and no less a luminary than Francis Edgeworth was writing
on the topic late in the last century.4
In Australia, the main need to model the seat-votes relationship arises in relation to State redistrib­
utions in South Australia, because of the requirement of subsection 83(1) of the Constitution Act,
1934 that an attempt be made to produce a redistribution that is ‘fair’. In the past, analysts have
tended to place reliance on the assumption, popularised by David Butler in the early post-war
Nuffield studies, that seats will change hands in approximately the same numbers as if the swing
were uniform.5 At federal elections, this assumption is now plainly empirically false, and in any case
deterministic models of the seats-votes relationship have long since been subjected to definitive
criticism. A recognition of their inherent defectiveness has led to the development, notably by
Andrew Gelman and Gary King, of sophisticated probabilistic models.6 These have been applied in
the Australian context by Simon Jackman,7 and have been embodied in a software package, JudgeIt,
developed by Gelman and King, which won the American Political Science Association’s 1992
Research Software Award. In the redistricting processes of the early 1990s, JudgeIt was used
…in roughly half the US states by or on behalf of the federal courts, Democrats,
Republicans, third parties, nonpartisan organizations, minority groups, members of
congress, state legislators, the U.S. Justice Department, other government officials, and
private citizens.8
It is somewhat depressing that these important technical developments, which have put modelling
of the seats-votes relationship on a defensible footing, have been almost totally ignored by policy
makers in Australia; and that one instead continues to find references to uniform swing models, and
even the ‘cube law’, as providing a benchmark for ‘fair’ boundaries.
Handling queues effectively9
In the field of applied mathematics (specifically operations research), there exists an area of study
known as queueing analysis. It deals with situations in which a large number of clients have to be
provided a service in a systematic way, using limited resources optimally. It has been applied to
situations as diverse as the management of traffic at busy airports; the design of immigration
queueing arrangements; the appropriate sizing of telephone exchanges and communications
networks; and the efficient management of assembly lines. The system which we use in polling
4 F Y Edgeworth, ‘Miscellaneous Applications of the Calculus of Probabilities’, Journal of the Royal Statistical Society,
Series A, vol. 61, 1898, pp. 534-544.
5 David Butler, ‘The Relation of Seats to Votes’, in R B McCallum and A Readman, The British General Election of
1945, Oxford University Press, London, 1947.
6 Andrew Gelman and Gary King, ‘Estimating the Electoral Consequences of Legislative Redistricting’, Journal of the
American Statistical Association, vol. 85, no. 410, June 1990, pp. 274-282; Andrew Gelman and Gary King, ‘A
Unified Method of Evaluating Electoral Systems and Redistricting Plans’, American Journal of Political Science, vol.
38, no. 2, May 1994, pp. 514-553.
7 Simon Jackman, ‘Measuring Electoral Bias: Australia, 1949-93’, British Journal of Political Science, vol. 24, 1994, pp.
319-357.
8 Andrew Gelman and Gary King, JudgeIt: A Program for Evaluating Electoral Systems and Redistricting Plans, version
2.12, October 31, 1994, mimeographed, p. 4.
9 In the discussion which follows I am drawing on text which I wrote for a working paper circulated to the first
meeting (held in Stockholm in 1995) of the Election Management Bodies of the Member States of the International
Institute for Democracy and Electoral Assistance.
70
Electoral Research: Spread Over Many Fields
places at federal elections and refer to as ‘bank style queueing’ is in fact a very efficient arrangement,
known technically as a single-entry multiple server queue; it is no accident that it is now used by
Woolworths in their express lanes at supermarkets.10
The insights of queueing analysis do not appear to have been applied in any comprehensive or
readily accessible way to the specific problem of providing efficient polling facilities at minimal
cost, and decisions on polling place staffing and numbers tend to be made using a combination of
experience and intuition. If either of these fails in a particular case, the consequence can be either
excessively long waiting times at polling places, or excessive staff costs. Since polling costs tend to
be a major component of overall election costs, the potential savings associated with moving to
optimal arrangements could be considerable.
The sorts of questions to which queueing analysis can provide answers include the following
●
What is the most efficient way of organising the layout, voter flow pattern, staffing and
provisioning of a polling place, given the legal and practical constraints in place?
●
What will be the average time that voters will have to wait in order to be served?
●
What will be the impacts on (i) average waiting times; and (ii) staff costs, of an increase or
decrease in staffing, or of a reorganisation of the polling place?
As electoral administrations diversify and modernise their activities, the areas in which such
questions need to be answered are becoming broader. Essentially the same issues arise, for example,
when one is designing a telephone call centre of the type which the AEC is now using extensively at
election times. If telephone or Internet voting are ever to be seriously considered, the same sorts of
issues will arise in relation to questions of network and server capacity.
Statistical quality control
Legislation providing for citizen initiated referendums often prescribes that a relatively large
number of signatures must be provided in support of an initiative. For example, a bill introduced
some years ago to provide for such referendums in the ACT would have required signatures from
5% of the number of eligible electors at the previous election: around 10 000 people. In general it
will not be practicable to check the bona fides of such a large number of signatures, and statistical
quality control techniques, based on the drawing of inferences from a random sample, will
therefore be required. A similar problem may arise if a determination has to be made of whether a
political party which has applied for registration has a prescribed number of members.
The problem is not a particularly complex one, but must be approached within a probabilistic
framework: rather than being able to know for certain whether or not a petition (or party
registration application) is compliant, we must determine an acceptable probability of making an
erroneous decision - typically 5% or 1% - and then determine, on the basis of that probability, a
minimum number of verified signatures in the sample which will have to be reached to enabled the
petition or application to be accepted. This can be derived from the fact that the proportion of
invalid signatures in a random sample follows the hypergeometric probability distribution. For
example, if 10 000 signatures are required, and a petitioner provides 12 000, and 1200 chosen at
random are checked, then at least 1030 of the 1200 must be shown to be valid in order for one to
be 99% certain that the petition complies with the requirement.
10 For a discussion of some of these issues, see E S Buffa and J S Dyer, Essentials of Management Science/Operations
Research, John Wiley & Sons, New York, 1978, chapter 8.
71
Electoral Research: Spread Over Many Fields
Election night data analysis
The results at federal elections in Australia are known quicker now than ever before: unless an
election is a true cliffhanger, who has won government will be clear about two hours after the close
of the polls in the eastern States. This was not always the case. Prior to 1980, early results were
notoriously unreliable, as it was impossible to tell whether they were providing evidence of a
particular level of swing, or simply constituted an unrepresentative sample of the electorate.
Statisticians, again, led the way in providing solutions to the problem. Ross Cunningham used
previous data of partial counting in each division to produce ‘bias curves’ which could be used to
correct for bias, producing much more useful early figures. His methodology was applied with
striking success in the Ten Network’s 1980 federal election coverage.11 In 1986, he undertook an
analysis for the AEC which demonstrated that even greater improvements could be achieved by the
implementation of the now well-known ‘matched booths’ technique. That step, taken in 1990,
constituted the most important quantum improvement in the analysis of election night data in
Australia.12
Demographic projections for the purposes of redistributions
The introduction in 1984 of the ‘future enrolment equality’ requirement as a major criterion
governing federal redistributions has led to a new need for accurate demographic forecasts. These
have been produced using techniques the sophistication of which has steadily increased over time.
The AEC now uses a cohort-component method, developed in consultation with demographers at
the Australian Bureau of Statistics, which is widely accepted as the most accurate population
projection methodology.
Other prospective areas of research
As technology continues to develop, the areas of research of which electoral administrators ought to
be aware are becoming ever broader. The following spring to mind as fields which are worthy of
further investigation.
●
Developments in the field of optical character recognition, especially handwriting. As long as
Australia retains full preferential voting, any prospect of the introduction of electronic scanning
of ballots is likely to depend on progress in this area.
● Use
of biomechanical identification systems for prevention of impersonation. Procedures of this
type are already in widespread non-electoral use - for example on the doors at the AEC’s Central
Office in Canberra - and it is certainly worthwhile for electoral administrators to keep the practi­
cability, costs and benefits of such systems under review.
●
Electronic signature matching. Mechanisms by which signatures on declaration votes could be
electronically compared with digitised signatures captured at the time of enrolment would have
the potential to revolutionise the preliminary scrutiny of declaration votes.
●
Applications of cryptology to Internet electoral processes. If Internet voting is to become a reality,
it will inevitably have to involve the use of state-of-the-art data encryption techniques. Electoral
administrators will increasingly have to be prepared to acquire an understanding of issues such as
electronic signature definition and public key cryptography.
11 R B Cunningham and K W J Malafant, ‘Forecasting Outcomes of Australian House of Representatives Elections on
Election Night’, The Mathematical Scientist, vol. 7, 1982, pp. 105-14.
12 Michael Maley and Rodney Medew, ‘Some Approaches to Election Night Forecasting in Australia’, Australian
Journal of Political Science, vol. 26, 1991, pp. 51-62. For a discussion of approaches adopted to election night
forecasting in a number of different countries, see the special edition of The New Zealand Statistician, vol. 25, no. 2,
December 1990.
72
THE BENEFITS AND DIFFICULTIES OF COLLABORATION BETWEEN
ELECTORAL ADMINISTRATORS AND ANALYSTS
73
8
Fostering Cooperation Between Electoral Authorities And Analysts
Gerard Newman1
Background
The Department of the Parliamentary Library is one of five separate departments dedicated to
providing services to the Commonwealth Parliament. The other departments are: Department of the
Senate, Department of the House of Representatives, Department of the Parliamentary Reporting
Staff (soon to be renamed Department of Parliamentary Communication Services) and Joint House
Department.
The Parliamentary Library’s mission is to support the parliamentary process by providing Senators
and Members with quality information services, analysis and advice. The Parliamentary Library sees
its role as being the principal provider of such services to the Commonwealth Parliament as a
whole. The library’s clients are the 224 members of the parliament, their staff, parliamentary
committees and parliamentary staff. The library has a budget of $15.1million and a staff of some
184 (mostly employed at professional levels).
The client service arm of the library is the Information and Research Services Program, which is
divided into seven subject groups. These are:
Economics, Commerce and Industrial Relations;
Foreign Affairs, Defence and Trade;
● Law and Bills Digest;
● Politics and Public Administration;
● Science, Technology, Environment and Resources;
● Social Policy;
● Statistics.
●
●
The library satisfies client requirements by providing both an ‘on demand’ client service - where
clients contact the library with specific questions, and a pro active service - when publications and
material are prepared in anticipation of client interest. In 1998-1999, the library responded to some
32 000 information and research requests and produced 75 separate publications, not including
212 Bills Digests and twelve issues of an economic indicators publication.
Information Requirements
The library’s requirements for information and data reflect the Parliament’s information and
research demands as manifested to the library. Some of these requirements are discussed below:
●
Accuracy - The information provided to members is often used by them in the chamber or in
some other public forum, or is used to develop public policy options. Thus it is imperative that
any information provided by the Parliamentary Library must be as accurate as possible.
1 The views expressed in this paper are those of the author and do not necessarily reflect those of the Department of
the Parliamentary Library.
74
Fostering Cooperation Between Electoral Authorities And Analysts
●
Timeliness - Members operate in a time sensitive environment where information that is
provided late is often worthless. The focus of much of the library’s work is to meet deadlines with
some form of response to the client’s question. The library does not have the luxury of being able
to devote unlimited time to preparing the definitive answer.
●
Free from bias - Members are constantly being bombarded with fact and opinion from sources
that have an interest in promoting a particular point of view. This is a legitimate part of the
democratic process but makes it important for Parliament to have access to information and
advice that is free of bias. The Parliamentary Library is one of only a few providers of such
information and advice.
●
Confidentiality - Members operate in an environment where confidentiality is of paramount
importance. When seeking information on behalf of a client the library never divulges the client’s
identity. Similarly the library never divulges any aspect of a client’s request to another client.
●
Responsiveness - With the vast amount of information and material that members have to deal
with, it is important that the responses provided by the library are suited to the client’s
requirements. If a client only wants a brief fact sheet then it is critical that the library responds
accordingly; similarly a request for a detailed policy options paper must be responded to in kind.
Being responsive to the client’s requirements can sometimes mean that the library is asked to
provide slanted advice or material that supports a particular point of view.
The above are some of the more important requirements for information and advice that Parliament
places on information providers like the Parliamentary Library. As electoral authorities face similar
demands from their clients (including the various parliamentary libraries) these requirements
should be very familiar. In addition to the above there are some further requirements for
information and data that parliamentary libraries place on electoral authorities. These are:
● Accessibility
- The Internet has lead to a quantum leap in our level of accessibility to electoral
information. However, problems with the regularity of updates to Internet data and the
availability of some Internet sites can be frustrating, particularly when user expectations have
been raised. Access to electoral information on the Internet, however, should not preclude access
to information by other means, including email, CD-ROM and hard copy.
●
Comprehensiveness - Our demand for electoral information, particularly election result data, is
virtually insatiable and must place a considerable burden on electoral authorities in meeting this
requirement. Whether it be the results of an election at the polling place level, the distribution of
preferences of an excluded Senate candidate, or the names of candidates in some forgotten byelection, our expectations are that these and other requests for information will be able to be
answered by the appropriate electoral authority.
●
Consistency - One of the problems in dealing with data is the need for consistency, both
internally and over time. Nothing destroys the perceived accuracy of data than a lack of
consistency.
●
Cost - With the ‘user pays’ principle pervading all aspects of government administration it is no
longer possible to expect electoral authorities to make vast amounts of information freely
available. However, in the true traditions of special pleading, electoral authorities are asked to
look favourably on requests for data.
75
Fostering Cooperation Between Electoral Authorities And Analysts
Examples Of Data Use
In an attempt to demonstrate to electoral authorities how electoral data can be used, some examples
are now provided of the recent usage of such data by the Parliamentary Library.
Thematic maps - Members of parliament are particularly interested in the demographics and
voting behaviour of their constituency. Demographic data is available at the Census Collection
District (CCD) level from the five-yearly Census of Population and Housing conducted by the
Australian Bureau of Statistics. CCDs are the smallest geographic spatial area for which census data
are available and are the building blocks for larger geographic areas eg electoral divisions. However,
demographic data by itself is only part of the picture and needs to be used in conjunction with
voting data. Unfortunately voting data is only available at the polling place level and some means
has to be found to convert voting data to collection districts or vice versa. The solution to this
problem is to be found in the catchment area analysis conducted by the Australian Electoral
Commission (AEC) after each federal election. From the marked roll generated on polling day the
AEC is able to estimate the number of voters from each CCD voting in particular polling places.
The AEC has been kind enough to supply this data to the Parliamentary Library for research
purposes. The data is then used by the library to estimate the voting behaviour for each CCD in
Australia. The results can be displayed in the form of a thematic map and can be used with the
census data for statistical analysis.
Electoral redistributions - With electoral redistributions being a frequent feature of the political
landscape, there is considerable interest in the electoral consequences of such redistributions.
Indeed the first question asked when redistributions proposals are announced is ‘What are the
electoral consequences?’. Electoral authorities by virtue of their role in the process are reluctant or
indeed unwilling to be involved in this type of analysis and it is left to bodies like parliamentary
libraries, the political parties or electoral analysts to undertake the task. The Parliamentary Library
utilises two different methodologies for this analysis – both rely on data supplied by the AEC. The
first method involves the allocation of polling places to the new divisions and the subsequent
aggregation of the polling place results and pro-rated non-ordinary votes to obtain the new
divisional results. The second method is more sophisticated and utilises the catchment area data
referred to above. With electoral authorities using computer based mapping systems for the
drawing of electoral boundaries and basing the enrolment calculations on CCDs, a listing of the
CCDs in the new divisions is an important by-product of the process. This listing can be used to
estimate the voting pattern of the redistributed divisions.
Cooperation
The above examples provide an illustration not only of how electoral data can be used by analysts
but also the level of cooperation between electoral authorities and analysts, in this case the AEC and
the Parliamentary Library. The level of cooperation, both formally and informally, between the two
types of organisations is high and hopefully of mutual benefit. Some examples of cooperation and
the benefits that can accrue are now outlined:
● Dissemination
of information - Analysts can assist electoral authorities in the dissemination of
electoral information. They can relieve authorities of some of the burden in providing electoral
information and can provide a filter for information requests.
76
Fostering Cooperation Between Electoral Authorities And Analysts
●
Collaboration - Electoral authorities and analysts can collaborate on joint projects that can be
mutually beneficial. The Electoral Atlas produced jointly by the AEC and the Parliamentary
Library is a good illustration of the fruits of collaboration.
●
Overcoming administrative restrictions - Electoral authorities by the very nature of their place
in the political process are restricted in the type of analysis and research they can undertake.
Analysts on the other hand operate in a much freer environment and can conduct the research
that authorities are unable to perform. The above-mentioned use of AEC data to estimate the
effects of electoral redistributions is a good example of this.
● Feedback
- Analysts can provide a valuable source of informed feedback and critical comment
that can assist electoral authorities in their role. It can be lonely on the ‘pedestal of impartiality’
and analysts can assist electoral administrators by providing an objective sounding board for
ideas.
●
Facilities access - Organisations like parliamentary libraries can assist electoral authorities by
making facilities more readily accessible. For each federal redistribution the AEC makes available
to the Parliamentary Library a copy of the electoral boundaries mapping software used to conduct
the redistribution. The Parliamentary Library has assisted the AEC by making this facility
available to members.
Conclusion
The above are just some of the many examples of successful cooperation between electoral
authorities and parliamentary libraries. Cooperation can provide mutual benefits to all those
involved in the process. The level of cooperation between the two types of organisations is already
at a high level and can only be enhanced by occasions such as conferences.
77
9
Developments In The Provision Of Electoral Information
Andrew Hawkey and Michelle Davy
There are clear benefits for electoral authorities in building a sound relationship with electoral
analysts. Elections are the only significant way the ‘political will’ of the people can be expressed.
The work of electoral analysts increases public awareness and understanding of the election, which
in turn leads to positive outcomes such as process transparency and increased participation in
various ways. For an electoral authority, it is crucial that elections are not only conducted in a fair
and accurate manner, but that they are seen to be so. Elections that meet these criteria assist in
producing public confidence in the democratic process and the legitimacy of the elected body.
Modern technology has facilitated a new level of access to more complex election information for
the analyst and the interested individual. It is also changing the way electoral authorities provide
information. In recent years, all Australian electoral authorities have produced election Internet
sites. At the 1999 Commonwealth Referendum, the traditional tallyroom was replaced by a ‘virtual’
tallyroom, affecting not only the process but also the ‘culture’ of elections.
Analysts’ findings can also provide support for changes to existing electoral systems. For example,
the Hare-Clark electoral system has been enhanced on a number of occasions since its inception
partly as a result of analysts’ comments.
Available Electoral Resources
Electoral authorities produce a variety of electoral data and information from various sources,
including enrolment, electoral boundary and election data. Legislation and policy, which differ
between jurisdictions, regulate the provision of this data.
Enrolment data is provided under law to members of parliament, candidates, parties and the
public. The amount of data and the method of providing that data varies depending upon the
legislation.
Election information ranges from roll close enrolment figures to details of nominations to distri­
bution of preferences. However most attention centres on election night results. Information
provided on that night usually includes electorate figures, percentage counted, polling place figures,
two party preferred results and electorate or polling place swings.
All authorities publish election reports, which include core election statistics and results.
Authorities also produce historical comparisons and other information papers relating to various
issues. In addition, specific requests for particular data are dealt with, where legislation and office
resources allow.
Trends In Data
Demand to improve the mode and speed of delivery through modern technology is changing the
presentation of election information.
78
Developments In The Provision Of Electoral Information
Following the 1998 Tasmanian State election, results were available on the Internet within
15 minutes of figures being verified for downloading.
Parliamentary Reports are sent to the printer in electronic form, having been fully formatted and
structured within the office. Any of the information contained in these reports can therefore be
provided to analysts on disk, CD-ROM or sent by email. Prior to 1994, these reports were only
available in hard copy.
Internet
The use of the Internet is ever increasing and – as was the case at the recent Victorian election –
becoming an integral part of election administration as well as information distribution. At this
election, each returning officer was linked electronically to the central office. All election databases
were stored centrally, which facilitated a consistent and efficient approach and enabled prompt and
accurate production of election information. The Victorian Electoral Commission also developed a
special Internet application, which enabled parties to receive specifically detailed election results at
their preferred venues on election night and during the scrutiny period.
The Internet provides the opportunity to give even greater access to historical election information
such as lists of members of previous parliaments, candidates, scrutiny sheets and sample ballot
papers from previous elections. This information will provide a valuable tool for analysis and
comparison with current trends.
The increased use of Internet facilities may initially bring added pressure to electoral office
infrastructure and human resources. However once the infrastructures are established, increased
efficiencies, such as a reduction of the use of telephone inquiry services, will result.
Mapping
The last few years have seen the development of computerised mapping systems for use in electoral
boundary redistributions. This technology incorporates geocoding of electoral and geographical
boundaries with census collection districts (CCDs).
Other data such as polling place results can be overlaid to visually combine boundary and other
political information, providing an innovative tool for presenting information.
Computerised Scrutiny
The development of computerised scrutiny programs provides a new range of data for authorities,
analysts and parties to examine. Using these applications, elector preferences marked on ballot
papers are data entered into the program. Once all ballot papers preferences are entered, the
program conducts the full scrutiny, identifying and moving every ballot paper where necessary.
These programs have the advantage of providing an audit trail of the scrutiny, in printed and
electronic form.
Computerised scrutiny programs are used in Senate elections, and the New South Wales, Western
Australian and South Australian upper house elections.
79
Developments In The Provision Of Electoral Information
The data capture of the complete set of ballot papers cast at an election provides a new and
bountiful resource for election analysis. Research surveys such as informality, donkey voting, how
many preferences are cast in an ‘optional preferential’ election, cross party voting or voting trends
such as preferences given to women candidates, may now be feasible.
The availability of this new resource raises a number of issues. Firstly, every preference on every
ballot paper can be provided to the analyst. Each ballot paper can also be traced back to the original
polling place. This provides the electoral or party analyst with extraordinarily detailed information.
If, in the future, access were to be given to information as to who voted at particular polling places,
the ability to overlay this information with other sets of personal and political data could be
perceived as beginning to erode the ethos of the secret ballot.
Similarly, the ability to know all of the preferences of the ballot papers received by any candidate
could be analysed to determine the result of a recount following a casual vacancy. This could be
seen to open the door to accusations of manipulation of the democratic process. An example is if a
particular candidate (not necessarily one who would be successful) was requested not to stand at a
recount to enable another to win.
For those states that have ‘ticket voting’ as an option, where only the preferences from a relatively
small number of ballot papers need to be individually entered, there are significant cost and time
saving advantages. For other jurisdictions such as Tasmania, the absence of a ticket vote combined
with optional preferential and Robson Rotation of candidates’ names on ballot papers, would
considerably increase the cost of scrutinies and could delay the election result.
Use Of Data
Electoral roll data is arguably the most consistently up-to-date database of Australian citizens and
where they reside, and therefore is much sought after for a variety of other purposes, including
medical research, law enforcement and social surveys. The primary reason for the creation and
maintenance of the electoral roll is for the conduct of elections, and its use for other purposes raises
privacy issues. Provision for non-electoral purposes is prescribed in legislation, or at the discretion
of the authority, if permitted by electoral and privacy legislation and policy.
Generally, there are no formal restrictions on the use of election data. There may be benefits in
having some mutually agreed guidelines. Ideally, it is considered that an electoral authority should
be entitled to have the information it provides to analysts used in a responsible and accurate
manner. In turn, the analyst should be entitled to accurate, relevant and timely information.
All published election figures are usually sourced to the electoral authority. If data is not properly
analysed and leads to improper conclusions this could not only damage an authority’s reputation,
but also it could possibly lessen the integrity of the whole democratic process.
Building Our Relationship
It is important to give a high priority to developing a positive relationship with election analysts.
Given the complexities of and differences between electoral systems, it is particularly important that
analysts and the media present correct information and election implications. A factor of concern is
that quite a number of media commentators may be covering their first election.
80
Developments In The Provision Of Electoral Information
A number of strategies have been adopted to develop this relationship. Informal information
sessions are usually held prior to major electoral events to provide timetable details, a run-down of
the advertising and public awareness campaign, the types of and access to election information and
statistics throughout the event, and various contact people within the organisation. This is also an
opportunity for the media to request particular statistics. For example, in response to a request
from Tasmanian newspapers, the Tasmanian Electoral Office (TEO) provided daily figures and
graphs of returns of postal votes during the postal period for each council at the 1994 local
government elections.
In addition, informal training sessions are held from time to time on new and current election
processes, and their possible implications. For example, in the weeks prior to the 1998 Tasmanian
State election polling day, the TEO conducted sessions for journalists explaining the Hare-Clark
process, how to interpret election night results and discussing possible implications resulting from
the change in the number of members in each division from seven to five.
Post-election meetings provide an opportunity for feedback from media and analysts, to receive any
new requests for information arising from the election, and to outline when post election
information (eg the election report) will be available. Following the 1999 Tasmanian upper house
elections, where one member was elected by a margin of only 65 votes, a request for information on
any previous ‘close’ upper house elections was received. Research found that the margin was the
third smallest in magnitude, and the second smallest with respect to the percentage of the formal
vote.
Informal ongoing relationships with analysts could be further enhanced by such actions as the
periodic release of information on electoral activities and dates (eg quarterly publishing of
enrolment figures).
The TEO also supports conferences as another positive step in opening up communication and
understanding of each other’s resources and information needs.
81
10
Living In Interesting Times
Jenni Newton
South Australians have been interested in electoral matters for a long time.
South Australia introduced the secret ballot in 1856 – a month after Victoria and only the second
colony to do so. By the time that the various colonies voted to federate in 1899 and 1900, women
in South Australia as well as in New Zealand and Western Australia could have their own votes
recorded.
More recently, after the South Australian State election of 1989, in which the Liberal Party received
52% of the two party preferred vote but failed to win a majority of the seats, the Constitution Act
was changed to require a review of the electoral boundaries after each State election, and - perhaps
more importantly - the changes required the South Australian Electoral Districts Boundaries
Commission to take votes into account when making any boundary changes.
The South Australian Electoral Districts Boundaries Commission is required to draw State electoral
boundaries, as far as practicable, in such a way that they are:
…fair to prospective candidates and groups of candidates so that, if candidates of a
particular group attract more than 50 per cent of the popular vote (determined by
aggregating votes cast throughout the State and allocating preferences to the necessary
extent), they will be elected in sufficient numbers to enable a government to be formed.
[Constitution Act (SA) s.83 (1)]
Once the Electoral Districts Boundaries Commission (EDBC) was required to check the electoral
effect of its redistribution two things happened: electoral data became available in more useful
form, and debate about the whole process of redistribution encouraged research.
Three redistributions have been undertaken since the change to the Act (1991, 1994 and 1997­
1998) and each of the Commission Reports has included very detailed comments on the
Commission’s assumptions and conclusions. The relatively short time-frame between redistrib­
utions, and the Commission’s willingness to publish detailed explanations have meant that there is
now a continuing dialogue about how best to fulfill the requirements of the Act. An idea expressed
in one Report can be followed up and explored in time for the next hearings.
Although it may be felt that the EDBC sits a bit outside play as far as electoral research for the
jurisdictions is concerned, it would be wrong to underestimate its impact on electoral research in
South Australia during the 1990’s.
With regard to collaboration between the electoral authorities and researchers or commentators,
that collaboration implies working together on a project, aiming to achieve some common
objectives - and perhaps some individual ones as well. This author has collaborated on at least two
ventures with the State Electoral Office (SEO) here in South Australia. The first was a study of
donkey and reverse donkey voting at State elections, and the second was compilation of a complete
82
Living In Interesting Times
set of all candidates who have ever stood for election to either House of the South Australian
Parliament. Opportunities to work on collaborative projects have not presented themselves very
often, although this may not be the case in other states or systems.
There has been a lot of co-operation. From the author’s point of view this has been fairly one-way in
that a lot more has been asked of the SEO than it has asked of the Parliamentary Library. Some of
this work has been of use to the SEO too – for example a summary of electoral legislation since the
1850’s, which was conducted by the Library for a select committee some years ago.
Finally, both analysts and SEO staff have worked together quite separately to assist another body –
the EDBC, in particular at their 1991, 1994 and 1997-1998 enquiries. The reason why this work is
seen as a kind of collaborative effort is that although quite different roles were adopted, the same
aim was shared, namely to provide information that might help the Commission to achieve that
very difficult requirement of drawing boundaries which could produce a fair result at a subsequent
election.
The SEO may outline the benefits they believe they receive from working with analysts, but the
benefits received by the author can be noted.
First and probably of most value is the access to data that is not available to people working outside
the electoral administration. An example is the examination of actual ballot papers after an election
to look at donkey voting. This enabled a fairly common assumption held by State MPs to be tested,
namely that being higher on the ballot paper than one’s major-party rival confers an advantage of
up to 2%. The SEO interest may have been in whether there was a consistent level of interference
by the electoral law (which requires a full statement of preferences) on voters’ actual preferences.
What was found was a highly inconsistent level of donkey voting across the seats investigated.
A second advantage of working closely with the SEO comes from free and open communication.
This has worked in the author’s favour many times. Knowing that data exist can often be a great
springboard. Ten or fifteen years ago when the final two candidates in a seat were not represen­
tatives of the ALP or the Liberal Party there was no recount to determine an ALP: LIB two party
preferred vote in that seat. Later, recounts were done but not published; now recounts are done and
published (that applied to 11 of our 47 seats at the most recent election). In the meantime, just
knowing that recounts were done, and being able to have access to unpublished data, was a
tremendous help later when it came to accurately representing the effects of a boundary change on
affected electorates.
So far only work that has been done in conjunction with the SEO has been mentioned. This is
partly because members of a state parliament are more interested in research which uses state data.
But the experience of requesting and working with Australian Electoral Commission (AEC) data has
been most interesting. First of all the Divisional Returning Officers (DROs) have been fantastic. At
the 1993 federal election the AEC required a two party preferred vote to be counted out for each
booth. This was exciting because the intention was to show how the voters in the proposed State
seats had voted at that federal election of 1993 (specifically because it had already been calculated
how these same voters had voted at the State election of 1989 on the proposed boundaries, but it
was not felt that the 1989 election was any guide as to how South Australians might vote at a State
election due later in 1993). Two party preferred results for each booth, rather than for each
enormous federal division could potentially make the estimates much more accurate when chunks
of one state electorate needed to be moved into another. The other reason why these figures were
83
Living In Interesting Times
enticing, was that for the very first time it would be possible to assess an assumption that
preferences would have been allocated in the same pattern at each booth as they were across an
entire electorate. These figures could show how far out the estimates had been in the past.
Each Divisional Returning Officer was telephoned before and after the federal election, and every
one of them sent unpublished two party preferred counts, by booth, for their divisions, a gesture
which was much appreciated. It may be that the interest in the two party preferred booth figures
was also of interest to the DROs. The AEC would have explained the importance of counting out a
two party preferred figure for each booth, but it must have involved a considerable increase in work
for the staff out at the booths on election night. To have an outsider showing interest in the figures
may have helped to reassure staff that the extra effort was important.
A less rewarding episode occurred when it was indicated that the AEC had data showing voters’
suburbs of residence with the booth at which they lodged a vote. This data could throw light on
another assumption that had been used, namely that electors use their local booth (and therefore
moving the booths in an electorate would be a good proxy for moving electors). The AEC in
Adelaide was contacted and the request passed to their media liaison office, but the media people
did not understand what was wanted and there was no access to the research or technical people
who could have helped. No response was received to a written request which was frustrating. South
Australia was placing enormous emphasis on the fairness of electoral redistributions, even ordering
new boundaries after every State election, but no access to data was available which could throw
light on one of the key assumptions of the estimation process. Some time later it transpired that the
SEO had the data and could run off a copy. The data showed that about 80% of electors voted at
their local booth - more in country areas.
The main conclusion to be drawn from this experience is that the more contact analysts have with
electoral administrators and research staff, the more we will be able to help each other. We are not in
competition with each other, nor are state results and research irrelevant to the federal situation
(and vice versa). On the contrary, the work that has been done for the EDBC here could be helpful
to people in the federal arena or in other states. Conferences are one of a number of ways that can
increase communication between research and administration people. Individuals need to speak to
each other; if we go through intermediaries it can be expected that the message will be distorted or
have someone else’s priorities imposed on it.
What about the future? There is so much exciting work waiting to be done. Wafting in the wind
here in South Australia is a collaborative project reviewing our electoral history, but it seems that
the change to counting votes by data input processes may open up a whole new range of
possibilities. Will analysts (perhaps in collaboration with SEO staff) be able to ask for stored data
files from say the 1997 State election? It may be possible to look at the results at one election
according to different methods of counting the votes. For example, what if South Australia did
adopt a West-German style Mixed Member Proportional (MMP) top-up system? What if a New
Zealand style MMP count was tried? What if an optional preferential vote was used instead of full
preferential voting?
This work is relevant because South Australian members may not be content to have their
electorates changed radically every three or four years. It is too disruptive and downright dangerous
for some of them. There may be a move towards multi-member electorates before too long, and that
will generate some very interesting work.
84
Living In Interesting Times
It may be that the process of gaining access to actual voting data will take a while, although this
should not be a difficulty. The difficulty will lie in the fact that two sorts of queries are received
from members. The first kind of request relates to electorates throughout the State, and usually the
member would be happy to waive the rule that the work is kept confidential and would allow
distribution of the results to other members or to a pool of electoral research. Anticipatory work is
included in this category, questions such as: What’s it worth to be a sitting member? What’s the
donkey vote worth? How did the voters in the proposed State electorates vote at the recent federal
election?
The second kind of query is more political, in that it relates more to an individual member’s own
interests or agenda, and the member is not likely to want the results widely distributed. The
confidentiality of this work would not allow the results to be contributed to a pool of electoral
research, although it must be said that this work is generally so restricted in scope that it would not
be of great interest to other researchers. (Queries could include: What percentage increase would
my party need in order to get an extra member elected to the Upper House? What have been the
results for a given geographical area at the last few State and federal elections? If the current public
opinion poll is correct, which seats would my party be likely to lose at an election?)
Returning to the possibility of having access to voting data on disc or CD-ROM. Perhaps the SEO
will see the data just as information, on a more detailed basis but comparable with the figures
currently published after each election. On this reasoning, if data could be accessed, it could be
used for any query at all, and on this basis the SEO would probably be willing to release the data to
political parties as well.
On the other hand, the SEO could be happier not to release the data but to allow analysts to use it
in collaboration with office staff. Estimating the result for a given election if a different method of
counting had been used might be one area where this collaboration would work well. But collabo­
ration would limit the use of the data to those requests from MPs (or anticipatory work) where the
question relates to electorates throughout the State. It is doubtful if collaboration could be
requested on a more political query which related to a single member or a single party: eg a query
such as how optional preferential voting might advantage independents. Collaboration on a query
of this sort could be difficult.
If analysts could have access to voting data, for some queries but not for others, would that place
the SEO in a difficult position in relation to the Parliament and the parties? And if given the data
but only for a price, would the cost be prohibitive? Who would pay?
All of these questions will more easily be resolved in a climate of goodwill, and if results are shared
as much as possible. In South Australia, we are certainly off to a good start.
There is an ancient Chinese curse: ‘May you live in interesting times’. We have been blessed to live
in electorally interesting times.
85
86
TECHNOLOGY – IMPLICATIONS FOR THE INDUSTRY
87
11
Geographic Information Systems In Political And Electoral Work
1
Rod Medew
Introduction
The area of computer science is moving ahead at a greater rate than the industrial revolution, and
the possibilities for the use of computer technology in problem solving now seem to be endless. It is
impossible to say with any certainty what technological innovations we will be faced with 12
months from now. Nowhere has prognostication been so rapid and so explosive as in the field of
computer technology. Within this field the future has been adopted and accepted more rapidly
than in any other form of change in this century.
One expanding field of technology is that of Geographic Information Systems (GISs), which are
increasingly being used by electoral bodies for the planning of electoral boundaries. There are now
vast amounts of geospatial data in the world, which can assist in solving spatial allocation problems,
and can be applied in data analysis that can be used strategically both politically and in the electoral
world. In 1994 President Clinton established the Federal Geographic Data Committee to develop a
National Spatial Data Infrastructure (NSDI), which when complete will contain the world’s most
comprehensive spatial data warehouse. The question we face is that of how can we use these types
of data more effectively.
The GIS is a computer system capable of assembling and storing geographically referenced data; but
it also should be thought of as including the people who operate the equipment. Karl Marx would
arguably have seen the GIS as a technological whole: data, computer, information, all intensely
interactive with humans, and deterministic. There has been a substantial growth in GIS use, and the
software itself is constantly increasing in its value for planning and problem solving. As more data
become available to us, we must discover methods and technologies that will assist us in converting
them into useful information. This paper argues that GIS and related technologies will assist us in
the analysis of large volumes of data, allowing us to better understand electoral behaviour, and to
better manage the terrestrial process of electoral delimitation.
Management Information Systems
In the electoral and political areas, organisations continue to collect vast amounts of data, which
contain diverse forms of information. Continuing technological innovations make it possible to
store this amount of data, and to have built-in query tools that end users can use with no computer
programmer input. These intuitive ‘front ends’ that connect to large databases have been named
Management Information Systems (MISs), allowing users to print reports and do complicated
searches on the database. In its simplest form, the MIS is a sophisticated database with a click-on
screen (controlled by a mouse) that can be used by people with little computer knowledge.
However in the area of political, social and electoral research, it is important that these databases
have a spatial dimension, as all electoral governance is axiomatically spatially referenced. On a dayto-day basis, electoral practitioners make decisions based on spatial consequences. The GIS thus
complements the MIS.
1 The views expressed herein are purely those of the author and do not necessarily reflect the views of the Australian
Electoral Commission.
88
Geographic Information Systems In Political And Electoral Work
Definition Of A GIS
In its basic form the GIS system consists of a computer, and a program that will display, manipulate
and analyse spatial data in the form of maps. ‘The GIS engine can be described as a collection of
computer hardware, software and geographic data that is designed to efficiently capture, store,
integrate, update, modify, analyse, create and display geographically referenced data for use in
business process’ David Maguire (1995). The ability to display maps and add features, such as
polling places and census data, clearly has a number of benefits that may help lead us to a utopian
electoral environment. The use of GISs and their application to make life easier for electoral and
political researchers is already here; we are simply not using it!
One of the main advantages of using GIS technology and the corresponding spatial data is the
physical reality that is represented: objects appear to be very real and reflect social, legal, economic
and political realities. This is especially necessary for work in redistricting. It is through the use of a
GIS that we can codify the social reality constructed from our direct experience in the physical
world.
Redistricting
To date the only practical application of GISs in the electoral field in the Australian context has
been in redistricting. The redistricting criteria vary between the States and Territories of Australia
and the Commonwealth. However in most cases the primary redistricting criterion is equality of
representation, expressed demographically by the number of electors in each electoral district. The
criterion is typically applied to both present and future demographic patterns within each electoral
division. The Commonwealth Electoral Act 1918 (CEA) prescribes that at the time of a redistri­
bution all federal divisions must have enrolments within 10% of the State average enrolment.
Further, the CEA specifies that in general an attempt must be made to ensure that, 3 years and 6
months from the time of the redistribution, the number of electors in each electoral division will
not deviate from the State average enrolment by more than 3.5%. There are other criteria which are
subordinate to the main mathematical criteria: community of interest, means of communication
and travel, physical features and area, and current boundaries. The electoral divisions are
constructed from small geographic units - Census Collection Districts (CCDs) - and for each CCD,
data are available on the current and projected number of electors, and other demographic and
census variables. The use of census data is often valuable when comparing the number of electors
with a population cohort group, such as Australian citizens 18 years and over. In terms of the other
criteria for redistricting, Walmsley (1985) has argued that social indicators may assist in
determining communities of interest.
Passive Redistricting
The use of GIS technology can greatly assist the user in determining electoral boundaries which
satisfy the above criteria. The GIS data consist of the CCDs, amalgamated to form the current
electoral boundaries. By moving a number of CCDs from one proposed division to another, not
only will the user be able to see quickly the physical change to the proposed division, but also the
effects of the shift on the mathematical constraints which must be satisfied. Use of census data can
also assist in assessing community of interest, and transport costs. This method of enabling the user
to modify and analyse the result, I have named passive automated redistricting. The Australian
Electoral Commission (AEC) uses MapInfo as a base, and has developed such a system. The system
89
Geographic Information Systems In Political And Electoral Work
runs on a relatively simple PC and has been successfully used by the AEC in federal redistricting. It
is clear that by using such a system alternative proposals can be identified and analysed, much
more efficiently and quickly than was previously the case. Some might see a downside to this: it is
now easier for the user to create a number of alternative proposals all of which are mathematically
and geographically correct, which has the potential to force committees to agonise over which of
the multiple solutions they will propose. Prior to the use of GISs, the user laboured until a single,
and in most cases, final solution was calculated.
The use of GIS technology also has the additional advantage of being cartographically correct, and
consequently maps can be produced with details such as roads, rivers and cultural features. Such is
the case with the system, Electoral Distribution and Mapping System (EDAMS) built by the Land
Information Centre (LIC) in Bathurst, NSW. The system has been designed to meet the
cartographic requirements of redistricting within Australia, as well as passive redistricting as
described above. Clifford (1998) argues that redistributions consist of two fundamental parts, ‘the
first being the creation of electorates and the second the production of maps. The former need is
met by EDAMS providing on-screen spatial decision support through the display of a compre­
hensive spatial data set consisting of topographic and cadastral data’. In order to provide the
cartographic requirements of redistricting committees, the system developed by LIC displays a
detailed map created from the topographic and cadastral data sets maintained by the state mapping
authorities. The system consequently has the ability to produce high quality map sets during
redistricting.
Non Passive Redistricting
Both the AEC system and EDAMS are passive in their use, and require human intervention in order
to determine new electoral boundaries. The system pioneered by Horn used a non-passive
approach to redistricting (Horn, 1995). Horn codified the electoral redistricting criteria, and
developed algorithms based on the principles of stepwise transformation, which involve subjecting
a starting plan to a series of small changes until a final solution has been resolved. A random
number initiates the system, and the user identifies the optimisation and compactness criteria. The
system creates the best solution based on the criteria identified by the user. While the measure of
compactness has little place in Australia, it has been and is an increasingly used criterion in the
United States of America. Horn codified the compactness measurement described by Morrill
(1991), and intended to use it as a defence against gerrymandering. While the system did automat­
ically draw reasonable boundaries that fulfilled the mathematical criteria, it did not allow for the
spatial geographic realities such as mountains and rivers and redistricting committees were forced
to modify the automated boundaries. However Horn’s pioneering work in the field of redistricting
remains unparalleled, and little work has been devoted to the subject of non-passive redistricting
work since the late 1980s. In this area there is scope for electoral and political research.
Electoral Administration
GIS technology has the potential for use beyond redistricting. In Australia, voter registration or
elector enrolment is compulsory for all Australian citizens aged 18 years and over. The number of
people enrolled continues to grow. In 1901 only 984 754 people were enrolled (Jaensch 1995)
compared to a total of 12.2 million enrolled for the November 1999 Referendum. As the numbers
of eligible electors increase, there is an increasing need to ensure all eligible cohorts are enrolled. It
is here that the GIS can be of great assistance to electoral administration.
90
Geographic Information Systems In Political And Electoral Work
A significant technological aspect of GIS is the emerging ability for users with little exposure to
databases and codifying business rules to be able to load data into a database and perform advanced
analysis. Such data can represent, for example, vacant house changes in enrolment patterns,
obtained from any number of sources. In the Australian context these data are collected at the
lowest possible geographic level such as CCD. The data can then be represented in a pictorial form,
or detailed analysis can be performed using built-in tools. Such data and representation would be of
invaluable benefit to any electoral manager in his or her day-to-day work.
Another application is the identification of optimal locations for polling places, using locationallocation algorithms to ensure the deployment of polling places over the electoral division to
satisfy a geographically based pattern. In other words, the GIS can be used to assist the user to
locate the polling place in the most appropriate location to serve the maximum number of voters.
The area the voters are attracted from is named the ‘catchment area’, and represents the geographic
area (in terms of CCDs) from where the maximum number of voters is attracted to a particular
polling place. Again, most of the research work in this area has been explored by Horn in his work
on the Interactive Territorial Assignment model (ITA) developed by CSIRO in the late 1980s.
However few if any electoral applications have been implemented using GISs in the area of electoral
administration other than the redistricting process.
GISs can also be used to display time-series data cartographically. One can observe demographic
growth at a low geographic level, and in the case of Australia it is possible to use a cadastre to
display this information. This information is invaluable in terms of planning for the electoral
administrator.
Political Applications
Political researchers are primarily interested in identifying change, and constructing models where
they can draw qualitative conclusions. It is here that the GIS is an extremely powerful tool. Using
data from the census, the catchment areas of polling places, and results of previous elections, the
user can target certain demographic groups. For example, one may be to correlate informal votes
with certain language groups. If such a correlation is discovered then possible solutions may be the
targeting of information campaigns in that particular language.
Of greater importance to the political researcher is correlating voting patterns with socio-economic
data. The results of such analysis will assist political parties in campaigning strategies and targeting.
It will also structure their input into suggestions for delineation of electoral boundaries. There is a
growing use of GISs in this area both domestically and internationally.
The GIS is also helpful in producing ‘what if’ scenarios. Dynamically changing voter patterns and
electoral results, and observing the consequential results, is a very valuable tool to researchers, and
if planned correctly can embody predictive elements that may shape political policies.
For some years now the reporting of polling night by the media and especially the television media
has become more sophisticated as a result of data provided by electoral organisations (Maley and
Medew 1991). It is now possible for the media to receive results by individual polling places. In
some cases, there are already moves to use GIS technology to display the results on a map, using
catchment areas, and to correlate these results with pre-determined census data. The provision of
such services via the Internet may overtake this to some extent. Nonetheless it remains simply an
idea that is achievable but to date no real research has emerged.
91
Geographic Information Systems In Political And Electoral Work
Finally, the GIS may be used to solve ecological inference problems associated with voting
behaviour. To date this has required very complex algorithms (King 1997) and large computing
power. The GIS is well situated to assist in solving these problems and identifying individual
behaviour from aggregated or ecological data.
Conclusion
The use of GIS technology in electoral administration in Australia remains limited to redistricting.
This paper has argued there are many more applications that can be beneficial to electoral bodies,
political parties and other users of electoral and political data such as the media. One of the main
reasons that electoral organisations should embrace GIS technology is the possibility for end users
with little computing knowledge to analyse data. The cost of GIS technology, coupled with the
availability of spatial data, ensures that GIS technology can be easily set up and used quickly.
The requirements for a sophisticated, multi purpose, information display vehicle that presents large
amounts of data in a simple, easy to understand manner is fulfilled by GIS technology and the basis
is a map. The future for GISs in this area has tremendous potential, and while it may not be
everyone’s ideal it remains grossly unexplored.
References
1. Clifford, E. (1998), Plotting Elections - EDAMS ‘Electoral Distribution and Mapping Systems’, AURISA Conference
1998, Canberra.
2. Commonwealth Electoral Act (1918), Canberra: Commonwealth Government Printer.
3. Eco, U. (1989), The Aesthetics of Chaosmos, Harvard University Press, New York.
4. Horn, M. (1995), ‘Solution Techniques for Large Regional Partitioning Problems’, Geographical Analysis, Vol 27
No 3, Ohio State University Press, Ohio.
5. King, Gary (1997), A Solution to the Ecological Inference Problem, Princeton University Press, New Jersey.
6. Jaensch, D. (1995), Election! How and why Australia votes, Allen & Unwin, St Leonards.
7. Maguire, D. (1995), ‘Planning Your Business GIS’ in GIS For Business: Discovering The Missing Piece In Your
Business Strategy, GeoInformation, Cambridge UK.
8. Maley, M. & Medew, R. (1991), ‘Some Approaches to Election Night Forecasting in Australia’, Australian Journal
of Political Science, Vol 26 No 1.
9. Morrill, R. (1991), ‘Making Redistricting Models More Flexible and Realistic’, The Operational Geographer, 9, 2-9.
10. Walmsley, D. J. (1985), ‘Community of Interest - A Geographical Perspective’ Research Report 3/85 Community of
Interest Australian Electoral Commission, Australian Government Publishing Service, Canberra.
92
Why Data Definition Is More Important Than Technology For Electoral Research
12
Why Data Definition Is More Important Than Technology For
Electoral Research
Antony Green
Electoral research data and data definition is more important than technology. Technology may
determine how and what data is stored, but in the end it is the storage of the data in the first place
which is the issue in electoral research. For research, knowledge of the data available is the key
question, not the technical aspects of its storage.
As probably the only person in Australia paid to analyse elections as an occupation, I am able to
bring a unique perspective to this question. Probably no one else is such a constant user of the
election results data produced by electoral offices around the country though, with no involvement
in campaigning or polling, limited use is made of the electoral rolls. The capacity to access a suite of
database programs with which to re-analyse election results, and the experience of dealing with the
different formats, both logical and physical, used by the various electoral offices, adds a particular
dimension to my argument.
My previous career was in the computer industry, first as a programmer, and later as a data analyst.
It is the latter which has been particularly important in my work, both for the ABC and for the
NSW Parliamentary Library, for which a range of electoral publications has been produced.
In this article, the future for electoral research is examined from the perspective of a data analyst.
The use of computer technology to capture electoral data is really only relevant to research to the
extent to which the items of data have had to be defined in a logical database structure.
The term database is used here loosely, as many different approaches have been used to implement
computer technology. The larger electoral offices have developed substantial database driven
systems, where the unique identifiers for each item of data, and the relations between data, must be
specified within the structure of the database. Many of the smaller electoral offices have used
systems based on Excel spreadsheets. However, whatever the technology and software used, the
first task everybody has had to undertake is the definition of all the data to be stored, and the
logical relationships between data files.
As an example, in the pre-computer era, it was feasible to refer to each polling place by name. With
computers, a more unique method of identification is required. The Australian and Victorian
Electoral Commissions, with greater responsibility for roll maintenance and therefore an interest in
using the roll to site polling places, go to considerable effort to uniquely identify venues within a
state. In most other states, where simply recording the result has been the primary task of computer
systems, uniquely numbering polling places within electorates has been sufficient. But either way,
this process of identifying a data item called a Polling Place has been undertaken, and a comput­
erised method of identifying it developed.
For electoral researchers, the most important information is knowledge that data has been
collected. In the pre-computer days, the only data available was the published Statistical Returns.
Today, those returns are more likely to be summaries of the complete data set of results. The AEC
93
Why Data Definition Is More Important Than Technology For Electoral Research
has made good use of CD-ROMs to make the more detailed results available (and saved consid­
erably on paper), but with limited demand, it is not surprising that State electoral bodies have not
followed this approach.
Having gone to considerable effort to use technology to improve the efficiency of their internal
procedures, as well as make the election process itself more efficient, electoral offices should not
ignore the fact that researchers outside of their office have a legitimate interest in the data they have
captured. And to their credit, every electoral office in the country has been willing to assist in
providing access to results in computer readable format.
Most electoral offices have available in computer readable format, a range of data that researchers
are not necessarily aware was captured. The question is, what should electoral bodies do to make
legitimate researchers aware of the data that is available?
It is recommended that all electoral offices have available documents that outline the data that is
captured and stored. In the case of offices that use large database products, this may be a document
containing file layouts or data dictionaries. In smaller systems, it could be simple examples of Excel
spreadsheets or Word documents that are available.
These sorts of documents normally reside in the files of an electoral office’s computer experts,
rather than in the hands of public information officers. But any researcher trying to make use of
large volumes of computer results would have to be familiar with the process of defining data for
storage.
Recommendation 1
That electoral offices have available for examination documents that set out the data
sources available from the electoral office. Depending on the sophistication of the
database products used, this may include formal data dictionaries and file layouts,
examples of spreadsheet and word processor documents, and relevant source
documents used.
Issues of Privacy
While it should be possible to find the details of data storage, an electoral office should not release
all of the data they store. For obvious reasons, there are privacy considerations that are important
when examining data about electors.
To date, most privacy issues that electoral bodies have had to deal with have concerned access to
private data stored on the electoral roll, such as gender, age and occupation. No doubt the growing
push to implement electronic or Internet voting will open whole new areas of potential privacy
concerns in the electoral process.
Today, manual procedures mean it is impossible to draw a connection between an elector and their
vote, and it critical that in implementing new forms of voting this break in the link between voter
and vote is maintained. Even the simple computerisation and scanning of electoral rolls has
generated data that raises potential issues for privacy and electoral administration.
While on the surface, information on where or how someone voted appears innocuous, it could be
used by an unscrupulous people or parties to organise multiple voting. It is pleasing that electoral
94
Why Data Definition Is More Important Than Technology For Electoral Research
bodies seem to have adopted the approach of the Australian Bureau of Statistics (ABS) in only
making information of this type available aggregated at CCD level.
Recommendation 2
Where for privacy reasons, an electoral office will not make generally available certain
parts of their data sets for reasons of privacy, that such restrictions be included in the
data sources documentation.
Electronic Ballot Papers
In recent years, the use of computers to carry out the complex quota preferential counts for the
Senate and state Legislative Councils has resulted in the storage of a data not previously available.
Previously, many administrations destroyed the ballot papers after a certain period of time because
of the simple storage problem. (Jurisdictions that use countback methods to conduct by-elections of
course retain the ballot papers.)
Sheer impracticality made it impossible previously to carry out any research on these ballot papers.
Now data from the ballot papers has been captured, a new field of research has suddenly opened
that was not previously possible.
Using computers to conduct the count of course raises a whole new range of problems to do with
scrutiny. There was a dispute with regard to making this data available in relation to the Melbourne
City Council elections of March 1999. As a count back system is used to fill casual vacancies, the
Council was concerned that making the data available could lead to manipulation of nominations
for a count back. However, in terms of ensuring that the count is conducted correctly, making this
sort of data available is the only way that candidates and parties can scrutinise the conduct of the
count.
The ballot paper question is a useful example of how electoral offices should deal with future
expansion of the use of computers. The capture of more data raises new questions concerning
whether new data should be made available. At times, this may even require changes to legislation
to deal with new scenarios created by the switch from manual to computerised procedures.
Recommendation 3
The capture of new data creates new possibilities for fields of research. Electoral
offices should consider possible outside uses of any new data captured, and adminis­
trative and legal restrictions on the use of this data.
How to Make the Data Available
Over the years, many operating systems and software packages have come and gone. What today is
industry standard, tomorrow is incompatible with the software everyone else is using.
However, the one thing that is still accessible is simple flat files of data. Text files with simple
strings of text, each line of text representing a record in the database file, with fields in each record
separated by a delimiter. It is often very easy to generate these files in exactly the same format as
described on the data dictionaries made available to researchers.
95
Why Data Definition Is More Important Than Technology For Electoral Research
It is pleasing that the AEC has recently decided to start including these simple text files on their
CD-ROMS of election results. This is an important step in making the raw election result data more
readily available to outside users and this approach is recommended to other electoral offices.
Recommendation 4
Electoral offices should look at the simplest method to make election data available.
Apart from the simplest packages, it is unlikely that researchers will use the same
database products as electoral offices, so data should be made available in the
simplest format (such as simple strips of data files) for transfer to outside systems.
Dealing with Requests from the Computer Illiterate
All of the above recommendations are fine for researchers who come to electoral offices with some
idea of what to do with data. Many requests are also received from minor parties and students with
simpler requests for whom the provision of data in the fully dissected form of a database would be
beyond their need, and probably beyond their ability to cope.
It is suggested that the equivalent of an electoral office’s statistical returns should be available in
simple spreadsheet or word processor format. This is effectively what the AEC have done with their
CD-ROM, and while a smaller demand for state election results can be expected, it would not be
difficult for most tables published in statistical returns to be re-produced in simple spreadsheet or
word processor format.
This raises the perennial problem of versions of software, and whether users can read disks. It is
suggested that disks in the format of the most commonly used packages (usual Excel and Word) are
produced, and users left to deal with conversion problems.
As with statistical returns, some small fee should probably be applied to such requests. However,
the fee should be set more at a level to discourage frivolous requests than on the basis of full cost
recovery.
96
13
Elections And Technology – Implications For The Future
Phillip Green
Introduction
Technology has dramatically changed the administration of elections in Australia and the rest of the
world. In the last 20 years or so electoral roll maintenance has progressed from rolls kept on
handwritten cards and ledgers to rolls kept on national on-line databases. Office automation has
progressed from typing pools using manual typewriters and multi-coloured carbon-copies to
networked personal computers using word processors that check your spieling as you type.
Take away our desktop computers, printers, databases, phones, faxes, email, the Internet, intranets,
scanners, bar code readers and counting machines and most electoral administrations today would
not be able to run an election effectively in the time limits required by modern legislation. This
technology, much of it unheard of 20 years ago, is now indispensable in the modern electoral
world.
In order to discuss the implications of technology for the future of the electoral world, and for
electoral research, it is necessary to discuss the present uses of technology, both for elections and in
the wider world. In many ways the future of electoral technology has already arrived. There are
many established and emerging technologies that have not yet been fully harnessed for electoral
purposes.
It is arguably a good thing that the electoral world is not right at the cutting edge of new
technology. Adoption of new technology often carries with it considerable risks. Elections, partic­
ularly parliamentary elections, are crucial events in the functioning of our democratic society, and
electoral administrations have a duty to ensure that elections function smoothly and that risks are
minimised. Consequently there is a lot to be said for caution when new technology is being
considered.
Within this framework, this paper will discuss:
●
the different types of technology currently being used for electoral purposes in Australia;
●
possible new applications of technology for electoral purposes;
●
International IDEA’s Elections and Technology project;
●
how these uses of technology can be harnessed for electoral research;
●
a vision of the ideal electoral world.
Electoral Technology Today
Many casual observers, whose only experience of the electoral process is voting on polling day,
come away with the impression that elections are low-tech events. In fact, the low-tech nature of
the voting and the counting processes are the exception rather than the rule – these days most
Australian election processes behind the scenes make extensive use of technology.
97
Elections And Technology – Implications For The Future
Electoral rolls
The compilation and maintenance of electoral rolls are significant applications of electoral
technology. Electoral rolls tend to be kept on mainframe databases, which are often networked and
available for on-line update and inquiries.
The power of modern database technology means that electoral rolls are more than just lists of
names and addresses. They contain a wealth of data that can be manipulated for many purposes.
Subject to privacy restrictions, names, addresses and other personal data such as date of birth can
be used to produce targeted mailing lists or samples for medical and other research. Numbers of
electors enrolled for particular locations can generate a range of useful statistics. Combined with
other information, such as census data and voting statistics, enrolment statistics can reveal much
about enrolment and voting behaviour. Enrolment data is crucial to today’s one-vote, one-value
redistribution processes.
The address component of the electoral roll can also be a very powerful tool. The address files
maintained by the electoral authorities are among the most complete and accurate address records
in the country. Electoral authorities are replacing the old door-knock methods of keeping the rolls
up to date by using database information to identify addresses to target with direct mail. Electoral
roll databases can facilitate targeted enrolment review processes by identifying addresses where no
one is enrolled or addresses where too many people appear to be enrolled. They can also identify
areas of particularly high growth or enrolment turnover, so they can be targeted by doorknocks or
mail review. Databases of other agencies, such as Australia Post’s mail redirection service, are also
being used to target addresses of moving electors.
Electoral roll databases also produce a range of printed and electronic off-shoots. These include
printed and electronic versions of the rolls for members of parliament, political parties, researchers
and government agencies, as well as in-house applications, such as rolls for public display and
certified lists for use in polling places.
Rolls used in polling places and elsewhere to record names of electors who have voted also generate
useful information. In most, if not all, Australian jurisdictions, printed rolls are scanned using OMR
(Optical Mark Reading) scanning technology, to produce electronic lists of voters. These are
sometimes supplemented by lists of other kinds of voters (such as postal voters) generated by bar
code readers or data entry. This voting data can then be used to analyse voting behaviour, such as
identifying where voters vote – at locations near their home or elsewhere.
The potential for use and misuse of enrolment and voting data is high. Many people are concerned
with the privacy aspects of making enrolment information available to users outside the electoral
authorities, such as political parties, MPs and other agencies. For example, it is not widely known
that the Commonwealth Electoral Act was amended in 1999 to provide that all federally registered
political parties and federal MPs are entitled to receive the dates of birth of all electors.
It is important for the integrity of the electoral process that electoral roll data is sensitively used, to
avoid the possibility that large numbers of people will fail to enrol because of privacy concerns.
Election management systems
Parliamentary elections are huge logistical exercises. In every electoral district, hundreds of people
have to be employed, thousands of ballot papers have to be printed and distributed, a vast array of
98
Elections And Technology – Implications For The Future
materials has to be purchased, stored, distributed and retrieved, and dozens of polling premises
have to be located and hired.
Keeping track of the logistics, particularly working out how many items are required and where
they are needed, is an ideal application of technology. More and more, electoral authorities are
using linked databases to automate as much of the process as possible. Election result databases and
redistribution programs can be linked to election management systems to automate the purchasing
and distribution process, by identifying how many votes each polling place can expect to receive at
the next election.
To facilitate the effective use of these election management databases, there is generally a complex
network of hardware and communications. As election systems usually include sensitive data such
as personal details and election results, the level of security needed is high. With geographically
spread locations, electronic communications between offices are crucial.
Electoral boundary delimitation
Another powerful use of technology for electoral purposes is for boundary delimitation (or redistri­
bution). Rod Medew from the Australian Electoral Commission covers this topic in these
proceedings in ‘Geographic Systems in Political and Electoral Work’.
Office automation
A use of technology that has become so commonplace as to be unremarkable is office automation.
The impact of office automation on electoral management should not be underestimated - it has
revolutionised it. Word processors, spreadsheets, simple database programs, presentation software,
email and Internet access have dramatically increased productivity in the electoral workplace.
The word processor, in particular, is a huge time-saver. Instead of sending material to typesetters
and printers, a process that used to take weeks or months while proofs were ferried backwards and
forwards for corrections, these days high quality printed material can be produced on our desktops
and printed in our offices in minutes. In the ACT, most of our election forms and training materials
are produced on our desktops (and customised by our election management database) and printed
in-house. As recently as 4 years ago, these same materials were routinely sent to professional
typesetters and printers.
Scanning
Various scanning technologies are increasingly being used for electoral purposes. As mentioned
previously, certified lists of voters are used at election times in polling places and elsewhere to
record the names of electors who have voted. They are then are scanned using Optical Mark Reader
technology and used to produce lists of voters which serve to identify multiple voters and non­
voters.
Bar codes printed on various items such as postal ballot envelopes are scanned by bar-code readers
to record names of voters and to identify return-to-sender mail. Australia Post has just introduced
bar-coded mail addressing to reduce mail costs and increase mailing efficiency, a new use of
scanning that will be of considerable benefit to electoral processes, particularly as more and more of
our business is being conducted by mail – particularly local government elections and industrial
ballots.
99
Elections And Technology – Implications For The Future
Images captured by scanning also have significant electoral uses. Electoral enrolment forms are
currently imaged and stored on disk, which can be manually retrieved and printed out on call. In
the near future, it is hoped that images such as signatures from enrolment forms will be available
on-line on our PCs, which will greatly increase the ability to perform signature checks on various
electoral forms, such declaration vote envelopes. In the longer term, scanning products may be able
to automatically compare two signatures to identify possible fraud.
The next major use of scanning is likely to be Intelligent Character Recognition (ICR). ICR is
currently being used by other industries to capture information from forms, such as exam results,
stock inventory forms and tax forms. The potential of ICR for electoral forms is obvious, given that
so many electoral processes use forms.
Electronic vote counting
Australia is only just putting its toe in the water with electronic vote counting. While the USA with
its first-past-the-post system has been using electronic voting and vote tallying systems for many
years, Australia’s preferential voting systems have to date proved difficult to automate.
At the present time all Australian parliamentary ballots are cast using hand-written ballots and
counted (at least to first preferences) by hand. A major step forward in recent years has been the
process adopted for complex single transferable vote (STV) multi-member elections such as the
1997 election of delegates to the Constitutional Convention, the 1998 Senate election and recent
elections for the upper houses in New South Wales, Western Australia and South Australia. In these
cases, after the ballots had been counted by hand to first preferences and ticket votes identified and
set aside, the preferences for individual candidates on the non-ticket votes have been typed into
computer systems by data-entry operators. Computer programs have then been used to calculate
the results of the scrutiny.
This process has led to significant savings in the time taken to produce results, and increased the
accuracy levels of the counting process to reduce the likelihood of errors in the count.
Even though most ballots are still counted by hand, there is considerable use made of technology
behind the scenes to compile and report the results of the count. All Australian jurisdictions use
computer results packages of various levels of complexity to compile and report election results.
The latest development in this area has been the publication of election night results live on the
Internet, with the 1999 Constitutional Referendum dispensing with a physical tally room and
making sole use of the Internet to publish the results of the Referendum to the nation. By all
accounts this appears to have been a resounding success.
Some election night information packages used by electoral authorities simply report compiled
statistics and percentages, with analysis of the statistics left to the media and political commen­
tators. Other authorities, such as the Australian Electoral Commission and the ACT Electoral
Commission, go further and use statistical analysis techniques such as matched polling place swing
analysis, to provide more information of use to analysts.
Computer programs are also used for more complex multi-member single transferable vote systems
to perform the calculations during the scrutiny. The Tasmanian and ACT electoral authorities use a
spreadsheet program to conduct their Hare-Clark scrutinies. This system relies on hand sorting of
ballot papers, but takes the pain out of the complex mathematics that are applied to the physical
results.
100
Elections And Technology – Implications For The Future
The Internet
Another burgeoning technology that is increasingly being used for a range of electoral applications
is the Internet. Every Australian electoral authority now has a web site, each with a wide range of
information and services available. As discussed above, many jurisdictions have recently moved to
publish election results live on-line on election night.
The Internet is an ideal way for disseminating electoral information, particularly information of use
to targeted groups such as party and candidate information manuals and electoral education
products. The Internet is now a viable alternative to telephone information hot-lines, and allows
users to find information in more detail than is feasible by phone, and they can get it when they
want it, in or outside business hours. One huge advantage of publishing information on the
Internet is the ability to keep information constantly up to date.
Multimedia applications accessed via the Internet or CD-ROM or at electoral education centres are
ideal ways of delivering electoral education to both school age and adult clients. By providing
sound, colour and movement, as well as interactive hands-on learning tools, these applications are
already revolutionising the way electoral education is taught. This will continue to change as
technology advances.
The Internet also allows clients to down load forms, such as enrolment forms and postal vote
applications, which in other circumstances they would have to obtain by physically attending an
electoral office or post office, or waiting for them to be sent by post. This facility has significant
implications for the franchise, as it increases the chances of electors meeting close of rolls and close
of poll deadlines.
The possibilities of the Internet are only just beginning to be explored by electoral authorities. On­
line transactions such as enrolment and postal vote applications are probably only a matter of time.
Electoral roll enquiries on the Internet are also not far away. The New Zealand electoral authorities,
for example, currently allow electors to confirm whether they are correctly enrolled by using the
Internet. On-line voting is also a possibility in the future.
There are a few notes of caution to sound here. First, there are many hurdles to overcome regarding
the verification of identity and security before enrolling and voting on the Internet become feasible.
Second, privacy rights of individuals should be kept in mind when providing access on-line to
enrolment records. Third, Internet use is still the domain of a minority of electoral clients, partic­
ularly the better-off, better educated urban dwellers. Consequently, the Internet, at least in the
short term, should only be looked on as an addition to methods of communications, not as a
replacement for other measures.
Information call centres
Technology is also being put to good use in election inquiry call centres. Sophisticated on-line
systems are now being used to allow call centre operators to log and answer calls from the public.
These systems allow large numbers of casual staff to answer complex enquiries, through the use of
clever knowledge-based systems, thereby improving service to the public and reducing the
demands on those permanent staff with their knowledge stored in their heads.
These same systems can be mirrored on Internet sites to allow Internet-savvy users to conduct their
own enquiries on-line.
101
Elections And Technology – Implications For The Future
Possible New Applications Of Technology For Electoral Purposes
Technology is so ingrained in electoral management that it has become indispensable. Looking into
the future at what new applications of technology might be made for electoral purposes, it appears
most likely that change will build on technology that is already being used.
In particular, election management systems are becoming more integrated, so that the various
different components of the electoral process are becoming part of one system, rather than a set of
unrelated systems. For example, by the next ACT Legislative Assembly election due in October
2001, we expect to have a fully integrated election management system. Consequently, the
nominations system, which records the names of candidates, will be linked to ballot paper design
and printing, the list of candidates for public release, the funding and disclosure system, polling
place forms, the on-line inquiry system, the Internet site, the Hare-Clark counting system, the
election results package – in fact, anywhere candidates’ names are used in the election process.
The Internet can only increase as a tool for electoral purposes. If security, privacy and identity
problems can be overcome, just about every electoral transaction could be conducted over the
Internet, from enrolment to voting to displaying the results, and everything in between.
Smartcards are a relatively new technology that could have significant electoral uses. Read/write
smartcards that can identify individuals and record facts like ‘this elector has now voted in this
election’ have obvious applications. Bio-identification systems built into smartcards, combined with
rigorous issuing procedures, could see smartcards as the answer to minimising the possibility of
voting fraud.
Increased use of scanning technology could continue to improve election management processes
related to data capture and data retrieval. Most electoral data is currently typed into computer
systems – election results and enrolment forms being large-scale examples. Greater use of intelligent
scanning systems could see the amount of manual data capture reduced dramatically, while at the
same time increasing accuracy. The ability to call up images of original documents on-line also has
the potential to improve services, particularly where signature verification is required. As
computing power and data storage capacity increase, these kinds of applications will become
commonplace.
Bar code technology is also likely to be used more and more. Many electoral processes involve
processing incoming and outgoing mail and forms. Bar codes promise to remove much of the
drudgery from these tasks, at the same time improving speed and accuracy.
The ‘big one’, particularly for those of us with complex multi-member preferential systems, is
whether technology can be used for the voting and/or the counting process – the last bastion of
manual electoral processes. At present, the use of technology is limited to using manual data input
to convert preferences on handwritten ballot papers into election scrutiny outcomes.
The ACT Electoral Commission is committed to recommending to the ACT Legislative Assembly a
proposal to use technology to increase the speed and accuracy of the October 2001 election count.
The options being considering include voting on an electronic interface in polling places, and/or
making use of scanning or other forms of data entry to automatically capture preferences recorded
on paper ballots. The Commission has opted for solutions capable of being used in polling places
for the 2001 election, having taken the view that Internet or telephone voting would not be feasible
in the short term.
102
Elections And Technology – Implications For The Future
Electronic vote-capture devices are in widespread use overseas, however in almost all cases these
are used in first-past-the-post systems, and are therefore simpler than would be needed for our
more complex preferential systems. However, the advantages of doing away with paper ballots are
many, and the time may come in the not too distant future when almost all Australians will vote
electronically in one way or another.
One form of electronic voting that might be feasible is telephone voting. The telephone is being
used increasingly for electronic transactions such as banking, bill paying and purchasing. While the
telephone might not be practical for voting using the ACT’s Hare-Clark system with Robson
rotation and 40 or more candidates per election, it would be practical to use it for simpler elections
or referendums. Indeed a simple yes/no question like the recent Commonwealth Constitutional
referendum would be an ideal candidate for telephone voting, provided the issues of security and
identity are addressed.
International IDEA’s Elections And Technology Project
The Administration and Cost of Elections (ACE) Project is a major project undertaken by the
International Institute for Democracy and Electoral Assistance (International IDEA), the United
Nations and the International Foundation for Electoral Systems. It is an Internet site
(www.aceproject.org) and CD-ROM that is intended to be ‘the first global information resource ever
created about alternatives in election administration’.
A new module currently being included in the ACE Project – Elections and Technology. This will
aim to cover not only the range of technology applications that can be used for electoral purposes,
but also the principles behind selecting, purchasing and using technology in a manner appropriate
to particular circumstances. It is hoped this will be available by the middle of 2000.
The ACE Project team is seeking contributions from writers of files or case studies on specific
elections and technology issues. The ACT Electoral Commissioner is the Australian contact.
Electoral Technology And Research
Technology is increasingly being used for all aspects of the electoral process. How can this
technology be of use to electoral researchers?
At present the main technological contribution made by electoral authorities to researchers is the
provision of election result data and electoral redistribution data in electronic form. As the
discussion above has indicated, there are many other kinds of data being collected and used for
internal election management purposes. In order to tap into this wealth of information, researchers
will need to identify what sources of information would be of use to them, and negotiate access to
that data.
The most obvious aspects of technology that could assist electoral research are the increasing
amount of electoral data that is available in electronic form and the accessibility of that data. The
Internet in particular has made it so much easier to conduct electoral research than before. Electoral
legislation, Hansard debates, election results, election ‘lore’ and procedures are increasingly
becoming accessible on the Internet. The amount of information available can only increase as
electoral authorities expand the available options on their websites.
103
Elections And Technology – Implications For The Future
The recent adoption of computerised STV scrutinies and future moves to computerise other voting
and counting procedures will open up research potential that has not been possible before. For the
first time, it will be possible to look at preference voting beyond the formal count. If the data could
be made available, researchers could look at the full range of preferences cast by voters and explore
a whole range of ‘what ifs’. What if the rules were changed, how would that effect the outcome?
How many voters switch between parties when allocating preferences, and how many stick to the
party line? Could we get a better handle on the level of donkey voting? If data was available for a
system where countback was used to fill casual vacancies, researchers could work out who would
be elected to fill any potential vacancy.
Other research fields outside political science might be able to tap into the vast amount of data
stored in election management and electoral roll systems. Samples of names and addresses are
already being used for medical research purposes, particularly samples stratified by age ranges (a
use of electoral roll data permitted by the Commonwealth Electoral Act). Electoral roll data could
be used by demographers and genealogists, particularly where changes are shown over time.
As many electoral processes involve dealing with large numbers of people, they are ripe for study
by the various behavioural sciences. For example, mass communication campaigns are regularly
conducted, and different communication methods are being trialled at various places and at various
times. Queues are often a function of election processes, and the study of queuing behaviour can be
of benefit to and can learn from electoral events.
Much of this research potential is still largely untouched. As researchers become more aware of the
possibilities, there may be more use made of electoral resources. Making information available to
researchers can be a resource-intensive undertaking for electoral authorities, and electoral
authorities will need to be convinced of the benefit of putting resources into making more
information available.
A Vision Of The Ideal Electoral World
A key issue for the electoral world of the future is the fate of the polling place.
New, comparatively low-cost, voting solutions like the Internet and the telephone by their nature
do not need a polling place venue. If polling place voting can be replaced by these electronic
methods, supplemented by postal voting or extended pre-polling at a limited number of locations
by those unable to vote electronically, the voting world of the future could be very different.
Such a change could have a profound effect on both voting behaviour and political campaigning
techniques. At present, political campaigns are geared towards attaining maximum impact on one
particular day – polling day, when around 80% of electors vote. Doing away with polling day will
change this political dynamic, since it will force political campaigners to reach their peak earlier
and then to maintain that momentum for the duration of the polling period. While it is arguable
that campaigners should be doing that now – given the large number of pre-poll votes being cast –
it is also possible that the party machines will resist such a dramatic change.
It is also possible that removing the political institution of voting at local polling places on polling
day will change the dynamics of the voting process. Electors voting in the privacy and isolation of
their own home may well vote differently than if they were taking part in the public ritual of polling
day. It is already a known and accepted fact that voting patterns of postal and pre-poll voters differ
from those of ‘ordinary’ voters on polling day. Whether this would change Australian politics for
better or for worse would be difficult to judge.
104
Elections And Technology – Implications For The Future
A significant worry, if we did abolish polling day, is the Australian tendency to leave everything to
the last minute. There would be a real risk that a national Internet and/or telephone voting system,
designed to take 14 million votes over a three week period, could be deluged with 12 million votes
on the last day. Electoral administrators of the future will need to plan for this national quirk.
It is possible that the Australian electoral process of the future may face one of those crucial
moments in history. If we go down one path, Australia may abandon polling places and polling
days in favour of virtual voting in the privacy of our own homes or at voting terminals in shopping
centres. Down the other path, Australia may opt to stick with polling places, but transform them
with new technology.
In this alternative future, the new-age polling place might work like this.
The elector walks in off the street and up to a terminal, much like an automatic teller machine, and
inserts her smartcard. The smartcard registers that the person is entitled to vote for a particular
electorate, and checks that she has not voted before. A camera on top of the terminal checks the
person’s facial features, compares them to the details recorded on the card and verifies the person is
the person named on the card.
(Of course, the smartcard has been issued to the person following an identity confirmation check
and her enrolled address is linked to the national address register shared by all Government
agencies.)
A virtual ballot paper appears on the screen (Robson rotated of course). The screen is tilted at 45
degrees and the details shown on it are not visible by anyone else around her. The elector marks
her preferences by touching the screen, and the screen prompts her to confirm her choice, at the
same time ensuring the vote is formal (although she could cast an informal vote if she wants to). A
help button gives the person any information she wants. Other buttons display each registered
voting ticket if she wants to see them. Upon confirmation, the smartcard is electronically marked to
indicate that the person had voted in today’s election, so that it can not be used again.
The smartcard pops out of the terminal, and a message thanks the voter for her participation. She
takes the card, nods to the pair of polling officials who are there to help those few people who need
it and walks out. The pair of scrutineers looks very bored.
Meanwhile the elector’s vote has travelled down the network connection to the central tally
computer. It joins the many thousands of other encrypted votes coming in from all corners of the
electorate, some of them from dedicated data lines, some from shared phone lines, some from
satellite feeds, some from mobile phone communication towers.
The polls close, the computer program kicks in and 30 seconds later the election result is in.
Around the country, unemployed psephologists and pundits sigh and look at the results on their
Internet screens, while successful politicians celebrate and unsuccessful ones mourn, at least
thankful that the agonised waits of the past are no more.
Fact or fantasy? Only time will tell.
105
●
What is the core for research - past, present and future?
E L E C T O R A L
●
What are the boundaries between electoral and political inquiry?
THE CORE AND THE BOUNDARIES
●
What are the benefits and opportunities for cooperation between electoral
administrators and analysts?
●
Where is technology taking the industry and can it be used to further electoral research?
Research Series
South Australian State Electoral Office
June 2000
SA STATE ELECTORAL OFFICE
Editors: Jane Peace and Janet Taylor
R E S E A R C H
CONFERENCE PAPERS
Foreword: Elizabeth Ho
Keynote Address: Colin A Hughes
Papers: Nicholas Economou and Brian Costar,
Antony Green, Phillip Green, Andrew Hawkey
and Michelle Davy, Dean Jaensch, Michael
Maley, Alan McRobie, Rod Medew, Gerard
Newman, Jenni Newton, Campbell Sharman,
John Wanna, George Williams
PA P E R S
S E R I E S
The papers contained in this volume explore
some of the issues relating to an agenda for
electoral research in Australia from the
perspective of both administrators and
analysts, questions such as:
C O N F E R E N C E
R E S E A R C H
Collaborative ventures, changing technology
and an increasing acknowledgement of the
role played by analysts in their commentary
on the electoral processes has led to renewed
debate on the need for dialogue with the
electoral jurisdictions and investigation of the
basis on which such interaction can occur.
ELECTORAL RESEARCH – THE CORE AND THE BOUNDARIES
Electoral Research
The Core And The Boundaries
South Australian State Electoral Office
Editors: Jane Peace and Janet Taylor