Prior to the 2007 federal election trial of electronically assisted voting, some electors who are blind or had low vision were only able to vote at federal elections by having another person assist them to complete their ballot papers. While this enabled these electors to participate in elections, it meant that their votes were not secret and independent.

The committee has also looked at a number of broader considerations including:

• the number of people who participated, or who might have otherwise benefited from the technology adopted for the trial;
• the usability of the voting system and its possible use by others;
• alternative voting systems; and
• planning and consultation by the Australian Electoral Commission (AEC).

The committee’s evaluation of the trial relies heavily on material prepared by the AEC, including the AEC’s own review and an evaluation undertaken by a contractor on behalf of the AEC. In addition to this material, the committee has drawn on submissions to the 2007 election inquiry and experiences in other jurisdictions.

Another data source included in the contractor’s evaluation report of the trial put the number of potential electors who are blind or have low vision at around 158,000.⁴

It is important to recognise that not all electors who are blind or have low vision would utilise electronically assisted voting facilities should they be provided at future elections. It has been noted that many electors who are blind or have low vision are comfortable with the range of options already available to them including casting an assisted vote at a pre-poll centre or ordinary polling booth, or, for those with limited vision, completing postal votes using electronic magnifiers in their own homes or in other locations.⁶

Prior to the 2007 federal election, electronically assisted voting for electors who were blind or had low vision had been possible at selected pre-poll voting centres for ACT elections in 2001 and 2004 and at special ‘e-centres’ for the Victorian state election in 2006. An assisted voting system has also recently been developed in Tasmania and was provided for use at a single location in a recent election. The systems used in each of these jurisdictions differ and they are all different to the system trialled at the 2007 federal election.

Assisted electronic voting has been available at three elections (2001, 2004 and 2008) in the ACT. The system is used in five pre-poll voting centres, which are open for 3 weeks before polling day and on election day as ordinary polling places. At the 2004 election more than 28,000 votes were cast electronically.

Electronically assisted voting for electors who are blind or have low vision will continue to be provided by the Victorian Electoral Commission at future state elections. However, the Victorian Government is yet to determine whether it will expand access to other groups such as those with a print disability.¹³

Electronically assisted voting for electors who are blind or have low vision was trialled at the 2007 election for the Legislative Council. Only two electors cast votes using the system at the one pre-poll centre where the facility was available.¹⁴

In practice, some electors who are blind or have low vision choose to vote by postal vote, and seek the assistance of friends or relatives in completing the ballot paper. Electors who have low vision and are able to utilise electronic magnification equipment may be able to cast a secret and independent vote using such equipment if it is available.¹⁷

The 29 locations used for the trial were open for 14 days in the leadup to the election and on election day. Of the sites used for the trial, only six (Kooyong, Ballarat, Shepparton, Warragul, Geelong and Gilles Plains), were newly-created pre-poll voting centres, having not been used as pre-poll centres in previous federal elections. These were usually in disability service centres where they had not been established previously.¹⁸

Table 4.1 – 2007 election assisted electronic voting trial locations and votes cast

Source: Sheridan and Associates, Evaluation of the electronic voting trial for blind and sight impaired electors at the 2007 federal election: Final evaluation report (2008), pp 26–27.

The cost of the trial was estimated to be $2.2 million (table 4.2). Based on the total cost and number of votes cast, the average cost per vote cast was $2,597.²⁰ Had voter turnout been at the higher end of expectations (1,550) the average cost per vote would still have been in the order of $1,425 per vote. This compares to an average cost per elector of $8.36 at the 2007 federal election.²¹

Table 4.2 – 2007 election assisted electronic voting trial estimated costs

Note: (a) Communication costs are for the formal communication strategy. Additional expenditure was incurred in demonstrating machines — which generate free radio, television and newspaper coverage.

Source: Australian Electoral Commission, Report into Electronically Assisted Voting at the 2007 Federal Election for Electors who are Blind or have Low Vision (2008), pp 61–62.

The committee notes that in considering the choice of sites for the trial, the objective of maximising participation was subject to satisfying a range of other criteria:

• At least one centre should be located in each capital city;
• Any other centres should be located in disability service centres where suitable premises are available as these centres are generally located near public transport and the majority of the target group are familiar with them;
• Rural areas should have representation; and
• Expected voter turnout should be such that the trial can be adequately evaluated in terms of system suitability and demand.²²

While only 1.5 per cent of survey respondents expressed that they were dissatisfied or very dissatisfied with using the electronic voting machines, the independent evaluation noted that these users were:

• likely to be older than average;
• almost half as likely to be a computer user;
• more than 50 per cent more likely to be a screen user;
• likely to spend marginally more time in travel to the location;
• over three times less likely to be satisfied with ease of use of the electronic voting machines;
• less than half as likely to be satisfied with the clarity of the audio instructions;
• less than a third as likely to be satisfied with the clarity of the screen instructions;
• less than half as likely to be satisfied with the usability of the screen; and
• 25 per cent less likely to be satisfied with the privacy that they had in using the electronic voting machine.²⁴

Submissions to the committee from electors who are blind or have low vision and who had used electronically assisted voting at the 2007 federal election were positive, with many electors reiterating views previously put to the committee of the value that they placed in being able to cast an independent and secret vote, many for the first time in their lives.²⁵ A selection of these comments is presented in box 4.1.

The trial was also supported by a number of organisations representing the blind or vision impaired community including Vision Australia, Blind Citizens Australia and the Human Rights and Equal Opportunity Commission.²⁶ These organisations supported making electronically assisted voting a permanent feature at future federal elections. Blind Citizens Australia noted that:

Feedback from our members tells us that the opportunity was greatly appreciated by many. We believe that electronically assisted voting should be introduced as a permanent measure with at least one polling booth made accessible in every polling station at the next federal election so that more of the 500,000 Australians who are vision impaired or blind can exercise this critical right.²⁷

A statement of requirements was issued to two contractors who had been involved in developing the Victorian and ACT electronically assisted voting systems. This was done under an abbreviated procurement process that specified a number of features that were different to the systems previously used in these jurisdictions but allowed some flexibility in delivering a solution that met the AEC’s requirements including:

• A system that can accommodate full preferential voting for the House of Representatives, proportional representation for the Senate and caters for a referendum if necessary;
• The requirement for a printed ballot paper in barcode format, and the module to decode and print the contents of those barcodes; and
• Hardware which includes but is not limited to the computer or processor itself, a computer screen suitable for vision impaired electors (which may include touch screen capability), an input device suitable for blind and vision impaired voters (including tactile indicators on the device) and headphones.

The voting system was audited by a contractor accredited with the National Association of Testing Authorities. The contractor was asked to ensure that the voting system met the following criteria:

• Resistant to malicious tampering;
• Free from malicious source code;
• Presents an accurate representation of votes cast in the printed record without gain or loss; and
• Does not allow the association of a voter with the vote cast.²⁸

The audit contractor made the following findings and certified that the voting system complied with the specified criteria:

• that the system design includes features that provide the level of security required by the AEC;
• that the AEC conducted its testing of the electronic voting machine (EVM) with due diligence;
• no evidence was found of malicious source code in the EVM;
• There were no errors detected in tests for security, accuracy and compliance of the system; and
• that risks identified in this report have been avoided or minimised to a level that would allow the EVM to comply with AEC requirements regarding security, accuracy and voting functionality.²⁹

Some of the key differences between the system used at the 2007 federal election from that provided by Software Improvements at ACT elections included:

• The inclusion of a printer that produced a machine-readable barcode for insertion into a declaration envelope, with no votes recorded on the machine. The declaration votes were then forwarded to the relevant division for decoding and counting. For ACT elections, there is no printed output, with votes stored on the machine;
• As no votes were stored on the machines, votes were counted using paper ballot papers, which were produced at the relevant divisional office using barcode readers. For ACT elections, votes cast electronically are stored on machines and downloaded directly into the count early on election night;
• The use of a telephone-style numeric keypad to enter choices to visual or audio prompts. For ACT elections a numeric keypad is also used, but choices are made by selecting arrow keys, rather than choices being assigned to specific numbers; and
• Only English audio instructions were available. For ACT elections, voters could choose to receive instructions in up to 12 languages.

The AEC recognises the value of an independent and secret vote to all electors, including those who are blind or have low vision. However, the AEC also noted that:

The high cost of the trial must be balanced against this important principle. The provision of facilities of this type on a large scale, if not matched by a significant level of take-up, would ultimately give rise to costs which would, in an era of scarce resources, impact on the services provided to other electors.³⁰

Some inquiry participants supported a more general rollout of assisted electronic voting:

• Blind Citizens Australia considered that electronically assisted voting should be introduced as a permanent measure with at least one polling booth made accessible in every polling station at the next election;³²
• Vision Australia noted that the 29 sites used for the 2007 federal election trial constituted only 0.36 per cent of the over 8,000 polling places used at the election and considered that the number of polling places with electronically assisted voting facilities be increased each election until all electorates have at least one polling place which has e-voting capability; and
• The Human Rights and Equal Opportunity Commission considered that electronically assisted voting should be made available in as many locations as possible and at least in every electorate.³³

Firstly, as electors who are blind or have low vision are more likely to be older, there is likely to be a general and continuing reluctance to use a computer to cast a vote, although this is expected to decline over time.³⁵ A community educator with the Royal Blind Society for South Australia told the committee that:

Our client database, for example, shows the average age of our clients is around 78, 79 years of age and, because of that advanced age, they are predominantly females, because females live longer than males. Everything that I have read in the way of suggested answers to these problems revolves around quite complex technology. If I can use my mother as an example: like myself, she is legally blind. She is well and truly into her 80s. She has never sat in front of a computer in all of her life. If she were to sit in front of a computer, she is unable to see the keyboard, so she is not even able to respond to any audio prompts and things like that.

I would put to the committee that this is the situation of most blind people in this country. Nearly half the people we deal with are over 80 years of age. So when you talk about electronically assisted voting—I am not talking from the organisation here; I will probably get a smack on the wrist for this—personally, I think it is a bit of a nonsense to expect most blind people to sit down and operate a computer in a polling booth.³⁶

Secondly, a significant proportion of people with low vision are able to utilise other lower-cost technologies, such as electronic magnification, to assist them to make an independent and secret ballot.³⁷ The same representative from the Royal Blind Society for South Australia told the committee that:

Most people imagine that, in order to call yourself blind, it must follow that you cannot see. Again, this is a very common misunderstanding out there. Most blind people can see. According to the social security system and the medical profession—in fact, according to the World Health Organisation—I am a blind person, but I can see. I am more comfortable with a term like ‘partially sighted’. I might meet the standards to be classified as a blind person but, with mobility, I have very few problems. But, again, this is the case with the majority of blind people. When you are talking about the majority of blind people being elderly, obviously they are more frail, more cautious with their movements, less adventurous with their independent movement, and that sort of thing, so there are other factors as well.

… The majority of blind people cannot and do not want to use a computer; in fact, they are totally intimidated by computers. For the partially sighted population, it has been suggested that electronic magnification is a solution, and it is: it is a great solution. Unfortunately, this equipment costs, for an effective unit, about $4,000.³⁸

Greater availability of electronic magnifiers may be one way that the AEC can facilitate a secret and independent vote for vision impaired electors. The AEC told the committee that:

For the 2007 election, electronic magnifiers (Closed Circuit TVs) were available at a number of the 29 trial sites. These were either hired or loaned for the trial period, and they were utilised by some electors to cast their votes. Purchasing and storing these machines for a 2 week voting period every three years is not practical, or cost effective. The AEC could consider making this equipment available at more locations in the future, subject to appropriate hire equipment being available, but this technology provides no aid to voters who are completely blind.³⁹

However, there are a number of proposals that have been put to the committee that appear to provide a more sustainable basis for continuing with a limited form of assisted electronic voting in the future. These will largely rely on maintaining or lowering fixed costs in combination with increasing participation levels.

It is difficult to determine the likely final fixed costs of continuing electronically assisted voting given the uncertainty over the administration costs of running the program within the AEC, the extent of competitive pressures on potential providers and the development and hardware costs as technology changes.

There are two main options for expanding participation — increasing the proportion of electors who are blind or have low vision utilising the voting machines and/or expanding eligibility for participation to include other groups including those with a print disability.

At the 2006 Victorian State election, 199 electors cast a secret and independent vote at one of six pre-poll voting centres using electronically assisted voting. The 2007 federal election trial utilised five of these sites, four of which were Vision Australia premises. While participation at the five sites was up by 41 per cent overall there were also two sites where participation declined (table 4.3).

Table 4.3 – Assisted electronic voting for electors who are blind or have low vision: Number of votes cast at the 2006 Victorian state election and 2007 federal election

Further, there is no indication, apart from Victoria and the ACT, that electronically assisted voting will be provided by other jurisdictions in the near future. In NSW for example, a parliamentary committee examining the conduct of the 2007 state election did not support the implementation of electronically assisted voting at this stage, instead recommending that the NSW Electoral Commission further examine possible options.⁴⁴

The Human Rights and Equal Opportunity Commission supported removing restrictions on eligibility, noting that print disability was not restricted only to those who are blind or have a vision impairment, but also included:

• Australians who cannot complete a secret ballot using pencil and paper by reason of physical disability; and
• people who (by reason of intellectual or learning disabilities, or other language or literacy difficulties) cannot effectively use written instructions in completing a ballot paper, but could have effective access to a secret and independent ballot through being able to have their input read back to them electronically.⁴⁷

The AEC told the committee that if eligibility were to be extended to electors with a print disability then, subject to funding, it would support up to 20 additional sites for electronically assisted voting in locations yet to be determined, depending on the demographics of the target audience.

The committee therefore does not consider that electronically assisted voting for electors who are blind or have low vision should be made a permanent feature of federal elections at this time.

Recommendation 4

Given the high average cost per vote of $2,597 for electronically assisted voting compared to an average cost per elector of $8.36 at the 2007 federal election and a concern that participation will not increase to sustainable levels, the committee recommends that electronically assisted voting for electors who are blind or have low vision should not be continued at future federal elections.

The committee recognises that for some electors who have low vision, casting a secret and independent vote could be achieved using aids such as electronic magnifiers. The committee considers that electors who have low vision may benefit from the provision of such alternate facilities in accessible locations and should be able to do so where practicable.

Recommendation 5

Assisted voting provisions in the Commonwealth Electoral Act 1918 give people who are blind or have low vision the opportunity to seek assistance from a person appointed by them in casting a vote at federal elections and referenda. Electors who have low vision may benefit from the provision of electronic magnifiers. The committee recommends that the government provide sufficient resources to the Australian Electoral Commission for the deployment of electronic magnifiers at sites where there is likely to be demand from electors who have low vision.