Ballot paper ordering at Scottish council elections

The ordering of candidates on the ballot paper does not affect voters’ ability to locate and vote for their preferred candidate

• it took voters no longer to locate their preferred candidate on the ballot paper when the names were ordered other than A to Z
• we found that voters rarely noticed or thought about the order of candidates on the ballot paper

While our research could not replicate the exact voter experience of completing ballot papers at a real election, we did clearly find that the ordering of candidates on ballot papers did not affect voters’ ability to navigate and complete the ballot paper in the way they intended.

This was the case regardless of whether they were looking for a specific candidate on the ballot paper or were voting along party lines.

Do voters notice the order of candidates?

During the interviews with voters, the order of candidates on the ballot paper was rarely raised, with only 15 out of 118 participants doing so spontaneously.

Even when specifically asked to look at and comment on the order of the candidates on the ballot paper, only a minority of voters noticed whether their ballot paper had the candidates in alphabetical, reverse-alphabetical or random order.

Can the order of candidates make it easier or harder for voters to find their preferred candidate(s)?

Our research found that the order of the candidates had no impact on participants’ ability to find and vote for their preferred candidate(s).

We asked 70 of the 118 research participants to find and vote for a specific fictitious candidate on the ballot paper.

20 out of the 70 participants failed to give their vote to this candidate. However, this was never because they were unable to locate the name. The reasons given were that they had forgotten the name of the candidate, and even though it was a test, some preferred to vote for a candidate from the party they usually support, or did not want to give a vote to a candidate they knew nothing about.

We asked all of the participants in our eye tracking research to find a specific candidate on the ballot paper and timed how long it took them to find that candidate on the differently ordered ballot papers. We found little difference in the time taken to find the candidate based on the order of the ballot papers, with it taking participants less than three seconds on average across all ballot papers.

The data in table one shows that, on average, voters took slightly longer to find the candidate in the Z to A order and less time to find the candidate when the candidates were ordered by lot.

However, this is likely to be because participants tended to read from the top to the bottom of the ballot paper and is therefore related to the position of the specified candidate on the test ballot paper.

The specified candidate was higher up the ballot paper order in the version ordered by lot, and further down in the version ordered Z to A.

Do voters have a preference for how the candidates are ordered on the ballot paper?

When specifically asked, a small majority of the voters participating in our research preferred an A to Z order (60 out of 118 participants) although over a third (41 out of 118) expressed no preference because they, in general, did not think it would make any difference to people’s ability to navigate and complete the ballot paper.

Some participants expressed a view that a change to the traditional A to Z order might raise questions about how the order was decided and whether it was fair.

Reasons for preferring A to Z

Participants who expressed a preference for A to Z order did not tend to have a strong preference, nor did they think it would make it much easier or quicker for people to find their preferred candidate.

The main reasons given by participants expressing a preference for A to Z were that A to Z is a commonly used and therefore familiar system for voters. Consequently it may be considered more transparent and – by extension – ‘fair’.

Some also expressed a belief that it would be easier or quicker for other voters to find the preferred candidate on the ballot paper

However, the belief that it would be easier to find a candidate when the names are set out in A to Z order is not substantiated by our research, as most participants did not notice the order.

Where the order of candidates on the ballot paper was other than A to Z, or where the participants did not realise it was A to Z, there was some suspicion and speculation amongst a few participants as to how the order had been arrived at and whether it had been designed to advantage parties or candidates nearer the top of the ballot paper.

Given the potential for any random or Z to A ordering to potentially appear as a non-random order (say with a particular party’s candidates grouped at the top of the ballot paper) the concerns from participants noted above indicate there will be a need for transparency about how any order is arrived at, with information easily available for voters to dispel any concerns.

Reasons for preferring Z to A

Of the very small number of participants who expressed a preference for Z to A ordering (7 out of 118), the reasons seemed to be less about the concept of a reverse alphabetical approach and more to do with the coincidental aspects of the Z to A ballot paper they had been given, such as candidates from their preferred party being nearer the top.

Reasons for preferring ordering by lot

Amongst the few participants who preferred the ballot papers to be ordered by lot (9 out of 118), reasons offered included a belief that a lack of order would encourage people to look more carefully at the full list of candidates, and a belief that it might be fairer for candidates.

Understanding of the STV voting system

Even when voters were able to find their preferred candidate on the ballot paper without difficulty, our research found that many voters still made errors in completing the ballot which, in an election, would lead to their vote not counting at all or not counting in the way they intended. These included:

• voters marking all their choices with an X or a tick
• participants not realising they could vote for more than one candidate from the same party or that they could vote for more than one candidate
• a small number confusing independent with independence and/or the UK Independence Party

Our findings demonstrate that there is an ongoing and pressing need for us, Returning Officers and poll staff, political parties and candidates to continue to educate voters on the STV voting system.

Findings from specific groups of voters

We included participants with English as a second language and people with learning difficulties. We also included people with low literacy levels and people with sight loss.

Our findings with these groups were no different to participants overall, in that they did not notice the order of candidates on the ballot paper and it had no impact on their ability to complete the paper.

English as a second language

Participants in our research included ten voters with English as a second language. Some said they found the ballot papers easy to complete and some said they found it hard.

However, where participants found it hard to complete the ballot paper, it did not relate to the order of candidates’ names, but to confusion or uncertainty about how to vote using the Single Transferable Vote system, such as whether to use an ‘X’ or how many candidates they could vote for.

Voters with learning difficulties

We tested the order of ballot papers with twelve voters with mild to moderate learning difficulties and, again, we found no evidence that the order of candidates impacted on their ability to complete the ballot paper.

However, some participants spoilt their ballot paper by voting using multiple ticks or Xs.

As with other participants, people with learning difficulties tended to express a preference for an A to Z ballot when specifically prompted to give an opinion.

Two participants with dyslexia thought that an A to Z ballot would be easier for people with dyslexia to complete, although both had completed non-A to Z ballots with no difficulty.

Two participants with learning difficulties said that the list of names on the ballot was very long and this had made voting harder for them.

Voters with low literacy levels

We found no evidence that ballot order had an effect on how easy people with low literacy levels found it to complete the ballot paper. However, we did find that some spoilt their ballot by using multiple ticks or Xs.

The term ‘independent’ caused some issues for a small number of participants with low literacy levels, with some confusing ‘independent’ with ‘independence’.

Voters with sight loss

Nine voters who participated in our research had significant sight loss.

In order to replicate the polling station experience as closely as possible, large print ballot papers were provided for reference, as they would be in the polling station at an election. These are to assist voters to locate their preferred candidate, before voting on a standard sized ballot paper.

We did not supply the Tactile Voting Device (TVD), which a small number of blind and partially sighted voters use in the polling station. A number of concerns have been raised about its usability at STV elections and we did not want to conflate these issues with our findings related to ballot ordering.

We found no evidence that the order of names had any effect on their ability to complete the ballot paper.

Some participants with sight loss spoilt their ballot papers by putting multiple crosses or ticks, but, this did not relate to ballot order.

We did note that the participants with sight loss seemed less likely to look at the instructions, which may have been because they were trying to minimise the amount of text they had to read.

Participants with sight loss tended to say they had no preference about the order. However, two indicated a preference for A to Z with one saying that she would find A to Z easier because she relied on her memory to find her preferred candidate on the ballot paper – although she had nonetheless found the name of the fictitious candidate (on the Z to A ballot) as required and had not noticed the order at the time.

What we learnt from accessibility groups

While the findings from our research did not identify any specific evidence to suggest that ballot paper order would have a negative impact on disabled voters’ ability to find their preferred candidate, we recognise that this was based on a small number of participants in our research.

We also looked at the views submitted by accessibility groups and those representing disabled people in response to the Scottish Government’s consultation on electoral reform and invited further comment from these groups. Inclusion Scotland and RNIB provided further information for us to consider.

Whether responding directly to us, or to the Scottish Government, the majority of accessibility organisations raised concerns about the potential impact of any change to ballot ordering on disabled people.

Many highlighted that it was helpful for them – particularly for those with sight loss – to be able to familiarise themselves with the order of candidates on the ballot paper before going to vote at a polling station.

Concerns were raised about whether any alternative system may make this more difficult to do. Respondents suggested an A to Z order might give disabled people more certainty about where the candidate(s) they favoured were located on the ballot paper ahead of them voting.

Some respondents also highlighted that any additional uncertainty about how to find their preferred candidate(s) on the ballot paper might lead to an increasing reliance on polling station staff to assist disabled voters in completing their ballot paper. They believed that this would compromise the secrecy of the vote.

Respondents also highlighted that parties often produce election material containing an image of the relevant ballot paper, in order to assist voters to locate their candidate(s) on the ballot when they come to vote. They thought it may be confusing for voters if that order does not match the real ballot paper.

Respondents suggested that an early decision about the ballot paper order would be helpful in enabling parties to produce campaign material which reflects the actual order of candidates on the ballot paper. It was acknowledged that two versions of the ballot paper would make this more difficult to achieve.

RNIB Scotland provided feedback from the RNIB Scotland Committee on the specific options being considered. They found that the majority of Committee members favoured the current system of A to Z ordering as it was familiar to them and they thought any change could disproportionately impact on voters with sight loss.

Of the two alternative options for ordering presented, their Committee members preferred the option of ordering by lot as it provided more certainty on the order in which candidates would appear on their ballot paper. Alternating A to Z and Z to A ballot orders also raised questions for them about the provision of accessible materials in polling stations, as there would need to be two versions of the large print ballot paper available for blind and partially sighted voters’ reference.

While outside of the scope of our research, some respondents with sight loss suggested that it would be easier to find their preferred candidates on the ballot paper if they were grouped by party.

What else we learned from our research with voters

While not necessarily within the scope of the research we did find some additional issues arising which may inform the design of any future ballot papers.

How do voters typically navigate the ballot paper?

Our eye-tracking data enabled us to better understand how voters navigated the ballot paper and identify which sections they paid more attention to.

Participants who used the paper ballot, and also those who participated in the eye-tracking research, told us that, unless they were looking for the specific name of a candidate, the party emblems were the most useful section of the ballot paper to help them find their preferred candidate rather than candidate or party name.

The data from our eye-tracking research told us that participants spent almost as much time looking at party emblems as they did the names and addresses of candidates (7.9 seconds looking at emblems as opposed to 9.6 seconds looking at names and addresses).

The instructions and party names were looked at for the least amount of time (4 and 3.9 seconds respectively).

The interest in party emblems from voters perhaps explains the fact that there was a perception amongst some participants that it was unfair that independent candidates did not have an emblem.

Electoral administrators and their suppliers told us that they could implement and deliver any of the alternative ballot orders being proposed.

However, they were concerned that any changes may lead to the following:

• confusion for voters if two different versions of the ballot paper were used across a ward
• pressure on the electoral timetable which may delay the printing of ballot papers and the issue of postal votes
• increased costs which councils would need to meet

Electoral administrators would want to see prescribed in legislation the detailed rules for administering any alternative order.

What electoral administrators said about alternating A to Z and Z to A ballot papers

The electoral administrators we spoke to questioned how ballot papers would be issued in the polling station. Their strong preference was to have one book of ballot papers in which the order alternated rather than two books of ballot papers in which each ballot followed the same order. This approach would be reliant on polling staff alternating between two books as they issued the papers.

Returning Officers were concerned that voters may be confused if two versions of the ballot paper were in circulation and may even question the veracity of one version, thereby undermining confidence in the poll.

This was seen as a particular concern in the case of postal voters where two people in the same household may receive differently ordered ballot papers.

It was noted that particular consideration would need to be given to the instructions in postal ballot packs under this option.

Returning Officers also identified potential issues over the procedure for re-issuing ballot papers, either in the polling station or to postal voters (where a voter has unintentionally spoilt their ballot paper or postal ballot statement).

The voters would have familiarised themselves with one order and may receive a differently ordered version as a replacement.

Returning Officers were clear that the election rules should explicitly require the re-issued ballot to replicate the order of the initial ballot issued. This may complicate the compilation of the corresponding numbers list, which polling staff are required to complete to record which ballot paper has been issued to which voter. This is an important safeguard against potential fraud.

Returning Officers also noted that the two differently ordered ballot papers could slow down the process of printing ballot papers (as two sets of data would need to be sent to the printer), and also for proofing ballot papers.

Two differently ordered ballots could also impact the speed and accuracy of a manual STV count at a by-election, where count workers would need to familiarise themselves with the two different ballot orders.

What electoral administrators said about drawing the order by lot

Electoral administrators thought that if a change had to be made to the ballot paper, in general, the drawing of order by lots may present less difficulties for the administration of the poll as all ballots would be in the same order.

Electoral administrators did, however, highlight the need for clear rules and consistency on how the drawing of lots was completed, including details of those who would have a right to observe the process and how any objections to the process could be recorded.

Their preference was for prescription via the election rules or by direction from the Elections Convener.

However, some suggested that a good practice approach could be set out in our guidance.

Electoral administrators raised concerns that in local government areas with large numbers of wards the drawing of lots may take up to one day to complete for all wards in that area.

This could put pressure on the election timetable by delaying the printing of ballot papers and any subsequent issue of postal ballot packs.

This would further limit the time available for postal voters to receive, complete and return their postal ballot packs in time for their vote to be counted.

What electoral administrators said about changing the order of ballots

In general, Electoral administrators were clear that any change, regardless of what it was, would need to be reflected across all formal election notices including the notice of poll, tendered ballot papers, polling station posters and large print ballot papers.

Some Returning Officers believed that the STV ballot paper was already complex enough for voters without risking any additional complexity with a change to ballot order.

Electoral administrators also noted that any suggestions of alphabetical bias on ballot papers were a concern for parties rather than voters and might be better managed through more vigorous local campaigning to familiarise voters with the different candidates.

Electoral administrators also told us that any changes, particularly if adopting an alternating A to Z and Z to A ballot, would increase printing and potentially counting costs. Councils currently meet all local government election costs – other than the costs of e-counting – from their own budgets.

They were keen to establish whether the Scottish Government would be meeting any additional costs that were incurred as the result of any changes.

We invited views from all the parties currently represented in the Scottish Parliament. We received comments from party officials from the Scottish Conservative Party, the Scottish Liberal Democrats and the Scottish National Party.

There was no clear consensus amongst parties about whether the system should change and, if so, in what way.

This may be because not all parties stand multiple candidates in the majority of Scottish council wards and so they do not consider the perceived problem of alphabetical order effect as an issue for them.

Only the Scottish National Party expressed a strong desire for change. They believed that the current system should be changed to alternating ballots of A to Z and Z to A.

They argued that this would mitigate the effect of alphabetical order effect whilst also giving earlier certainty as to the order of their candidates on the ballot paper.

In their view this would enable earlier planning for election material and vote management strategies across wards.

They believed that the drawing of order by lot would not mitigate any ballot order effect (by merely replacing the alphabetical advantage with the lot drawing advantage).