Advance voting pilots evaluation

The pilots were well run by Returning Officers. There were no notable issues on the advance voting days or on Thursday 5 May and voters were satisfied with their experience of voting. However, the results indicate that the opportunity to vote in-person ahead of polling day does not, on its own, boost turnout significantly. This is not unexpected; we know that people’s motivation to vote or not to vote is driven by several factors and is not solely linked to convenience or choice of method. 

While the turnout of early voters was low, the option was welcomed by those that used it and it does offer an additional choice for voters. We cannot judge, from the evidence of the pilots, what impact advance voting, if introduced, would have on turnout over time. 

The experience of the pilots provides some useful information on how advance voting could work in practice. For example, the use of electronic registers was successful, allowing the register for any area within the local authority to be accessed via a single device. However, the small number of pilots and the selection of locations in which they were tested means there is still work to do to develop any future policy and to understand how it could be implemented. 

Proposed changes to voting processes should be designed to deliver a likely benefit to voters, maintain the security and integrity of the system, and be realistically deliverable by electoral administrators. Our evaluation has identified several specific areas that need to be addressed if a further roll out of advance voting is considered:

• Location and number of venues –The selection of venues would be crucial to any implementation of advance voting and has the potential to cause concern to voters and candidates. Any plans for further use of advance voting should carefully consider how to select the right number of venues to offer a consistent and useful option to voters while being realistically deliverable by Returning Officers. Returning Officers should also have some flexibility to reflect the circumstances of their areas. For example, larger geographical authorities or areas with less accessible transport links are likely to need more sites. Increasing the number of venues would require an assessment of how they would be staffed. For example, it is clear from the Bridgend pilot that the requirement to staff around 20 advance voting locations over several days was extremely challenging.
• Number and choice of advance voting days – there is no clear evidence from the pilots of different days being more popular with voters (the low number of votes cast makes it difficult to draw a clear conclusion). There is, however, some evidence that the choice of days can have an impact on the administration of the poll, depending on how they interact with the existing preparations for a Thursday polling day. The challenges presented are not at all insurmountable but may require additional electoral administrator staff capacity to address them, depending on the days chosen.
• Costs – Given the low turnout among early voters, it is important that any future changes are informed by an assessment of the costs and benefits of wider implementation, including the use of electronic registers to support advance voting, which was the largest area of additional costs in the pilots. This should include an assessment of any development costs that would not be needed again and whether economies of scale could be realised. The costs of the pilots cannot in themselves be taken as a clear indication of the likely costs of any future roll out of advance voting.
• Resilient electoral services – we know that many elections teams already face increasing challenges in delivering well run polls within existing core staffing and budgets. It is therefore important that existing capacity is considered if any roll out is planned. There is the potential for specific policy decisions (more days, more venues) and the circumstances of polls (combination, greater turnout) to put more or less pressure on elections’ teams. 
• Raising public awareness – any wider roll out should be supported by additional, more coordinated and centrally-funded public awareness activity. This would underpin ongoing local awareness raising activities (such as were undertaken at these pilots). 
• Time to plan for changes – changes concerning the running of elections should be agreed and in place six months before the poll to allow for adequate implementation by electoral administrators and so voters can be informed about what they need to do. Planning for any future roll out of advance voting should ensure that at least that minimum period is available.

Few voters chose to vote early

A small number of voters chose to cast their vote at the advance voting polling stations. The three pilots with a single advance centre had similar results, with 0.2 – 0.3% of registered polling station voters casting a vote on the advance days. In Bridgend, where a number of usual polling stations were open, the proportion was slightly higher at 1.5%. 

The level of turnout is broadly in line with the previous tests of advance voting carried out in the early 2000s. In the final set of pilots in 2007, the proportions ranged from 0.5 – 7.0%. These earlier results are not directly comparable as the specific approaches being piloted varied. For example, of the five pilots in 2007 only one had two days of advance voting; the other four had between four and nine days. Most also had more than one voting centre and several were not piloting advance voting for the first time.

It is difficult to draw conclusions from the small number of pilots in May 2022. The data could suggest that a larger number of polling stations are more likely to attract voters than single, central venues. However, other factors could be driving the difference, including the local elections themselves and how competitive voters perceived them to be. 

We can also draw no conclusion on any impact from the days chosen for voting. Caerphilly and Torfaen’s advance voting days were on Saturday and Sunday and they saw similar results to Blaenau Gwent where voting was on Tuesday and Wednesday.

Overall, the results indicate that the opportunity to vote in-person ahead of polling day does not, on its own, boost turnout significantly. This is not unexpected as we know that people’s motivation to vote is driven by a range of factors, not solely linked to convenience or choice of method.

We surveyed a sample of those who chose to vote early across the four pilot areas immediately after they had voted: 75% told us that they would have voted at the elections even if there was no advance voting option. 97% of these early voters said that they always or sometimes vote at local elections – suggesting that early voters were the already engaged, regular voters at local elections.

When we asked a representative sample of eligible electors in the pilot areas about whether the ability to vote over multiple days would affect their likelihood to vote: across the four areas between 27 and 37% said that it would make them more likely but around two-thirds said it would make no difference or that they were not sure. 

In order to test whether any changes in turnout in the pilot areas were linked to the advance voting option, we identified four control areas – local authorities with standard local elections and similar demographics and past turnout to the pilot areas  (see Background section for full details).

All of the pilot and control areas recorded a drop in turnout between 2017 and 2022 strongly suggesting that these differences are unrelated to the piloting activity and reinforce the point that the method of casting votes is far from the sole driver associated with turnout. Overall turnout in Wales was down compared to the last local elections in 2017 – from 42% to 39%. 

The low levels of advance turnout in the pilots should also be considered against the take up of electoral changes in the past. In 2001, at the first UK general election after postal voting on demand was introduced, 3.9% of the electorate were issued with a postal vote. This was a small increase on the 2.1% from 1997 when the rules were different. However, by 2019, 17.2% of the electorate were issued with a postal vote and it is now an important route by which a large proportion of the electorate choose to vote. We cannot judge, from the evidence of the pilots, what impact on turnout advance voting, if introduced, would have over time.

Some pupils were able to vote at school

The Bridgend pilot also saw a polling station opened for one day (Tuesday 3 May) in a local school. This was only for pupils at that school, who were registered to vote, to cast an early vote. 

Advance turnout was higher at the school than the pilot average. Fourteen of the 76 pupils who were registered and eligible chose to cast a vote at the school polling station – 18% compared to the 1.5% of eligible polling station voters who voted early. 

As with the wider population, those who chose to vote early were satisfied with the experience. We asked pupils, who did not vote at all at the elections, why they had chosen not to. One third said it was because they did not know enough about how to vote and a fifth said they were not registered. Others also cited a lack of information about the election as a whole and about the specific candidates.

More awareness of the pilots may not have led to higher advance turnout

Returning Officers faced two significant challenges in raising awareness of the opportunity to vote in-person before polling day: a low level of engagement with the local elections; and the need to target campaigns and not mislead people in neighbouring authorities (or in Bridgend’s case, in the non-piloting wards).

Details of the practical challenges of the latter are set out below and would be much less of an issue in the event of a wider roll out.

However, at polls which attract less public interest it will be an inherent challenge to effectively raise awareness of changes to the method of voting. 

We carried out representative surveys of the eligible voting population across the pilot and control areas. One survey was carried out in February/March 2022, to provide baseline data, with a second completed after polling day on 5 May. In this research we asked if people thought they would be able to vote in person over multiple days (Yes/No/Don’t know). There were changes in claimed awareness between our initial survey and our survey after polling day but the pattern is not clear cut. 

The proportion of people who said you could not vote in person over multiple days increased in both the pilot and control authorities. The increases were smaller in the pilot areas suggesting some effect from the public awareness campaigns. Torfaen was the only pilot area where the proportion who correctly thought voting over multiple days was an option increased between February/March and May. 

All pilot areas had a higher level of awareness of the option to vote early than the Wales average (shown on the chart as a blue line – 14%). All of the control areas had lower than average levels of awareness.

Figure 1 Public awareness of the advance voting pilot

Separately we asked those in the pilot areas if they were aware that they could vote in person before the usual polling day on 5 May. Claimed awareness ranged from 22 – 30% with the rest either saying they were not aware or were not sure. 

We also asked people who said they voted on Thursday 5 May why they chose not to vote early. Nearly half (46%) said it was because they did not know they could (30% said they preferred to vote on the ‘usual day’). 

The 46% of polling day voters who said they did not vote early because they did not know they could is a good indication of the low levels of awareness of the pilot. However, it is a less strong predictor of alternative behaviour – we should not assume that all or many of that 46% would have voted early if they were aware of the pilot. 

This is supported by evidence from the 2007 pilots. In that case awareness was higher: half of survey respondents across the piloting areas said they were aware of the advance voting option. But levels of turnout were very similar (with the exception of one area) to the 2022 pilots.

It is important to maximise the public’s awareness of changes such as these, but it is also unlikely that greater levels of awareness in this case would have resulted in very significant jumps in either advance turnout or overall turnout.

People had a range of reasons for choosing to vote early

In interviews outside the advance voting centres, we asked early voters to select, from a list of options, the main reason they had chosen to vote early: 45% said it was because they were going to be busy on Thursday 5 May, 14% said it was ‘more convenient’ and 10% said it was because they lived near to an advance opening polling station. 

In our separate online survey, those that claimed to have voted early were asked to pick the reason that best explained why. The table below shows the results for the combined pilot areas.

Reasons for choosing to vote in advance

The level of voting was consistent throughout the advance voting days

Notable patterns in when people chose to vote would tell us about how any wider roll out of advance voting could work most effectively. 

However, the hour-by-hour analysis shows a largely even distribution of votes across the different opening hours of the pilots. There are no significant patterns except for consistently lower levels in the last hour of opening across all four pilots.

There are no real differences between the weekend and weekday pilots. 

Figure 2 Cumulative distribution of votes

Early voters were positive about their experiences

The majority of these voters had a positive experience when voting and welcomed the opportunity presented by the pilots. Nearly all (99%) of those we interviewed outside polling stations said they were satisfied with the experience of voting early. 

When asked to say why they were satisfied, the main reasons selected were that it was quick to vote (76%), easy to vote (64%) and convenient to vote (45%). Around three-quarters (77%) said they would be very likely to vote early again if there was an opportunity to do so and a further 15% said they would be quite likely.

There was limited wider impact on public attitudes but perceptions were positive

We also looked at the potential wider impact on public attitudes through our surveys in the pilot and control areas. 

Overall, given the small numbers of people who chose to vote in-person and early, there is limited scope for notable impacts on public opinion. However, this also means that the absence of a change in attitude does not mean that the option to vote early would not have an impact if more people were aware of it and/or used it.

• Positively, in all areas satisfaction with the process of voting increased between the initial survey and the follow up, but with no significant difference between pilot and control areas. Satisfaction was also high with 92 – 93% of voters saying they were satisfied with the process.
• We asked if people thought the option to vote in-person on multiple days would make voting more or less convenient. The results show a small difference between the pilot and control areas, with a larger proportion of respondents in the pilot areas saying it would make it more convenient. The pilot areas also all show an increase in perceived convenience between the two surveys (only one of the control areas show an increase). These differences are all small and we should be cautious in drawing conclusions. 
• In order to assess any impact on public perceptions of the security of the poll we asked people whether they thought the ability to vote across multiple days would make voting more or less secure. There were no significant differences between the pilot areas and control areas in our surveys. People were most likely to say they thought it would make no real difference (54% across Wales said this in our post-election survey and a further 10% said they did not know). Overall, when asked if voting at a polling station was safe from fraud and abuse the vast majority in both the pilot and control areas said that it was. The results were in line with the overall finding for Wales where 85% of people said it was safe (10% said don’t know and 5% said it was unsafe).

Early confirmation of the scope of any change is important

The legislation allowing for the pilots was in place by March 2022. Returning Officers told us that, although there was no impact on the eventual running of the pilots, they would have preferred the details to be confirmed earlier. This would have avoided the need to carry out some preparations at risk. The practical impact in this case was offset by good communications between the pilot authorities and Welsh Government officials. This meant that there were no surprises in the legislation and that planning could continue before it was in place. However, this would be less manageable if the change applied to all local authorities across Wales.

Changes concerning the running of elections should be agreed and in place six months before the poll to allow for adequate implementation by electoral administrators. Planning for any future roll out of advance voting should ensure that at least that minimum period is available. 

The local public awareness campaigns used a wide range of channels to reach potential voters

Some of the challenges in engaging the public in this type of change are set out above, and campaigns run by the pilot authorities should be considered against those limitations.  

Returning Officers and their teams told us that they used a range of approaches to inform people about the pilots, specifically:

• Social media – mainly Twitter, Facebook and Instagram posts
• Local radio adverts 
• Household notification letters – these are letters sent out in February to confirm the registration details for each property in an area
• Local news articles and press releases

There were also other methods including council websites, billboards, post campaigns and ‘Promo’ or ‘Ad Vans’, which could be positioned in higher profile spots in the area. All of the pilots began their campaign in February in order to allow enough time to carry out a wide range of activities before the advance polling days. Information about the pilot was also featured on the poll cards in each area as required by the pilot legislation.

In our public opinion survey after the election we asked everyone who said they were aware of the pilot where they had seen information about it. The most common sources were ‘Council website’ (29%), ‘social media’ (21%), ‘leaflet/flyer from the council’ (16%) and ‘news website’ (15%). 

When we asked people who had chosen to vote early where they found out about the pilot, the main answer was the poll card (38%). A fifth of voters (19%) were made aware by a leaflet or newsletter from the local authority and 15% said they found out from their friends and family.

We also asked people how easy it was to find information about different aspects of the election. The perceived ease of finding information on the advance voting option (41%) was lower than other aspects such as information on parties/candidates (59%) and information on what the election was about (64%). 

The evidence suggests that some channels were more effective than others at reaching people. However, we cannot draw conclusions as the overall level of awareness was low and engagement with the pilots was primarily from more regular voters. 

If this change is rolled out nationally it is likely that all of the channels used in the pilots would be important to spread awareness with different sections of the public. However, they should be supported by larger, national activity.  A national campaign can use higher impact measures (such as TV), with consistent messaging across all local authorities and greater reach among the population as a whole. This would help to increase the likelihood of cut-through for the campaign messages beyond existing, regular voters.

The use of electronic registers was important to the smooth running of the pilots

The pilots each worked with the supplier Modern Democracy, using their system which provides electronic registers in the polling station on tablet devices. This allows for people to be marked as having voted and for registers for any area within the local authority to be accessed via a single device. 

In the three pilots with a single advance voting centre, this avoided the need for large paper based copies of the full local authority register to be held in each centre and for individual voters to be quickly found on the register when they came to collect a ballot paper. This also allowed Returning Officers to efficiently manage one of the integrity challenges that comes with multiple polling days – the risk of double voting (an elector attempting to vote more than once). We are not aware of any issues with double voting in the pilots.

Returning Officers and their teams felt that the pilot could potentially have been delivered without the IT elements. This was particularly the case for Bridgend where advance voting was happening in usual, individual polling stations. 

However, taking forward advance voting in one or more centralised locations per area without using electronic registers could put more pressure on Returning Officers and increase the risk of problems arising. The single-centre pilots reported that in the absence of the electronic registers it would have been a less good process for voters – and potentially very difficult for them to manage with higher turnout. It would have introduced challenges for administrators associated with the paper-heavy processes. This would have been a particular issue where advance voting is happening immediately before polling day as there would not be enough time to manage aspects of a fully paper-based process. For example, marked register data from advance voting centres would need to be transferred to the paper registers held in each ‘normal’ polling station. There is also the potential for other benefits to be realised from the use of electronic registers alongside supporting advanced voting including accuracy of records.

The hardware and software solution used in the pilots was a success – there were no significant problems on any of the polling days and more minor issues were quickly resolved. However, there are some lessons to be drawn from the pilots:

• Time for development and set up: the pilots were arranged to tight timescales and this included the work needed to develop the electronic registers system for the circumstances of the pilot. Additional time in this case would have made the process more comfortable but also potentially allowed for a smoother ‘back-end’ process for administrators through better integration with their existing electoral management software (EMS) systems. The tight timings lead to a functional, but not optimal, solution to linking the two systems. There were also some issues in managing two different piloting requirements between the single-centre approach and that taken in Bridgend. The requirements were not the same and more time to understand the differences would have helped to avoid some of the issues that needed to be dealt with during set up.  
• Training and support: for a pilot, bespoke training is often put in place that may not be sustainable or needed in the event of a wider roll out. In this case many administrators felt that more training than necessary was provided for polling staff. Many polling station staff have existing full or part-time jobs. Asking them to attend training during work hours could therefore be unworkable for some. Some pilots also felt that there was an overly cautious approach to other set up work, such as operational testing. While they acknowledged it was much better to have too much than too little, they clearly felt that it could be scaled back in the event of a roll out. 

The selection of venues was based on several considerations

For the three pilots using a single location, which was not usually a polling station, there were several factors considered when choosing the venue:

• Location: Returning Officers wanted venues which were central to the area, not just geographically but also in terms of transport links. A reasonable level of usual footfall was also recognised as a benefit, as was a venue which was already recognisable to people or where people may need to go to for a reason other than voting. Easy parking was also sought for those travelling from further afield.
• Accessibility: the venues needed to be fully accessible 
• Security: the Returning Officer needed to be confident in either being able to store ballot boxes securely within the venue itself (the council offices in Torfaen and Caerphilly) or to be able to transport them safely and securely to another site between each polling day (Blaenau Gwent moved the boxes from the Learning Zone to the nearby council offices).
• Space: each centre needed to have sufficient space to set up a number of polling station desks and to manage the flow of voters effectively. The space required varied by the size of local authority (and therefore the likely number of voters) but also by the approach taken by Returning Officers. For example, Torfaen and Blaenau Gwent opted to use two desks in their centre while Caerphilly used six desks.

The number and location of venues used in each area should be carefully considered

The factors above make choosing suitable venues more challenging than selecting locations for standard polling stations as they limit the Returning Officers’ choices. In the pilots it was relatively easy to find a single venue in an area, but it could become more challenging to find a greater number. 

There is no evidence from these pilots that the use of a single location in the local authority area significantly deterred voters who lived further away or had less easy access to the advance voting centre. However, this may be because of the low overall numbers of votes being cast early. In Blaenau Gwent, for example, the number of early votes across the different wards varied from one to eight. This is not enough variation to show a pattern. The advance voting option also mostly attracted engaged, regular voters whose tolerance for making longer or more complicated journeys will be greater than the average. 

In addition, the pilot local authorities are relatively small, geographically, compared to many areas in Wales. If advance voting was being rolled out widely there would need to be consideration of how many venues were needed in each area and their locations, taking account of overall size and transport connections. This should also take account of anticipated turnout as a single venue can cope with low levels of well-distributed turnout but may struggle with much higher turnout including concentrations at specific times of day. 

No complaints or negative feedback were received by Returning Officers about the locations of the centres in the pilots. We also received no comments on this aspect in our survey of candidates. However, there is potential for any future choice of venue to cause concerns among political parties and candidates. In any roll out, Returning Officers would need to ensure that they engage with local parties and candidates about the location(s). This process could be modelled on the existing approach to reviewing polling districts and polling places where the Returning Officers is required to follow a transparent process which involves proactively seeking views, publishing representations received and being clear on how decisions were reached. 

Access to usual polling stations was also a challenge in Bridgend

Bridgend aimed to use their regular polling stations in the low turnout wards selected for the pilot. However, in several cases this was not straightforward as access was needed for three days rather than the usual one. Schools in particular, which often need to close during polling, were reluctant to shut for three days. The two school locations which usually house polling stations could not be used at these elections and temporary portable buildings were put in place instead. This added work for the elections team as these venues then need to be set up to ensure they are accessible, etc.

If the approach taken in Bridgend was to be used more widely and particularly if it was used across an entire local authority then there could be challenges in finding polling stations. 

The days of the week chosen for advance voting could have an impact on ease of administration

The two approaches, of voting at the weekend before the poll and on the weekdays immediately ahead of polling day, were delivered without any significant issues. However, some of the administrators involved said that advance voting on the Tuesday and Wednesday before a Thursday election brought some additional pressures. 

Firstly, it largely removes contingency time for dealing with issues arising on the advance voting days ahead of a Thursday polling day. As above, in the event of issues with the electronic marking of registers, it would have been difficult to fall back to the full paper process for the Thursday.

Secondly, polling on the Tuesday and Wednesday has more overlap with some of the usual activities that administrators need to complete ahead of a Thursday election. This puts additional pressure on core elections teams which are already under strain and are often relatively small, with limited capacity to absorb new demands on their time. The teams in the pilots managed this issue well but that was partly achieved by notable overtime for core staff. This was particularly the case for the Bridgend pilot where managing the opening of over 20 polling stations across three days required significant extra work. 

In considering the use of advance voting in the future  it would be crucial to ensure that core elections teams have the capacity and resilience to manage the additional requirements. Where teams are already under strain, any new demands on their time creates a risk to the running of the election. This should also take account of wider factors that could have an impact, such as more complicated polls. 

Additional polling staff and some further training was needed

Many more staff were required in Bridgend where more sites for advance voting were available. In the single advance voting centres the numbers required were largely determined by the overall approach to managing the flow of voters, e.g. Blaenau Gwent and Torfaen used two polling station desks and Caerphilly used six.

This meant that overall numbers varied from six additional staff in Blaenau Gwent and Torfaen to 14 in Caerphilly and over 100 in Bridgend.

The single-centre pilots reported no issues with finding appropriate staff. They also felt that the number of staff was sufficient to manage the work on the advance voting days. 

It was unsurprisingly more of a challenge for Bridgend. While they recruited enough staff for the advance stations it was not easy to do so, and they were sceptical that it would be possible to find sufficient staff if all of their polling stations had been open for three days. Many polling station staff have jobs, including in local government. They need to be released from those existing roles and the more days that are required, the more difficult that can be.

Bridgend used two shifts to manage the staff workload with each covering a morning or afternoon on the advance voting days (one shift then also worked the full Thursday polling day). Their approach also took account of some detailed logistical challenges, for example the same workers would close a station on the Tuesday evening as opened the same station on the Wednesday morning (and the same for the Wednesday to Thursday transition). This aimed to minimise any issues around access to sites and/or equipment.

Across all the pilots, as the job was very similar to a standard election, the approach to training was also close to that normally employed for polling staff. Some further training (separate from the IT training mentioned above) was made available to staff in the advance voting centres to address the specific pilot challenges (such as managing marked registers across the days, and the sealing of ballot boxes). In all cases the pilots told us that this was a manageable additional requirement.

There were no issues with the integrity of the poll in pilot areas

There are several potential integrity challenges presented by managing voting over multiple days. These include, as set out above, ensuring ballot box security overnight and managing the risk of double voting risk. 

Returning Officers and their teams managed these risks well, working alongside local and force SPOCs. No concerns in relation to the integrity of the polls in the pilot areas have been raised with the Returning Officers, the police or by political parties and candidates.

Cost

The table below sets out the costs of the pilot scheme, broken down into the main categories of spend, as provided to Welsh Government by the pilot authorities. The data represents the costs they have incurred directly (such as in relation to some of the IT provision) or have provided to local authorities to fund aspects of the pilot. 

As with any pilot it is difficult to draw clear conclusions from these figures about the eventual cost of any wider roll out. This is for several reasons: 

• Local authorities have taken varying approaches to how they have recorded costs (e.g. whether core team staff costs are included, which pieces of equipment are being charged for) 
• Pilots will attract development costs that may not apply in the same way or at all for a roll out; 
• The local authority areas involved are not representative of all areas; 
• Some future costs may be missing from the pilot (e.g. national public awareness campaigns); 
• There may be economies of scale to be realised nationally, despite the increase in costs that would follow from taking forward the policy in more areas.   

* This includes the costs incurred directly by local authorities for ongoing use of software licences plus the additional cost borne directly by Welsh Government for pilot-specific activities.

For the reasons set out above, the costs of the pilots cannot be taken as a clear indication of the likely costs of any future roll out of advance voting. However, particularly given the low turnout among early voters, it is important that any future changes are informed by an assessment of the costs and benefits of wider implementation, including the use of electronic registers to support advance voting, which was the largest area of additional costs in the pilots.

Pilot scheme

The Welsh Government established a framework for electoral modernisation in Wales in July 2021. These pilots were one of the first initiatives under that framework. Welsh Government worked with volunteer local authorities to shape the pilots and agree the specific approach each area would take. The legislation required for the pilots was published in March 2022.  

Local Returning Officers ran the processes in each area, and Welsh Government oversaw the pilot scheme as a whole

Our evaluation

The law says that we have to publish an independent evaluation of the pilot scheme within three months of the election, and there are specific questions that we have to consider:

• Whether the turnout of voters was higher than it would have been if the scheme had not applied
• Whether voters found the procedures provided for their assistance by the scheme easy to use
• Whether the procedures provided for by the scheme led to any increase in personation or other electoral offences or in any other malpractice in connection with elections
• Whether those procedures led to any increase in expenditure, or to any savings, by the authority

We collected information from different sources to make sure that our evaluation of the scheme is thorough and robust. This included:

• Representative surveys asking people in each local area what they thought of the scheme – one carried out in February/March 2022 as a baseline measure and a second carried out after polling day in May 2022
• ‘Exit interviews’ with early voters after they had cast their vote
• Data on turnout at the elections including the number of early voters and when they cast their vote 
• Views and evidence from Returning Officers and electoral administrators who ran the pilots 
• Information about how much it cost to run the pilot scheme

Control areas

In order to help assess whether any changes in things like turnout or public opinion can be linked to the pilots rather than other factors we identified a set of control areas. These are local authorities with similar demographics (age and ethnicity profile, unemployment, etc.) and similar levels of previous turnout but where no pilots are taking place. This allows us to judge whether changes are only seen in the pilot areas or in both pilot and control.

The table below shows the pilots and their matching control area.

Supporting evidence

• Winter tracker tables
• Post-poll tables (pilot specific)

We worked with the research agency Strategic Research and Intelligence. They produced a report containing a summary of the findings from exit interviews carried out at the advance voting centres. If you would like to receive a copy of this PDF report by email, please send us a request. This report is available in English.

Opens in new windowRequest a PDF

Previous evidence

Between 2000 and 2007 there were a number of pilot schemes which provided opportunities for people to vote in person at advance voting facilities before polling day. 

The final round of pilots took place at the May 2007 local government elections in England. Five local authorities held pilot schemes providing advance voting facilities on various days before polling day, 3 May 2007.

Our evaluation of all these previous pilot schemes found that use of advance voting facilities has been limited and mainly confined to those already predisposed to vote.

Our evaluation of the 2007 pilots highlighted a number of findings which remain relevant for any consideration of advance voting and support our findings from the 2022 pilots in Wales. However, these pilots were not directly comparable to the more recent ones as the specific approaches varied. For example, only one pilot had two days of advance voting, the other four had between four and nine days. Most pilots also had more than one voting centre and several were not piloting for the first time.

Turnout

There was a limited impact on turnout. The percentage of voters casting ballots at advance voting centres ranged from 0.5% - 7%. In our surveys 74% of early voters said they would have voted anyway. There was no significant correlation between turnout levels and the different opening times of the voting centres.

There was also no pattern in terms of a preferred day for advance voting: in two of the five pilots Wednesday was the most popular day. In the other three it was Monday, Thursday and Friday.

Voter experience

Feedback from local stakeholders and public opinion research conducted for the Commission in the pilot scheme areas suggests that electors generally welcomed the increased convenience of advance voting. Those who voted early found the process easy to use.

Management and cost

The processes used to run the advance voting pilots were similar to those used at standard polls and presented limited additional challenge for Returning Officers.

The costs of the pilots varied depending on the nature of the specific pilot. For example, more voting centres and/or more voting days would attract a higher cost.

Security

There were no allegations of personation or other electoral offences reported in relation to the pilots. Risks such as double voting and the storage of ballots were identified and well managed.