As political participation, government services, and social interaction increasingly go digital, are we ready for i-voting – remote online voting – through a few clicks on a phone or laptop? Manu Savani and Justin Fisher look at what makes British voters willing to take up i-voting
A number of countries, including Austria, the Netherlands, Norway and the UK, have trialled i-voting but not taken it up. Germany banned online voting in 2009; South Africa halted potential moves towards it in 2021 because of concerns about fraud. Canada, Estonia and Switzerland are relatively rare examples where i-voting is being used.
Most of those trials took place some 10–15 years ago. The world is a different place now. When it comes to government services, we are increasingly going ‘digital by default’. Political participation and social lives are played out in the digital world with greater frequency. In the UK, at least, now seems a good time to reappraise the merits of i-voting.
Covid introduced new arguments around elections and public health, too. In the UK, the 2020 local elections that were postponed under pandemic restrictions might have been able to go ahead had i-voting been an option. In a timely report on The Future of Voting, the UK’s Electoral Commission revisited public attitudes towards i-voting. The report revealed public interest in i-voting, most people citing convenience as its main selling point. But the report also identified trade-offs and concerns, including around security, privacy, and confirmation that the vote had indeed been received.
Elections that were postponed under pandemic restrictions might have gone ahead, had i-voting been an option
We conducted two studies of British voters, with the aim of extending and deepening our understanding of attitudes towards i-voting.
First, we looked at whether voters care about who is in charge of the election. Does it matter whether it is an independent body or a private sector firm?
Secondly, we investigated whether the chance to test a prototype voting app changes opinions about i-voting.
Yes, it does. We conducted a survey experiment with a representative sample of 1,855 UK voters. Our experiment measured how willing respondents were to take up i-voting if it were available, on a scale of 0 (not at all) to 10 (very willing). We assigned respondents randomly to one of three groups, who each received a different informational vignette.
The first (control) gave some background on what i-voting is, and explained that it was being considered in the context of the pandemic. The second and third vignettes mentioned the advantages of not needing to travel and reducing public health risks. The second message specified ‘an independent body like the Electoral Commission’ would run the online ballots; the third referred instead to ‘a well-regarded tech company’ running the ballot.
We expected the positive frames would increase willingness to i-vote, with a stronger effect for the public sector treatment message. Respondents did indeed prefer the independent body, rather than a tech firm, to be in charge. Even with the additional positive frames, mentioning the tech firm seems to have put voters off, as there was no significant increase in willingness to i-vote compared with the control group.
Our survey found that voters would prefer an independent body – rather than a tech firm – were in charge of i-voting
Unsurprisingly, we also found that users positively associated political engagement and trust with willingness to vote online. Prior exposure to digital risks was important. The more a person was already using the internet daily, and exposed to online risks – with personal banking, for example – the more likely they were to respond positively to the public sector treatment.
The control group reported a mean willingness of 6.5 on the 0 to 10 scale. Those in the public sector treatment reported a mean of 6.9. The brief additional information gave rise to statistically significant but modest effects.
The online survey experiment generated robust causal inference when it came to testing attitudes towards the management of i-voting. But the i-voting itself remained an abstract concept. What happens when we give potential voters the opportunity to interact with a new voter interface on their phones and laptops?
With our computer science colleague Fotios Spyridonis, we designed a voting app that mimics a British ballot paper. We invited 32 London-based participants to our user testing lab. They ‘voted’ with the app on their smartphones, a tablet, and a laptop, before returning to a traditional voting slip and ballot box. Debrief interviews and responses to a UX survey generated rich insights.
Two-thirds of our participants said they would opt for i-voting if it were available, based on its convenience, ease of use, simplicity and accessibility. On the 0 to 10 scale, our participants now gave an average score of 8.4.
Participants evaluated the app positively, particularly in terms of design, layout, and navigation. Gaining experience of the app led to participants either maintaining (if they had high initial willingness) or improving (if they were initially less keen) their willingness to vote.
Respondents wanted to see stronger security features, and greater transparency about potential data risks
If i-voting were to become an option in future, respondents wanted to see stronger security and authentication features. They also wanted campaigns to inform voters about how i-voting works, and transparency about data risks, actors involved, and security measures in place.
Recent news that the UK Electoral Commission faced a cyber attack on the voter registration database shows that significant hurdles exist for i-voting. The research agenda is ripe for more work exploring which voter groups might benefit most from the option of i-voting, how design features could address security concerns, and how to ensure the voting experience remains ‘special’ and ‘significant’. Mixed-methods approaches combining experiments with rich conversations with voters, online with in-person testing, hold real promise.