How behavioural insights get people to the ballot box

The US is gearing up for an election, and it’ll be watched the world over. But while most people will focus on the end result, many in the civic engagement space might find more to interest them in the journey — specifically, the question of who votes, and who doesn’t.

ideas42, a non-profit design and consulting firm and Apolitical’s partner on our behavioural insights for civic engagement field guide, uses insights from behavioural sciences to address complex social problems. Its work with government includes co-founding the US's former Social and Behavioural Sciences Team, and behavioural design teams with New York City and Chicago.

Among the social problems ideas42 addresses is that very question of voter turnout, explored by the Nonvoter Innovation Lab.

“I think 2020 certainly makes for a high salience election and a moment, from a behavioural perspective, to capture people’s attention and create a better democracy infrastructure,” said Ted Robertson, Managing Director at ideas42.

The lab conducted extensive work on turnout during the 2018 US midterm elections. So what have they learned about getting voters to the ballot with behavioural insights?

Being a good voter

In Minnesota, a partnership between ideas42 and the office of the Secretary of State explored “reframing what it means to be a good voter”.

Through interviews and surveys with eligible voters in the state, alongside other research, the team identified a range of behavioural barriers to voting that they could seek to address.

Among them were a “mental model” defining what it meant to be good voter — that voting once every four years, and missing the midterms, is enough.

Many also described “social pressure” from friends and family as central to their decision making. Yet, as ideas42’s analysis puts it: “In midterm elections, where attention and excitement around voting is often lower than in presidential elections, voters may feel less pressure to participate.”

And while having a sense of voting as integral to one’s identity makes people more likely to vote, the team found that “‘being a voter’ isn’t always top of mind for many people when they consider their central identities and values”.

ideas42’s two experiments sought to overcome these barriers.

Working with the Office of the Minnesota Secretary of State, the team isolated two subsets of voters and sorted each randomly into three groups. Groups were sent postcards containing different messaging.

On each subset, the experiment tested the impact of a postcard with behavioural insights-informed messaging against a more simple reminder postcard, and no postcard at all.

One set of messaging focused on “positive social pressure”, targeting new voters who had registered for the first time, or updated their registration in 2016 and then voted in that year’s presidential election.

This group’s experimental postcard referenced their decision to vote in 2016 and figured 2018 as an opportunity to continue their “voting streak”, as well as suggesting “that whether or not a person votes in any given election is public information”.

This messaging increased turnout by 0.46 percentage points and generated 860 new votes.

The other targeted “infrequent voters”, and sought to play on their “Minnesota identity”. The behavioural science-informed postcard in this experiment “activate[d] aspects of Minnesotan identity that align with a sense of civic duty (shared responsibility, community, and togetherness) and suggest[ed] that voting is an expression of that identity.”

The increase in turnout among the group that received this card was not statistically significant. But ideas42’s write-up of the experiment noted that turnout rose by 0.43 percentage points among the group who got the more simple reminder to vote, and that this simple reminder included prominent Minnesotan imagery.

Another experiment aimed at younger voters saw ideas42 partner with Democracy Works. This civic technology organisation is the creator of TurboVote, an online platform that guides users through the voter registration process and updates them about elections in their area.

Young people were sent a range of different text messages via TurboVote, testing different behavioural insights approaches to encourage them to vote, leading up to the primary and general elections in 2018.

Alongside “social accountability” messages, which emphasised that voting histories are public records, “plan-making” messages also saw a significant increase in turnout at the general election.

These latter messages “prompted people to plan when they would vote, consider where they would be coming from, and determine how they would get to their polling site.”

These are just some of the lab's findings — more can be found via the "further reading" links at the bottom of this page.

Design matters

For Robertson, there’s a key takeaway from the lab’s work that’s broader than any specific deployment of behavioural insights.

“The number one overall insight is that design matters, that how you build your civic processes will either increase or, conversely, dampen, engagement,” he said.

There’s a key lesson here, for governments around the world to bear in mind: “We all know now from the advances of technology in the consumer world, that having things be frictionless matters,” Robertson added.

“It’s time that we really take that to heart in terms of government and civic processes. And it’s not easy. They didn’t make the iPhone overnight.”

Meanwhile, in terms of lessons for 2020, one stands out: “On civic engagement in particular and voting, it is about much more than being easy, it is about being meaningful,” Robertson said.

“When you ask me about things from 2018 that we want to expand for 2020, that’s number one, how we understand and apply behavioural science around identity and meaning.”

(Picture credit: Unsplash)

Further reading:

More information on some of the Nonvoter Innovation Lab's work can be found through these links:

Boosting Turnout in Minnesota's 2018 Midterm Elections

Getting Environmentalists To The Polls

Young Voters In The 2018 Midterm Elections