One town hall and how many minds changed

More than one in four Calgary voters who joined a single telephone town hall before Alberta's October 19, 2026 referendum say it changed their vote or helped them decide.

Set beside twenty years of research on what persuades people, that result says something bigger. Most persuasion tools barely work. The one that does, two-way conversation, has always had two weaknesses: it reaches people one at a time, and it's only as good as whoever makes the call. A telephone town hall fixes both.

A real vote, with ballots already cast

Converso ran a telephone town hall for South Calgary households ahead of the vote. Over 3,600 people joined one live call, and everyone heard the questions, the objections and the answers in real time.

Three days after the vote, we surveyed attendees. 220 responded, and 82% of them had voted. Every answer describes a decision already made, not an intention.

What moved

27% said the town hall changed their vote or helped them decide on the nine policy and constitutional questions. On separation, the most entrenched question on the ballot, one in ten moved.

We count people who walked in unsure and walked out decided as persuaded. We won't focus on the one-third who said they felt surer of their choice, although confident voters are the ones who turn out and repeat the argument to others.

The lower separation figure reflects how few people were still open. In August, 16% of likely voters were unsure on the nine questions on average, against 4% on separation. Applied to all nearly 3,700 participants, these rates mean about 800 voters moved on the nine questions and about 300 on separation, in one hour on one evening.

How that compares to the research

Two decades of controlled experiments on what changes minds line up closely with what we saw in Calgary.

Ads, mail and brief contact barely register. Kalla and Broockman's 2018 review of 49 field experiments found an average effect of about zero in general elections.

Real conversation works, and it lasts. In a 2016 study in Science, Broockman and Kalla found that one ten-minute, two-way conversation at the door changed attitudes for about one person in ten, for three months or more. Calgary's separation result matches that rate, on the campaign's most hardened question.

It works by phone. In 2020, the same method by phone added about 3.1 votes per 100 completed calls. Per person reached, that's roughly 100 times more effective than the ads, mail and short, one-way scripted calls a typical campaign relies on.

So do town halls. Abernathy and colleagues (2019) found that an hour on a telephone town hall improved how people rated the host, with gains two to three times larger among those who started out unfavourable. In randomized experiments, Minozzi, Neblo, Esterling and Lazer (2015) found that online town halls with sitting legislators changed policy views, trust and vote choice.

One honest note: Calgary's numbers are self-reported by people who chose to attend and to respond, and there's no control group. That's why we read them against the experiments. The experiments show that conversation causes change. Calgary shows the same pattern, at a similar size, on a live ballot.

Same conversation, far more people

Door knocking and kitchen-table meetings work for the same reason town halls do: a real person raises an objection and gets a real answer. What limits them is arithmetic. The 2020 phone experiment needed a team of trained volunteers just to complete about 700 conversations.

A telephone town hall takes that limit away. In Calgary, 3,600+ lines heard every question and every answer, including objections people hadn't thought to raise themselves. The conversation carries on after the call, too. Lazer and colleagues found that town hall participants went on to discuss the issue with others, adding up to 145% of the number who took part.

In-person town halls can persuade too, but they mostly reach people who already show up: studies of public meetings find that attendees skew older, male and homeowning. The people least likely to engage in traditional politics are the most interested in taking part when they can do it from home, and they tend to be more open to persuasion.

The right messenger, the same message, every time

One-to-one calling also has a quality problem. It only works when the caller has a real conversation: listening, handling the hard objection and getting the facts right. That depends entirely on who's on the line.

In practice, most volunteer and paid phone banks fall back on short, scripted calls, because that's the only way to get through the call list. Those calls are exactly the kind of brief contact Kalla and Broockman found had about zero effect in general elections. Even when callers try for a real conversation, volunteers get tired or nervous and paid callers read lines they don't own. Training helps, but quality still varies from one call to the next.

A town hall takes that variation away. The most credible person you have, whether that's the mayor, the minister, the bargaining committee chair or the CEO, gives every listener the same answer in the same words. When the hard objection comes up, they answer it once, in front of everyone.

Credibility does real work on its own. Arthur Lupia's research on California ballot measures found that voters who knew where a trusted source stood voted much like voters who had studied the measures in depth.

What it means for your organization

• Government: Consult beyond the room. A hall holds 80 people, while a call holds thousands and gives you a record of who took part.
• Political parties and campaigns: Leadership races, nominations, by-elections and referendums are where minds are still open, and in Canada the hard part is reaching people. A rural riding can cover thousands of square kilometres, and a national campaign spans six time zones. One live call puts your candidate in a real conversation with members and voters across all of it, in either official language, and holds their attention far longer than an ad or a mailer.
• Unions, associations and companies: Before a ratification vote or a major change, answer every objection live, in front of everyone.

The question isn't how many people you informed. It's how many came away thinking differently, and whether they heard it from the person they trust most.