Skip to content
Citizen Infrastructure

Deliberative polling

An ordinary opinion poll captures what people think off the top of their heads. Deliberative polling, pioneered by James Fishkin at Stanford’s Deliberative Democracy Lab, asks a better question: what would people think under good conditions — if they had the time, the information, and the chance to talk it through?

Three problems with ordinary polls

Section titled “Three problems with ordinary polls”

The method exists because raw polling is unreliable in three ways:

  • Rational ignorance — most of us sensibly tune out the details of ballot measures and candidates, because we have jobs, families, and lives. Not caring isn’t the issue; not having time is.
  • Phantom opinions — people will give an opinion on things they know nothing about, because admitting “I don’t know” is embarrassing. In a famous 1975 experiment, George Bishop asked the public about the “Public Affairs Act” and got firm agreement and disagreement — about a law that did not exist.
  • Echo chambers — we increasingly hear only news that suits us, so the person next to us may have a completely different picture of reality.

The method

Section titled “The method”

A representative microcosm of the population (selected by sortition) takes a survey, then receives balanced briefing materials vetted across the spectrum, then spends days in small-group discussion and plenary Q&A with experts, then takes the survey again. The finding is the change between the two — what people conclude once they’re actually informed and have heard each other out. The Lab has run this in 50+ countries across 100+ projects.

The results are often striking: at America in One Room (2019), positions softened on the hardest issues and the warmth each side felt toward the other rose measurably. Other notable runs:

  • Mongolia became the first country to legally require a national deliberative poll on any constitutional amendment before it can be considered (a 2017 law; the first poll convened 669 citizens).
  • Texas utility companies ran deliberative polls with their customers in the 1990s; the consumers backed renewables, and Texas went from near-last to first in US wind power.
  • Deliberations have also shaped policy toward the Roma minority in Bulgaria and electoral reform in South Korea.

How it differs from a citizens’ assembly

Section titled “How it differs from a citizens’ assembly”

Both are deliberative mini-publics, but they lean different ways: a deliberative poll is usually larger, focused on measuring informed opinion (often to inform a decision or a vote), while a citizens’ assembly is usually smaller and deeper, focused on producing concrete recommendations. New AI-assisted platforms (the Stanford Online Deliberation Platform) aim to run the small-group format at far larger scale. See also the epistemic case for democracy.

Sources

Section titled “Sources”