← all systems · see it in The System
Public opinion
Public opinion, or popular opinion, is the collective opinion on a specific topic or voting intention relevant to society.
Properties published scienceour proposal
Sourced fields, and our own readings.
| born at | booked: late 17th century as a political force; antecedents from the 12th-13th centuries |
| born precision | century |
| boundary | Inside: what a population thinks on public questions taken together: the distribution of individual attitudes on an issue, the climate of opinion people perceive, what they express or keep silent, the salience of issues (what the public thinks most important), and the publics that form around issues. Outside: the events and issues themselves, the news media, platforms and social networks that carry information, the elites and governments that try to shape it and that read it, the pollsters who measure it, and the elections and policies it feeds. |
| nested level | Built from individual attitudes held by people (human L6: thoughts, feelings, emotions) and expressed by them (speech, votes, posts); runs on people, on the mass media's public picture (mass-media L7) and on social networks; one of the inputs a government reads (government L2) and one of the forces of a democracy's public sphere. |
| scale | social |
Measures (22) published science
| London coffeehouses, 1739 | 551 coffeehouses |
| share of British men able to vote, 1900 | about 40 % |
| Literary Digest 1936 poll size | 2.3 million voters polled (postcards) |
| poll margin of error | 3 % at 95% confidence |
| sample for a 1% margin | about 10,000 people |
| poll response rate | about 10 % (recent years) |
| Chapel Hill study respondents | 100 residents |
| agenda-setting studies since 1972 | more than 400 studies |
| national discussion added by news articles | ~62.7 % of a day's volume per policy area |
| King et al. treatment period | 1 week |
| content analysed for echo chambers | >100 million pieces of content |
| Bail et al. exposure | 24 messages a day for 1 month retweets |
| Bail et al. pretreatment respondents | 1,652 (901 Democrats, 751 Republicans) |
| polls in the cross-national error study | >30,000 national polls (351 elections, 45 countries, 1942-2017) |
| GLES rolling cross-section window | 55 days before the 2021 German federal election |
| Global Climate Scheme surveys | 40,680 respondents in 20 countries |
| support for the Global Climate Scheme | three quarters of Europeans; half of Americans share |
| attitude transmission survey | 1,905 students; 4,000 relatives and friends people |
| Norwegian twin sample | 1,987 twins |
| posts analysed when the Kitagawa silence broke | 14.5 million social media posts over six months |
| reverse-backlash panel | 24 countries, 1989-2017 country-years |
| bounded-confidence consensus threshold (Deffuant) | 1/2 tolerance above which all agents reach one opinion |
Open contentions (8) open contention
Where sources or readings disagree. Both stay on the map.
- C-2964 When a minority holds an opinion, does it fall silent or win?
booked: The map keeps the competing readings; none is ruled.
the other reading: the spiral of silence: the minority silences itself; a committed minority at critical mass overturns the convention; majority rule: the initial majority always wins without bias - C-2975 Is public opinion stable and coherent, or volatile?
booked: The map keeps the competing readings; none is ruled.
the other reading: Almond-Lippmann: unstructured, incoherent, highly volatile (foreign policy) - C-2976 Does public opinion drive policy, or does policy drive opinion?
booked: The map keeps the competing readings; none is ruled.
the other reading: opinion drives policy (signals, electoral accountability); policy drives opinion (expectations from social policy; positive returns); the thermostat: opinion moves against policy (reverse backlash) versus backlash (radical-right success worsens attitudes); elites manipulate opinion (Habermas); presidents act autonomously on non-salient issues (Druckman and Jacobs) - C-2978 Where do people's opinions come from?
booked: The map keeps the competing readings; none is ruled.
the other reading: mass media as the main force (communication studies); vertical transmission from parents (but outside influences predominate for left-wing views); substantial heritability of ideology (twin studies); two-step flow through influentials versus lateral spread among ordinary people (Watts and Dodds) - C-2979 Do published polls create a bandwagon?
booked: The map keeps the competing readings; none is ruled.
the other reading: bandwagon effects exist (the older view, and the 1884 term); underdog effects (less evidence); polls as a device for influencing opinion (Hitchens) - C-2980 Are opinion polls getting less accurate?
booked: The map keeps the competing readings; none is ruled.
the other reading: conventional wisdom: polls are getting worse; falling response rates (about 10%) as a growing source of error - C-2981 When did public opinion first become a force?
booked: The map keeps the competing readings; none is ruled.
the other reading: medieval fama publica, with legal weight from the 12th-13th centuries; the term itself, Montaigne 1588; Habermas: the 18th-century bourgeois public sphere; a central force in British politics only by the 1820s - C-2982 What is public opinion?
booked: The map keeps the competing readings; none is ruled.
the other reading: Blumer: a public is those discussing an issue at one time; many publics; polling cannot measure the public; Habermas: formed in rational public debate, now highly susceptible to elite manipulation; Tönnies: a social function in societies equivalent to religion's in communities
Cross-scale patterns seen here (19) our proposal
Shapes that recur when the same kind of problem is solved at different scales, and how each shows in Public opinion. Each links to every other system that shows it.
- runaway from an unstable balance (structural)
Majority rule (Galam): the 50/50 point is unstable and the others stable, so from any start off balance all agents converge to the initial majority; the spiral of silence is the same runaway in expression (the minority's silence makes the majority louder, which makes the minority quieter). - same role, different builds (structural)
One job, reading what a population thinks, done by unrelated builds across time: medieval fama publica as legal proof, coffeehouse talk and papers, newspaper straw tallies (1824), mass postcard polls (Literary Digest), scientific sampling (Gallup, Crossley, Roper), and sentiment analysis of online p - templating (structural)
The scientific poll travelled as a module: Jean Stoetzel founded IFOP in Paris in 1938 after meeting Gallup, the first European survey institute, and Gallup opened a UK subsidiary that called the 1945 election; the same sample-and-weight method is used by private firms and governments around the wor - the whole sets the local reference (structural)
The climate of opinion (the whole) is the reference each person measures their own view against: willingness to speak depends on the perceived distribution, and people form opinions from what they believe their group's prevalent opinion is. Each reads it locally (through their own ties, feeds and me - Umwelt: each system carves its own slice of one shared medium (structural)
Each channel carves its own slice of one shared issue space: platforms differ (news consumption more segregated on Facebook than Reddit), echo chambers each hold their own picture, and Blumer's publics form around one issue each. The set of nine channels (news, platforms, talk, family, elites, polic - causal ceiling (partial)
No member can see the whole distribution: each reads the climate only through a quasi-statistical sense fed by their own ties and media, so perceived and actual opinion can be far apart (pluralistic ignorance; a false vision). Blumer adds that polling cannot measure the public as he defines it. Part - coordination-before-mapping (partial)
In the opinion-dynamics models (voter, Ising, majority rule) agents reach consensus through local imitation alone, with no map of the whole. Partial: real people also hold a perceived climate (a rough map), so the models show the map-free floor, not the whole system. - lightning (partial)
A stored difference released: decades of media silence on the Kitagawa abuse broke in 2023 once foreign coverage spread, and 14.5 million posts followed in six months; in experiments a committed minority at critical mass consistently overturned an established convention. Partial: the release ran ove - not-diluting wins (partial)
A component that never changes while the rest mix: one zealot in low dimension pulls everyone to its opinion; inflexible agents and a committed minority at critical mass overturn established behaviour. Partial: the fixed share is held by commitment, not by a density law. - same shape, opposite sign (partial)
One process, a published poll standing shifting vote choice, run in both directions: the bandwagon (back the candidate shown winning) and the underdog (sympathy votes for the party shown losing), with less evidence for the second. - same toolkit, independent builds (partial)
In 1936 Gallup, Crossley and Roper each built far smaller, scientifically based samples and all three called the result the 2.3-million-postcard Literary Digest got wrong. Partial: shared statistical toolkit, contemporaries; independence of the builds not documented in the source read. - smooth to sharp in finite time (partial)
In bounded-confidence models the time to reach the stationary state diverges at the thresholds where opinion clusters split: a smooth dependence that becomes singular at a finite parameter value. Partial: a model property, not yet seen in measured opinion. - symmetry breaking (name the cut) (partial)
Opinion dynamics as order out of disorder: from a disordered start with no preferred opinion, voter and Ising-type dynamics end in consensus on one side, the side chosen by fluctuation or initial majority. Partial: the models name the cut; measured cases not read. - The Payment to Entropy (partial)
Every hand-off from population to published number pays a toll: response rates are down to about 10% (a yield of about 0.1 from contact to answer), the rest is replaced by weighting with varying success, and a biased return (Landon voters keener to mail postcards) cost the Literary Digest its 1936 c - trade one property for another (partial)
Poll precision is bought with sample size: a 3% margin at about a thousand, 1% needs about 10,000; and a 2012 study found a trade-off between democratic and administrative decision-making. Partial: two properties and a dial, but the kinds spread along it are not described. - who opens the door (partial)
At the crossing from private view to public number the receiver builds the door: the pollster's question opens it ahead of time, and the door shapes what comes through (people asked about things they know nothing of give pseudo-opinions to please the questioner; Blumer: polling cannot measure the pu - namespace completeness (the map predicts its gaps) (surface)
Galam's general sequential probabilistic model contains the binary opinion models (voter, majority rule, Sznajd) as special cases: a general map with the known models as points in it. Surface: no empty slot was predicted and later filled. - nested ladder (surface)
Hurwitz and Peffley's hierarchy: core values ground postures, which shape issue positions. Surface: an ordering of bases, not a ladder where each rung implies the ones below. - outside-to-inside (surface)
Group opinion outside the person becomes an inner learning signal: conflict with the group triggers a prediction-error response in the rostral cingulate zone and ventral striatum that predicts conforming; parents' attitudes come first and pass to children. Surface: shown at the level of the person,
Valence: what it trades, and with whom (22) published science
Each exchange with a neighbour, read from both sides: who gains, who pays, and what carries it.
- Radio via sys:human-society · Public opinion mixed
- Social norm via sys:human-society · Public opinion mixed
- President of the United States via sys:human-society · Public opinion mixed
- Supreme Court of the United States via sys:human-society · Public opinion mixed
- Human (Homo sapiens) via sys:human-society · Public opinion gain
- Government via sys:human-society · Public opinion mixed
- Internet via sys:human-society · Public opinion mixed
- Human society (from local community up to societies) via sys:human-society · Public opinion gain
- Social network via sys:human-society · Public opinion mixed
- Mass media via sys:human-society · Public opinion mixed
- Political party via sys:human-society · Public opinion mixed
- Democracy via sys:human-society · Public opinion gain
- Election via sys:human-society · Public opinion gain
- Journalism via sys:human-society · Public opinion gain
- Newspaper via sys:human-society · Public opinion gain
- Information warfare via sys:human-society · Public opinion loss
- Misinformation via sys:human-society · Public opinion loss
- Opinion poll via sys:human-society · Public opinion mixed
- Advocacy group (interest group) via sys:human-society · Public opinion mixed
- Voting via sys:human-society · Public opinion gain
- Renewable energy via sys:public-opinion · Public opinion even
- Truth via sys:human-society · Public opinion mixed
Partners and parts published science
Built from / builds (10)
- Opinion leader
- Individual attitudes and opinions
- Publics (one per issue)
- The perceived climate of opinion
- IRA influence campaign in the 2016 US election
- The press and the state over what may be printed (licensing, censorship, suppression)
- Parties, governments, media and interest groups competing to frame issues for the public
- Misinformation against correction for what people believe
- Abolitionist narratives and novels against anti-Tom novels (1850s)
- Interest groups and public opinion competing for politicians' votes
Sits inside / holds (3)
expressed by (5)
- Internal Coordination/Regulation
- Sensing/Scanning
- Uncertainty Reduction/Learning
- Social Learning / Public Information Use
- Self-Monitoring
floor filled by (18)
- Human (Homo sapiens)
- Government
- Internet
- Human brain
- Human society (from local community up to societies)
- Social network
- Mass media
- Democracy
- Election
- Journalism
- Newspaper
- Opinion leader
- Opinion poll
- Individual attitudes and opinions
- Publics (one per issue)
- The perceived climate of opinion
- Advocacy group (interest group)
- Renewable energy
Kinds (8)
- Opinion poll
- Publics (one per issue)
- The mass (Blumer)
- Fama publica (medieval public repute)
- Public opinion on foreign policy
- Online public opinion
- International (world) public opinion
- Pseudo-opinion
Notes published science
- News really does make people talk: political scientists persuaded 48 small news outlets to publish on topics they approved, on dates picked at random, and each burst of articles raised national Twitter discussion of that policy area by about 62.7% of a normal day's volume, some 13,166 extra posts in the week (src:king-2017).
- Hearing the other side can push you further away: Republicans paid $11 to follow a bot that retweeted 24 liberal messages a day for a month came out substantially more conservative than before (src:bail-2018).
- The physics of a crowd making up its mind: in Galam's majority-rule model a 50/50 split is an unstable balance, so whichever side starts even slightly ahead ends up taking everyone; and the 'voter model' that physicists use for opinions began life in 1973 as a model of competing species (src:castellano-2009).
Layers
The ways it takes in its world, where they meet, and how it acts back, each from L1 (where it meets the world) up.
public opinion: from the issues of the day to policy and power (reader's view)
- L1: The issues of the day and the population that holds views on them: events, policies and conditions that affect people
- L2: Predispositions: core values, partisanship and attitudes, learned from parents, grandparents and peers and partly heritable
- L3: Exposure and transmission: news agendas and frames, influentials and personal networks, platforms and their feeds, elite rhetoric and misinformation
- L4: The perceived climate of opinion: the quasi-statistical sense of majority and minority, social desirability, conformity to the group, heuristics
- L5: Expression and measurement: speaking out or falling silent, public talk and posts, polls and votes
- L6: The aggregate opinion: the distribution of views on an issue, its salience and mood, consensus or fragmentation, stability or volatility
- L7: Opinion acting on power: signals to policymakers, electoral accountability and legitimacy; policy acting back on opinion
the boardroom (shared floors)
- L4: The perceived climate of opinion: each person's quasi-statistical sense of majority and minority, social desirability, conformity to the group (a prediction-error signal in the rostral cingulate zone and ventral striatum), heuristics
- L5: Expression and measurement: speak out or fall silent, talk and post, answer polls, vote
- L6: The aggregate opinion: the distribution on an issue, salience, mood; consensus or fragmentation; publics forming and dissolving per issue
- L7: Opinion on power: signals to policymakers, electoral accountability, legitimacy
way in: Family, peers and inheritance
- L1: Parents, grandparents and friends (1,905 students and 4,000 of their relatives and friends surveyed)
- L2: Vertical transmission from parents, indirect from grandparents; peer similarity; for left-wing views, outside influences predominate
- L3: Predispositions partly heritable: twin studies find substantial heritability of ideology
way in: Foreign media and international bodies
- L1: International coverage (BBC, CNN) of an issue domestic media had been silent on
- L2: International bodies: a UN working group's press conference
- L3: Domestic uptake filtered by loyalty: the loyal were slower to mention the issue
way in: Misinformation
- L1: Rumours, fiction, governments and politicians, vested interests
- L2: A changed media landscape, the Internet included, spreads it
- L3: Continued influence in memory: retractions often fail and can backfire
way in: News media: agenda and frames
- L1: Events and issues the media identify and publicise
- L2: The media agenda: what is newsworthy, how and when it is reported; frames chosen
- L3: Salience and frames passed to audiences: more coverage, more importance; emphasis frames select aspects
way in: Platforms and feeds
- L1: Posts and shared content on platforms (more than 100 million pieces on controversial topics across Gab, Facebook, Reddit, Twitter)
- L2: Feed algorithms: suggestions like what users already saw; inflammatory content pushed for engagement
- L3: Homophilic clusters: like-minded peers reinforce a shared narrative (echo chambers)
way in: Policy as lived
- L1: Policies, above all social policy, and the goods and services people get from the state
- L2: Normative expectations built by what policy delivers
- L3: The thermostatic reaction: part of the electorate moves against the direction of political change
way in: Politicians, governments and leaders
- L1: Advertising and rhetoric from politicians and others concerned with opinion
- L2: Leaders setting their own agenda and the standards they are judged by (the presidency)
- L3: Delegative heuristics: cues from authorities (media, president) taken in place of full information
way in: Published polls
- L1: Published poll findings in election coverage
- L2: Perceived popularity of parties and candidates
- L3: Bandwagon (follow the winner) or underdog (sympathy for the loser): contested
way in: Talk and influentials
- L1: Conversation in shared places: coffeehouses (551 in London by 1739), clubs, public meetings, now personal networks
- L2: Influentials: media reach opinion leaders, who reach the public (two-step flow)
- L3: Lateral spread among ordinary people, as likely as influentials to move opinion when people are easily influenced
way out: Pressure to conform
- L1: Conduct conforms to social norms
- L4: Fear of isolation for the perceived minority
- L6: The prevalent opinion of the group
way out: Public talk and posts
- L1: Posts on platforms, read by media and by others
- L5: Expression activated by coverage: national discussion of a policy area rises
way out: Signals to policymakers
- L1: Government policy and party strategy moved
- L5: Read through polls on salient matters (crime, the economy)
- L7: Preferences and likely votes signalled
way out: Votes
- L1: Votes cast by a widening electorate; parties compete for it
- L5: Intention measured before the vote
- L7: Voting intention
Sources (25) published science
Spiral of silence (Wikipedia, via Wikimedia Enterprise snapshot 2026-11-01, version 1382937075) (behind its links)
Wikipedia: Opinion poll (behind its links)
Castellano, Fortunato, Loreto 2009, Statistical physics of social dynamics, arXiv:0710.3256v2 (2009) (behind its links)
Agenda-setting theory (Wikipedia, via Wikimedia Enterprise snapshot 2026-11-01, version 1379562573) (behind its links)
Bail CA et al. (2018) Exposure to opposing views on social media can increase political polarization. PNAS 115:9216-9221, doi:10.1073/pnas.1804840115 (behind its links)
Jennings, Wlezien 2018, Election polling errors across time and space, Nature Human Behaviour 2:276-283, doi:10.1038/s41562-018-0315-6 (PMID 30936537) (behind its links)
Social norm (encyclopedia) (behind its links)
Jessee, S., Malhotra, N. & Sen, M. Is the Supreme Court veering rightward? The ebb and flow of representation. PNAS Nexus (2026). doi:10.1093/pnasnexus/pgag060 (behind its links)
Voting (Wikipedia; Wikimedia Enterprise snapshot 2026-11-01) (behind its links)
Wikipedia: Renewable energy (Wikimedia Enterprise snapshot 2026-11-01) (behind its links)