Why Philip Ernest Converse Is Famous for Discovering Ignorance

Philip Converse’s 1964 essay, “The Nature of Belief Systems in Mass Publics,” is the reason many political scientists hesitate to call mass political views ideologies. His essay relied heavily on the 1956–1960 American National Election Studies data to argue that the vast majority of mainstream voters do not hold stable, cohesive ideological belief systems (such as strict liberalism or conservatism) and to demonstrate that individual opinions often fluctuate widely from one wave of a survey to the next. Its claim was not that ordinary people are stupid. Ignorance means lacking knowledge about a specific topic, while stupidity means lacking basic intelligence or refusing to apply facts one already knows. Clearly, both are evident in American politics, but benign ignorance dominates. Converse explains why.

What Converse meant by a belief system

Converse refused the muddy word “ideology” as his main term. He defined a belief system as “a configuration of ideas and attitudes in which the elements are bound together by some form of constraint or functional interdependence.”

Constraint is the operational heart of his argument. In the static case, it is predictive: if you know one of a person’s positions, how well can you guess the others? American politics is highly polarized. The two major parties, Democrats and Republicans, hold opposing bundles of beliefs. Once these beliefs were well-defined; now they appear more scattered. Nevertheless, voters continue to adopt these entire packages instead of mixing random ideas.

Imagine a person who supports strict federal laws banning assault rifles. With that single piece of information, you can guess with high confidence that this person also supports abortion access and reproductive rights, favors government regulations on carbon emissions to fight climate change, and supports higher taxes on corporations and wealthy individuals.

Converse argued that high static constraint is rare. It is usually found only among “elites” (politicians, activists, and highly educated voters) who think ideologically.

In the dynamic case, constraint is revision: if one element changes, do the others have to move? If a political party or elite movement changes its stance on an issue, a person with high dynamic constraint will update their beliefs to match the new package. The change in one opinion “constrains” or dictates changes in others to maintain psychological consistency.

Converse concluded that for most ordinary citizens, dynamic constraint is very weak. When surveyed over several years, he discovered that the average person’s opinions flipped back and forth randomly. He concluded that many citizens do not hold stable, deeply rooted political beliefs but instead express what he called “nonattitudes.”

Constraint, he said, has three sources:

• Logical — formal implication (rare in real politics).

• Psycho-logical — the learned feeling that “this has to go with that.”

• Social — packages assembled by parties, churches, unions, and commentators, then taken as a lump.

He was explicit that wide-ranging ideologies almost never rest on formal logic. Elites experience the package as logical because they live inside the argument. Mass publics mostly receive the social package, or do not receive it at all.

The primary thesis is that, as you move down the information scale from elites toward the inattentive public, range shrinks, constraint decays, and stability collapses. He put the fade-out of genuine ideological constraint somewhere near the 90th percentile of political information in the 1950s sample. Above that line, liberalism and conservatism behave like systems. Below it, they mostly do not.

Five levels of conceptualization

The most-cited table in Converse’s 1964 essay is not about issue correlations. It is about how people talked when asked, in open-ended interviews, why they liked or disliked parties and candidates.

Ideologues plus near-ideologues were about one-eighth of adults and a slightly higher share of voters. The modal citizen was a group-interest reasoner: Democrats are for workers and the poor; Republicans are for business. That was the 1950s operating calculus.

This is why later scholarship, especially Donald Kinder’s 2006 retrospective, treats group-centrism as Converse’s constructive finding, not only his skeptical one. If abstract ideology is scarce, people still have a usable heuristic: whose side is this? Race, religion, class, nation, sex, and party identity do that work. Kinder’s 2006 retrospective called group-centrism the overlooked affirmative half of the 1964 essay.

Nonattitudes

The other explosive result was temporal instability. On many policy items, the same people, asked the same question two years apart, looked almost as if they were drawing answers from a coin flip. Converse’s “black-and-white” model said the public is a mixture of two types:

• a small set with real, stable attitudes;

• a large set with nonattitudes — no stored opinion, so the way a survey is framed creates an answer on the spot.

Party identification was relatively stable. Abstract ideology and many specific issues were not. That is the origin of the later joke that the average voter has “opinions” the way a weather vane has a destination.

What Converse’s essay was not saying

Three misreadings are common.

It did not say elites are philosopher-kings. Elite constraint can be social and dogmatic. Converse noted that the same binding that makes an ideology predictable also makes it rigid.

It did not say the public is random on everything. Group loyalties, some moral issues, and party identity can be quite stable. The claim is about wide-ranging, abstract, cross-issue systems.

It did not say democracy is impossible. It said the textbook picture of citizens who deduce policy from Locke, Marx, or Burke is an illusion. Recently, young people have been interviewed on the street and asked, “Why do you prefer socialism?” A common response is, “I believe everybody should all just get along.”

What later research kept, and what it cut

Kept. The elite–mass gap in constraint is one of the most replicated findings in public-opinion research. People who follow politics, know facts, and care about outcomes generally have more cohesive, well-defined, and internally aligned sets of policy positions. Those who do not follow politics closely generally do not. Kinder and Kalmoe’s Neither Liberal nor Conservative (2017) is the major modern restatement: after decades of polarization among activists and officeholders, most Americans are still not liberals or conservatives in Converse’s sense. They will pick a label. The label often tracks party and group more than a doctrine. Real ideologues remain concentrated among the politically engaged.

John Zaller’s The Nature and Origins of Mass Opinion (1992) is the main micro-theory built on Converse. People sample from the considerations they have recently received. If they have few considerations, answers wobble. If they are partisan and attentive, they accept messages from their side and reject the other, which looks like constraint but is often partisan reception rather than a privately deduced system.

Cut or qualified. Christopher Achen (1975) demonstrated that the apparent instability of public opinion in single survey items stems from question vagueness and measurement error rather than fickle voters. Building on this, Stephen Ansolabehere, Jonathan Rodden, and James Snyder (2008) showed that aggregating multiple related survey items into indices allows random errors to cancel out, revealing stable and ideologically constrained public preferences. Nie, Verba, and Petrocik once argued that constraint rose after the 1950s as politics polarized. Later work found that a good deal of the apparent rise was sorting: liberals became Democrats and conservatives became Republicans, so party and ideology lined up more tightly, even if many individuals still could not explain the package.

Affective polarization since the 1990s—dislike of the other party—may superficially look like ideological constraint. Hating “the left” or “MAGA” will make a person’s answers line up without requiring the person to be able to teach a seminar on liberalism or conservatism.

In Andrew Heywood’s political-theory framework, affective polarization is distinct from ideological constraint, meaning that voters can feel intense dislike for opposing parties without actually aligning on specific policy issues. While ideological constraint refers to how closely a person’s political attitudes conform to a coherent, structured ideology (such as traditional liberalism or conservatism), affective polarization is driven by social identity, emotional hostility, and an “us versus them” mentality.

How this maps onto Heywood’s boxes

Converse is the reason our Heywood-by-the-Numbers exercise had to use proxies.

Heywood’s eleven types are elite-level belief systems: they have range, names, lineages, and ways of managing internal tensions. Converse’s mass public mostly does not live there.

That is why:

• 25% “liberal” and 37% “conservative” are not 25% and 37% ideologues. Many of those labels are group badges, mood words, or party synonyms.

• Pew’s four “anchor” typology groups—38% of adults—are the closest published stand-in for people whose answers hang together. Even they are not all Converse ideologues.

• The 7% Leftward Progressives and 9% No Apologies Right are where you should expect something like a belief system: abstraction plus constraint plus engagement.

• Faith First Conservatives (12%) often have a religious system that then spills into politics. That can be high constraint without high political information. Converse allowed social sources of constraint; a church can bind issues the way a party can.

• Socialism at 39% favorable is a word evaluation, not a socialist belief system. Converse would treat that number as almost worthless until you knew whether the same people wanted public ownership, how they ranked it against growth, and whether the answer survived a second interview.

• Fascism as a mass U.S. ideology fails Converse’s test twice: tiny numbers and no evidence of a constrained doctrine in the public.

How many religious fundamentalists align with conservatism? This is a Converse-style question that needs nuance. Alignment can be social constraint (evangelical churches and the GOP bundled the issues) rather than psycho-logical constraint (the person derived school prayer, tax cuts, and Israel policy from one axiom). Both count as constraint. Only the second is what people usually mean by “articulating what they believe.”

Young adults, in Converse’s terms

The under-30 findings sit inside this framework rather than beside it.

Converse already knew that conceptualization tracks information, education, and involvement, not age as such. Age matters because those inputs are lower early in the life cycle. Pew’s knowledge gaps by age, the 19% who follow news closely, and the 19% who use none of the major sources are measures of the information scale on which constraint decays.

The correct Converse reading of young adults is not “they have no beliefs.” It is:

• They often have intense group and moral attitudes (gender, climate, Palestine, borders).

• Those attitudes may be stable inside a domain.

• They often lack horizontal constraint across domains and vertical constraint from a named ideology down to trade-offs.

• When two cherished pieces collide—post-growth versus energy abundance, women’s categories versus gender self-ID, free speech versus harm—the system has no ranking rule. That collision is the test Converse cared about.

A 22-year-old progressive in 2026 can be a near-ideologue who uses a newly redefined vocabulary—systemic, intersectional, and decolonize—and whose place on the ideological team is secure. That represents considerably more ideological abstraction than 1956’s “nature of the times,” the classic concept describing voters who evaluate political parties and candidates based purely on whether current conditions, such as the economy or war, are going well or poorly.

The same test applies to a 22-year-old nationalist in 2026 who uses terms such as sovereignty, reclamation, or re-migration.

The State of the Public Opinion Debate

After sixty years, public opinion research has converged on several core premises. Scholars generally agree that political elites exhibit greater ideological constraint than the mass public, that single-item survey measures overstate cognitive instability, and that partisan and group identities command more loyalty than abstract doctrine. Furthermore, partisan sorting has made political parties appear far more ideological than the median voter.

The remaining debate centers on how much substantive structure underlies this statistical noise. The Conversean orthodox view maintains that most survey responses reflect non-attitudes mediated by group affect. Conversely, revisionists argue that citizens possess two or three durable ideological dimensions, such as economic or cultural orientations, that standard surveys simply fail to measure accurately. Both camps concede that highly integrated ideological frameworks are held only by a small minority.

Consequently, quantifying adherence to strict ideologies like liberalism, socialism, or fundamentalism yields vastly lower estimates than self-identification metrics, word-favorability thermometers, or latent class models like Pew typologies. For the vast majority of the public, the valid unit of analysis is not a coherent ideology but rather a composite of group loyalties, domain-specific leanings, and partisan identification.