A Gallup poll conducted for the Lloyd’s Register Foundation has captured what a decade of intensifying political violence and institutional alienation has produced: one in ten American adults now names politics as the single greatest threat to their personal safety. The finding is unprecedented in scale and cross-party uniformity, placing the United States alone among 140 countries in ranking political conflict above personal health as a daily danger. The country where the institutions meant to manage conflict are themselves the source of insecurity is now measuring that fact directly.

What the numbers show

The 10% figure represents a steady climb from 6% in both the 2021 and 2023 survey waves. The U.S. effectively tied with Hungary for the top spot — Hungary recorded 9% in polling conducted months before its 2026 national election under sustained democratic pressure — followed by Serbia, Latvia, and Lithuania at 6%, and Slovakia, Russia, Myanmar, Spain, and Israel each at 5%. The global average across all 140 countries surveyed was 1%. No other country came close.

Benedict Vigers, a senior global news writer at Gallup, described the U.S. result as “unprecedented.” “No other country ranked higher. It’s 10 times the global average,” he said. “It speaks to how charged the political environment is and has been for many years in the United States that people are now saying that the political situation is not just an issue nationally but is a seeable daily risk to their safety as they go about their normal lives.”

The survey found no statistically significant differences by age, gender, education, or political affiliation — Republicans, Democrats, and independents all cited politics as their top concern at similar rates. The absence of partisan divergence means the fear is diffuse and ambient, not something any single political group claims uniquely. In 2025, a third of U.S. adults told a separate Gallup survey that politics and government was the country’s most important problem. Fifty-eight percent said the government “does not care about them at all” — an 11-percentage-point jump from 2021 that placed the U.S. among the most alienated populations in the OECD. In 2024, 5% of Americans named the United States itself as its own greatest enemy, the highest proportion on record for that question since Gallup began asking it in 2001.

How the perception forms: from events to data

A process map of how this perception forms and is measured reconstructs a chain with five main links. Political-violence events occur — the three assassination attempts against Donald Trump, the fatal shooting of conservative activist Charlie Kirk at Utah Valley University during a Turning Point USA event on September 10, 2025, the January 6, 2021, attack on the U.S. Capitol, and a documented surge in threats against election administrators and local officials. Media organizations report on these events, selecting and framing them for national audiences. Individuals consume these reports or experience political violence directly, and the signals accumulate. That accumulation converges in a perception-formation step, where an individual’s sense of personal safety is synthesized from multiple streams. The respondent then volunteers an answer when Gallup asks an open-ended question, and the response is coded into category labels. “Politics” emerged as the top-of-mind category for 10% of U.S. respondents in 2025.

The process map identifies three converging streams at the perception-formation node. Stream A is the actual trajectory of political-violence events — what the survey’s authors called “a decade-long trajectory.” Stream B is the institutional-trust deficit: when 58% of U.S. adults report that the government “does not care about them at all,” the institution responsible for managing political conflict is perceived as indifferent to it, and the political environment therefore feels ungoverned. The 5% who named the United States itself as its own greatest enemy in 2024 are not expressing distrust — they are expressing that the state has become an external threat. Stream C is media amplification: repeated dramatic coverage makes political-violence examples highly salient, crowding out lower-drama personal risks like health conditions or accidents. Cable-news business models structurally reward partisan outrage over de-escalation; the mid ring of institutions that could mediate is instead structurally aligned with escalation.

The 6% baseline that held stable in both 2021 and 2023 — despite the Jan. 6, 2021, Capitol attack falling between those measurement points — reveals a critical temporal pattern: individual high-profile events do not produce immediate measurement shifts. The violence accumulates slowly in public consciousness; the perception it feeds is a stock, not a flow. The cumulative weight, not any single incident, constitutes the ambient signal that eventually registers. The jump from 6% to 10% between 2023 and 2025 suggests accumulation reached a threshold where the perception shifted from a minority concern to a measurable one-in-ten phenomenon.

What “politics” actually means in these answers

The three respondent quotes published with the poll show how these streams converge differently for different individuals. A 54-year-old Republican man named “the state of polarisation of the American people.” A 42-year-old independent man said he saw the biggest risk in “people that are fanatical about their position and that they could be violent.” A 69-year-old Democratic woman said the political situation “has a lot to do with the fact whether if I am safe or not. I am concerned where we’re going or might go.”

The three quotes span party lines and age cohorts but share a grammar of threat: each respondent describes other Americans — neighbors, fellow citizens, people encountered in daily life — as potential sources of danger. The specific fears differ; the underlying orientation does not. All three treat the political environment as a space in which personal safety is no longer guaranteed.

Vigers said “almost all respondents who named politics as their top risk referenced polarization, either by blaming the opposing party or by citing a heightened sense of polarisation.” None of these respondents describes a political risk the way a policy analyst would use the term. They describe ambient social conflict — a generalized perception that the people around them are becoming dangerous because of what they believe. “Politics” is the label the survey assigns; “the state of polarisation of the American people” is what respondents are actually naming. The taxonomy treats politics as a discrete category; the responses suggest it is a container for a broader perception of social fracture that does not map cleanly onto existing risk categories. The open-ended format, while revealing, also compresses a multidimensional perception into a single label — meaning the 10% figure may understate the scope of political-conflict anxiety, because some respondents who feel this way may have categorized their concern differently when faced with the prompt.

Bottlenecks in the chain

At the perception-formation step, the media environment overrepresents dramatic political-violence events, creating a highly available but skewed risk signal. Official, accessible data about actual personal risk from political violence is largely absent from the daily information diet. Respondents construct their risk assessment from the dominant salience source without a countervailing calibration signal.

At the survey-administration step, Gallup’s capacity constraint is structural: the survey must achieve a nationally representative sample within a fixed budget spread across 140 countries. The U.S. sample size is not disclosed in the source material, and the reported finding of “no statistically significant differences” across age, gender, education, and party could reflect genuine uniformity or insufficient power to detect variations — the source does not allow a reader to distinguish.

Why bipartisan fear has no built-in brake

The data describes a condition in which all sides perceive themselves as threatened by the same dynamic. The 42-year-old independent does not fear Democrats or Republicans specifically; he fears “people that are fanatical about their position and that they could be violent.” The threat is generalized.

In a bilateral conflict, the presence of fear on both sides can create conditions for negotiation. But the World Risk Poll finding describes something different. A Republican who softens partisan rhetoric does not reduce a Democratic respondent’s sense that fanatical people on the other side could be violent, and vice versa. The feedback loop has no self-correcting mechanism built into it. When all sides perceive threat, no party’s de-escalation relieves the others’ sense of danger. There is no “other side” to negotiate with, because the perceived threat is the dynamic itself.

The 6% to 10% increase across four years, on an unprompted open-ended item, suggests the dynamic is accelerating rather than stabilizing. The U.S. effectively tying with Hungary — a country under sustained democratic pressure — places the American result in a peer group of nations whose democratic institutions are under measurable strain. The 10x global average multiple is not a curiosity; it is a signal that the U.S. has entered a category of countries where political conflict is experienced as a personal safety concern at population scale.

The institutions that would need to fix this are inside the problem

The most consequential finding from a stakeholder assessment is not about any party’s advantage but about something missing: the healer role. The decade of political violence cited in the survey has produced injury that settlement cannot close. The 54-year-old Republican’s “state of polarisation” framing names a wound, not a negotiable position. No national truth-and-reconciliation process or restorative-justice initiative addresses these injuries at scale. Without the healer role, each cycle of political violence leaves accumulated grievance that feeds the next.

That missing role exists within a surrounding community that is structurally unable to supply it. The mid ring — political institutions, media, election systems, religious congregations, employers — is itself a party to the conflict. Any institution perceived as aligned with one party forfeits credibility with the other. Election administrators and local officials are under active threat, meaning the referees themselves are under attack. The outer ring — federal institutions, democratic norms — faces a different obstacle: deep American distrust of centralized authority, a cultural norm that systematically reduces the credibility of courts, Congress, and executive agencies as neutral referees.

Courts, which nominally serve as the arbiter of political disputes, have become a partisan flashpoint. Federal judicial confirmations are now party-line votes. The Supreme Court’s legitimacy ratings have fallen below 50% in major national polls. The partisan capture of judicial appointments has converted the arbiter institution into a partisan prize, hollowing out its credibility as a neutral third side. An arbiter institution that is itself a conflict participant cannot credibly mediate the dynamic it is part of.

Election administrators — the local officials who run the mechanics of democratic contestation — face escalating personal threats. A referee under personal threat cannot reliably enforce rules. The decentralized U.S. election-administration system, with more than 10,000 independent jurisdictions, fragments whatever protective capacity exists into manageable pieces too small to address the systemic problem.

Civic bridge-building organizations exist — Braver Angels, Living Room Conversations, the National Governors Association’s bipartisan forums — but have not demonstrated impact at a scale that shifts ambient threat perception. The media environment that drives the salience of political conflict operates through algorithmic amplification of outrage on private platforms, a scale that voluntary civic programming cannot match. First Amendment constraints mean the government cannot regulate the polarizing content that contributes to the dynamic; bridge-building must be voluntary and non-coercive, which limits its reach. The question is not whether these organizations have a viable model — they do — but whether they can reach a scale that changes threat perception at the population level. Current evidence suggests they have not.

Other third-side roles are active but strained

The witness role is active — the Gallup-Lloyd’s Register Foundation poll itself documents the human cost across three waves, and media reporting on political violence performs the witness function. But the 6% to 10% climb across four years suggests witnessing alone is not constraining escalation. Visibility creates accountability costs only if someone acts on the visibility. The referee role is active but strained: the decentralized U.S. election-administration system of over 10,000 independent jurisdictions fragments capacity, and election officials face escalating personal threats that make impartial enforcement dangerous. The peacekeeper role is partially active but reactive: Secret Service, Capitol Police, and local law enforcement attempt to interpose, but three assassination attempts and a fatal shooting indicate containment is failing. Several other roles — provider (meeting security and recognition needs), teacher (building conflict-handling skills), mediator, equalizer — are needed but unfilled, and the institutions that would need to fill them are themselves entangled in the partisan dynamic.

Escalation signals to watch

The trend line itself is the primary escalation marker: 6%–6%–10% across four years signals acceleration, not stabilization. The conflict is shifting from prevention territory into containment territory. The bipartisan distribution of fear with no self-correcting mechanism means the feedback loop has no internal de-escalation path. The U.S. sharing the top ranking with Hungary places the American result in a peer group of nations whose democratic institutions are under strain, a peer-comparison marker suggesting the path from this state to institutional erosion is empirically documented.

Hardening rhetoric compounds the pattern. The 42-year-old independent’s description of opponents as “people that are fanatical about their position and that they could be violent” frames the other side as potential attackers, normalizing preemptive threat-orientation. The 58% government-alienation figure — an 11-percentage-point jump from 2021 — signals accelerating withdrawal of legitimacy from the arbiter role, removing the primary non-violent channel for grievance resolution. Cumulative political violence — each assassination attempt, each killing, each threat — becomes a new grievance in the absence of a healer role. The three assassination attempts against Trump and the killing of Charlie Kirk are not discrete events but linked inputs into a compounding system. Without structured acknowledgment, the cycle continues.

Four interventions and their limits

The most promising intervention targets the missing healer role directly: post-violence community processing after political violence, with local truth-and-reconciliation commissions piloted at municipal and county levels before any national scaling. The Carter Center or Kettering Foundation, operating alongside interfaith coalitions and restorative-justice practitioners, could design a process that breaks the cycle where each political violence event leaves accumulated grievance that feeds the next. This would provide structured acknowledgment and reduce the likelihood of grievance-driven retaliation — the single intervention most directly responsive to the decade-long injury pattern the poll documents.

Three other interventions address roles at the prevention and containment clusters. Bridge-builder scaling would fund nationally organized cross-partisan encounter, modeled on Braver Angels and Living Room Conversations but at population-changing scale. Teacher institutionalization would embed depolarization and conflict-handling skills in K-12 civics education through state boards of education. Referee support would provide federal protection guarantees for election workers through the Department of Justice. Each faces significant feasibility constraints: American political norms against centralized civic programming mean any large-scale bridge-building requires either a bipartisan congressional mandate or an unprecedented private foundation consortium; the politicized education environment makes curriculum adoption uncertain; and the decentralized election-administration system of over 10,000 independent jurisdictions means no single federal intervention can upgrade referee capacity systemically. Third-side intervention-effectiveness is rated low across all recommended interventions because the systemic drivers — media incentives, polarization as identity marker, delegitimized institutions — are deeper than any single role activation can reach.

What to watch

The questions a reader can carry to the next story:

  1. Does direct proximity to political violence drive the 10% figure, or does media consumption drive it nationally? Because direct exposure to political violence remains statistically rare for most Americans, the poll is primarily measuring what the media environment has made salient. That does not make the fear less real — salience drives perception — but it changes the diagnosis from widespread direct victimization to ambient anxiety fueled by a cumulative signal of dramatic events and eroded trust. If geography-concentrated, interventions can target affected communities. If media-driven, the structural constraints on content regulation make intervention harder.

  2. Would the bipartisan pattern hold if independent voters were sorted by ideology rather than party label? The current data may mask asymmetries within the independent category. Asymmetric threat perception would trigger the equalizer role, which is currently dormant.

  3. Does the symmetric threat perception persist under different political conditions — after a decisive election outcome, for instance — or does it crystallize into asymmetric perception? If contingent on the current moment, intervention windows may be tied to specific political cycles.

  4. Are existing bridge-building organizations reaching a scale that could shift ambient threat perception, or do they operate below the threshold where population-level impact becomes measurable? If below threshold, the gap is not about incremental funding but about whether the model itself can work at population scale.

  5. Would the poll’s finding hold if the question specified physical violence versus psychological threat? If primarily psychological threat, the equalizer and peacekeeper roles recede in priority; the healer and teacher roles rise. The open-ended methodology cannot distinguish.

  6. Are the survey’s internal mechanics — sample size, fielding dates, exact question wording — available to calibrate the statistical significance of the trend? The published material does not disclose these, meaning the 4-percentage-point shift cannot be independently assessed for sampling error. This coverage gap means the trend is suggestive but not confirmed.

This analysis describes what the sources document and what the survey data support. The characterization of institutional roles and structural constraints reflects an analytical framework applied to the reported facts, not a claim about intent.

Analytical techniques used in this piece

This analysis applies the methods below. Each links to a short, plain-English explainer you can read and reuse.

Process Mapping
Lays out a process end to end — steps, hand-offs, and bottlenecks.
Red-Team Assessment
Models a capable adversary probing a plan for the seams they would exploit.
The Third Side
Takes the vantage of the surrounding community that has a stake in resolving a conflict (Ury).