Accelerated transgressions in the second Trump presidency

Bright Line Watch February 2025 survey

Since regaining the pres­i­den­cy, Donald Trump is already chal­leng­ing America’s con­sti­tu­tion­al order. In the weeks since he took office, Trump launched a blizzard of executive actions that have not only shifted policy, but, in many cases, chal­lenged key prin­ci­ples of con­tem­po­rary American democracy (see appendix for a review).

In this context, we fielded parallel surveys of 520 political sci­en­tists (whom we refer to as “experts” below), 40 experts on online mis­in­for­ma­tion (whom we refer to “mis­in­for­ma­tion experts” below), and a rep­re­sen­ta­tive sample of 2,750 Americans (whom we refer to as “the public” below). These surveys, which we refer to as the February 2025 survey, were fielded from January 31 — February 10, 2025.

Our key findings are the following:

Assessments of demo­c­ra­t­ic performance

• Assessments of the overall performance of American democracy on a 0–100 scale have fallen to the lowest levels observed since we began tracking this measure in 2017: 53 among the public and 55 among experts.

• Out of 30 principles of democratic performance, the percentage of experts who say the US mostly or fully meets the standard plummeted by more than 30 percentage points on six principles and declined by 10–20 points on eleven others.

• Republican ratings of U.S. democracy increased slightly (59 on our 0–100 scale), but even Trump’s co-partisans perceive significant declines since November 2024 in the extent to which legal investigations of public officials are free from political influence, campaign donations influence public policy, and the patriotism of political adversaries is respected.

Threats to democracy

• Among Republicans, a majority (55%) of those whose allegiance is primarily to Trump rather than to the GOP say “Having a strong leader who does not have to bother with Congress” is a good way of governing the country.

• Political science experts regard the pardoning of January 6 offenders and firings of executive branch officials as the gravest threats to American democracy in the first weeks of the Trump administration, followed closely by the roles and influence of Elon Musk and his Department of Government Efficiency.

• Experts rated Kash Patel being confirmed as FBI director and the White House revoking media credentials for journalists as events that were highly likely to occur and threatening to democracy — with 63% assessing the threat as extraordinary. 

• Large majorities of Republicans approve of Trump administration actions that are rated by experts as threatening to democracy, whereas even larger majorities of Democrats disapprove.

Forecasts of future threats to democracy

• The set of events that experts identified as most likely to occur which pose the greatest threat to democracy include the confirmation of Patel as FBI director (now official) and the Trump administration revoking the media credentials of journalists covering the White House.

• After adjusting for a bias toward pessimism, forecasts of negative future events for democracy correspond closely between political science experts and forecasters on the Metaculus prediction platform.

Support for aggres­sion and violence

• Democrats are more supportive of aggressive action – including violence – against corporate CEOs in the name of economic justice than are Republicans. They are also more supportive of aggressive action against CEOs than against Republicans.

• Younger Democrats are more willing than older Democrats to endorse violence against CEOs or in the name of economic justice.

• Members of both parties are similarly supportive of aggressive action against members or leaders of the other party; support for property crimes or threats and harassment online is higher than support for physical violence. 

State of American democracy

We have asked expert and public respon­dents to rate the overall per­for­mance of U.S. democracy on a 0–100 scale since 2017 and 2018, respec­tive­ly. Our February 2025 survey also asked par­tic­i­pants to report the rating they would expect to give the United States on the same scale in 2027. The figure below shows the time series (solid lines) and projected future ratings from our expert sample, the public overall, and for the public sample broken out by partisanship. 

Expert ratings of American democracy have turned sharply downward. Since Bright Line Watch began its surveys, the experts have con­sis­tent­ly rated U.S. democracy well above the public overall and members of both major parties. From 2017–2024, expert assess­ments varied between 61 and 70, reaching a low of 61 in October 2020. They dropped only slightly after the 2024 election, declining from 69 in October 2024 to 67 in November 2024. However, the drop among experts from November 2024 to February 2025 was pre­cip­i­tous; average ratings have fallen to the lowest level we have observed since 2017 — an average of 55 out of 100. The experts predict further decline in the next two years, pro­ject­ing a rating of 47 in 2027. 

To put these ratings in context, the ratings that experts provide now for American democracy are com­pa­ra­ble to their assess­ment of Mexican democracy in October 2022. Their forecast for 2027 would put U.S. democracy between experts’ 2022 assess­ments of the Philippines and Brazil.

We also see the largest decrease in public ratings of U.S. democracy observed since we began col­lect­ing data in early 2018. Previously, public assess­ments had been remark­ably stable, with gains among one party typically offset by losses among the other — precisely the pattern we saw, for instance, between October and November 2024 (Democratic ratings declined and Republican ratings increased). Between November 2024 and February 2025, however, Republican assess­ments of democracy were stable, whereas ratings among Democrats plummeted from 60 to 49. Looking forward to 2027, the public foresees little change (from 53 to 54) — Democrats expect further decline (to 45) but Republicans project an increase (to 65).

Performance on democratic principles

As in past surveys, we asked expert and public respon­dents to rate the extent to which specific demo­c­ra­t­ic prin­ci­ples are currently upheld in the U.S. Each respon­dents was asked to rate a random subset of 9 items drawn from a list of 31.¹ 

The left panel of the figure below shows the per­cent­age of respon­dents who regard the U.S. as fully or mostly meeting the standard in question (as opposed to meeting it partly or not at all). The right panel shows the change (in per­cent­age points) from November 2024 to February 2025 in the pro­por­tion of the respon­dent group in question who say the U.S. meets that standard. The markers indicate respon­dent groups: political science experts (green) as well as Democrats (blue) and Republicans (red) from our public sample. (A version of this graph that presents overall expert and public assess­ments is provided in the appendix.)

We offer several obser­va­tions about these results. First, experts rate U.S. demo­c­ra­t­ic per­for­mance highest in the areas of elections (for example, that they are free of fraud, that all parties are allowed to compete, and that can­di­dates disclose infor­ma­tion about how they will govern) and freedom of expres­sion (for example, that citizens can express their opinions and that rights to free speech and to political protest are protected). Although public con­fi­dence on these prin­ci­ples lags behind that of the experts, we observe rel­a­tive­ly little polar­iza­tion between Democrats and Republicans on these items.

Second, con­sis­tent with the decline they perceive in U.S. democracy overall, the share of experts rating the U.S. as mostly or fully meeting the standard in question dropped by 30 per­cent­age points or more on six demo­c­ra­t­ic principles: 

• Government agencies do not punish political opponents (49 per­cent­age points)
• The press operates without inter­fer­ence (47 per­cent­age points)
• Government effec­tive­ly prevents private actors from engaging in polit­i­cal­ly-motivated violence or intim­i­da­tion (36 per­cent­age points)
• The leg­is­la­ture checks executive authority (35 per­cent­age points)
• Law enforce­ment inves­ti­ga­tions of public officials are free from political influence (34 per­cent­age points)
• Government sta­tis­tics are free from political influence (31 per­cent­age points) 

Declines of 10–20 per­cent­age points were observed for eleven other items among experts.

Next, experts, Democrats, and Republicans regard norms of demo­c­ra­t­ic discourse and tolerance to be in tatters. Fewer than 10% of experts and fewer than one-third of Democrats or Republicans rate the U.S. as mostly or fully meeting the standards of seeking common under­stand­ing of facts, seeking com­pro­mise with opponents, or respect­ing the patri­o­tism of political adversaries.

Finally, although Republicans rated U.S. demo­c­ra­t­ic per­for­mance more pos­i­tive­ly after Trump’s election win, they perceive more declines than improve­ments across the 30 prin­ci­ples we measured from November 2024 to February 2025. Three of these declines are sta­tis­ti­cal­ly sig­nif­i­cant: that law enforce­ment inves­ti­ga­tions are not polit­i­cal­ly com­pro­mised (12 per­cent­age points), that opponents’ patri­o­tism is not ques­tioned (13 points), and that campaign con­tri­bu­tions do not influence policy (10 points). On these items, even Republicans appear to recognize a sub­stan­tial decline in the early days of Trump’s second pres­i­den­cy (though we cannot rule out that Republicans are respond­ing to messages about the Biden admin­is­tra­tion from Trump and his allies).

Threats to democracy

We asked our experts to assess the impact of a series of recent events on American democracy. Respondents were first asked whether the event would benefit, threaten, or not affect American democracy. Those who selected benefit or threat were then asked about the extent of such an effect. The set of events that experts were asked to consider was the following:

• Donald Trump pardons nearly all indi­vid­u­als charged in con­nec­tion with the Jan. 6 Capitol attack, including violent offenders.
• Donald Trump fires at least twelve inspec­tors general without the 30-day notice to Congress required by the Inspector General Act.
• The Trump admin­is­tra­tion fires Department of Justice officials who worked on criminal inves­ti­ga­tions of Trump’s actions as president.
• Donald Trump and Secretary of State Marco Rubio moving to shut down the US Agency for International Development (USAID).
• Elon Musk serves as a close adviser to President Trump and the Department of Government Efficiency.
• Tech CEOs donate millions to Trump inau­gu­ra­tion and take actions intended to curry favor with the new administration.
• DOGE employees obtain access to federal personnel and payment systems.
• Donald Trump issues an executive order seeking to deny cit­i­zen­ship to people born in the U.S. whose parents are not lawful permanent residents.
• Donald Trump dis­miss­ing members of the Equal Employment Opportunity Commission and National Labor Relations Board before the expi­ra­tion of their terms.
• The White House budget office orders a pause to all grants and loans disbursed by the federal government.
• The Trump admin­is­tra­tion withdraws security pro­tec­tions for high-profile figures, such as Dr. Anthony Fauci, Gen. Mark Milley, and John Bolton, who have publicly clashed with Trump and face credible threats to their safety.
• The Senate confirms Pete Hegseth as Trump’s defense secretary.
• Donald Trump launches a cryp­tocur­ren­cy memecoin that is initially valued at billions of dollars.
• Meta replaces its third-party fact-checking program with a “Community Notes” system in the United States.
• Donald Trump orders the con­struc­tion of a facility for holding as many as 30,000 migrants at the naval base in Guantanamo Bay, Cuba.
• Meta relaxes its content mod­er­a­tion guide­lines in the United States.
• Donald Trump declares a national emergency at the U.S.-Mexico border, citing threats from cartels, criminal gangs, and other illicit activities.
• Joe Biden grants a “full and uncon­di­tion­al” pardon of his son Hunter Biden, absolving him of existing federal charges and shielding him from future pros­e­cu­tion for federal crimes.
• Donald Trump revokes Joe Biden’s executive order on address­ing risks from arti­fi­cial intel­li­gence (AI).
• Joe Biden grants a pre­emp­tive pardon to five members of his family besides Hunter Biden, shielding them from future pros­e­cu­tion for federal crimes.
• Donald Trump delays enforce­ment of the so-called TikTok ban after taking office.
• Donald Trump seeks to use the Alien Enemies Act to deport foreign gang members.
• Joe Biden grants a pre­emp­tive pardon to Dr. Anthony Fauci, retired Gen. Mark Milley, and members of the House of Representatives committee that inves­ti­gat­ed the Jan. 6, 2021 attack on the Capitol, shielding them from future pros­e­cu­tion for federal crimes.
• Joe Biden states that the Equal Rights Amendment to the U.S. Constitution should be con­sid­ered legally part of the Constitution.
• Ex-US Senator Bob Menendez (Democrat — New Jersey) sentenced to 11 years in federal prison for accepting bribes from foreign governments.

The left panel of the figure below shows expert ratings of the perceived benefits and threats to democracy. Events are listed in descend­ing order based on the per­cent­age of experts who rate them as a threat to democracy. (The right panel shows the per­cent­age of respon­dents who rated the scenario as neither a threat nor a benefit to democracy.)

Agreement on the threat to democracy was greatest for the January 6 pardons, which 95% of experts rated as a threat and which most rated as an extra­or­di­nary or serious threat (45% and 38%, respec­tive­ly). The numbers were almost the same for Trump’s firing of inspec­tors general (93% threat; 42% extra­or­di­nary, 40% serious) and Department of Justice inves­ti­ga­tors (92% threat; 60% extra­or­di­nary, 25% serious). 

Nearly as many of the experts perceived threats to democracy in the administration’s moves to shut down the U.S. Agency for International Development (89%), Elon Musk’s role in the Trump admin­is­tra­tion (88%), efforts by tech CEOs to seek influence (86%), the payment and personnel system access granted to DOGE (84%), Trump’s executive order denying birthright cit­i­zen­ship (83%), the dismissal without cause of officials from inde­pen­dent agencies (82%), the administration’s attempt (blocked in federal court) to suspend payment of federal grants (82%), and Trump’s sus­pen­sion of security for critics (80%). 

Finally, majori­ties of the experts iden­ti­fied threats to democracy in the con­fir­ma­tion of Pete Hegseth as Secretary of Defense (74%), Trump’s launching of a memecoin ahead of his second term (72%), recent changes at Meta (64% for ending fact-checking, 60% for loosening content mod­er­a­tion rules), the move to house undoc­u­ment­ed migrants at Guantanamo Bay (61%), and Trump’s dec­la­ra­tion of an emergency at the nation’s southern border (55%). Fewer perceived threats in Trump’s revo­ca­tion of Biden’s executive order on AI security (44%), his decision to suspend enforce­ment of leg­is­la­tion banning Tik Tok (42%), and his plan to use the Alien Enemies Act to deport foreign gang members (36%). 

Clustered near the bottom of the figure are a number of actions from the closing days of the Biden admin­is­tra­tion, some of which are regarded as more threat­en­ing than ben­e­fi­cial, even if not at levels com­pa­ra­ble to Trump admin­is­tra­tion actions. Many experts, for example, regard Biden’s pardons of his son, Hunter (47%), and of other family members (42%) as threats to democracy. In both cases, though, narrow majori­ties regarded the pardons as neither threat­en­ing nor ben­e­fi­cial. Similarly, Biden’s assertion that the Equal Rights Amendment (ERA) was part of the U.S. Constitution was largely regarded as neither a threat nor a benefit for American democracy (66%); those experts who perceived an effect were divided over whether the statement posed a threat (14%) or benefit (20%). 

Experts were most divided over Biden’s decision to pardon potential targets of ret­ri­bu­tion from Trump like Anthony Fauci – 23% rated the pardons as a benefit to democracy while 27% rated them as a threat. The only event we surveyed that was clearly regarded as ben­e­fi­cial for democracy (82%) was the sen­tenc­ing of former U.S. Senator Bob Menendez (D‑NJ) to 11 years in prison for bribery and espionage. 

Threats to democracy and public approval

We asked our public sample whether they approved or dis­ap­proved of the actions taken by Biden and Trump that experts rated as potential threats or benefits to democracy. The figure below plots the per­cent­age of Democratic and Republican par­tic­i­pants in our public sample who strongly or somewhat approve of the events in question (on the vertical axis) against the per­cent­age of experts rating each action as a serious or extra­or­di­nary threat to democracy (hor­i­zon­tal axis).

Two patterns are apparent. The first is a stark partisan divide: Democrats over­whelm­ing­ly approve of Biden’s actions, with approval ranging from 60% to 91%, while Republicans’ approval is much lower, ranging from 6% to 35%. On actions taken by Trump, the reverse is true – 70% or more of Republicans approve of every action surveyed whereas Democrats dis­ap­prove (usually with super­ma­jori­ties above 80%). Republican approval of Trump’s actions is somewhat lower for those that experts rate as less threat­en­ing to democracy, but the rela­tion­ship is modest. Second, experts rate Trump’s actions as more threat­en­ing to democracy than Biden’s. Six were rated as serious or extra­or­di­nary threats to democracy by more than 40% of experts. Biden’s actions, by contrast, were regarded to be serious or extra­or­di­nary threats by less than 10% of expert respondents. 

Comparing expert ratings of technology-related events

For a subset of recent events related to tech­nol­o­gy and online mis­in­for­ma­tion, we sep­a­rate­ly surveyed experts on online mis­in­for­ma­tion recruited from an invi­ta­tion-only email list main­tained by experts in the space. The figure below compares democracy threat/benefit ratings between our political science expert sample and the mis­in­for­ma­tion expert sample (excluding BLW panel members from the latter group to avoid double-counting; a version of the figure showing the per­cent­ages of each group who rate the event as neither a threat nor a benefit to democracy is provided in the appendix).

As the figure shows, expert assess­ments were similar between the two groups. Misinformation experts were more likely to see Meta’s decision to loosen content mod­er­a­tion rules as a threat to democracy (78%) than were political science experts (60%). They were also somewhat more likely to see Trump revoking the Biden AI order as a threat and less likely to see Musk’s advising role and tech CEO influence-seeking as threats.

Assessing the like­li­hood of future events

Following our practice in past surveys, we asked our expert sample to rate the prob­a­bil­i­ty of a set of future events. In this survey wave, however, we coor­di­nat­ed with the online fore­cast­ing platform Metaculus, which uses cash prizes and awards to incen­tivize accurate forecasts, to solicit forecasts to compare with those of our political science experts. We presented our experts and the Metaculus fore­cast­ers with the same set of potential events along with back­ground infor­ma­tion and criteria for how each forecast would be resolved.² Our fore­cast­ing community within Metaculus is hosted here. (The data reported below are only those Metaculus forecasts received during the expert survey period; see appendix for the dis­tri­b­u­tion of response times for each item for experts and Metaculus forecasts.)

The potential political events for which we solicited like­li­hood estimates is here:

• Will the Department of Justice or Education open a civil com­pli­ance inves­ti­ga­tion into whether diversity-related programs at Harvard con­sti­tute illegal dis­crim­i­na­tion or pref­er­ences before 2026?
• Will the United States and Denmark announce formal nego­ti­a­tions over the possible transfer of sov­er­eign­ty of Greenland to the United States before 2026?
• Will the National Archivist certify and publish the Equal Rights Amendment to the Constitution by June 30, 2025?
• Will the White House revoke the media cre­den­tials of reporters from two media outlets before 2026?
• Will Elizabeth MacDonough, the current Senate par­lia­men­tar­i­an, be removed or resign by June 30, 2025?
• Will the Department of Justice announce an inves­ti­ga­tion or pros­e­cu­tion of a Democrat who served as president, vice president, con­gres­sion­al leader, whip, or impeach­ment manager before 2026?
• Will President Trump formally invoke the Insurrection Act before 2026?
• Will President Trump formally invoke his authority under the Constitution to adjourn Congress before 2026?
• Will leg­is­la­tion be approved by at least one chamber of Congress in 2025 that partially or fully exempts at least one executive branch depart­ment, the CIA, or EPA from the Freedom of Information Act?
• Will leg­is­la­tion be approved by at least one chamber of Congress in 2025 that places new rules, limits, or require­ments on early voting, absentee voting, or voter reg­is­tra­tion in federal elections?
• Will Kash Patel be confirmed by the Senate as FBI Director by June 30, 2025?
• Will eight or more of the fifteen heads of the executive depart­ments be serving in an acting role on December 31, 2026?
• Will a federal court rule that the Trump admin­is­tra­tion violated the Impoundment Control Act in attempt­ing to per­ma­nent­ly withhold or cancel con­gres­sion­al­ly allocated funds by September 27, 2025?
• Will customs duties revenue to the federal gov­ern­ment increase from $80 billion in fiscal year 2023 to $160 billion in fiscal year 2026 (in 2023 dollars)?
• Will at least one announced Trump Cabinet nominee other than Matt Gaetz be withdrawn or rejected by the Senate before July 1, 2025?

The figure below reports median estimates from political science experts and Metaculus fore­cast­ers on a 0–100% scale. (We also asked experts and fore­cast­ers to rate a list of non-political questions from Metaculus that are listed in the appendix; these are included in a version of this graph provided in the appendix.)

The political science experts rate many of the events as more likely than do Metaculus fore­cast­ers, including an inves­ti­ga­tion of a leading Democrat (67% expert versus 40% Metaculus), Trump invoking the Insurrection Act (51% expert versus 20% Metaculus), an executive branch agency being exempted from FOIA (50% expert versus 15% Metaculus), and Trump adjourn­ing Congress (40% expert versus 8% Metaculus). Experts rated several of these events as having an approx­i­mate­ly 50% prob­a­bil­i­ty, which often indicates a lack of infor­ma­tion in binary fore­cast­ing tasks. Metaculus fore­cast­er ratings were more widely dis­trib­uted over the scale, with lower ratings on most items. 

These results are con­sis­tent with our previous finding that experts tended to over­es­ti­mate the prob­a­bil­i­ty of negative events for democracy. We therefore follow the procedure developed in that report to adjust expert forecasts taking this pessimism bias into account:

After this adjust­ment, we see a very close cor­re­spon­dence between the estimates provided by experts and those provided by fore­cast­ers (which are cor­re­lat­ed at 0.66). The most notable dif­fer­ences are for the con­fir­ma­tion of Kash Patel as FBI director (which took place after the study period but was seen as increas­ing­ly likely during the period of data col­lec­tion), and for leg­isla­tive action to tighten rules around voting. In both cases, Metaculus fore­cast­ers believe these events are more likely to occur than do our political science experts.³

We also asked our experts to assess the potential threat or benefit to democracy of twelve of these events. The figure below shows those responses.

Overwhelming majori­ties of experts perceived threats to democracy from Trump invoking the Insurrection Act (95% – with 63% assessing the threat as extra­or­di­nary), a leg­isla­tive move to exempt an executive branch agency from FOIA (94%), Trump invoking con­sti­tu­tion­al authority to adjourn Congress (92%), the Justice Department opening inves­ti­ga­tions into a Democratic leader (91%), con­fir­ma­tion of Kash Patel as FBI director (89%), the White House revoking media cre­den­tials (85%), or leg­isla­tive action to restrict ballot access (83%). 

Smaller majori­ties also would regard the removal of the Senate Parliamentarian (58%) or an inves­ti­ga­tion into diversity-related programs at Harvard (51%) as demo­c­ra­t­ic threats. 

By contrast, a plurality (48%) would regard pub­li­ca­tion of the ERA by the National Archivist as a demo­c­ra­t­ic benefit, though 17% rate it as a threat and 34% perceive no effect. Finally, a strong majority of 83% would regard a court ruling that Trump violated the Impoundment Control Act as a benefit to democracy.

The figure below helps evaluate threats to democracy by combining their estimated like­li­hood with the perceived threat to democracy. Events in the lower left of the figure are the least likely to occur and pose the smallest threat — most notably, ERA cer­ti­fi­ca­tion and nego­ti­a­tion with Denmark over Greenland. Events in the top right of the figure, by contrast, are the most likely to occur and pose the greatest threat to democracy. This set includes the con­fir­ma­tion of Patel as FBI director and the Trump admin­is­tra­tion revoking the media cre­den­tials of jour­nal­ists covering the White House.

s

Support for a non-demo­c­ra­t­ic political regime

Using questions from the World Values Survey, we asked our public sample about their support for democracy versus rule by experts and strongman rule. Specifically, par­tic­i­pants were asked to rate the following systems of gov­ern­ment as very bad, fairly bad, fairly good, or very good:

• Democracy (“A demo­c­ra­t­ic political system”)
• Expert rule (“Having experts, not the gov­ern­ment, make decisions according to what they think is best for the country”)
• Strongman
  • Original version: “Having a strong leader who does not have to bother with Congress and elections”
  • Modified version: “Having a strong leader who does not have to bother with Congress”

Prominent studies (here, here, here, and here) have tracked support over time for the original strongman question. But the legit­i­ma­cy of elections is powerful even in author­i­tar­i­an countries, whereas leg­is­la­tures are often mar­gin­al­ized. We therefore sought to measure how many people would do away with the leg­isla­tive check on executive authority — the precise issue at stake in many current disputes over actions taken by the Trump admin­is­tra­tion — by ran­dom­iz­ing the wording of the strongman question.

Because strongman rule is explic­it­ly per­son­al­is­tic (i.e., it favors the con­cen­tra­tion of power in the hands of an indi­vid­ual and dimin­ish­es the role of plu­ral­is­tic insti­tu­tions), we are also par­tic­u­lar­ly inter­est­ed in its appeal to the subset of Republican party iden­ti­fiers who regard them­selves as primarily sup­port­ers of Donald Trump rather than of the Republican Party as a whole. As we note in the appendix to this report, people who say they primarily support Trump now comprise about 60% of the Republicans whereas “insti­tu­tion­al­ist” Republicans who say they primarily support the party have fallen to about 40%. 

The figure below shows the per­cent­ages of each partisan group – with Republicans broken out by their primary alle­giance – regarding each regime type as “fairly good” or “very good” rather than “fairly bad” or “very bad.”

Encouragingly, all groups rate democracy more favorably than other regime types (the same pattern we saw in our January-February 2021 survey; support for each regime type remains stable across partisan groups). We also note that all groups regard expert rule less favorably than democracy but more favorably than either version of strongman rule. Notably, although disdain for expert rule might be asso­ci­at­ed with Republican critiques of main­stream science and insti­tu­tions of higher education, approval of expert rule is only mod­er­ate­ly lower among Republicans (57% overall) than among Democrats (64%). 

Among Republicans, however, we see important dif­fer­ences when we dif­fer­en­ti­ate between Trump sup­port­ers and insti­tu­tion­al­ists, espe­cial­ly when we ask them to evaluate a strongman leader who continues to hold elections but sidelines Congress. Compared to insti­tu­tion­al­ist Republicans, Trump sup­port­ers are somewhat less likely to rate democracy favorably and more likely to view strongman rule pos­i­tive­ly. In total, 55% of Trump sup­port­ers endorse a strong leader who doesn’t have to bother with Congress and 45% endorse a strong leader who doesn’t have to bother with Congress or elections, compared to 44% and 30% of insti­tu­tion­al­ist Republicans, respectively. 

This form of strongman rule, with regular elections but a weakened leg­isla­tive check on the executive, is precisely what scholars warn against when they describe fears of the U.S. slipping into “com­pet­i­tive author­i­tar­i­an­ism” — a pattern con­sis­tent with episodes of demo­c­ra­t­ic erosion in other countries. The fact that it attracts majority support from Trump’s base is therefore deeply concerning.

Public support for aggres­sive political action and violence

How much do Americans support, or at least tolerate, violence, threats, and harass­ment against their political adver­saries? In November 2021, we joined this long-running debate among political sci­en­tists, which gained new attention after the January 6, 2021 attack on the Capitol. At the time, we found that support for violence was lower than past research had suggested. The share of partisans who would endorse physical violence to advance partisan goals was, in most instances, in the low single digits. Any support for violence is troubling, but the levels we found were much less than prior research and media coverage had suggested.

In December 2024, the shooting of UnitedHealthcare CEO Brian Thompson returned attention to public support for political violence. Surveys and news media accounts reported high levels of support for Luigi Mangione, Thompson’s alleged murderer among left-leaning members of the public. In our most recent survey, we therefore compared support for violence against corporate leaders with support for partisan violence.

Specifically, par­tic­i­pants in our public sample were randomly assigned to one of two sets of questions that probed their support for behaviors ranging from online harass­ment to physical attacks. The question sets differed in who took the action (actor), who or what was the target of the action, and on the goals of the action, as sum­ma­rized in the table below:

Survey par­tic­i­pants ran­dom­ized to see questions on…

	Participant action	Anti-corporate action
Actor	A member of survey par­tic­i­pan­t’s party (Democrat or Republican)	“an ordinary person”
Target	A leader or elected official from the other main party (Republican or Democrat).	“a CEO of a large corporation”
Goals	“their political goals”	“economic justice or labor rights”

Once assigned to receive questions focused on partisan action or on anti-corporate action, par­tic­i­pants were presented with questions in three distinct categories:

Threats and harassment

• Is it ever OK for [actor] to send threat­en­ing and intim­i­dat­ing messages to [target] or is it never OK?
• Is it ever OK for an [actor] in the public to harass a [target] on the Internet in a way that makes the [target] feel fright­ened or it is never OK?

Rationale for violence

• What if a [target] enacts harmful or exploita­tive policies? Would violence (the use of physical force to harm someone) be justified then or is it never justified?
• What if a [target] utilizes his/her wealth to influence an election or political decision? Would violence (the use of physical force to harm someone) be justified then or is it never justified?
• What if a [target] engages in financial crimes or cor­rup­tion? Would violence (the use of physical force to harm someone) be justified then or is it never justified?

Severity of law-breaking action

• Is it ever justified for [actor] to commit non-violent mis­de­meanors (for example, tres­pass­ing during a protest) in advancing [goals] these days or is it never justified?
• Is it ever justified for [actor] to commit non-violent felonies (for example, causing sig­nif­i­cant property damage during a protest) in advancing [goals] these days or is it never justified?
• Is it ever justified for [actor] to commit violent felonies (for example, assault during a con­fronta­tion with counter-pro­tes­tors) in advancing [goals] these days or is it never justified?

In each case, if the survey respon­dent indicated that aggres­sive actions are ever OK or justified, they were asked a follow-up question probing their degree of openness to aggres­sive action or violence – for example, “How often is it OK to send threat­en­ing messages?” or “How much is it justified to commit violent felonies in advancing political goals?” The follow-up questions them­selves always included a “Never” or “Not at all” response option. We regard a survey par­tic­i­pant as sup­port­ive of aggres­sive action only if their response to both the initial and follow-up question indicated openness to that action. We also follow what we regard to be best practice in esti­mat­ing support for violence by excluding responses from inat­ten­tive survey par­tic­i­pants (16% of all respon­dents).⁴ 

We find no evidence of increas­ing support for political violence over time (see appendix). We thus focus on how support for violence varies by target, actor, and the action in question.

The figure below compares support among attentive Democratic and Republican par­tic­i­pants for threats and harass­ment, for law-breaking actions of varied severity, and for violence in response to various alleged offenses. The left side of the figure shows support for aggres­sion against opposing partisans or actions taken to advance political goals. Support is similar across parties on most items, and higher for trans­gres­sions against property and for threats and harass­ment versus physical violence against a person.

Republicans are slightly more sup­port­ive of online harass­ment (11%) and of sending threat­en­ing messages to oppo­si­tion partisans (12%) than are Democrats (7%). On the questions about the use of physical force to harm an opposing partisan leader, we observe no sta­tis­ti­cal­ly sig­nif­i­cant dif­fer­ences between partisan groups regard­less of the rationale offered. In the context of political protests, Democrats are more inclined to support non-violent mis­de­meanors (for example, tres­pass­ing; 33%) than are Republicans (18%), and more likely to support non-violent felonies (for example, sig­nif­i­cant property damage; 9% for Democrats versus 4% for Republicans). Support for violent felonies during political protests is low among both groups – 3% for Democrats, 2% for Republicans – and the groups are not sta­tis­ti­cal­ly dis­cernible on that question.

The right side of the figure compares support for aggres­sive action and violence against CEOs between partisan groups. Here we observe con­sis­tent­ly higher support among Democrats than Republicans. Democrats are more sup­port­ive of online harass­ment (16% to 7% for Republicans) and threats (14% to 7%) against CEOs as well as the use of physical force against CEOs who have imple­ment­ed harmful or exploita­tive policies (20% to 6%), used their wealth to influence politics (18% to 4%), or committed financial crimes (13% to 6%). 

In the wake of the Thompson killing, media attention focused on support among Gen Z and Millennials for violence as ret­ri­bu­tion for perceived economic injustice, par­tic­u­lar­ly on the ide­o­log­i­cal left. The next figure plots the same data, but breaks out 18–34 year old respon­dents from those 35 years old and older.

Differences by age cohort and limited when we examine support for aggres­sive and violent actions against partisan opponents. We do, however, see sub­stan­tial dif­fer­ences between people who are 18–34 years old and those who are older in terms of their will­ing­ness to endorse aggres­sive and violent actions against “a CEO of a larger cor­po­ra­tion” and those taken in the pursuit of “economic justice or labor rights,” par­tic­u­lar­ly among Democrats. In total, 23% of 18–34 year old Democrats regard violent felonies as at least sometimes justified in the pursuit of economic justice, compared to just 7% of Democrats age 35 or older. Younger Democrats are also more willing to endorse the use of violence if a CEO has pursued harmful or exploita­tive policies (38% versus 14% for older Democrats), used their wealth to influence politics (29% versus 14%), or has engaged in financial crimes or cor­rup­tion (28% to 8%).⁵ 

The next figure plots the data dif­fer­ent­ly, showing com­par­isons within partisan group according to whether the target of action is an oppo­si­tion partisan as opposed to a CEO. Among Democrats, support for aggres­sive action and violence against CEOs is far higher across the board than support for the same actions against Republicans. Most notably, even on the items that specif­i­cal­ly describe physical violence (“the use of physical force to harm someone”), Democrats are far more likely to support such action against CEOs than against Republican leaders who have taken the same actions – harmful or exploita­tive policies (20% to 11%), using wealth to influence politics (18% to 8%), or engaging in financial crimes or cor­rup­tion (13% to 5%). Among Republicans, by contrast, dif­fer­ences asso­ci­at­ed with the target of action are small and, where they are sta­tis­ti­cal­ly sig­nif­i­cant, support for aggres­sion against Democrats is higher than support for action against CEOs (11% to 7% for threat­en­ing messages).

To sum up, we find that members of both parties support aggres­sive action against the other party at about the same level, and that support for trans­gres­sions against property or non-physical threats and harass­ment are higher than support for physical violence. Democrats, however, support aggres­sive action – including violence – against CEOs in the name of economic justice at far higher levels than Republicans, and they support aggres­sive action against CEOs at higher levels than they support action against their Republican opponents. Younger Democrats, in par­tic­u­lar, are more willing to support aggres­sive action and violence against CEOs and in the name of economic justice than are their older counterparts.

Appendix

Bright Line Watch conducted its twenty-fourth survey of academic experts and its twenty-first survey of the general public from January 31–February 10, 2025. Our public sample consisted of 2750 par­tic­i­pants from the YouGov panel who were selected and weighted to be rep­re­sen­ta­tive of the U.S. adult pop­u­la­tion. We also surveyed 520 political science experts across a diverse range of subfields. Our email list was con­struct­ed from the faculty list of U.S. insti­tu­tions rep­re­sent­ed in the online program of the 2016 American Political Science Association con­fer­ence and updated by reviewing depart­ment websites and job placement records from Ph.D. programs in the period since.

All estimates shown in the report used weights provided by YouGov. Our expert sample is unweight­ed because we do not collect demo­graph­ic data to protect anonymity. Error bars in our graphs represent 95% con­fi­dence intervals. Data are available here.

Summary of recent events threatening democratic norms and practices

Key areas of gov­er­nance affected in the first weeks of the Trump admin­is­tra­tion include:

• Rule of law. President Trump pardoned and commuted sentences for January 6 rioters (including those convicted of violent crimes and seditious con­spir­a­cy) and the Department of Justice directed federal pros­e­cu­tors to tem­porar­i­ly withdraw cor­rup­tion charges against New York City Mayor Eric Adams in return for coop­er­a­tion on immi­gra­tion enforce­ment. The admin­is­tra­tion also began inves­ti­ga­tions into thousands of FBI personnel who were assigned to the January 6 inves­ti­ga­tions, withdrew security pro­tec­tions from Trump critics who face demon­strat­ed security threats, and issued an executive order calling on federal agencies to provide the White House with lists of private and non-profit sector insti­tu­tions rec­om­mend­ed for federal inves­ti­ga­tion, breaking the post-Watergate norm of targeting decisions being insulated from White House influence. Together, these actions signal that Trump will protect those who break the law on his behalf and punish those who enforce the law against his interests.
• Power of the purse. The Constitution grants to Congress the authority to appro­pri­ate funds for gov­ern­ment spending, but the admin­is­tra­tion has claimed the authority both to make spending com­mit­ments not autho­rized by Congress in an effort to pressure federal employees to resign and to reduce or eliminate expen­di­tures appro­pri­at­ed by Congress if they do not align with the administration’s goals.
• Executive branch personnel. Most federal gov­ern­ment employees are fireable only for cause. The admin­is­tra­tion fired more than a dozen pros­e­cu­tors from the Department of Justice because they had worked on pros­e­cu­tions of President Trump. The admin­is­tra­tion fired 17 Inspectors General, inde­pen­dent watchdogs who are tasked with iden­ti­fy­ing instances of waste, fraud, and mis­man­age­ment in the executive branch. The firings undermine account­abil­i­ty in the executive branch and defy civil-service pro­tec­tion laws. President Trump also dismissed leaders of inde­pen­dent agencies, including two members of the Equal Employment Opportunity Commission, a member of the National Labor Relations Board, and the chair of the Federal Elections Commission, although their terms were not completed. The courts have repeat­ed­ly found that pres­i­dents lack the legal authority to dismiss leaders of inde­pen­dent agencies. 
• Administrative procedure. The Trump admin­is­tra­tion shuttered a major inde­pen­dent federal agency, USAID, again citing policy diver­gence and ignoring Congress’s authority to establish and authorize gov­ern­ment agencies and their programs, and threat­ened similar actions against the Department of Education. The admin­is­tra­tion also sought to confer control of sensitive personnel and payment systems to rep­re­sen­ta­tives of the so-called Department of Government Efficiency under unclear legal authority.
• Individual rights. President Trump issued an executive order denying the cit­i­zen­ship rights of people born in the United States that are unam­bigu­ous­ly granted under the 14^th Amendment to the Constitution.

31 democratic principles

• Government officials are legally sanc­tioned for misconduct
• Government officials do not use public office for private gain
• Government agencies are not used to monitor, attack, or punish political opponents
• All adult citizens enjoy the same legal and political rights
• Government does not interfere with jour­nal­ists or news organizations
• Government effec­tive­ly prevents private actors from engaging in polit­i­cal­ly-motivated violence or intimidation
• Government protects indi­vid­u­als’ right to engage in unpopular speech or expression
• Political com­pe­ti­tion occurs without criticism of opponents’ loyalty or patriotism
• Elections are free from foreign influence
• Parties and can­di­dates are not barred due to their political beliefs and ideologies
• All adult citizens have equal oppor­tu­ni­ty to vote
• All votes have equal impact on election outcomes
• Elections are conducted, ballots counted, and winners deter­mined without pervasive fraud or manipulation
• Executive authority cannot be expanded beyond con­sti­tu­tion­al limits
• The leg­is­la­ture is able to effec­tive­ly limit executive power
• The judiciary is able to effec­tive­ly limit executive power
• The elected branches respect judicial independence
• Voter par­tic­i­pa­tion in elections is generally high
• Information about the sources of campaign funding is available to the public
• Public policy is not deter­mined by large campaign contributions
• Citizens can make their opinions heard in open debate about policies that are under consideration
• The geo­graph­ic bound­aries of electoral districts do not sys­tem­at­i­cal­ly advantage any par­tic­u­lar political party
• Even when there are dis­agree­ments about ideology or policy, political leaders generally share a common under­stand­ing of relevant facts
• Elected officials seek com­pro­mise with political opponents
• Citizens have access to infor­ma­tion about can­di­dates that is relevant to how they would govern
• Government protects indi­vid­u­als’ right to engage in peaceful protest
• Law enforce­ment inves­ti­ga­tions of public officials or their asso­ciates are free from political influence or interference
• Government sta­tis­tics and data are produced by experts who are not influ­enced by political considerations
• The law is enforced equally for all persons
• Incumbent politi­cians who lose elections publicly concede defeat
• Universities, busi­ness­es, and pro­fes­sion­al orga­ni­za­tions operate inde­pen­dent­ly of gov­ern­ment pressure

Ratings of U.S. democratic performance

Threats and benefits to democracy

Response times to forecast items

Non-political forecast questions

• Will real housing prices in the US increase more in 2025 compared to 2024?
• Will any senior exec­u­tives of these major public companies be impris­oned for corporate crimes before 2026?⁶
• Will the 12-month per­cent­age change in the US Consumer Price Index be lower in November 2025 than it was in November 2024?
• Will there be any reported human-to-human trans­mis­sion of highly path­o­gen­ic avian influenza H5N1 globally before 2026?
• Will the number of deaths from antibi­ot­ic-resistant infec­tions per year double by December 31, 2025?
• Will the UK housing market crash before July 2025?
• Will the U.S. employ­ment to pop­u­la­tion ratio fall below 50% before 2027?
• Will the World Health Organization designate a new COVID-19 Variant of Concern before July 1, 2025?
• Will the United States score in the bottom 50% for academic freedom in 2025 (according to the AFI)?
• Will Freedom House report that internet freedom has declined every year until 2025?

Comparing expert and Metaculus forecasts (all events)

Comparing raw and pessimism-adjusted median expert forecasts

Trump’s growing hold on the Republican Party

Since late 2022, we have asked each respon­dent who iden­ti­fies as a Republican (or leans toward that party) whether they regard them­selves as more a supporter of the Republican Party or more a supporter of Donald Trump.

Through the summer of 2023, more iden­ti­fied primarily with the party. By early 2024, when Trump had locked up the party’s pres­i­den­tial nominee, a narrow majority iden­ti­fied primarily with Trump, and the share of Trump-over-party Republican respon­dents has grown steadily since, sta­bi­liz­ing at around 59% in the past the most recent surveys.

Support for political aggression and violence

Election confidence and legitimacy

1. In past surveys, our list had included 30 prin­ci­ples. The full list is provided in the appendix. The item added for the first time in this survey wave was: “Universities, busi­ness­es, and pro­fes­sion­al orga­ni­za­tions operate inde­pen­dent­ly of gov­ern­ment pressure.”
2. To motivate par­tic­i­pa­tion on Metaculus, the most accurate fore­cast­ers from that platform qualify for $2,500 in prize money.
3. The appendix includes a figure showing the dis­tri­b­u­tion by date of forecasts from political science experts and from the Metaculus fore­cast­ers on each item. The expert forecasts were mostly collected in a few days after our main survey invi­ta­tion while the Metaculus forecasts are more uniformly dis­trib­uted across the data col­lec­tion period. As a result, the Metaculus sample enjoyed an infor­ma­tion­al advantage on items where salient facts were changing over time – for example, on the Patel confirmation.
4. We embedded two items in our survey that were designed to identify inat­ten­tive respon­dents who were not reading questions and respond­ing carefully. The 16% of respon­dents who failed to respond to both questions correctly were clas­si­fied as inat­ten­tive. The appendix includes a figure that compares estimated levels of support for violence across inat­ten­tive as well as attentive survey par­tic­i­pants. We place less credence in the estimates from the inat­ten­tive subset, however, for the reasons outlined by Westwood, Grimmer, Tyler, and Nall (2022).
5. We note that the rela­tion­ship between age and support for violence among Democrats may be con­found­ed by ideology. Specifically, younger Democrats are also ide­o­log­i­cal­ly more liberal than are older Democrats. However, when we predict support for violence in a logit model using self-reported ideology (a seven-point scale) and age group (18–34 years old or not), age group is a sta­tis­ti­cal­ly sig­nif­i­cant predictor of all eight potential actions against CEOs or in the name of economic justice (ideology is also sig­nif­i­cant on 7 of 8 estimates). In short, younger Democrats are more willing to endorse violence even when we take ideology into account.
6. As outlined in the res­o­lu­tion criteria, the list of exec­u­tives includes those employed by companies listed in the following major indices at any point in 2021 or later: S&P 500, FTSE 100, DAX 40, CAC 40, FTSE MIB, and Nikkei 225.