American democracy at the start of the Biden presidency
Bright Line Watch January-February 2021 surveys

U.S. democracy passed a milestone on January 20, 2021 with the inau­gu­ra­tion of Joe Biden and Kamala Harris. Donald Trump’s pres­i­den­cy presented unprece­dent­ed chal­lenges to this country’s demo­c­ra­t­ic norms and practices. His role in a violent insur­rec­tion, sparking a second impeach­ment has fueled further debate over how to preserve American democracy and prevent further erosion. In this context, Bright Line Watch conducted its 14th expert survey and 11th public survey.

Bright Line Watch began its surveys in February 2017 during the first weeks of the Trump admin­is­tra­tion. The current surveys, which were fielded January 28-February 8, 2021, seek to measure the state of American democracy in the first days of the Biden admin­is­tra­tion. Yet, the country still lives in the shadow of the Trump legacy. As our results indicate, loyalties and antipathy toward Trump — whose Senate impeach­ment trial began imme­di­ate­ly after the surveys were fielded — continue to shape the views of citizens and of gov­ern­ment officials. 

We asked our public sample about the impor­tance they place on democracy, about the legit­i­ma­cy of the 2020 election results and their con­fi­dence that votes were cast and counted fairly, and about how they regard Donald Trump. We also presented them with a candidate-choice survey exper­i­ment designed to measure the effect of can­di­dates’ positions on the 2020 election and of Trump’s impeach­ment on support for Republican can­di­dates for office. Finally, we asked respon­dents to con­tem­plate the dis­so­lu­tion of the United States — a prospect that has been broached increas­ing­ly fre­quent­ly in recent months — and measured their com­mit­ment to the union. 

We asked our experts to evaluate 16 proposed reforms that are intended to improve the quality of American democracy. We also asked them to evaluate a number of events that occurred between November and January in terms of their impor­tance, their (ab)normality, and the potential threat they pose to democracy. As in previous surveys, we asked both experts and the public to assess the quality of U.S. democracy overall and to rate its per­for­mance on 30 distinct demo­c­ra­t­ic prin­ci­ples. In this wave, we also asked respon­dents to rate the impor­tance of each principle to democracy. 

We report a number of key findings below from the public and experts we surveyed:


  • Partisan differences on the 2020 election and on legal and political accountability for former President Trump are profound. Democrats trust the election results, support disqualifying Trump from holding office, and believe he should face criminal prosecution. Republicans distrust the election results and favor moving on without consequences for Trump. Independents are split.

  • A candidate choice experiment reveals cross-party consensus on government spending for COVID relief but stark polarization over certification of the presidential election and impeachment, with Republicans punishing Republican candidates for crossing the party line on either issue.

  • When presented with a proposal for their region to secede from the United States, almost one in three Americans (29%) are willing to entertain the prospect. Republicans (33%) support secession more than Democrats (21%), but Democrats are more amenable to secession than Republicans in areas where they tend to hold power.


  • The experts overwhelmingly favor a set of reform proposals to expand voting participation, tighten campaign finance regulation, and modify how electoral districts are configured and how votes are cast. They also favor abolishing the Senate filibuster and imposing term limits on Supreme Court justices but reject compulsory voting.

  • We find that prior expert forecasts about the election and the presidential transition, which reflected deep skepticism about President Trump’s likely actions, were largely on target. 

  • Experts rate the January 6 insurrection and President Trump’s pressure on state-level officials to overturn the election as among the most abnormal and important events of the Trump presidency. Our experts overwhelmingly regard these events and the votes by a majority of Republican lawmakers in Congress not to certify the presidential election results as grave or serious threats to American democracy.

Experts and public

  • Overall estimates of democratic performance rose among experts since October 2020.

  • Ratings of democracy reversed between those who approve of President Trump and those who disapprove of him. Approvers had rated U.S. democracy higher than disapprovers since 2017, but that pattern inverted in January 2021. As a result of these offsetting changes, public ratings of U.S. democracy did not change.

  • Those who approve of Trump have increased the priority they place on judicial independence relative to prior surveys. Those who disapprove of Trump downgraded the importance of protections for free speech, preventing politically motivated punishment by government agencies, and allowing all parties to participate in elections.

Public survey

Election confidence

As in our October and November 2020 surveys, we asked public respon­dents how much con­fi­dence they had in the election results. In this survey wave, Bright Line Watch also partnered with CivicPulse to survey 212 local gov­ern­ment pol­i­cy­mak­ers from across the country on election con­fi­dence. Both the public and local officials were asked “How confident are you that the votes for the 2020 pres­i­den­tial election were counted correctly in [your county, your state, the US overall]?”1 Our goal was to test whether local officials — who played a key role in the 2020 election — would be better informed than the public about the integrity of the electoral system.

We note several key findings. First, across both the public and the CivicPulse sample of local pol­i­cy­mak­ers,2 Democrats exhibit greater con­fi­dence in electoral admin­is­tra­tion than do Republicans or inde­pen­dents at every level of gov­ern­ment (county, state, national). Second, in both samples, con­fi­dence in electoral admin­is­tra­tion is higher at levels of gov­ern­ment closer to home. Confidence in county-level and state-level vote counts are far higher than con­fi­dence in aggre­ga­tion at the national level among inde­pen­dents and, par­tic­u­lar­ly, among Republicans. 

Local pol­i­cy­mak­ers are also more confident than the public that votes were counted correctly. This is mar­gin­al­ly true among Democrats, although con­fi­dence is so high among the Democratic public sample that the potential for dif­fer­ence is limited. Among Republicans and inde­pen­dents, we observe sub­stan­tial­ly greater con­fi­dence among the CivicPulse sample of public pol­i­cy­mak­ers than among the public, but this dif­fer­ence is measured versus a baseline of low con­fi­dence. For instance, 42% of Republican public pol­i­cy­mak­ers express con­fi­dence in the integrity of the national-level results compared to only 22% of Republicans in the public.

Trump’s liability and accountability

Our public survey also asked if the former President should be dis­qual­i­fied from holding office and whether he committed any crimes before or while holding office. The figure below plots these results, which were collected imme­di­ate­ly prior to Trump’s second Senate impeach­ment trial, by respon­dent party affil­i­a­tion. For both dis­qual­i­fi­ca­tion and crim­i­nal­i­ty, responses from Democrats and Republicans are dia­met­ri­cal­ly opposed. Virtually all Democrats support dis­qual­i­fi­ca­tion and think Trump committed crimes before or while in office (94% and 93%, respec­tive­ly) whereas only 14% and 16% of Republicans, respec­tive­ly, agreed. Among the 93% of Democrats who believe Trump to be crim­i­nal­ly liable, 95% want him pros­e­cut­ed (thus, 88% of Democrats overall). By contrast, opinion is divided about whether Trump should be crim­i­nal­ly pros­e­cut­ed among the few Republicans who think he committed crimes: 52% (only 8% of Republicans overall) say that he should be held account­able whereas another 36% (or 6% overall) believe the country should move on. Independents — who comprise 21% of our sample — are split between the partisan groups, with 49% favoring Trump’s dis­qual­i­fi­ca­tion and 50% saying Trump committed crimes. Among the latter group, the vast majority (78% — or 39% of all inde­pen­dents) want Trump prosecuted.

Candidate choice experiment

These results reveal profound partisan divides over the election and the forms of account­abil­i­ty that Donald Trump should face. However, they do not reveal the extent to which these con­sid­er­a­tions factor into the choices voters make among can­di­dates. In par­tic­u­lar, they do not indicate whether Republicans who break from the majority of their party will face pun­ish­ment for upholding the integrity of the election or voting to hold Trump account­able in the impeach­ment process. It is also important to determine the relative impor­tance of these matters versus other concerns (e.g., COVID).

To better under­stand these choices, the public survey included a candidate choice exper­i­ment in which respon­dents were asked to choose between a series of paired hypo­thet­i­cal can­di­dates “in an upcoming election.” Each candidate was defined by a profile con­sist­ing of a set of personal attrib­ut­es and stances, including their name (which signaled both gender and race/ethnicity), their par­ti­san­ship, and their positions on gov­ern­ment spending for COVID relief and trans­porta­tion infra­struc­ture, cer­ti­fi­ca­tion of the 2020 pres­i­den­tial election results, and Trump’s impeachment. 

Because Democrats uniformly support cer­ti­fi­ca­tion and impeach­ment, we only tested support for cer­ti­fi­ca­tion and impeach­ment among hypo­thet­i­cal Republican can­di­dates who might plausibly take different positions. Thus, a respon­dent might be asked to choose between a white Republican woman who favored both spending bills, supported cer­ti­fi­ca­tion, and opposed impeach­ment and a black Republican man with similar positions on the spending bills who opposed both cer­ti­fi­ca­tion and impeach­ment. Each respon­dent chose among three pairs of can­di­dates whose attrib­ut­es (except par­ti­san­ship) were randomly assigned. The design allows us to estimate the effect of each attribute in the can­di­dates’ profiles — for example, their race, their gender, or their stance on impeach­ment — on the prob­a­bil­i­ty that the respon­dent would support him or her. The design also allows us to estimate how much the effect of candidate attrib­ut­es on vote choice differs between different types of respon­dents. For example, we can estimate the dif­fer­ence in effects of sup­port­ing cer­ti­fy­ing the election results between Democrats and Republicans.

The figure below shows the marginal effect of a candidate’s race/ethnicity (Black, Latino, or white), gender (woman or man), and policy positions (sup­port­ing COVID relief, trans­porta­tion spending, election cer­ti­fi­ca­tion, and impeach­ment) on the like­li­hood of pre­fer­ring one Republican candidate over another for Democrats, Republicans, and inde­pen­dents.3 The estimate for each item, which is known as the Average Marginal Component Effect, can be inter­pret­ed as the change in like­li­hood of voting for that candidate, all else equal, if the candidate moves from the baseline category to the category listed in the figure. Thus, the estimates for Black show the dif­fer­ence in the prob­a­bil­i­ty of sup­port­ing a candidate who is Black rather than one who is white. Most impor­tant­ly, the estimates for policy position show the effect of shifting from oppo­si­tion to support on each item — e.g., from opposing impeach­ment to sup­port­ing it.

Starting with demo­graph­ics, we can see that Democrats are about 4 per­cent­age points more likely to favor a Black candidate over a white candidate, other things equal, whereas Republicans are about 3 points less likely to select a Black candidate (we find no mea­sur­able effect for inde­pen­dents). The effects are similar for Latinos relative to whites though smaller in magnitude (+2 points for Democrats and ‑1 point for Republicans). On gender, we find Democrats slightly favor female can­di­dates (2 points) while Republicans and inde­pen­dents slightly favor males (-1 and ‑2 points, respectively). 

Policy effects dwarf demo­graph­ic ones. Spending on COVID relief attracts strong bipar­ti­san support. Republicans favor a candidate who supports a $500 billion appro­pri­a­tion over one who opposes it by 11 points, inde­pen­dents by 12 points, and Democrats by 18 points. Infrastructure spending is less con­sen­su­al but it is not polarized. Democrats favor can­di­dates who support infra­struc­ture spending (11 points). Republicans and inde­pen­dents are indif­fer­ent between a candidate who supports infra­struc­ture spending and one who does not.

In sharp contrast to these rel­a­tive­ly bipar­ti­san responses, positions on Trump drive huge partisan polar­iza­tion. Democrats are vastly (29 points) more likely to support can­di­dates who affirm cer­ti­fi­ca­tion than one who opposed it. Republicans, by contrast, favor can­di­dates who oppose cer­ti­fi­ca­tion over those who support it by 11 points. Backing impeach­ment increases support among Democrats by 20 per­cent­age points, but decreases support among Republicans by 30 points. These penalties under­score the political risks for Republicans in Congress who buck their party base and express oppo­si­tion to Trump’s challenge to a demo­c­ra­t­ic election.

The responses of inde­pen­dents suggest ambiva­lence on these crucial items. They lean toward can­di­dates who supported cer­ti­fi­ca­tion (by 2 points, although the effect does not reach sta­tis­ti­cal sig­nif­i­cance) but disfavor those who supported impeach­ment (-7 points). In its intended voting behavior, at least, this pivotal con­stituen­cy appears not to have embraced the narrative that the election was stolen, but to endorse the idea that the country should move on rather than use impeach­ment to hold Trump account­able for the January 6 attack on the Capitol.4


We asked respon­dents, for the first time, their views about a scenario in which the United States would break up into more than one country. Secession is a genuinely radical propo­si­tion. Until recently, we would have regarded it as too marginal to include in a survey. But state leg­is­la­tors in Mississippi and Texas and state GOP leaders in Texas and Wyoming have openly advocated secession in recent months, prompting us to design two survey items to gauge per­cep­tions of this idea.5 We caution that these survey items reflect initial reactions by respon­dents about an issue that they are very unlikely to have con­sid­ered carefully.

Our first question is generic and framed in terms of reducing conflict:

Some people say the divisions within our country have grown so deep that we would be better off dividing into more like-minded regions that would govern them­selves sep­a­rate­ly. Do you support or oppose the idea of the United States dividing into more than one nation?” 

Overall, 29% of respon­dents supported the dis­so­lu­tion of the country into like-minded regions (10% strongly, 19% somewhat). Support was highest among Republicans (35%) and inde­pen­dents (37%) compared to only 21% among Democrats. Though most Americans reject the prospect of secession, these results at least suggest that a sub­stan­tial minority of people do not instinc­tive­ly reject the idea.

The results above may reflect the wording of the question, which frames dis­so­lu­tion as a way to mitigate conflict. We therefore also asked about secession in a more concrete manner that outlined the geo­graph­i­cal and political reality respon­dents would face as a result of secession: 

Would you support or oppose [your state] seceding from the United States to join a new union with [list of states in new union]?” 

We con­struct­ed five prospec­tive new unions and inserted the relevant states for respon­dents into the question wording above. For example, a par­tic­i­pant from California in our survey would be asked about joining a new union along with Washington, Oregon, Hawaii and Alaska. These sets are provided below:

  • Pacific: California, Washington, Oregon, Hawaii, and Alaska
  • Mountain: Idaho, Montana, Wyoming, Utah, Colorado, Nevada, Arizona, and New Mexico
  • South: Texas, Oklahoma, Arkansas, Louisiana, Mississippi, Alabama, Georgia, Florida, South Carolina, North Carolina, Virginia, Kentucky, and Tennessee
  • Heartland: Michigan, Ohio, West Virginia, Illinois, Indiana, Minnesota, Wisconsin, Iowa, Missouri, North Dakota, South Dakota, Kansas, and Nebraska
  • Northeast: Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, Connecticut, New York, New Jersey, Pennsylvania, Maryland, Delaware, and the District of Columbia

Support for secession under the specific hypo­thet­i­cal unions format is illus­trat­ed in the map below.6 The support levels are, again, non-trivial — as high overall as in the like-minded regions question. The three regions with coast­lines exhibit the most openness to splitting from the union, with the South and West at 33% overall and the Northeast just behind at 32%. The land­locked Heartland and Mountain regions come in at 24% and 28%, respectively.

Support also cor­re­sponds with regional partisan context. In the Pacific and Northeast regions, both of which are deep blue and could be expected to be dominated by the Democratic Party (or its post-secession descen­dants), Democrats favor secession most, followed by inde­pen­dents and Republicans. In the deep red Mountain and Southern regions, that pattern is reversed with Republicans most amenable to secession. In the Heartland, a col­lec­tion of mostly red states that also includes purple Michigan, Minnesota, and Wisconsin, inde­pen­dents are the group most inclined toward secession. 

The unwill­ing­ness of respon­dents to reject secession outright is wide­spread and context-dependent. Republicans express greater support for secession overall than Democrats, but Democrats are more amenable to secession than are Republicans in regions they dominate. 

Democracy versus alternatives

We asked our public sample about their support for four broad types of political rule, which we summarize as:

  • expert rule (“Having experts, not the gov­ern­ment, make decisions according to what they think is best for the country”), 
  • strongman (“Having a strong leader who does not have to bother with Congress and elections”), 
  • military (“Having the army rule”), or
  • democracy (“A demo­c­ra­t­ic political system”).

The figure below shows the per­cent­ages of each partisan group regarding each regime type as “fairly” or “very good” rather than “fairly” or “very bad.” Among Democrats, Republicans, and inde­pen­dents alike, democracy is the most favored regime, though Republicans and inde­pen­dents demon­strate less uniform support than do Democrats (81% and 82% versus 94%, respec­tive­ly). Relatively small shares of each group favor strongman rule and even smaller numbers would welcome military rule — and impor­tant­ly, the dif­fer­ences across partisan groups on these options are sta­tis­ti­cal­ly insignif­i­cant. Strikingly, Democrats (63%) find rule by experts far more palatable than do Republicans (36%), a finding that may reflect past demands by their party to give greater priority to experts in the response to the COVID-19 pandemic. On the whole, although Republicans show slightly less com­mit­ment to democracy than Democrats, they show no greater affinity for non-demo­c­ra­t­ic options and less appetite for rule by experts.

Expert survey

Reform proposals

We asked our experts to assess 16 prominent reform proposals that aim to improve the quality of American democracy. Many of the proposals are drawn from the Our Common Purpose project that was coor­di­nat­ed by the American Academy of Arts and Sciences. Each expert par­tic­i­pant was asked to rate eight randomly drawn proposals. The full set of state­ments describ­ing each reform is in the Appendix. The figure below shows expert support and oppo­si­tion for each proposal.

Experts over­whelm­ing­ly support the proposed reforms. Of the 16 we tested, majori­ties of our expert respon­dents strongly supported 9 and strongly or mod­er­ate­ly supported 15. The only proposal that did not garner majority support was com­pul­so­ry voting.

The proposals fall into distinct cat­e­gories. The largest group aims to increase voter par­tic­i­pa­tion, par­tic­u­lar­ly among tra­di­tion­al­ly mar­gin­al­ized groups. Increasing flex­i­bil­i­ty on when and how ballots can be cast (95% support), guar­an­tee­ing suffrage rights to ex-felons (91%), same-day reg­is­tra­tion (91%), moving Election Day to a national holiday (87%), and pre-reg­is­ter­ing young voters (85%) all attracted support from more than four in five experts. By contrast, com­pul­so­ry voting was supported by only 29% of experts, perhaps reflect­ing dis­com­fort with the manner by which it tries to increase par­tic­i­pa­tion as well as recent research on the range of unin­tend­ed con­se­quences it can generate, including dis­il­lu­sion­ment with democracy itself.

Another group of proposals seeks to reduce the influence of large indi­vid­ual and corporate campaign donors in American elections. These include increased trans­paren­cy on the source of donations, providing public campaign funding, and amending the Constitution to impose greater restric­tions on private spending. All these garner strong support among the experts (98%, 87%, and 88%, respectively).

A third set of proposals focuses on the rules for con­vert­ing voter support into rep­re­sen­ta­tion. At the top of this list is requiring states to establish non-partisan redis­trict­ing com­mis­sions to reduce partisan ger­ry­man­der­ing (95% support). Next, at 84%, is support for switching to a system in which the president is elected by direct popular vote instead of by the Electoral College.7 Also in this category are two electoral reforms, ranked-choice voting (78% support) and elim­i­nat­ing the require­ment for Members of Congress to be elected from single-member districts (73%), both of which aim to open paths to electoral success for can­di­dates other than those who can prevail in either Democratic or Republican primary contests. 

In turn, enlarging the House of Representatives (64%) would increase the ratio of rep­re­sen­ta­tives to citizens, allowing for a more fine-grained mapping of rep­re­sen­ta­tives’ char­ac­ter­is­tics onto con­stituent preferences.

The last two proposals focus on the conduct of gov­er­nance rather than elections. The first would limit the period for which federal judges could serve on the Supreme Court to 18 years (77% support), guar­an­tee­ing a vacancy on the Court every two years. This proposal seeks to reduce the stakes for high court appoint­ments and thereby cool the attendant politics both during elections and in the day-to-day operation of the Senate. The next would eliminate the 60-vote require­ment to suspend debate in the Senate, elim­i­nat­ing the fil­i­buster and effec­tive­ly returning the chamber to majority rule (74%). 

Democratic nightmares (which came true, which did not)

In our November 2020 survey, the experts offered prob­a­bilis­tic forecasts about a set of potential events in the weeks following the election through inau­gu­ra­tion. With the new admin­is­tra­tion in place, we can now assess the accuracy of these forecasts. On the whole, the experts did pretty well. Most of the events they rated as low prob­a­bil­i­ty did not take place:

  • President Trump would be sworn in for 2nd term (2%);
  • A faithless elector (5%) or the House of Representatives (7%) would decide the electoral outcome; 
  • At some point during the tran­si­tion period the country would have an acting president under 20th Amendment (10%);
  • At least 1 state would abandon the popular vote as its mode of deter­min­ing electors (10%), would send competing slates of electors to Washington (15%), or would miss the December 14 safe-harbor deadline for deter­min­ing its slate of electors (25%);
  • The Vice-President and Speaker of the House would dispute accep­tance of Electoral College votes (20%) or President Trump would resign before January 20 and be pardoned by his successor, President Pence (27%).

At the other end of the prog­nos­ti­ca­tion spectrum, the experts also assigned high like­li­hood to a number of events that did come to pass. President Trump did not attend Joe Biden’s inau­gu­ra­tion (90%) and refused to  concede that he had lost the election (84%). Trump also commuted the sentences of his political allies Mike Flynn and Paul Manafort (80%), blocked funding for the Biden tran­si­tion (80%), and refused to cooperate with the Biden tran­si­tion team after the Electoral College vote (78%).

But the experts’ pre­dic­tions were far from perfect. They rated a number of items as likely that did not occur, par­tic­u­lar­ly with regard to President Trump’s use of the pardon authority. The experts expected pre­emp­tive pardons for Rudolph Giuliani (80%), close Trump family members (74%), and Trump himself (56%). In the aftermath of the November election, they also expected President Trump to fire the FBI director before he left office (70%) and to destroy incrim­i­nat­ing documents (82%). None of these events came to pass (to the best of our knowledge, in the case of document destruc­tion). The reason may be that advisers persuaded Trump that such moves would backfire.

Some forecasts were ambiguous. Thirty-five percent of experts said one or more legal disputes over the election would be decided by the U.S. Supreme Court. In fact, the Court, of course, declined to hear cases brought by Trump with alle­ga­tions of electoral malfea­sance. Though it did not issue a decision, the Court’s (non-)action was critical to the election outcome. Nor did President Trump meet the forecast threshold of firing more than 100 civil servants between the election and his departure from office (60%). But he did act in late October to eliminate job pro­tec­tions for thousands of civil servants and tried to shift thousands of his own political appointees into civil service status, to protect them from being replaced by the Biden admin­is­tra­tion. Another forecast had to do with whether President Trump would attempt to negotiate legal amnesty for himself and his family after the election (70%). He did not do so publicly, but insider reports indicated that explored this possibility.

We close with dis­cus­sion of an item that warrants con­sid­er­a­tion less for its match to expert forecasts than for its exis­ten­tial impor­tance to American democracy — whether Republican Senate leader Mitch McConnell and House Minority Leader Kevin McCarthy would fail to acknowl­edge Joe Biden as the legit­i­mate winner of the election even after Biden was inau­gu­rat­ed in January 2021. Our experts rated this extreme event as rel­a­tive­ly unlikely (20%) and it did not ulti­mate­ly come to pass. Though McConnell and McCarthy failed to actively oppose efforts to overturn the election results, they ulti­mate­ly met the deadline, attending President Biden’s inau­gu­ra­tion and acknowl­edg­ing him as the legit­i­mate president since. 

Political events

We then asked experts to rate the impor­tance and abnor­mal­i­ty of 21 events that took place since our previous survey in November 2020. The impor­tance of an event was judged on a 5‑point scale from “Unimportant” to “Important” and its normalcy was scored on another 5‑point scale from “Normal” to “Abnormal.” The figure below visu­al­izes the rela­tion­ship between these variables. Consistent with patterns over the course of Trump’s pres­i­den­cy, many events of the past couple of months occupy the important and abnormal quadrant (top-right corner) of the graph. The two events of Biden’s pres­i­den­cy included in our survey, “Biden signs 17 executive orders imme­di­ate­ly after taking office to reverse Trump policies” and “Biden requests waiver allowing recently retired Army general to serve as Defense Secretary,”  were judged to be rel­a­tive­ly normal occur­rences in a pres­i­den­tial administration.

The events that our experts deemed most important and most abnormal are in the grey square in the figure. Of these events, pro-Trump insur­rec­tion­ists’ invasion of the U.S. Capitol out­stripped all others in its perceived abnor­mal­i­ty and impor­tance. Not far behind is a cluster of events, all of them related to attempts to overturn the election: Trump pres­sur­ing Georgia officials to “find” enough votes to overturn state election results, state attorneys general request­ing that the Supreme Court inval­i­date election results in Pennsylvania, Wisconsin, Michigan, and Georgia, and 147 Republicans in Congress sup­port­ing objec­tions to the cer­ti­fi­ca­tion of President Biden’s election. 

Threats to democracy

To further unpack the sig­nif­i­cance of these events, we asked experts to rate the severity of the threat they posed to democracy. Unsurprisingly, more than 90% of experts viewed the items that scored highest across the (ab)normality-and-importance dimen­sions as either a moderate, serious, or grave threat.

One item that our experts rated as abnormal was also one that few viewed as posing much of a threat to democracy: the House impeach­ing President Donald Trump for a second time. In total, 93% of our academic experts rated the two impeach­ments them­selves as pre­sent­ing little or no threat to U.S. democracy (7% sensed a moderate threat; none said it was a serious or grave threat). In the wake of Trump’s second acquittal, some prominent Republicans have voiced a different per­spec­tive, pre­dict­ing that the impeach­ments would create a spiral of partisan ret­ri­bu­tion. For instance, Senator Lindsey Graham suggested that Vice-President Kamala Harris could be impeached if Republicans retake the House of Representatives for having expressed support for Black Lives Matter pro­test­ers in summer 2020.

Public and experts on the quality of American democracy

We asked both expert and public respon­dents to assess the quality of U.S. democracy overall. The figure below shows mean eval­u­a­tions of U.S. democracy on a 0–100 scale among experts (green line) and the public (purple line). It also sep­a­rate­ly plots responses among the public from Trump approvers (red line) and dis­ap­provers (blue line). Our October 2020 survey, conducted just before the election, marked a turning point. Across earlier surveys, ratings among our experts had drifted generally downward, from a mean of 69 in our first survey (February 2017) to 61 before the pres­i­den­tial election (October 2020). During that period, Trump approvers in the public rated demo­c­ra­t­ic per­for­mance sub­stan­tial­ly higher (around 60) than did Trump dis­ap­provers (around 50). In our November 2020 and January-February 2021 surveys, ratings of overall demo­c­ra­t­ic quality rose sharply among both our experts (from 61 to 64 and then to 66) and among members of the public who dis­ap­prove of Donald Trump (from 49 to 53 and then to 58). By contrast, ratings among Trump approvers dropped (from 60 to 55 and then to 47). These counter-trends among Trump approvers and dis­ap­provers effec­tive­ly cancel out; the overall rating among the public sample thus remains level at 54. 

We also asked the experts and the public to evaluate each of thirty distinct demo­c­ra­t­ic prin­ci­ples, on their impor­tance to democracy, and on how well the U.S. was per­form­ing on them. The Appendix provides the complete list of 30 prin­ci­ples as well as figures illus­trat­ing the assess­ments among the expert and public samples of the impor­tance of each to democracy and of per­for­mance. In those graphs, the public’s ratings are again presented sep­a­rate­ly for Trump sup­port­ers and opponents. Here, we note some items that show sub­stan­tial changes over time. We last measured each item’s impor­tance to democracy among the public in March 2019. Between the two surveys, overall estimates of impor­tance have remained largely stable. But among Trump approvers, per­cep­tions of the impor­tance of three prin­ci­ples has risen markedly:8

  • that elections should be free from foreign influence (+14 points), 
  • that political can­di­dates disclose infor­ma­tion about how they would govern (+11.8 points), and 
  • that the judiciary can limit executive power (+14.2 points). 

Among Trump dis­ap­provers, only one item rose sub­stan­tial­ly in impor­tance: that par­tic­i­pa­tion in elections is high (+12.7 points). Three dimen­sions expe­ri­enced a marked decline among Trump disapprovers:

  • that all parties are allowed to compete in elections (-11 points), 
  • that the gov­ern­ment protects freedom of speech (-10.2 points), and 
  • that gov­ern­ment agencies are not used to punish political opponents (-11.5 points). 

Political sci­en­tists have doc­u­ment­ed that people’s views about the health of the U.S. economy is highly sensitive to their par­ti­san­ship. In much the same way, the public’s views about democracy appear heavily influ­enced by whether one’s party wins or loses.  Contrary trends among Trump sup­port­ers and opponents reflect their reversal of fortunes in the November election and the transfer of executive power that followed. Trump opponents, whose narrow victory depended on record turnout, came to an increased appre­ci­a­tion for electoral par­tic­i­pa­tion. But their concern for pro­tec­tions for gov­ern­ment opponents ebbed. Trump sup­port­ers, by contrast, gained new appre­ci­a­tion for the judiciary’s ability to check the executive and a new suspicion of any foreign influence over U.S. elections.


Bright Line Watch surveys on the state of America’s democracy, January-February 2021

From January 28-February 8, 2021, Bright Line Watch conducted its four­teenth survey of academic experts, and eleventh of the general public, on the quality of democracy in the United States. Our public sample consisted of 2,700 survey par­tic­i­pants from the YouGov sample who were selected and weighted to be rep­re­sen­ta­tive of the U.S. adult pop­u­la­tion. We also surveyed 527 political science experts across a diverse range of subfields (5% of solicited invi­ta­tions). 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 tra­di­tion­al­ly unweight­ed because we do not collect demo­graph­ic data to protect anonymity. Error bars in our graphs represent 95% con­fi­dence intervals. Data are available here.

Both the expert and public samples in Wave 14 responded to a battery of questions about demo­c­ra­t­ic per­for­mance and impor­tance in the United States. Afterward, they were asked to evaluate the quality of American democracy overall on a 100-point scale. Experts were also asked to evaluate the quality of democracy in their state on the same 0–100 scale. Expert respon­dents were then asked to respond to three addi­tion­al batteries: one in which they were presented with a list of political events and asked to rate them on normalcy and impor­tance, another in which they were asked to evaluate proposals to improve the quality of American democracy, and a final set of questions about the degree to which select events posed a threat to democracy. 

16 reform proposals

  • Enlarge House of Representatives
    Substantially enlarge the House of Representatives through federal leg­is­la­tion to make it and the Electoral College more rep­re­sen­ta­tive of the nation’s population
  • Ranked-choice voting
    Introduce ranked-choice voting in pres­i­den­tial, con­gres­sion­al, and state elections.
  • Multi-member districts
    Repeal the 1967 law that mandates single-member districts for the House so that states have the option to use multi-member districts on the condition that they adopt a non-winner-take-all election model.
  • Redistricting com­mis­sions
    Federal leg­is­la­tion requiring fair con­gres­sion­al districts to be deter­mined by state-estab­lished inde­pen­dent citizen-redis­trict­ing commissions.
  • Campaign finance reg­u­la­tion
    Amend the Constitution to authorize the reg­u­la­tion of election con­tri­bu­tions and spending to limit the  undue influence of money in our political system.
  • Campaign finance trans­paren­cy
    Pass strong campaign-finance dis­clo­sure laws in all fifty states that require full trans­paren­cy for campaign donations, including from 501(c)(4) orga­ni­za­tions and LLCs.
  • Public funding of campaigns
    Pass “clean election laws” for federal, state, and local elections through mech­a­nisms such as public matching donation systems and democracy vouchers, which amplify the power of small donors.
  • 18-year Supreme Court terms
    Establish, through federal leg­is­la­tion, eighteen-year terms for Supreme Court justices with appoint­ments staggered such that one nom­i­na­tion comes up during each term of Congress.
  • Flexibility on when and how to vote
    Give people more choices about where and when they vote, with state-level leg­is­la­tion in all states that supports the imple­men­ta­tion of vote centers and early voting.
  • Election day on national holiday
    Change federal election day to Veterans Day to honor the service of veterans and the sac­ri­fices they have made in defense of our con­sti­tu­tion­al democracy, and to ensure that voting can occur on a day that many people have off from work.
  • Same-day voter reg­is­tra­tion
    Establish, through state and federal leg­is­la­tion, same-day reg­is­tra­tion and universal automatic voter registration.
  • Pre-register young voters
    Establish, through state leg­is­la­tion, the pre­reg­is­tra­tion of sixteen- and seventeen-year-olds and provide edu­ca­tion­al oppor­tu­ni­ties for them to practice voting as part of the pre-reg­is­tra­tion process.
  • Compulsory voting
    Establish, through leg­is­la­tion, voting in federal elections to be a require­ment of cit­i­zen­ship. All eligible voters would have to par­tic­i­pate or submit a valid reason for nonparticipation.
  • Suffrage for ex-felons
    Restore federal and state voting rights to citizens with felony con­vic­tions imme­di­ate­ly and auto­mat­i­cal­ly upon their release from prison.
  • National popular vote
    Change to a system in which the president is elected by direct popular vote, instead of by the Electoral College.
  • Eliminate fil­i­buster
    Change US Senate rules to eliminate the fil­i­buster, allowing a simple majority of senators to bring any proposal to the floor for a vote.


  • Trump pressures Georgia official to “find” enough votes to overturn state election results
  • Congress and Senate override Trump veto of military spending bill
  • All 10 living former defense sec­re­taries say U.S. military has no role in election dispute
  • Pro-Trump insur­rec­tion­ists storm the Capitol
  • 147 Republicans in Congress support objec­tions to the cer­ti­fi­ca­tion of President Biden’s election
  • White House forces Georgia U.S. Attorney to resign for not inves­ti­gat­ing election fraud claims
  • Trump issues Thanksgiving turkey pardon
  • Trump signs $900 billion stimulus and gov­ern­ment spending bill
  • Trump pardons Roger Stone, Paul Manafort, Steve Bannon, and Charles Kushner
  • Trump contacts 31 GOP officials in bat­tle­ground states to try to overturn election
  • State attorneys general request SCOTUS inval­i­date election results in PA, WI, MI, and GA
  • Biden requests waiver allowing recently retired Army general to serve as Defense Secretary
  • Cabinet sec­re­taries for Education, Transportation, DHS resign following riot at Capitol
  • Trump does not attend Biden inauguration
  • Trump admin­is­tra­tion brokers deal to normalize relations between Morocco and Israel
  • GOP con­gress­men reports col­leagues didn’t vote to impeach Trump out of fear for family safety
  • House of Representatives impeaches Trump for the second time
  • Trump autho­rizes con­struc­tion of women’s suffrage monument in D.C.
  • Trump races to strip job pro­tec­tions from budget analysts before transition
  • Biden signs 17 executive orders imme­di­ate­ly after taking office to reverse Trump policies
  • Pelosi talks to Joint Chiefs about blocking Trump from accessing nuclear codes
  • Top military leaders condemn violent invasion of the Capitol in joint statement
  • My Pillow founder encour­ages Trump to declare martial law during a White House visit
  • Social media companies deplat­form President Trump

Threats to democracy

  • President Trump declaring that he had won the election before the race had been called. 
  • President Trump refusing to concede defeat after being declared loser of the election.
  • President Trump refusing to commit before the election to a peaceful tran­si­tion of power. 
  • President Trump pres­sur­ing Ukraine to inves­ti­gate Joe Biden.
  • President Trump calling the press an “enemy of the people.”
  • The Commission on Presidential Debates can­celling the October 15 debate.
  • The appoint­ment of Amy Coney Barrett to the Supreme Court.
  • Pro-Trump pro­tes­tors storming the U.S. Capitol.
  • Trump pres­sur­ing Georgia’s Secretary of State to “find” enough votes to overturn the result in that state.
  • A majority of Republicans in Congress voting against the cer­ti­fi­ca­tion of Joe Biden’s win in the Electoral College.
  • The House of Representatives impeach­ing President Trump for a second time.
  • Violent protests in Portland, Oregon and the creation of the Capitol Hill Autonomous Zone in Seattle, Washington.
  • Media figures calling for President Trump’s removal from office.
  • The Supreme Court dis­miss­ing President Trump’s legal chal­lenges to the election results.
  • President Trump pardoning Roger Stone, Paul Manafort, and Steve Bannon.

Democratic prin­ci­ples

The foun­da­tion of Bright Line Watch’s surveys is a list of 30 state­ments express­ing a range of demo­c­ra­t­ic prin­ci­ples (the full list is provided below). Democracy is a mul­ti­di­men­sion­al concept. Our goal is to provide a detailed set of measures of demo­c­ra­t­ic values and of the quality of American democracy. We are also inter­est­ed in the resilience of democracy and the nature of potential threats it faces. Based on the expe­ri­ences of other countries that have expe­ri­enced demo­c­ra­t­ic setbacks, we recognize that demo­c­ra­t­ic erosion is not nec­es­sar­i­ly an across-the-board phe­nom­e­non. Some facets of democracy may be under­mined first while others remain intact, at least initially. The range of prin­ci­ples that we measure allows us to focus attention on variation in specific insti­tu­tions and practices that, in com­bi­na­tion, shape the overall per­for­mance of our democracy.

Bright Line Watch’s Wave 1 survey included 19 state­ments of demo­c­ra­t­ic prin­ci­ples. Based on feedback from respon­dents and con­sul­ta­tion with col­leagues, we expanded that list to 29 state­ments in Wave 2. We then reduced that set to a set of 27 state­ments for the Wave 3 through Wave 8 surveys. 17 of those 27 state­ments were included in Wave 1, and all 27 were included in Wave 2. We added one statement to the list in Wave 9. In Wave 12, we began asking respon­dents to assess two addi­tion­al principles: 

  • The law is enforced equally for all persons
  • Politicians who lose free and fair elections will concede defeat

The full set of state­ments is presented below and grouped the­mat­i­cal­ly for clarity. In the surveys, the prin­ci­ples were not cat­e­go­rized or labeled. Each respon­dent was shown a randomly selected subset of state­ments and asked to rate the per­for­mance of the United States on those dimen­sions. For the per­for­mance questions, both the public and experts were asked to rate 9 state­ments. In the impor­tance battery, respon­dents were asked to rate a random subset of 4 state­ments, up-weighting the 2 state­ments that were intro­duced in Wave 12 so that more respon­dents could rate the new statements. 


  • Elections are conducted, ballots counted, and winners deter­mined without pervasive fraud or manipulation
  • Citizens have access to infor­ma­tion about can­di­dates that is relevant to how they would govern
  • The geo­graph­ic bound­aries of electoral districts do not sys­tem­at­i­cal­ly advantage any par­tic­u­lar political party
  • Information about the sources of campaign funding is available to the public
  • Public policy is not deter­mined by large campaign contributions
  • Elections are free from foreign influence
  • Politicians who lose free and fair elections will concede defeat


  • All adult citizens have equal oppor­tu­ni­ty to vote
  • All votes have equal impact on election outcomes
  • Voter par­tic­i­pa­tion in elections is generally high


  • All adult citizens enjoy the same legal and political rights
  • Parties and can­di­dates are not barred due to their political beliefs and ideologies
  • Government protects indi­vid­u­als’ right to engage in unpopular speech or expression
  • Government protects indi­vid­u­als’ right to engage in peaceful protest
  • Citizens can make their opinions heard in open debate about policies that are under consideration
  • The law is enforced equally for all persons


  • Government does not interfere with jour­nal­ists or news organizations
  • Government effec­tive­ly prevents private actors from engaging in polit­i­cal­ly-motivated violence or intimidation
  • Government agencies are not used to monitor, attack, or punish political opponents


  • Government officials are legally sanc­tioned for misconduct
  • Government officials do not use public office for private gain
  • 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


  • 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


  • 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
  • Political com­pe­ti­tion occurs without criticism of opponents’ loyalty or patriotism

To measure perceived demo­c­ra­t­ic per­for­mance, the survey asked, “How well do the following state­ments describe the United States as of today?” Each respon­dent was then presented with 9 state­ments of principle, randomly drawn from the set above, and offered the following response options:

  • The U.S. does not meet this standard
  • The U.S. partly meets this standard
  • The U.S. mostly meets this standard
  • The U.S. fully meets this standard
  • Not sure

To measure the perceived impor­tance of each principle, the survey asked, “How important are these char­ac­ter­is­tics for demo­c­ra­t­ic gov­ern­ment?” Each respon­dent was then presented with 4 state­ments, randomly drawn from the set above, and offered the following response options:

  • Not relevant: this has no impact on democracy
  • Beneficial: this enhances democracy but not required for democracy
  • Important: if this is absent, democracy is compromised
  • Essential: A country cannot be con­sid­ered demo­c­ra­t­ic without this

Additional figures

Election confidence

Note on con­fi­dence in vote counts over time. In our October, November, and January/February surveys, Bright Line Watch included asked members of the public sample about their con­fi­dence that “… votes [in your county/statewide/nationwide] were cast as voters intended?” In the January/February wave, we also coor­di­nat­ed with CivicPulse to ask a similar question — us to our public sample, CivicPulse to a sample of 212 local and state-level public pol­i­cy­mak­ers — although the wording differed slightly, asking about con­fi­dence that (“… votes for the pres­i­den­tial election were counted correctly [in your county/statewide/nationwide]?” Thus, the CivicPulse wording specif­i­cal­ly referred to the pres­i­den­tial election, and it stip­u­lat­ed “correctly” rather than “as voters intended.”

In the January/February survey, Bright Line Watch split its sample, asking each variant to half our par­tic­i­pants. If response patterns were identical in each variant, it would have been possible to combine the data but, as it happens, the slight dif­fer­ences in wording appear to have affected responses. Among those who approve of President Trump, con­fi­dence was lower in the variant that referred to the pres­i­den­tial election. In the main report, we presented results from that variant, which allowed com­par­i­son between our public sample and the CivicPulse sample of elected officials. The full dis­tri­b­u­tions among each group are shown in the figure below.

The next figure presents results from the variant that Bright Line Watch used in October, November, and January/February, which allows over-time com­par­i­son of responses using con­sis­tent wording. The time series shows that, from November, just after the election, to January/February, there was little change, among either Trump approvers or dis­ap­provers, in con­fi­dence in the vote counts at the state or national level. Approvers and dis­ap­provers had about identical levels of con­fi­dence on the eve of the election in October, then diverged sharply in November, and have remained separated by a wide gulf, par­tic­u­lar­ly regarding con­fi­dence in the national level count.

Rightful winner, legitimate results, support for the EC

Conjoint experiment by support for Trump

Political violence

Support for secession by Trump approval

Importance battery

Performance battery


  1. We instead asked half the par­tic­i­pants in our public sample, “How confident are you that votes nation­wide were counted as voters intended?” to cor­re­spond to the wording used in our October and November surveys. The dis­tri­b­u­tions of responses between people asked this version of the question and those asked specif­i­cal­ly about the 2020 pres­i­den­tial election using the wording above were generally similar, but Trump sup­port­ers reported less con­fi­dence when asked specif­i­cal­ly about the recent pres­i­den­tial election. The Appendix provides figures illus­trat­ing the dif­fer­ences between versions of the question and time series data for the version we used in our three most recent surveys.
  2. You can direct questions about the CivicPulse local pol­i­cy­mak­er sample to Nathan Lee, Managing Director, at, or visit
  3. Due to a coding error, the con­fi­dence intervals around the estimates in the figure initially released with this reports, in February 2021, were too small. (The point estimates were accurate.) The con­fi­dence intervals, and cor­re­spond­ing inter­pre­ta­tions in the text of the report, have been corrected in this current version [July 11, 2021].
  4. Recall that our survey closed the day before Trump’s second impeach­ment trial began in the Senate, so our results do not reflect exposure to any infor­ma­tion the trial revealed.
  5. Rush Limbaugh also briefly advocated secession before beating a hasty retreat from the position.
  6. For both the generic and the concrete secession questions, the response options were: Strongly support, Somewhat support, Somewhat oppose, Strongly oppose. For sim­plic­i­ty, we group responses in the map by support versus oppo­si­tion. The full data are available here.
  7. We asked an analogous question to our public sample and found that 80% of Democrats support moving to a national popular vote, but only 23% of Republicans and 36% of inde­pen­dents do.
  8. We note prin­ci­ples on which the share of respon­dents rating a principle as “essential” or “important” to democracy (as opposed to “not relevant” or “ben­e­fi­cial”) has changed by more than 10 per­cent­age points.