Tempered expectations and hardened divisions a year into the Biden presidency
Bright Line Watch November 2021 surveys

Joe Biden entered the pres­i­den­cy in January aiming to strength­en American democracy by deliv­er­ing popular policies and reducing partisan enmity. A year later, he convened a democracy summit to address the chal­lenges that rising author­i­tar­i­an­ism poses around the world. Here at home, however, the promise of unity has faded and our demo­c­ra­t­ic vul­ner­a­bil­i­ties remain.

In this report, we describe the findings from parallel surveys we conducted November 5–19, 2021 among political science experts and the public to gauge the status of our democracy and the prospects for reforms that might improve it.

Our key findings are:

  • Partisan divisions over the legitimacy of the 2020 election remain profound. Confidence in the 2022 elections is already more polarized than confidence in the 2020 election was in October of that year.

  • Democrats underestimate the commitment of Republican supporters to democratic norms and principles, and Republicans underestimate the commitment of Democrats. Correcting these misperceptions only modestly decreases support for illiberal actions intended to sabotage or undermine the other party.

  • Support for political violence has been overstated in prior surveys due to respondent inattention and questions that required respondents to select a single option disavowing violence. We correct for these issues and find reduced support for violence. Still, millions of Americans explicitly endorse political violence directed against the other party.

  • Experts and the public believe that fundamental changes are needed to make American government work for current times and that what’s most needed are policy and rule changes that don’t require constitutional amendments. 

  • Experts strongly prefer that Senate seats be apportioned to states by population rather than equally. This preference is shared by Democrats but opposed by Republicans and Independents. Overall, the public prefers the status quo.

  • Experts are relatively evenly divided about which of numerous problems facing American democracy is most severe, though they rank economic inequality, unrepresentative political institutions, and racial inequality highest on average.

  • A number of reforms to campaign and legislative rules and practices attract widespread expert support as beneficial to democracy. With few exceptions, however, experts think they are quite unlikely to be enacted.

  • Expert and public perceptions of the performance of U.S. democracy have changed little since June 2021.

Confidence in the 2020 and 2022 elections

We begin by examining public attitudes on a topic that has remained central to American political discourse since 2020: con­fi­dence in elections. In our November 2021 survey, we repeated a question we had asked pre­vi­ous­ly, asking respon­dents whether or not they thought Joe Biden was the rightful winner of the pres­i­den­tial election in November 2020. As the figure below indicates, just 63% of Americans said Biden was def­i­nite­ly or probably the rightful winner — an estimate that is sta­tis­ti­cal­ly indis­tin­guish­able from the 64% we found in our surveys in February and June 2021. Moreover, we continue to see a vast dis­crep­an­cy between the parties. 94% of Democrats say Biden is the rightful winner compared to just 26% of Republicans — a split that has also remained remark­ably stable since Biden took office. Despite the continued lack of evidence to support Donald Trump’s false claims of wide­spread fraud in the 2020 election, the public has not become more accepting of the validity of Biden’s victory over time.

Attacks on the prior election result may be affecting con­fi­dence in next year’s midterm election. Only 62% of Americans said they were very or somewhat confident that votes nation­wide will be counted as voters intend in the November 2022 election. This is slightly higher than the 59% who expressed con­fi­dence in October 2020, weeks before last year’s election, but divisions along partisan lines have deepened. Even though Donald Trump was already fre­quent­ly making false claims of wide­spread fraud back in 2020, our data showed similar levels of con­fi­dence between Democrats (66%) and Republicans (58%). By now, however, a partisan gulf has opened, with 80% of Democrats express­ing con­fi­dence compared to only 42% of Republicans. 

We observe a different pattern with regard to voter access. Overall con­fi­dence that all legally entitled voters will have access to the ballot has declined from a year ago, from 74% to 64%, while the partisan gap on this issue has narrowed slightly. In the run-up to the 2020 election, con­fi­dence in voter access was 18 per­cent­age points higher among Republicans than Democrats (87% to 69%). The gap is now 10 per­cent­age points, with 71% of Republicans (whose party has taken the lead in passing a wave of state-level restric­tions) and 61% of Democrats (whose party has opposed these restric­tions) confident that all eligible voters will be able to suc­cess­ful­ly cast a ballot in the November 2022 election for the U.S. Congress.

Experiment: Out-partisan views and support for anti-demo­c­ra­t­ic practices

Despite the current partisan divide in con­fi­dence in the election system, Americans may over­es­ti­mate the pro­por­tion of their partisan opponents who reject democracy itself. How deep are these mis­per­cep­tions and do they con­tribute to support for anti-demo­c­ra­t­ic practices? If so, what can we do about it?

Scholars have noted that partisans — Democrats and Republicans alike — tend to over­es­ti­mate the extremism of their political adver­saries and that such over­es­ti­ma­tion is asso­ci­at­ed with will­ing­ness to take, or support, extreme action oneself. For example, recent studies show that partisans who under­es­ti­mate their opponents’ support for demo­c­ra­t­ic prin­ci­ples are more likely to support anti-demo­c­ra­t­ic practices and vio­la­tions of demo­c­ra­t­ic norms. Similarly, partisans who over­es­ti­mate rival partisans’ support for violence report greater will­ing­ness to engage in violence. Correspondingly, exper­i­men­tal inter­ven­tions to correct these sorts of mis­per­cep­tions about partisan opponents can reduce expressed support for norm vio­la­tions and violence.

With these findings in mind, we sought to assess public support for practices that run contrary to a broad con­cep­tion of the public interest and can rea­son­ably be described as anti-demo­c­ra­t­ic. We also sought to determine whether providing survey par­tic­i­pants with accurate infor­ma­tion about their partisan opponents would mitigate support for anti-demo­c­ra­t­ic practices. 

Our design is modeled on research by Michael H. Pasek, Lee-Or Ankori-Karlinsky, Alex Levy-Vene, and Samantha L. Moore-Berg. First, a ran­dom­ized subset of our par­tic­i­pants engaged in an exercise to determine, and to correct, cross-party mis­per­cep­tion. We asked them to estimate what per­cent­age of their co-partisans and members of the other party would say that each of the following prin­ci­ples is important or essential to demo­c­ra­t­ic gov­ern­ment:1

  • Government agencies are not used to monitor, attack, or punish political opponents
  • Elections are conducted, ballots counted, and winners deter­mined without pervasive fraud or manipulation
  • Law enforce­ment inves­ti­ga­tions of public officials or their asso­ciates are free from political influence or interference
  • All adult citizens enjoy the same legal and political rights

These respon­dents were then provided with feedback con­trast­ing their own estimates of out-partisan 2 com­mit­ments to demo­c­ra­t­ic prin­ci­ples with actual data from a Bright Line Watch survey earlier this year in the following format:

You guessed that XX% of [out-partisans] say the following principle is important or essential to demo­c­ra­t­ic gov­ern­ment. Actually, our most recent survey found that YY% of [out-partisans] say this principle is important or essential to demo­c­ra­t­ic government.

As prior research suggested, respon­dents fre­quent­ly under­es­ti­mat­ed the com­mit­ments of out-partisans to demo­c­ra­t­ic prin­ci­ples. In total, 65% of respon­dents in the treatment group were informed that they had under­es­ti­mat­ed the demo­c­ra­t­ic com­mit­ments of their partisan adver­saries on all four items. The figure below contrasts the perceived demo­c­ra­t­ic com­mit­ments of Democrats and Republicans offered by co-partisans (round markers) and out-partisans (triangles) with the true values (squares). Blue points represent estimates offered about Democrats; red points represent estimates offered about Republicans. 

Participants from both parties under­es­ti­mate the demo­c­ra­t­ic com­mit­ments of people from both parties but these errors are espe­cial­ly great for partisan opponents. Across items, respon­dents under­es­ti­mat­ed the share of people who supported the impor­tance of demo­c­ra­t­ic prin­ci­ples by 13 per­cent­age points on average for co-partisans and by 36 per­cent­age points on average for oppo­si­tion partisans. For instance, 92% of Democrats said it was important that all adult citizens enjoy the same legal and political rights, but Democrats estimated that only 75% of Democrats would say so on average and Republicans estimated that only 51% would say so. 

After these par­tic­i­pants had been presented with accurate infor­ma­tion about support for demo­c­ra­t­ic values, all were asked about their support for practices that undermine the public good and could damage democracy. Specifically, we asked they agree or disagree that their co-partisans should take the following set of actions (from research by Pasek et al.):

  • should do every­thing they can to hurt the [other party], even if it is at the short-term expense of the country;
  • should do every­thing in their power within the law to make it as difficult as possible for [the other party] to run the gov­ern­ment effectively;
  • should redraw districts to maximize their potential to win more seats in federal elections, even if it may be tech­ni­cal­ly illegal;
  • should use the Federal Communications Commission to heavily restrict or shut down Fox News [shown to Democrats] / MSNBC [shown to Republicans] to stop the spread of fake news
  • it’s OK to sacrifice US economic pros­per­i­ty in the short-term in order to hurt [the other party’s] chances in future elections. 

The figure below shows the per­cent­age of control group respon­dents that agreed with each statement. We restrict these estimates to respon­dents whose prior answers indicate that they were reading questions carefully. (Our process for iden­ti­fy­ing such respon­dents is described below. As we show there, our estimates of support for illiberal or anti-demo­c­ra­t­ic behaviors are otherwise inflated by careless responses from inat­ten­tive respondents.)

First, support for co-partisans taking illiberal actions was dis­cour­ag­ing­ly wide­spread, indi­cat­ing once again that the normative com­mit­ments described in the exper­i­men­tal treatment are not always upheld in practice. Of those who did not receive the inter­ven­tion, 38% of Democrats and 40% of Republicans favor FCC action against dis­fa­vored news media networks. Similarly, 28% of Democrats and 39% of Republicans in the control group support doing every­thing possible to prevent the other party from governing effec­tive­ly. Lower, but still sub­stan­tial, shares of control group respon­dents favor hurting the out-party at the short-term expense of the country (17% of Democrats, 15% of Republicans), pushing the legal limit on partisan ger­ry­man­der­ing efforts (15% of Democrats, 13% of Republicans), or damaging the economy to enhance electoral prospects (13% of Democrats, 11% of Republicans). In general, support for these actions was similar across parties with the exception of whether their party should do every­thing possible to prevent the other party from governing effec­tive­ly, which we would expect to be more appealing to Republicans with their party out of power in Washington.

We now turn to assessing the effect of the exper­i­men­tal treatment on the full sample, which we report below on the five-point agree-disagree scale provided to respondents. 

Across the full battery of anti-demo­c­ra­t­ic actions, the effect of the exper­i­men­tal treatment was sta­tis­ti­cal­ly sig­nif­i­cant and in the expected direction, but sub­stan­tive­ly small (just 0.08 standard devi­a­tions). Providing accurate infor­ma­tion about support for democracy among partisan adver­saries reduced support for illiberal actions by co-partisans by 0.07 points on a five-point agree-disagree scale. (Full results are shown in the appendix.) We specif­i­cal­ly observed sig­nif­i­cant, but small, reduc­tions in support for actions intended to damage the other party (hurt the other party at the expense of the country, make it hard for them to govern, and sacrifice economic pros­per­i­ty to damage them). By contrast, the inter­ven­tion had no mea­sur­able effect on actions against unfa­vor­able media sources or gerrymandering. 

These results suggest that under­scor­ing the oppo­si­tion party’s shared com­mit­ment to respect­ing demo­c­ra­t­ic norms might help strength­en oppo­si­tion to illiberal actions by co-partisans, but any such effects would be quite limited in magnitude.

Experiment: Measuring support for political violence

Concerns about American democracy extend beyond illiberal actions taken by political elites to support threats, harass­ment, and even violence. How many Americans condone aggres­sive or violent acts committed by their political brethren? The issue attracted sub­stan­tial attention in the wake of anti-racism demon­stra­tions and counter-demon­stra­tions across the country in summer 2020 and espe­cial­ly after the January 6, 2021 riot at the U.S. Capitol. 

More recently, in November 2021, the House of Representatives censured Representative Paul Gosar (R‑AZ) after he tweeted a video animation depicting him killing Representative Alexandria Ocasio-Cortez (D‑NY) — but only two of Gosar’s Republican col­leagues supported the censure while 207 voted against it. How pervasive are such attitudes among the public at large? 

A question battery initially developed by the political sci­en­tists Nathan Kalmoe and Liliana Mason and repli­cat­ed by other researchers (including Bright Line Watch) appears to show alarm­ing­ly high levels of support for aggres­sive and violent actions toward opposing partisans. It consists of the following questions:

  • When, if ever, is it OK for [own party — Democrats/Republicans] to send threatening and intimidating messages to [opposing party — Republican/Democrat] leaders? (Never, Occasionally, Frequently, Always)

  • When, if ever, is it OK for an ordinary [own party — Democrat/Republican] in the public to harass an ordinary [opposing party — Democrat, Republican] on the Internet, in a way that makes the target feel frightened? (Never, Occasionally, Frequently, Always)

  • How much do you feel it is justified for [own party — Democrats/Republicans] to use violence in advancing their political goals these days? (Not at all, A little, A moderate amount, A lot, A great deal)

  • What if [opposing party — Democrats/Republicans] win the [2020/2024] presidential election? How much do you feel violence would be justified then? (Not at all, A little, A moderate amount, A lot, A great deal)

In spring 2020, 16% of Kalmoe and Mason’s respon­dents said their own party using violence to advance its political goals would be at least a little justified, a figure that increased to 21% if the other party won the pres­i­den­tial election that year. According to data they provided to us from research they conducted for their forth­com­ing book, Kalmoe and Mason find slightly higher levels of potential support for violence in more recent surveys. Most recently, they found that 17–26% of respon­dents did not wholly reject threats, harass­ment, and violence in a June 2021 survey. In par­tic­u­lar, the pro­por­tion of Americans who do not wholly reject violence was 24% for Republicans and 19% for Democrats in general and 28% and 25%, respec­tive­ly, if their party were to lose the next pres­i­den­tial election. Bright Line Watch similarly found in October 2020 that 26% of Trump approvers and 21% of dis­ap­provers were willing to condone violence in response to the other side winning the pres­i­den­tial election.

The political sci­en­tists Sean Westwood (a colleague of BLW’s John Carey and Brendan Nyhan), Justin Grimmer, Matthew Tyler, and Clayton Nall contend that top-line results from the original Kalmoe/Mason scale over­es­ti­mate support for violence because they lack precise def­i­n­i­tions and offer response options that inflate reported support for violence among inat­ten­tive respondents. 

Kalmoe and Mason report that their forth­com­ing book presents addi­tion­al analyses using different question formats and follow-up questions which also find more moderate will­ing­ness to support violence. In this report, however, we focus on the top-line figures, which have attracted sub­stan­tial attention and raised alarm. Our goal is to provide the most accurate estimate of support for threats, harass­ment, and violence. We inves­ti­gat­ed three ways in which the original question scale could lead to over­es­ti­ma­tion of support for aggres­sive actions per Westwood and his coauthors:

  • Response options: Response scales in surveys often span the full range of possible responses. For instance, a question measuring presidential approval on a scale ranging from strong disapproval to strong approval. But among the four or five choices provided by the original Kalmoe/Mason response scale, only one signals full rejection; the rest indicate acceptance of at least some aggressive actions. As Westwood et al. note, this design might inflate the proportion of respondents who appear to endorse violence, threats, and harassment. 

In our survey, half of respondents were presented with the original Kalmoe/Mason response options and half were instead asked two questions in a branching design. We first asked “Is it ever OK/justified to [action]?” with just two options, Yes or No. Respondents who said “Yes” received a follow-up question with response options that followed those provided in the original Kalmoe/Mason questions. (We classified an answer of “Never” “Not at all” as equivalent to answering “No” to the first question.)

In presenting our results below, we follow Kalmoe and Mason’s practice of distinguishing the percentage of respondents who fully reject aggressive actions or violence altogether from those who express any support for such actions.

  • Respondent attentiveness: Online survey respondents vary in the care and attentiveness they give to questions. Westwood et al. show that inattentive respondents might appear to endorse violence when they are instead answering randomly or carelessly. We therefore embedded two items in our survey that were designed to identify inattentive respondents who were not reading questions and responding carefully (see appendix for wording). The 18% of respondents who failed to respond to both questions correctly (e.g., failing to disagree with the statement that Barack Obama was the first U.S. president) were classified as inattentive.

  • Differing interpretations of “violence”: Two of the original scale questions refer to violence without specifying what that entails. Per Westwood et al., respondents endorsing “violence” might have actions ranging from destruction of property to assault to assassination in mind.

We probed three ways in which the way violence or other aggressive actions are described might affect public support. First, we created three versions of the question asking “How much do you feel it is justified for [own party] to use violence in advancing their political goals these days?” that explicitly differentiate between non-violent misdemeanors, non-violent felonies, and violent felonies: 

  • How much do you feel it is justified for [own party] to commit non-violent misdemeanors in advancing their political goals these days?”

  • How much do you feel it is justified for [own party] to commit non-violent felonies in advancing their political goals these days?”

  • How much do you feel it is justified for [own party] to commit violent felonies in advancing their political goals these days?”

Because these questions do not indicate which crimes are being committed, we also tested alternate versions that not only use the branching formats described above but spell out specific examples of each crime or explicitly define violence as using physical force to harm someone. For instance, the reference to “violent felonies” in the question above is changed to instead state “violent felonies (for example, assault during a confrontation with counter-protestors).” Similarly, we tested a modified version of the question about a presidential election victory by the other party in 2024 that not only uses a branching format but explicitly defines violence as “the use of physical force to harm someone.” 

We also sought to measure support for violence in less general and hypo­thet­i­cal terms by testing two new questions — one based on the Kalmoe/Mason framework and one created by Robert Pape and his col­leagues — that measure support for violence to “prevent Congress from cer­ti­fy­ing Electoral College votes on January 6” and to “restore Donald Trump to the pres­i­den­cy.” Two versions of each were tested using the same approach described above — one with a single response option and one with a branching scale and a def­i­n­i­tion for violence (“the use of physical force to harm someone”) added to the January 6 item. 

In total, we thus tested eight question items — the threats and harass­ment questions; questions measuring support for non-violent mis­de­meanors, non-violent felonies, and violent felonies; a question measuring support for violence if the other party wins the election; and questions measuring support for violence on January 6 and violence to restore Trump to the presidency. 

In the figure below, we sep­a­rate­ly report the pro­por­tion of attentive and inat­ten­tive respon­dents who think threats, harass­ment, crimes, or violence can sometimes be justified among two groups:

  1. Respondents who were provided the original response scale and descriptions of actions that omit examples and definitions

  2. Respondents provided a branching response scale along with an example or definition

We report three findings. First, fewer Americans endorse these actions when branching and specific ter­mi­nol­o­gy is provided. This pattern is con­sis­tent for each item, including threats and online harass­ment, where the only dif­fer­ence between groups is the set of response options (on those, we use identical wording to Kalmoe and Mason’s original scale). Second, frequency of seeming endorse­ment of political threats, harass­ment, law­break­ing, and violence is much higher among inat­ten­tive respon­dents across items. However, the pattern of greater endorse­ments among inat­ten­tive respon­dents (who represent 19% of the sample) is greatly reduced among those who received the branching response options and specific examples or definitions. 

These findings indicate that the inter­ac­tion between inat­ten­tion and a single response scale inflates estimates of support for threats, harass­ment, and violence (a problem that may affect many online surveys in which we are espe­cial­ly inter­est­ed in responses that deviate from one end of a response scale). In contrast, our best estimates of public support for political violence, threats, and harass­ment are sub­stan­tial­ly lower than past research found. When we exclude inat­ten­tive respon­dents, first ask attentive respon­dents a binary question, and provide them with a specific def­i­n­i­tion or example where appro­pri­ate, reported support is 9% for threats, 8% for harass­ment, 6% for non-violent felonies, 4% for violent felonies, 4% for violence if the other party wins the 2024 election, 4% for violence on January 6, and 5% for violence to restore Trump to the pres­i­den­cy. These figures are much lower than we estimated using the original scale in October 2020.3

These findings are in some ways reas­sur­ing, but it is important to note that they do not mean the risk of political violence is minimal. Unfortunately, even small numbers of people who encourage or engage in violence can have dangerous and desta­bi­liz­ing effects on our political system. To date, for instance, 719 people have been charged in the January 6 insur­rec­tion, which rep­re­sents less than 0.001% of the people who voted for Donald Trump in 2020.

Given the events that took place January 6, it is espe­cial­ly important to consider how support for political aggres­sion and violence varies by party. We further dis­ag­gre­gate the data by strength of respon­dent par­ti­san­ship given the influence of the party bases on the actions of politi­cians. The figure below thus reports separate estimates for Democrats and Republicans who identify strongly with their party and those who don’t. (These data represent estimates from the attentive subsample who answered branching questions that provided examples or def­i­n­i­tions where appro­pri­ate and thus provides con­ser­v­a­tive estimates of support for violence.)

As the figure shows, the parties differ notably on several dimen­sions. Democrats show higher support for non-violent mis­de­meanors and non-violent felonies than Republicans, espe­cial­ly among those who do not identify strongly with their party. In total, 35% and 13% of Democrats who do not identify strongly with their party support non-violent mis­de­meanors and non-violent felonies, respec­tive­ly, compared to 29% and 7% of those who do identify strongly. By contrast, Republicans support threats, violence if the other party wins the 2024 election, violence on January 6, and violence to restore Trump to the pres­i­den­cy at a higher rate than Democrats. Each of these levels of support is highest among Republicans who identify strongly with their party (18% for threats, 9% for 2024, 9% for January 6 violence, and 17% for violence to restore Trump compared to 12%, 2%, 6%, and 5%, respec­tive­ly, among those who do not identify strongly).

Fundamental changes to American democracy

Contemporary debate about American democracy includes serious recon­sid­er­a­tion of whether and how the structure of our system of gov­ern­ment is con­tribut­ing to potential erosion. Following a 2018 Pew poll, we asked both the public and experts whether they think the design and structure of gov­ern­ment serves the country well or if sig­nif­i­cant changes are needed to make it work for current times. We then asked respon­dents who favored change whether the best approach would be to seek to make changes that don’t require con­sti­tu­tion­al amend­ments, seeking to amend the Constitution, or creating a new con­sti­tu­tion alto­geth­er. The next figure shows results for both expert respon­dents and the public.

Large majori­ties of experts (74%) and of the public (62%) favor fun­da­men­tal change to the structure of American gov­ern­ment. Among reform sup­port­ers, a majority of experts and the public believe the best approach to making sig­nif­i­cant change is to pursue reforms that do not require amending the Constitution (45% of all experts, 33% of all members of the public). Relatively few in either group believe the best approach is to seek to amend the Constitution (22% of experts, 23% of the public) or to create a new charter alto­geth­er (7% of experts, 6% of the public). Yet, no single approach attracts majority support among either group overall. 

Disaggregating our public respon­dents by par­ti­san­ship reveals stark polar­iza­tion. In total, 62% of Republicans oppose change in the fun­da­men­tal structure of American gov­ern­ment, whereas 80% of Democrats support it.. Support is also higher among Democrats than Republicans for pursuing policies that do not require amending the Constitution (40% versus 24%) and for amending the Constitution (32% versus 12%). Few sup­port­ers of either party favor pursuing a new con­sti­tu­tion, though support is also higher for Democrats (8% versus 2%). 

Many observers are par­tic­u­lar­ly concerned about the malap­por­tion­ment of the U.S. Senate. We thus asked expert and public respon­dents whether they favor equal rep­re­sen­ta­tion for all states in the U.S. Senate or a system in which states with more people would have more senators. As the figure below illus­trates, experts and the public are divided: 72% of experts, but only 36% of the public, favor Senate rep­re­sen­ta­tion by population.

Given the partisan bias of the current Senate map, it is not sur­pris­ing that 83% of Republicans support keeping two senators per state. However, 75% of inde­pen­dents also prefer to retain the status quo and only a bare majority of Democrats, 56%, favors rep­re­sen­ta­tion by population. 

Prioritization of problems for democracy

To better under­stand experts’ assess­ments of what ails American democracy and how those problems might be fixed, we iden­ti­fied the following list of commonly mentioned problems facing American democracy and asked experts to rank their importance:

  • Economic inequality

  • Racial inequality

  • Unrepresentative political institutions

  • Partisan polarization

  • Misinformation and uncivil debate

  • Populism/illiberalism

  • Wokeness”/identity politics

The figure below shows the dis­tri­b­u­tion of impor­tance rankings for each problem sorted by average rank order. 

Unrepresentative political insti­tu­tions was most fre­quent­ly rated as the most important problem facing American democracy (26%). By average ranking, however, economic inequal­i­ty rates as most important because more respon­dents rated it as the second, third, or fourth most important problem. Relatively few experts (5%) rated racial inequal­i­ty as the most important problem facing American democracy but so many rank it second, third, or fourth that it rates as third most important by average rank.

On the whole, the rankings data reveal little expert consensus on problem impor­tance other than agreement that American democracy faces more pressing threats than “wokeness” and identity politics (ranked first by only 4% and last by 70%). Even factors ranked as rel­a­tive­ly less important — populism/illiberalism, mis­in­for­ma­tion and uncivil debate, and partisan polar­iza­tion — were ranked as the most important problem by 13–16% of experts, not far from the 26% who rated unrep­re­sen­ta­tive political insti­tu­tions first. (Respondents fre­quent­ly commented that the task was chal­leng­ing because the problems were closely related.)

Prospects for demo­c­ra­t­ic reform

We also presented experts with nine proposals for political reforms based on prominent debates among politi­cians, activists, and academics. Consistent with the reform strategy most favored by our expert respon­dents, none of the iden­ti­fied proposals requires amending the Constitution. These reforms focus over­whelm­ing­ly on the rep­re­sen­ta­tive­ness of political insti­tu­tions, which experts ranked as the second most important problem facing U.S. democracy on average. Though one – public funding for campaigns – aims to reduce the depen­dence of elected officials on campaign con­tri­bu­tions from wealthy donors and thus may address economic inequal­i­ty in political representation. 

We asked expert respon­dents to rate the expected effect each reform would have on American democracy, if adopted. We then asked respon­dents to forecast the like­li­hood, on a 0–100% scale, that each of the nine proposals would be adopted by 2030. 

The figure at left below illus­trates the results on expected effects, rated on a scale ranging from “extra­or­di­nary threat” to “extra­or­di­nary benefit.” Short summary descrip­tions of each reform proposal are provided in the middle of the figure. (The exact wording used in the rating task and fore­cast­ing exercise is provided in the appendix.) 

Despite their dif­fer­ences in rating the problems facing American democracy, the experts fre­quent­ly were in agreement on the benefits of many proposed reforms. More experts saw benefits than threats to democracy for each of the nine proposed reforms. Only one option – non-partisan primaries and two-round elections – was rated as ben­e­fi­cial by less than a majority of expert respon­dents. On seven of the nine proposals, more than 60% of the experts saw adoption as ben­e­fi­cial to democracy. 

Opinion was most favorable toward the adoption of non-partisan redis­trict­ing com­mis­sions, which 87% of respon­dents saw as a benefit to democracy and 65% regarded as seriously or extra­or­di­nar­i­ly ben­e­fi­cial. Only 3% of respon­dents regarded redis­trict­ing reform as a threat to democracy.

Shifting from our current, state-level winner-take-all system for allo­cat­ing Electoral College votes to a national popular vote for the pres­i­den­cy was nearly as well regarded by experts. Overall, 81% of experts perceive a benefit to democracy from a national popular vote system (58% serious or extra­or­di­nary) compared with 11% regarding such a reform as a threat. This support is con­sis­tent with recent evidence that Electoral College victories by popular vote losers (as occurred in 2000 and 2016) can damage the legit­i­ma­cy of the electoral system.

The next two most favored proposals would change the U.S. Senate. The first would grant statehood and thus two senators each to Puerto Rico and the District of Columbia, poten­tial­ly reducing the current Republican advantage in the chamber. The second would eliminate the fil­i­buster, an extra-con­sti­tu­tion­al provision that allows a minority of 41 senators to block most leg­is­la­tion. The next three items, all favored by solid majori­ties, address rules for elections. They include passing “clean election laws” for federal elections creating mech­a­nisms such as public matching donation systems and democracy vouchers, shifting to the use of multi-member districts in House elections, and using ranked-choice ballots in pres­i­den­tial and con­gres­sion­al elections. The last widely supported proposal would expand the House of Representatives to make it and the Electoral College more rep­re­sen­ta­tive of the nation’s pop­u­la­tion. A majority (54%) of experts saw this reform as ben­e­fi­cial for democracy. By contrast, only 43% favored the proposal to choose members of Congress through two-round elections in which non-partisan primaries are followed by runoff elections among the “top X” can­di­dates (two, four, five, etc.). Fifty-seven percent of experts either said this system would have no effect (42%) or that it would threaten democracy (15%).  

However appealing some of these reforms are to our experts, how likely are they to be adopted? To elicit the experts’ views of the reforms’ chances, we asked them to assess the like­li­hood that the following changes would take place by the year 2030, on a 0–100% scale:

  • Redistricting commissions are used in 25 or more states (currently used in 10);

  • States representing a majority of Electoral College votes join an interstate compact creating a national popular vote system;

  • The Senate is expanded via the addition of Puerto Rico and Washington, DC as states;

  • The Senate filibuster is eliminated for all legislation;

  • Public funding for congressional campaigns is available in three or more states (currently zero);

  • Multi-member districts are used in three or more states (currently zero);

  • Ranked-choice voting is used in three or more states (currently one);

  • The House of Representatives includes more than 435 members;

  • Non-partisan primaries and two-round elections are used in 10 or more states (currently five).

The figure at right above shows the dis­tri­b­u­tion of like­li­hood estimates across expert respon­dents, with the median estimate marked by an X. Two of the four most-favored reforms — the spread of redis­trict­ing com­mis­sions to half of all states, and the elim­i­na­tion of the Senate fil­i­buster — were rated as the most likely to take place in the next decade, though the median forecasts for these were only 30% and 50%, respec­tive­ly. Ranked-choice voting, which is currently used at the state level only in Maine but has come into wider use at the municipal level, is also rated as having a 30% median chance of adoption in three or more states by 2030. Experts were even more skeptical about the like­li­hood of adoption of the remaining of the reform proposals. The median forecast for the other six reform proposals was 20% (national popular vote, Senate expansion, public funding of campaigns, non-partisan primaries and two-round elections) or 10% (multi-member districts, enlarging the House of Representatives).


We continue to survey experts about the normality and impor­tance of events that have taken place in recent months. The ratings they provided are plotted on the figure below. It iden­ti­fies a number of events that experts rate as rel­a­tive­ly normal and important (e.g., the US with­draw­al from Afghanistan), normal and unim­por­tant (e.g., the Fourth of July party at the White House), events rated as rel­a­tive­ly abnormal but unim­por­tant (e.g., Sean Hannity praising Donald Trump for passing a cognitive test), and a number of events rated as both abnormal and important (e.g., Fox host Tucker Carlson endorsing “great replace­ment” theory). We are par­tic­u­lar­ly concerned about events in the shaded upper-right quadrant, which experts rated as mostly important to important and mostly abnormal to abnormal on average. 

We therefore zoom in on the upper-right quadrant of espe­cial­ly abnormal and important events in the figure below, which jux­ta­pos­es expert ratings of events from our most recent survey with those from our past surveys dating back to August 2018. As the figure indicates, six events were rated as highly abnormal and important in this most recent survey, including con­ser­v­a­tive lawyer John Eastman’s plan to overturn the results of the 2020 pres­i­den­tial election, the University of Florida pre­vent­ing pro­fes­sors there from tes­ti­fy­ing in voting rights cases, White House chief of staff Mark Meadows promoting the “Italygate” con­spir­a­cy theory about the 2020 election results, and Arizona Republicans in the state leg­is­la­ture stripping powers from the state’s Democratic Secretary of State. 

However, experts saw two events as espe­cial­ly note­wor­thy. Both concern the com­mit­ment of the Republican Party to the integrity of the electoral system. First, at least one in three declared GOP can­di­dates for the U.S. House or Senate in 2022 rejects the outcome of the 2020 pres­i­den­tial election. Second, 10 of 15 Republican can­di­dates for Secretary of State in five bat­tle­ground states question the legit­i­ma­cy of the 2020 election. These events suggest in turn that the party is increas­ing­ly embracing Donald Trump’s efforts to undermine the legit­i­ma­cy of American elections and that partisans in charge of admin­is­ter­ing elections could threaten their integrity from within. The embrace of false claims about 2020 by GOP Secretary of State can­di­dates was rated as espe­cial­ly abnormal and important — the first such rating for an event con­cern­ing the 2022 and 2024 elections. (Prior events that received com­pa­ra­ble ratings concerned Trump’s conduct in office and the effort to overturn the result of the 2020 election.) In total, 97% rated it as mostly abnormal or abnormal, including 82% who rated it as abnormal (the highest value on a five-point scale from normal to abnormal). Similarly, 96% rated it as mostly important or important, including 80% who rated it as important (the highest value on a five-point scale from unim­por­tant to important). 

Democratic per­for­mance

Finally, as in previous surveys, we asked each expert and public respon­dent to rate the overall per­for­mance of American democracy on a scale from 0 to 100. The figure below reports the average ratings for Democrats and Republicans in the public sample and for experts going back to February 2017.

Experts (green) con­sis­tent­ly rate American democracy more pos­i­tive­ly than the public (purple) does, but also broadly regarded Trump’s pres­i­den­cy as a threat to democracy. Democracy ratings among the experts accord­ing­ly increased after Trump’s defeat. However, expert optimism appears to have crested at this higher level; mean assess­ments were sta­tis­ti­cal­ly unchanged between June (68) and November (66).  Similarly, assess­ments among the public overall (purple) did not change sig­nif­i­cant­ly since June overall or for any partisan group. Democrats continue to rate U.S. democracy as per­form­ing better than Republicans, the opposite of the pattern seen during the Trump years. Those without a partisan affil­i­a­tion con­sis­tent­ly score American democracy lower than either partisan group.

The next figure breaks out per­for­mance ratings on each of 30 demo­c­ra­t­ic prin­ci­ples. (Full descrip­tions of each principle listed in the figure are included in the appendix.) The markers for each principle indicate the per­cent­age of expert (green) and public (purple) respon­dents who regard the United States as fully or mostly meeting the standard (as opposed to meeting it partly or not at all). Consistent with the overall ratings, the experts rate U.S. demo­c­ra­t­ic per­for­mance more pos­i­tive­ly than the public overall. Few excep­tions exist, however, such as voting rights being equally protected for all citizens, politi­cians operating with a common under­stand­ing on factual matters, and electoral districts not sys­tem­at­i­cal­ly favoring one party over the other.

The next figure shows per­for­mance assess­ments on the same 30 prin­ci­ples for the public sample only by respon­dent par­ti­san­ship. On some prin­ci­ples, middling overall public assess­ments hide stark partisan polar­iza­tion. Three patterns are noteworthy.

First, Democrats perceive higher levels of inequal­i­ty among citizens than do Republicans. For example, 77% of Republicans are confident that voting rights are guar­an­teed equally for all citizens compared with only 35% of Democrats. Similarly, 47% of Republicans, but only 27% of Democrats, believe legal and political rights more generally are equally protected for all citizens, and 34% of Republicans, but only 13% of Democrats believe the law is enforced equally for all citizens. 

By contrast, on matters other than citizen equality under law, Democrats are far more confident than Republicans in the basic operation of gov­ern­ment. For example, 71% of Democrats, but only 22% of Republicans, regard American elections as free of fraud. 38% of Democrats, but only 12% of Republicans, believe that gov­ern­ment sta­tis­tics are free from political manip­u­la­tion. 35% of Democrats, but only 13% of Republicans, trust that gov­ern­ment agencies are not used to punish political opponents.

Finally, respon­dents who express no partisan affil­i­a­tion tend to rate per­for­mance low, with assess­ments closer to the lower partisan group on polarized items, and in many cases lower than either partisan cohorts — for example, on whether the political system is open to all parties, whether the right to peaceful protest is protected, whether all votes have equal impact on election results, whether political can­di­dates disclose infor­ma­tion relevant to how they would govern, and whether the judiciary imposes an effective check on executive authority.


Bright Line Watch conducted its sixteenth survey of academic experts during November 5–19, 2021 and eleventh of the general public during November 10–19, 2021, on the quality of democracy in the United States. Our public sample consisted of 2750 survey par­tic­i­pants from the YouGov sample who were selected and weighted to be rep­re­sen­ta­tive of the U.S. adult pop­u­la­tion. We also surveyed 564 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 16 responded to a battery of questions about demo­c­ra­t­ic per­for­mance in the United States. Afterward, they were asked to evaluate the quality of American democracy overall on a 100-point scale.

How well do the following state­ments describe the United States as of today?

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

Additional Components of Expert Survey

Expert respon­dents were asked to respond to address batteries of questions about proposed political reforms and about current political events.

Full descrip­tions of proposed reforms

  1. Enlarge the 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.
  2. Ranked-choice voting: Introduce ranked-choice voting in pres­i­den­tial and con­gres­sion­al elections.
  3. 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.
  4. Redistricting com­mis­sions: Pass federal leg­is­la­tion requiring fair con­gres­sion­al districts to be deter­mined by state-estab­lished inde­pen­dent citizen-redis­trict­ing commissions.
  5. Public funding of campaigns: Pass “clean election laws” for federal elections that would create mech­a­nisms such as public matching donation systems and democracy vouchers.
  6. National popular vote: Establish an inter­state compact that would allow pres­i­dents to be elected by direct popular vote instead of the Electoral College.
  7. 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.
  8. Non-partisan primaries and two-round elections: Choose members of Congress in two-round elections in which all voters pick from a common set of can­di­dates and the “top X” (two, four, five, etc.) advance to a run-off if none wins a majority in the first round.
  9. Expand the Senate: Admit Washington, DC and Puerto Rico as states, expanding the Senate to 104 senators.

Political events

In this series of questions, we ask how normal or abnormal and how important or unim­por­tant recent political events are.

Is this normal or abnormal?

  • Normal
  • Mostly normal
  • Borderline normal
  • Mostly abnormal
  • Abnormal

Is this unim­por­tant or important?

  • Unimportant
  • Mostly unim­por­tant
  • Semi-important
  • Mostly important
  • Important
  1. GOP senators fil­i­buster Democrats’ election admin­is­tra­tion and voting rights legislation
  2. GOP-con­trolled state leg­is­la­ture in Arizona removes power over election-related lawsuits from Democratic Secretary of State
  3. Trump praises Ashli Babbitt, who was killed in the January 6 insurrection
  4. Conservative lawyer’s memo outlines six-point plan for VP Mike Pence to overturn the 2020 election
  5. About 1/3 of GOP 2022 Congressional can­di­dates have publicly endorsed false claims about the 2020 election outcome
  6. Texas enacts new law restrict­ing voting access over Democratic objections
  7. 10 of 15 Republican can­di­dates for Secretary of State in five bat­tle­ground states question the legit­i­ma­cy of the 2020 election
  8. Tucker Carlson claims the Biden admin­is­tra­tion’s immi­gra­tion policy is called “the great replacement”
  9. Florida, Texas, and other states ban the teaching of “critical race theory”
  10. 14 Republican con­gress­men write letter attacking Biden’s mental faculties
  11. White House Chief of Staff emailed acting attorney general theory that Italian firm rigged 2020 election
  12. Hannity com­pli­ments Trump on acing test intended to identify dementia and cognitive decline
  13. Conservative Political Action Conference (CPAC) Dallas attendees chant “lock him up” in reference to Dr. Fauci
  14. The singer Nicki Minaj tweets about her cousin’s friend’s supposed side effects from a COVID-19 vaccine 
  15. Associated Press inves­ti­ga­tion finds fewer than 200 cases of potential voter fraud out of 3.4 million ballots cast in Arizona
  16. Sen. McConnell encour­ages followers to get vac­ci­nat­ed against COVID-19
  17. Biden evacuates US troops from Afghanistan
  18. Supreme Court initially declines to block Texas abortion law allowing lawsuits against abortion providers after 6 weeks
  19. Biden mandates COVID-19 vaccine for all federal employees
  20. Biden opposes changing fil­i­buster rule in Senate to raise or suspend the debt ceiling
  21. SCOTUS Justice Breyer says he has not decided when he will retire
  22. Mitt Romney says “the politi­ciza­tion of the [COVID-19] vaccine is an outrage and frankly moronic”
  23. Biden signs bill to make Juneteenth a national holiday
  24. Larry Elder concedes to Gavin Newsom in California guber­na­to­r­i­al recall election
  25. Gov. Andrew Cuomo (D‑NY) resigns amidst sexual assault allegations
  26. Sen. Blackburn says infra­struc­ture bills are an attempt to enact socialism in America
  27. Biden hosts 4th of July party on White House South Lawn
  28. University of Florida prohibits political science pro­fes­sors from tes­ti­fy­ing as expert witnesses in a case chal­leng­ing a Florida voting rights law supported by Governor Ron DeSantis

Additional com­po­nents of public survey

Please indicate whether you agree or disagree with each statement below.

  • Strongly agree
  • Somewhat agree
  • Neither agree nor disagree
  • Somewhat disagree
  • Strongly disagree

Attention check 1: 

  • By law, abortion should never be permitted
  • In order to reduce the budget deficit, the federal gov­ern­ment should eliminate all welfare programs that help poor people
  • The federal gov­ern­ment should raise the minimum wage to $10
  • The federal gov­ern­ment should guarantee health insurance for all citizens
  • The federal gov­ern­ment should pass new rules that protect the right of workers to join labor unions
  • Barack Obama was the first president of the United States

Attention check 2: 

  • People convicted of murder should be given the death penalty 
  • World War I came after World War II 
  • Gays and lesbians should have the right to legally marry 
  • In order to reduce the budget deficit, the federal gov­ern­ment should raise taxes on people that make more than $250,000 per year 
  • The Affordable Care Act passed by Congress in 2010 should be repealed

Threats, harass­ment, and violence: 

  • When, if ever, is it OK for [own party — Democrats/Republicans] to send threat­en­ing and intim­i­dat­ing messages to [opposing party — Republican/Democrat] leaders? (Never, Occasionally, Frequently, Always)
  • When, if ever, is it OK for an ordinary [own party — Democrat/Republican] in the public to harass an ordinary [opposing party — Democrat, Republican] on the Internet, in a way that makes the target feel fright­ened? (Never, Occasionally, Frequently, Always)
  • How much do you feel it is justified for [own party — Democrats/Republicans] to commit non-violent mis­de­meanors in advancing their political goals these days? (Not at all, A little, A moderate amount, A lot, A great deal)
  • How much do you feel it is justified for [own party — Democrats/Republicans] to commit non-violent felonies in advancing their political goals these days? (Not at all, A little, A moderate amount, A lot, A great deal)
  • How much do you feel it is justified for [own party — Democrats/Republicans] to commit violent felonies in advancing their political goals these days? (Not at all, A little, A moderate amount, A lot, A great deal)
  • How much do you feel it is justified for [own party — Democrats/Republicans] to use violence in advancing their political goals these days? (Not at all, A little, A moderate amount, A lot, A great deal)
  • Thinking about the riot that took place at the U.S. Capitol on January 6 while Congress was in the process of cer­ti­fy­ing Electoral College votes… How much do you feel violence was justified to prevent Congress from cer­ti­fy­ing Electoral College votes? (Not at all, A little, A moderate amount, A lot, A great deal)
  • Thinking about the riot that took place at the U.S. Capitol on January 6 while Congress was in the process of cer­ti­fy­ing Electoral College votes… How much do you feel violence is justified to restore Donald Trump to the pres­i­den­cy? (Not at all, A little, A moderate amount, A lot, A great deal)

Additional result

Additional figures


  1. Respondents who are inde­pen­dents (i.e., who do not identify with or lean toward either party) were asked to provide estimates for each party in ran­dom­ized order.
  2. Out-partisan” refers to members of the other party from a given survey respondent.
  3. We cannot fully rule out the pos­si­bil­i­ty that inat­ten­tive respon­dents are also sincerely more sup­port­ive of threats, harass­ment, and violence than attentive respon­dents, but the huge reduction in endorse­ment we observe among inat­ten­tive respon­dents when the question format and wording are changed suggests that the dif­fer­ence is largely an artifact of question format. On this basis, we prefer estimates based on attentive respon­dents as best reflec­tive of overall support for violence, threats, and harassment.