Abstract
Do US voters prefer inexperienced candidates? Candidates who have never held elected office before have had greater success in recent presidential and congressional elections. However, it could be that voters prefer the type of anti-establishment rhetoric that such candidates use more than the lack of experience itself. We conduct a 2x2 factorial experiment that manipulates a fictitious congressional candidate’s experience and rhetoric toward the political system. Results from a nationally representative Qualtrics sample and two follow-up studies from Mechanical Turk show that respondents evaluate the candidate more positively when he uses anti-establishment rhetoric instead of pro-establishment rhetoric. Though the findings are mixed, we find weak and inconsistent evidence that respondent prefer inexperienced candidates to experienced ones. The results suggest that outsider candidates receive an electoral boost by using anti-establishment messaging, but that candidates’ political résumés matter less to potential voters.
Introduction
Do US voters respond favorably to political inexperience among candidates? Candidates who have never held elected office often characterize their lack of experience as an asset on the stump. Inexperience can signal to voters that the candidate remains uncorrupted by the dirty business of politics and is thus capable of bringing needed change to government. Inexperienced candidates seem to have gained electoral momentum in recent years. Most notably, Donald Trump won the 2016 presidential election with no prior experience in elected office. Inexperienced candidates have also performed better running for Congress. In open primary elections in 2016 and 2018, nearly 50% of the winning candidates had never held public office, a marked increase from earlier election cycles (Porter and Treul, 2019).
Yet it is unclear that inexperience itself is what appeals to voters. “Outsider” candidates lacking prior experience often employ anti-establishment rhetoric to court voters (Carreras, 2012). Anti-establishment rhetoric is campaign messaging that signals opposition to existing power structures or the political status quo. Inexperienced candidates might use this rhetoric to justify why power should be taken from current officeholders and given instead to novice challengers like them. Experienced candidates might eschew such rhetoric, partly because they have already found success winning office through existing power structures and partly because railing against the system that previously brought them to power could ring false to voters. It could be that the recent success of inexperienced candidates stems not from voters’ desire for inexperience, but from voters’ favorable appraisals of the kind of rhetoric that inexperienced candidates tend to employ.
To separate the effects of rhetoric and experience on voters’ evaluations of candidates, we conduct a 2x2 factorial survey experiment using a nationally representative sample. We manipulate whether a fictitious congressional candidate previously served in his state’s legislature or not, and whether the candidate provides a statement that includes pro-establishment or anti-establishment rhetoric. Results show that respondents react positively to candidates employing anti-establishment rhetoric, but react no differently to candidates with different levels of prior experience. We also conduct two follow-up studies using convenience samples from Mechanical Turk (MTurk). The first shows that the findings from the original experiment are robust to minor alterations in the rhetoric. The second shows respondents react no more favorably to anti-establishment rhetoric that is critical of parties than rhetoric that is critical of corruption, but does show that respondents view inexperienced candidates somewhat more warmly.
The findings help us to understand the appeal of outsider candidates. While the findings are mixed, the balance of evidence suggests that outsider candidates’ messaging strategies have a stronger effect on voter evaluations than their résumés. We cannot rule out that candidates win voter support by emphasizing their inexperience, but voters appear not to weight candidates’ background in politics consistently as a positive or a negative. However, in a period of low trust in government, running against the system appears to attract voter support.
Distinguishing experience from rhetoric
Candidates without prior experience in office are often labeled as outsiders in media coverage and occasionally in academic literature (e.g. Canon, 2011; Zaller and Hunt, 1995). However, the term “outsider” might also refer to candidates whose rhetoric is directed against existing power structures or “the establishment” (Carreras, 2012). Though inexperience and anti-establishment rhetoric often go hand-in-hand, candidates’ rhetoric need not be consistent with their experience (Carreras, 2012; Kenney, 1998). A candidate can speak and act as an outsider while holding office, though candidates can make a better case that they are outsiders if their actions within the institution match their rhetoric. Senator Bernie Sanders’ success in positioning himself as an anti-establishment outsider in the 2016 and 2020 Democratic primaries, despite his roughly 40 years of service in elected office, provides an example. Likewise, inexperienced candidates can use pro-establishment rhetoric. In 2016, former chief executive officer (CEO) Carly Fiorina competed in the Republican primary without employing the same kind of anti-establishment rhetoric as her fellow inexperienced rival Donald Trump. Therefore, we test separate hypotheses about the appeal of experience and rhetoric to voters.
Experience
We use experience to refer to whether a candidate has previously held elected office. Experience could refer to a background in a field with skills relevant to politics, like law (Roberds and Roberts, 2002), or to political experience outside elected office, as in campaign staffing. However, it is most often invoked to indicate prior elected officeholding—the classic “quality” candidate dichotomy (Jacobson, 1989). In fact, experience is one of the best predictors of candidate success in congressional elections.
Exactly why experienced candidates perform better has been the subject of some debate. Experienced officeholders have been successful in at least one campaign. Part of their advantage might come from skills necessary for campaigning, like fundraising or public speaking (Porter and Treul, 2019; Squire, 1992), or qualities that make them better public servants, like personal integrity or competence on the job (Buttice and Stone, 2012; Mondak, 1995). However, prior experience does not inevitably lead to a strong performance on the campaign trail or in office. Strategic entry—deciding to run or not based on the likelihood of victory—explains some of the success of experienced candidates at the ballot box (Jacobson and Kernell, 1983; Maestas and Rugeley, 2008). Successful inexperienced candidates also tend to emerge in contests where there is greater opportunity for victory (Canon, 1990; Roberds and Roberts, 2002).
However, the extent of citizen demand for inexperienced candidates remains unclear. Most of the available evidence for the experience advantage comes from analyses of election outcomes (Jacobson, 1989; Mondak, 1995; but see Buttice and Stone, 2012). Survey evidence exploring Americans’ expressed preferences for political experience is more limited. Existing studies show mixed to weak evidence that voters desire their candidates to have prior experience. Fridkin and Kenney (2011) asked voters to evaluate candidates in 21 real races for US Senate seats in 2006, finding that respondents tended to rate challenger candidates with greater experience slightly more positively, but that experience had no significant association with incumbent evaluations. Kirkland and Coppock (2018) found that fictitious mayoral candidates with prior political experience were preferred and seen as more competent—but only by Democratic respondents when the candidate’s party was not revealed. Such findings could imply that supply-side factors like strategic entry (Maestas and Rugeley, 2008) or fundraising ability (Porter and Treul, 2019) better explain the quality candidate advantage than voter preference. We test the hypothesis:
H1: Respondents evaluate experienced candidates more positively than inexperienced candidates.
Rhetoric
We use anti-establishment rhetoric to refer to campaign messaging that signals opposition to the political status quo. The success of politicians’ appeals using this kind of rhetoric depends on the “ability of the political actor to convince potential supporters that he indeed stands in opposition to, and is not part of, the entrenched power structure” (Barr, 2009: 32). Populists are known for employing anti-establishment rhetoric (Canovan, 1981; Oliver and Rahn, 2016), but such rhetoric is not necessarily populist (Barr, 2009).
There are good reasons to suspect that anti-establishment rhetoric would resonate with US voters. Trust in government has declined for nearly half a century (Hetherington and Rudolph, 2015; Hetherington and Weiler, 2018). Many Americans do not feel well represented by either major party (Oliver and Rahn, 2016). Presidential candidates of both parties have gained large and loyal followings using rhetoric that criticizes politics as usual. Congressional candidates for decades have run against the idea of Congress even as they cultivated loyal constituencies in their home districts (Fenno, 1978). Given widespread skepticism of the political status quo, candidates who employ anti-establishment rhetoric are likely to find a large and sympathetic audience.
It is possible that anti-establishment rhetoric would not be well received by US voters, depending on the target of the rhetoric. Implicit trust in government—the kind of trust associated with support for democratic institutions broadly rather than the current set of ruling elites—is more widespread than explicit trust in the US population, likely due to early childhood socialization (Intawan and Nicholson, 2018). Implicit trust is associated with system-justifying beliefs; people who implicitly trust government are more likely to perceive the political system as fair and legitimate. If citizens perceived candidates’ rhetoric as attacking the set of elites in power, they would likely react more approvingly than if they perceived the rhetoric to be attacking democratic institutions or the US political system in general. We test the hypothesis:
H2: Respondents evaluate candidates using anti-establishment rhetoric more positively than candidates using pro-establishment rhetoric.
Experimental design and data
To gather data on public attitudes toward inexperience and anti-establishment rhetoric, we posed a battery of questions to a nationally representative sample of respondents. Data came from the Politics in the Field at the University of North Carolina (P-FUNC) survey, a multi-investigator study of Americans’ social and political attitudes. The sample was recruited by Qualtrics, which maintains a panel of survey participants matching Census demographics. Respondents completed the survey online between 27 November and 20 December 2018.
The experiment follows a 2x2 factorial design. Respondents were randomly assigned to one of four treatment groups in which they read different vignettes about a congressional candidate. The vignette wordings evoked four candidate styles: an experienced establishment candidate, an inexperienced establishment candidate, an experienced anti-establishment candidate, and an inexperienced anti-establishment candidate. To capture experience, the candidate was presented as either a first-time candidate or a 20-year veteran of their state’s legislature. To capture rhetoric, the establishment candidate was presented as a party powerbroker and gave a statement about his consensus-building attitude toward politics. The anti-establishment candidate was presented as an independent maverick and gave a statement about his disdain for Washington corruption.
Full wordings for the candidate vignettes appear in Table 1. Respondents read vignettes in which the word “[REDACTED]” appeared instead of a candidate name or party name in order to remove potential biases of candidate characteristics (gender, race, etc.) or party affiliation from the experiment. After reading the vignette, respondents were asked to react using three outcome measures. First, respondents rated the candidate on a 0–100 feeling thermometer, with higher values indicating a warmer response to the candidate. Second, respondents indicated their likelihood of voting for such a candidate on a 5-point scale. Third, respondents indicated their likelihood of donating US$10 to the candidate's campaign on a 5-point scale. Exact question wording for the outcome variables is presented in section 1 of the appendix.
Qualtrics treatment wordings.
Note: Respondents were randomly assigned to one of four treatment groups in a 2x2 factorial design. Baseline condition wording presented in the left panel. Bolded text on the right indicates language manipulation moving to a condition with an inexperienced candidate. Italicized text indicates language manipulation moving to a condition with anti-establishment rhetoric.
Results
Figure 1 displays the predicted values of each outcome across treatments, while the accompanying ordinary least squares (OLS) regression results are provided in Table A1 of the appendix. We begin with the predicted values of respondent warmness toward the candidate, reported in Figure 1(a). Respondents felt lukewarm toward the experienced/establishment candidate (mean = 59.85 on a 100-point scale). Moving to the inexperienced/establishment candidate, respondent evaluations felt slightly, but not significantly, warmer (0.79, p = 0.60). When presented with candidates with establishment messages, respondents did not feel more warmly toward inexperienced candidates over experienced ones.
Candidate evaluations across treatment groups.
Turning to the treatments presenting a candidate with an anti-establishment message, we see a significant increase in respondent warmness. Compared to an experienced/establishment candidate, respondents rated an experienced/anti-establishment candidate nearly seven points higher on the feeling thermometer (6.92, p = 0.00). We see a similar finding in the fourth condition, with respondents giving a similar bump (7.32, p = 0.00) to the inexperienced/anti-establishment candidate compared to the experienced/establishment candidate. Respondents did not rate the inexperienced/anti-establishment candidate significantly more highly than his more experienced counterpart with the same message (0.40, p = 0.79). The anti-establishment message in the fourth condition also seems to drive the increase in respondent warmness when compared to the inexperienced/establishment candidate in the second condition; respondents felt significantly more warmly toward the former than the latter (6.53, p = 0.00).
Moving to Figure 1(b), we see a similar pattern emerge when respondents are asked their likelihood of voting for the candidate on a 5-point Likert scale. Again, we see little difference in respondents’ likelihood of voting for an experienced or an inexperienced establishment candidate (0.01, p = 0.85). However, respondents reported they were significantly more likely to vote for both the experienced/anti-establishment candidate (0.20, p = 0.00) and the inexperienced/anti-establishment candidate (0.26, p = 0.00) compared with the experienced/establishment candidate. Respondents also did not draw a noticeable distinction between anti-establishment candidates on the basis of experience (0.06, p = 0.35).
Figure 1(c) shows results when we asked respondents if they would consider making a US$10 donation to the candidate presented. Similar to the other two panels, experience does not make a difference in responses to the two establishment candidates (0.06, p = 0.42). Respondents expressed more willingness to donate to both the experienced/anti-establishment candidate (0.20, p = 0.01) and the inexperienced/anti-establishment candidate (0.33, p = 0.00) than to the experienced/establishment candidate. The responses on the basis of experience between the anti-establishment candidates were not significantly different either (0.13, p = 0.13).
On an exploratory basis, we observe whether treatment effects varied among respondents of different party identifications. Results are presented in Table A2 in the appendix. We find no effect of candidate experience on evaluations among Republicans, Democrats, or Independents. We find some variation in the strength of the effects of anti-establishment rhetoric across subgroups, though coefficient estimates are signed in the same direction for all outcome variables among all subgroups. Anti-establishment rhetoric caused Republicans, Democrats, and Independents alike to feel more warmly toward the candidate. However, only Republicans and Independents were significantly more likely to express willingness to vote for the anti-establishment candidate. Only Independents were significantly more likely to express willingness to make a donation.
To summarize, the candidate’s prior experience was not a determining factor in respondent evaluations. In cases when candidates used establishment and anti-establishment rhetoric alike, respondents did not evaluate the experienced candidate more favorably than the inexperienced candidate or vice versa. Likewise, the OLS regression results in Table A1 of the appendix show no significant main effects for the experience treatment across all outcome variables. Therefore, we find no evidence in support of Hypothesis 1. This finding is interesting in and of itself, as it indicates that neither experience nor inexperience draws voters to a candidate. However, respondents reacted more favorably to candidates who used anti-establishment rhetoric, regardless of experience level. In the OLS regression results, the main effects for establishment rhetoric were negative and significant across all three outcome variables. Regression analysis showed no statistically significant coefficient estimates on the interaction terms across all three outcome variables. Therefore, we cannot reject the null hypothesis that there is no difference of the effects of rhetoric at different levels of candidate experience. Overall, we find evidence in support of Hypothesis 2.
Follow-up studies
We conducted two follow-up studies using convenience samples of respondents recruited from Amazon’s Mechanical Turk platform in order to observe the robustness of treatment effects when the candidate's rhetoric is altered. While not a pure replication, the purpose of MTurk Study 1 was to approximate the original experimental conditions while streamlining and focusing the treatment language. The anti-establishment language in the original sample is much lengthier than the language describing candidate experience. Respondents might have overlooked the candidate’s experience in the context of the prompt. We also changed the anti-establishment rhetoric to focus more strictly on the candidate's connection to his party’s establishment. Using slightly different wording in the two samples carries the benefit of allowing us to be reasonably surer that results observed are due to underlying attitudes rather than reactions to specific language. Full vignette and question wording for the study is presented in Table A3 of the appendix.
We recruited 500 participants in June 2019. Descriptive statistics for the sample are presented in Table A4 of the appendix. OLS regression results, presented in Table A5 of the appendix, show that respondents are more willing to vote for candidates using anti-establishment rhetoric than candidates using establishment rhetoric. As in the Qualtrics sample, respondents in this study did not evaluate the experienced candidate more favorably than the inexperienced candidate or vice versa.
The purpose of MTurk Study 2 was to understand whether reactions to anti-establishment rhetoric change when different targets of the rhetoric are isolated. The treatment wordings for the Qualtrics sample describe the candidate as both pro-party politics and pro-compromise in the establishment treatment, while they describe him as both anti-party politics and anti-corruption in the anti-establishment treatment. In MTurk Study 1, the rhetoric focuses solely on party politics. For MTurk Study 2, we vary the target of the anti-establishment rhetoric (party politics or corruption) to determine whether they produce different effects on voter evaluations. Vignette and question wording is presented in Table A6 of the appendix.
We recruited 1000 participants in January 2020. Descriptive statistics for this sample are presented in Table A7 of the appendix. In short, we find that respondents react to anti-corruption rhetoric no more positively than they do to anti-party politics rhetoric, as indicated by the OLS regression results in Table A8 of the appendix. However, in this study we find mixed evidence that respondents evaluate inexperienced candidates more positively than experienced. Specifically, respondents felt more warmly toward the inexperienced candidate than the experienced one, though were no more likely to vote, donate, or volunteer for the inexperienced candidate. This finding directly conflicts with findings in the first two samples. The finding could be a function of the convenience sample or a statistical outlier. It may also indicate that under certain circumstances, US voters are more approving of inexperienced candidates. Future research could explore whether voter preferences for inexperience vary across choice contexts.
Discussion
The results suggest that candidates’ rhetoric more strongly and consistently affects voter evaluations while candidates’ résumés weakly and less consistently affect them. We consistently find that respondents on average feel more warmly toward and report being likelier to vote for candidates who employ anti-establishment rhetoric. The reward was modest, roughly 6–7 points on the 0–100 feeling thermometer and 0.2–0.3 points on the 5-point voting scale. The findings suggest that, in the absence of other information, messaging one’s opposition to the political status quo boosts candidates among potential voters. This finding expands on our understanding of messaging in today’s political context, suggesting that messaging that criticizes the political system writ large resonates with potential voters.
When it comes to experience, the picture is muddier. The clearest conclusion we can draw from the experiments is that average Americans do not favor experienced candidates; in no case did we find that respondents prefer them to first-time candidates. If anything, the results show that the level of experience had no consistent, significant effect on respondents’ evaluations in either direction. However, due to the findings in MTurk Study 2, we cannot rule out that voters prefer inexperienced candidates. Our findings fall in line with prior survey research showing that Americans have mixed opinions on the question of whether candidates should have prior political experience (Fridkin and Kenney, 2011; Kirkland and Coppock, 2018).
Our findings have limitations. Results are based on evaluations of congressional candidates. There is good reason to suspect that these results would not translate to voter evaluations of presidential candidates, given the different nature of the two jobs. Indeed, prior evidence suggests that Americans do weigh candidate traits like leadership differently when evaluating these two types of candidates (Hayes, 2010). Moreover, we only tested one type of prior experience (state legislative service). Americans may value other types of pre-congressional experience, like that of mayor or governor. Because we focus on pre-congressional experience, we cannot necessarily draw conclusions about how voters weigh incumbency. Additionally, we note that neither rhetoric type nor experience level consistently moved respondents toward donating or volunteering, suggesting that these factors alone are not sufficient to mobilize respondents to take higher-cost personal actions. Finally, because the 3 studies were conducted over the course of 14 months, it is possible that events playing out during the Trump Administration affected the results in unmeasured ways across the studies.
Though experience itself may give candidates little additional traction when appealing to voters, it would be difficult to conclude that it does not matter to election outcomes. It is quite likely that the campaigning and officeholding skills that experienced candidates possess give them a leg up when running for office. Experience might also be highly valued by party activists, donors, and other elected officials. These actors play an outsized role in recruiting candidates, culling the field in the primary by directing limited resources to certain campaigns, and signaling party support to average voters. However, when it comes to the qualities that potential voters desire in candidates, it seems Americans respond more favorably to outsiders based on their rhetoric rather than their résumé.
Supplemental Material
sj-pdf-1-rap-10.1177_20531680211034958 – Supplemental material for Inexperienced or anti-establishment? Voter preferences for outsider congressional candidates
Supplemental material, sj-pdf-1-rap-10.1177_20531680211034958 for Inexperienced or anti-establishment? Voter preferences for outsider congressional candidates by Eric R. Hansen and Sarah A. Treul in Research & Politics
Footnotes
Acknowledgements
We thank Meghan Condon, David Doherty, Mara Suttman-Lea, Andrew Tyner, and several anonymous reviewers for their helpful comments, as well as Marc Hetherington for facilitating the P-FUNC survey. The authors remain responsible for any errors.
Correction (June 2025):
Author note
A previous version of this paper was presented at the 2019 meeting of the Midwest Political Science Association. The Institutional Review Boards at Loyola University Chicago and the University of North Carolina at Chapel Hill reviewed the survey experiments described herein.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Supplementary material
Carnegie Corporation of New York Grant
This publication was made possible (in part) by a grant from the Carnegie Corporation of New York. The statements made and views expressed are solely the responsibility of the author.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
