The Racial Typification of Election Officials and Trust in Elections

Abstract

Who do Americans think runs their elections? This paper examines the racial typification of election officials, that is, how Americans perceive the racial composition of the people who administer elections and what drives those perceptions. Using survey data from a module in the 2024 Cooperative Election Study paired with county-level demographic data, we model perceptions of election official racial composition as a compositional dependent variable using seemingly unrelated regression (SUR) models. We find that partisanship is the driving predictor of racialized perceptions. Republicans perceive a substantially greater share of election officials as non-white, even after controlling for the actual racial demographics of respondents’ counties. We then show that these perceptions carry downstream consequences for election trust. Among white respondents, perceiving a greater share of election officials as Black is associated with significantly lower trust in local election officials and lower confidence that the 2024 election would be run fairly. These findings suggest that as America diversifies, public trust in elections may be undermined not by actual failures of administration, but by racialized perceptions of who holds administrative power.

Who do Americans think runs their elections? The answer, it turns out, depends less on reality than on the perceiver’s partisanship and worldview. In our survey data, the average American estimates that only 55 percent of local election officials are white—compared to approximately 90 percent in reality. ¹ This dramatic overestimation of the non-white share of election officials is not uniformly distributed: it is structured by political identity. Republicans perceive election officials as substantially more non-white than do Democrats, even after accounting for the actual demographics of respondents’ local communities.

This paper investigates both the sources and consequences of these racialized perceptions. We draw on the concept of racial typification. Racial typification is the cognitive process by which individuals associate racial groups with certain social roles or behaviors (Chiricos, Welch and Gertz 2004; Udani, Manion and Kimball 2024). A large body of work has documented how Americans racially typify crime, disproportionately associating Black and Latino individuals with criminal behavior (Chiricos and Eschholz 2002; Dixon and Linz 2000; Welch 2007). Udani, Manion and Kimball (2024) extend this logic to the electoral domain, showing that white Americans dramatically overestimate the proportion of illegal voters who are racial minorities and favor harsher punishments for voters of color. We build on this foundation to argue that racial typification extends beyond the “criminals” of election fraud to the administrators of elections. Just as Americans “other” lawbreakers (Udani, Manion and Kimball 2024), they may also “other” the people who run elections—perceiving them as more racially diverse than they actually are, in ways that reflect partisan and conspiratorial worldviews.

Our approach proceeds in two stages. First, we treat respondents’ perceptions of the racial composition of election officials as our primary dependent variable. Because these perceptions are compositional (the estimated percentages of Black, white, Latino, and other officials must sum to 100) we employ a seemingly unrelated regression (SUR) framework using log-ratio transformations of the compositional shares (Eastman et al., 2024; Philips, Rutherford and Whitten 2016; Philips, Rutherford and Whitten 2015). This allows us to model how partisanship, conspiracism, and local demographics jointly shape all four components of perceived racial composition simultaneously. Second, we examine the downstream association between these racialized perceptions and trust in elections. We show that among white respondents, perceiving election officials as more non-white is associated with lower trust in elections.

Together, the analyses tell a story with broader implications for a diversifying America. If white Americans are systematically less trusting of elections when they perceive non-whites to be in charge of administration, then the same dynamics that have eroded support for other public goods, from schools to parks to public transit (McGhee 2022), may threaten confidence in elections as well.

The Formation of Racial Perceptions

A substantial body of work has established that Americans routinely misestimate the sizes of racial and ethnic groups. Studies of racial innumeracy have shown that Americans often overestimate the size of Black and Latino populations and that these overestimates are politically structured (Alba, Rumbaut and Marotz 2005; Gallagher 2003; Nadeau, Niemi and Levine 1993; Sigelman and Niemi 2001; Wong 2007). Often, these distorted perceptions about group sizes are not random errors; they are systematically related to attitudes, with those who perceive larger minority populations tending to express more negative racial attitudes and greater support for restrictive policies (Ahler and Sood 2018; Alba, Rumbaut and Marotz 2005; Landy, Guay and Marghetis 2018; Nadeau, Niemi and Levine 1993; Zhirkov 2021).

These patterns of structured perception likely extend to how individuals view election officials. Election administrators nationally are not a highly visible group to the average voter. In the absence of concrete information, voters likely draw on an array of factors in forming their impressions of who elections administrators are, including local demographics, political cues, and broader group-centric stereotypes (Ahler and Sood 2018; Gallagher 2003; Nadeau, Niemi and Levine 1993). Given that election officials have become a central focus of contested narratives about election integrity since 2020, they are a particularly ripe target for politically influenced demographic perceptions. But what political influences shape Americans’ perceptions of who runs elections? We argue that three factors are likely to be important: partisan identity, conspiratorial thinking, and local demographic context.

We expect partisanship to be a key predictor of how people perceive the racial composition of election officials. Republican elites have repeatedly framed election fraud as a problem concentrated in urban areas with large minority populations (Cottrell, Herron and Westwood 2018; Dreier and Martin 2010; Smith 2017). President Trump’s claims of mass voter fraud—often targeting cities like Detroit, Philadelphia, and Atlanta—infuse an explicit racial dimension into narratives about election administration. To the extent that Republican voters internalize these elite cues, they may come to associate election administration itself with racial minorities, perceiving a larger non-white presence among the officials who run elections. Additionally, Udani, Manion and Kimball (2024) find that the racial typification of election fraud is more pronounced among white Americans, particularly those with conservative orientations, suggesting a broader pattern in which the electoral domain is racialized along partisan lines. This leads to our first hypothesis.

Hypothesis 1: Republicans will perceive a greater share of election officials as non-white compared to Democrats.

Conspiratorial thinking may independently shape racial perceptions of election officials. Individuals who believe that a few powerful actors secretly control society are predisposed to see hidden forces at work in public institutions. In the context of elections, this conspiratorial lens may amplify perceptions of out-group presence among officials—the perception that “they” are running things feeds into a broader narrative of institutional capture by illegitimate actors. The overlap between conspiratorial thinking and racial threat perceptions is well documented, and conspiracism has been shown to predict support for restrictive election measures (Chouhy, Lehmann and Singer 2023; Udani and Kimball 2018). Our second hypothesis follows.

Hypothesis 2: Individuals higher in conspiratorial thinking will perceive a greater share of election officials as non-white.

Finally, we expect that the actual racial demographics of respondents’ local context will shape their perceptions. Voters with limited knowledge of election officials nation-wide may extrapolate from the demographics of their own communities when estimating the composition of election officials (Wong 2007). People living in more racially diverse counties may reasonably perceive a more diverse set of election officials, reflecting at least some degree of accurate inference from local context. At the same time, the racial innumeracy literature suggests that the relationship between actual demographics and perceived demographics is often distorted, with perceived diversity frequently outstripping reality (Alba, Rumbaut and Marotz 2005; Gallagher 2003). These theoretical expectations lead to our third hypothesis.

Hypothesis 3: Respondents in counties with larger non-white populations will perceive a more diverse set of election officials.

Why It Matters: Racial Perceptions and Electoral Trust

Having established what we expect to drive racial perceptions of election officials, we now turn to why these perceptions matter. We argue that the racial composition of election officials, as perceived by voters, shapes public trust in elections. This argument draws on two complementary literatures: representative bureaucracy and racial threat.

Theories of descriptive representation and representative bureaucracy suggest that individuals are more likely to trust government services when they believe the officials administering those services are demographically similar to themselves. This pattern has been documented across a range of public institutions, from Congress to police departments to local councils (Choi and Hong 2021; Gay 2002; Hayes and Hibbing 2017; Kaslovsky, Rogowski and Stone 2021; Lemi 2021). In law enforcement, Black and Latino individuals report higher trust and perceived legitimacy when interacting with officers of the same race (Ba et al., 2021; Riccucci, Van Ryzin and Jackson 2018; Theobald and Haider-Markel 2009). Within election administration specifically, minority voters who interact with co-racial poll workers report higher confidence in the electoral process (King 2018; King and Barnes 2019), and descriptive representation among election officials can boost turnout (Coll 2024; Hayes et al., 2024). Gay (2002) describes a “spirals of trust” phenomenon, wherein descriptive representation generates a feedback loop of growing institutional confidence among represented groups.

Conversely, when voters perceive election officials as belonging to a racial out-group, racial threat dynamics may erode trust. Research on racial threat has consistently found that perceptions of large or growing minority populations are associated with negative attitudes toward those groups and the institutions associated with them (Alba, Rumbaut and Marotz 2005; Gallagher 2003; Wong 2007). If voters are predisposed to associate electoral wrongdoing with racial out-groups (Udani, Manion and Kimball 2024), it stands to reason that they may also distrust the officials perceived to be overseeing elections when those officials are seen as belonging to an out-group.

Notably, the relationship between racial perceptions and trust is likely bidirectional. Individuals’ institutional trust may act as a baseline against which perceptions and behavior are formed, functioning as an independent variable in itself (Hetherington 1998). Consequently, existing research on political trust suggests that in addition to being shaped by performance and prior dispositions, political trust also structures subsequent political evaluations (Citrin and Stoker 2018; Levi and Stoker 2000; Wilkes 2015), indicating a potential mutually reinforcing relationship between perception of trustworthiness and racial associations (Filindra, Kaplan and Buyuker 2022). Thus, voters who already distrust elections may be more inclined to perceive election officials as racially different from themselves, and perceptions of out-group officials may erode trust. We cannot fully adjudicate this endogeneity with our cross-sectional data. However, the association is substantively important regardless of the causal direction. If racialized perceptions of election officials and distrust in elections co-occur and reinforce one another, this dynamic has significant implications for democratic governance. We therefore present the following hypotheses as associational expectations:

Hypothesis 4: Perceiving election officials as racial out-group members will be associated with lower trust in local election officials.

Hypothesis 5: Perceiving election officials as racial out-group members will be associated with a lower likelihood of believing that the 2024 election will be run fairly.

Data and Research Methods

We use survey data from our team module in the 2024 Cooperative Election Study (CES). CES modules are national surveys administered to 1,000 respondents. The data we use in this paper are from the pre-election wave of this CES module, which was fielded from October 1 to November 4, 2024. The timing of the pre-election survey ensures that voter attitudes are reflective of general perceptions of electoral processes and election officials, independent of specific election results or immediate post-election sentiments. We supplement the individual-level survey data with county-level demographic and political variables, including the racial composition of each respondent’s county population, the percentage of the county population living in rural areas, and the 2020 presidential election margin in the county, to capture contextual factors that may shape perceptions as each covaries with both race and partisanship (Pew 2024) and thus might condition how respondents infer the race of election officials.

Measuring Perceptions of Election Officials

Our primary dependent variable is the perceived racial composition of election officials, measured using a compositional allocation question inspired by the method used by Udani, Manion and Kimball (2024) in studying the racial typification of illegal voters. The precise question posed to respondents follows.

“What percentage of local election officials in America, the people who run elections at the city, county or state level, would you say are… Black? White? Hispanic? Other?”

The numerical allocators provided by respondents for each racial category were required to sum to 100. The resulting four-part composition (percent Black, percent white, percent Latino, and percent other) constitutes our compositional dependent variable. ²

The distribution of responses, seen in Figure 1, suggests that the average estimations of the racial makeup of election officials, though loosely approximating the US population, depart meaningfully from reality. Notably, real demographic data from the Democracy Fund indicate that the actual number of local election officials who identify as white is around 90 percent, whereas the respondents estimate around 56 percent to be white. While our survey item is formally agnostic as to the specific level of election official respondents have in mind, whether that’s state or local, the apparent discrepancy is still substantial, considering the vast majority of election officials work at the local level.

Figure 1.

Distribution of responses to question, “What percentage of local election officials in America, the people who run elections at the city, county or state level, would you say are [Black/White/Hispanic/Other]?”

The large gap between perceptions and reality provides some initial evidence that, on average, Americans perceive the nation’s state and local election officials to be much less white than they actually are. As the existing work predicts, racial perceptions of large out-group size are often not benign but rather can be indicative of racial animus and threat (Nadeau, Niemi and Levine 1993; Wong 2007).

Figure 1 also displays perceptions broken out by respondent race. There is some evidence that respondents perceive slightly more co-racial officials. For example, Black respondents estimate that 30 percent of election officials are Black, compared to 22 percent among white respondents. And Latino respondents estimate 20 percent of officials are Latino, compared to 12 percent among white respondents. However, these differences are modest in magnitude, and all racial groups substantially overestimate the non-white share of election officials relative to available benchmarks.

Measuring Trust in Elections

To determine the effect of these perceptions on trust in elections, we use two dependent variables. The first is trust in local election officials, measured as an index of three questions developed by Brown et al. (2024):

“How much do you trust local election officials to … ?”

1. “… Follow the rules of election law”

2. “… Do their part to ensure a free and fair election”

3. “… Be ethical people”

Response options include “a great deal,” “a lot,” “a moderate amount,” “a little,” and “none at all,” and are coded on a five-point Likert scale where 5 = “a great deal” and 1 = “none at all.” We rescale the resulting 15-point additive index to range from zero to one. The three measures have a Cronbach’s alpha of 0.95, but, for robustness, we estimate our models on the three separate measures and include the results in the Supplemental Materials.

The distribution of responses to these three prompts, shown in Figure 2, reveals just how similarly respondents viewed the questions. For all three of the questions, a plurality of respondents trust local election officials “a moderate amount.” Around 25–30 percent of respondents trust election officials only “a little” or “none at all,” while around 40 percent of respondents trust them “a lot” or “a great deal.”

Figure 2.

Distribution of responses to question, “How much do you trust local election officials to …?”

The second dependent variable is belief that the 2024 election will be run fairly, measured using:

“Do you believe the 2024 election will be run fairly? I.e., will the results be honest and accurate?”

Response options include “yes,” “no,” and “not sure.” As our goal is to assess trust, we dichotomize this measure such that 1 = “yes” and 0 = “no” (with “not sure” grouped with “no”). ³

Figure 3 shows the distribution of responses to this question. A slim majority of respondents (51 percent to be exact) expect, heading into the 2024 election, that the election will be run fairly. This variable caputres broader trust in the overall electoral process, complementing the more targeted trust in local electoin officials questions. While most respondents trust local election officials a moderate amount or more, many remain skeptical of the fairness of the broader election. Including these two separate measures of trust allows us to assess whether perceptions of local election officials are associated not only with evalautions of those officials specifically, but also with confidence in the election broadly.

Figure 3.

Distribution of responses to question, “Do you believe the 2024 election will be run fairly? I.e., will the results be honest and accurate?”

Predictors and Controls

For our primary analysis of racial perceptions, our key predictors of interest are respondent party identification (seven-point scale), conspiratorial thinking, ⁴ and local demographic context (county-level percent non-white population from the 2024 American Community Survey, ⁵ county-level percent rural population from the 2020 decennial census, and county-level margin for the 2020 Republican presidential candidate). We also include controls for political interest, family income, education, age, gender (dummy: 1 = male), and respondent race (dummies for white and Black respondents, with other races as the reference category).

For our secondary trust analysis, we include the perceived percentage of election officials who are Black, Latino, and other race as the key independent variables (with white as the omitted reference category), along with the same set of individual-level controls: party identification, political interest, conspiracism, family income, education, age, and gender (Brown et al. 2024; Shino and Smith 2025).

Modeling Strategy

For the primary analysis, we estimate a seemingly unrelated regression (SUR) model. Because the perceived racial composition of election officials is a set of proportions that must sum to one, standard linear regression of each category independently is inappropriate—it cannot enforce the constraint that changes in one component necessarily imply offsetting changes in the others. To address this, we construct three log-ratio dependent variables using the perceived white share as the baseline category: ln $\left(\frac{\%Black}{\%White}\right)$ , ln $\left(\frac{\%Latino}{\%White}\right)$ , and ln $\left(\frac{\%Other}{\%White}\right)$ . These three equations are estimated jointly via SUR, which accounts for the cross-equation correlation in errors that is inherent in compositional data (Eastman et al. 2024; Philips, Rutherford and Whitten 2015, 2016). Predicted racial shares are then recovered via the softmax (inverse logit) transformation, and we employ a nonparametric bootstrap procedure with 1,000 simulations to construct confidence intervals around predicted compositions at different levels of the predictors.

For the secondary trust analysis, we estimate linear regression models on our trust in local election official index and logistic regression models on the election fairness dependent variable. These models are estimated separately by respondent race (white, Black, and Latino subsamples) to capture the theoretically expected in-group versus out-group dynamics. However, the Black and Latino respondent subsamples are quite small (154 and 87 observations in each group, respectively), and results should be interpreted with caution.

Findings

What Predicts Racial Perceptions of Election Officials?

We begin with what explains variation in how Americans perceive the racial composition of election officials. Table 1 presents the coefficient estimates from the SUR model, showing how our predictors relate to each log-ratio component of perceived racial composition (Black/White, Latino/White, and Other/White). Positive coefficients indicate that a predictor is associated with perceiving a higher share of that group relative to white officials.

Table 1.

Seemingly unrelated regression (SUR) models predicting logged racial perceptions of local election officials

​	Dependent variable:
	ln (% Black/% White)	ln (% Latino/% White)	ln (% Other/% White)
Party ID (R)	0.116*** (0.026)	0.063* (0.034)	0.113*** (0.041)
Conspiracism	−0.057 (0.051)	−0.028 (0.068)	0.006 (0.082)
Political interest	0.136** (0.064)	0.113 (0.085)	0.032 (0.103)
Family income	−0.034** (0.017)	0.007 (0.023)	0.013 (0.028)
Education	−0.097** (0.042)	−0.054 (0.055)	0.047 (0.067)
Age	0.005 (0.003)	−0.0003 (0.004)	−0.013** (0.005)
White	0.162 (0.161)	−0.198 (0.212)	−0.277 (0.258)
Black	0.685*** (0.202)	−0.492* (0.266)	−0.361 (0.324)
Male	−0.082 (0.110)	0.060 (0.145)	0.311* (0.176)
County % Non-white	0.011*** (0.004)	0.008* (0.005)	0.008 (0.006)
County % Rural	0.003 (0.003)	0.001 (0.004)	0.002 (0.005)
County 2020 GOP Margin	−0.002 (0.002)	−0.001 (0.003)	−0.001 (0.004)
Constant	−2.118*** (0.391)	−2.623*** (0.514)	−3.381*** (0.628)
Observations	748	748	748
R 2	0.080	0.005	0.020

Note: *p < 0.1; **p < 0.05; ***p < 0.01.

Consistent with Hypothesis 1, partisanship is a strong and consistent predictor of racial perceptions across all three log-ratio equations. Moving from the Democratic to the Republican end of the party identification scale is associated with perceiving a greater share of election officials as non-white relative to white—the effect is statistically significant for the Black/White ratio (p < 0.001) and the Other/White ratio (p = 0.006), and approaches significance for the Latino/White ratio (p = 0.064). This finding is consistent with the elite-driven racialization of election administration narratives, in which Republican leaders have repeatedly cast doubt on elections in heavily minority urban areas. Republicans appear to have internalized a perception of election administration as less white than it actually is. ⁶

In contrast, we find no evidence supporting Hypothesis 2. Conspiratorial thinking is not independently associated with more racialized perceptions of election officials. The coefficients are small, inconsistent in direction across equations, and fail to reach conventional levels of statistical significance. Recall that our dependent variable is the composition of election officials, not their intention to engage or not engage in conspiratorial election acts (like election fraud: discarding ballots, altering counts, etc. (Brown et al. 2026)). While other work finds that both left- and right-leaning individuals endorse conspiratorial views when they are connected to political losses (Enders et al. 2023), we do not observe that behavior extends to a broader belief about the racial identity of election administrators, absent an explicit connection to their administrative conduct. Taken together, these results suggest that racialized perceptions of election administration are structured primarily by partisan identity rather than generalized conspiratorial worldviews.

We find some evidence for Hypothesis 3: county demographic context shapes racial perceptions of election officials, but the effects are substantively small. The effects are especially concentrated between a tradeoff of perceived percentage of white versus Black election officials. In a mostly white county (5 percent nonwhite), the estimated percentage of white election officials is 74 percent (against 16 percent Black election officials, 7 percent Latino election officials, and 3 percent other race). If we shift this to a 65 percent nonwhite county, the estimated percentage of Black election officials is 25 percent against 61 percent white election officials (Latino and other at 9 percent and 4 percent, respectively). These changes, while statistically significant, necessitate drastic comparisons between counties to observe substantive significance. Otherwise, county-level rural population share, 2020 GOP vote margin, and other contextual indicators are not reliably associated with perceived composition. These results provide minimal support for Hypothesis 3 and suggest that respondents are not systematically updating based on their county demographic or political environments.

To interpret the magnitude of the support for Hypothesis 1, Figure 4 illustrates the substantive magnitude of these effects by displaying the predicted composition of election officials across levels of party identification. The predicted values were estimated holding all other variables at their means and setting respondent race to white and gender to male. Figure 4 reveals a clear relationship between party identification and perceived composition. Strong Democrats predict that roughly 73 percent of election officials are white, while strong Republicans predict only about 59 percent—a gap of 14 percentage points. The corresponding perceived share of Black officials rises from 17 percent among Strong Democrats to 27 percent among Strong Republicans. Bootstrapped 95 percent confidence intervals confirm that the composition shift across the partisan spectrum is statistically meaningful. Alternatively, there’s a null relationship between conspiratorial thinking and racial perceptions. Those lowest in conspiratorial thinking are no more (or less) likely to perceive a whiter set of election officials than those highest in conspiracism (65 percent versus 68 percent).

Figure 4.

Predicted perceived racial composition of election officials by party identification, from SUR model with bootstrapped predictions (1,000 simulations). All other variables held at their means; predictions for white male respondents.

Why It Matters: Racial Perceptions and Trust

Having established that racial perceptions of election officials are politically structured, we now examine the downstream association between these perceptions and electoral trust. We present this analysis as explicitly associational, acknowledging that perceptions and trust may be mutually reinforcing rather than strictly unidirectional.

Figure 5 presents coefficient estimates from the trust and fairness models side by side, with all predictors rescaled to range from 0 to 1 for comparability (full regression results on the original scales appear in Tables SM1 and SM2 of the Supplemental Materials). The right panel of Figure 5 (Table SM1 in the Supplemental Materials) shows that white respondents express significantly lower trust in local election officials when they perceive those officials to be Black (p = 0.013) or Latino (p = 0.033). For Black and Latino respondents, the racial perception coefficients are not statistically significant, though the small subsample sizes limit our ability to detect effects. The direction of effects for Latino respondents is suggestive of a positive relationship between perceiving more co-racial officials and trust.

Figure 5.

Association between perceived racial composition of election officials and belief in fairness of the election (logistic regression) (left) and trust in local election officials index (linear regression) (right), by respondent race.

Results are similar when we turn to perceptions of election fairness, presented in the left panel of Figure 5 (Table SM2 in the Supplemental Materials). White respondents who believe more Black officials run elections are significantly less likely to say the election will be fair (p = 0.002). Perceiving more Latino or other-race officials is not significantly associated with fairness beliefs among white respondents. For Black and Latino respondents, the racial perception coefficients do not achieve statistical significance in the fairness models, again likely reflecting the small subsample sizes.

As white respondents’ estimates of the percentage of Black election officials increase from 0 to 50 percent (assuming this trades off with their perception of white election officials, as Figure 4 suggests)—holding all other covariates at their means—the predicted probability of perceiving the election as fair declines sharply. A white respondent who perceives no election officials are Black has a predicted probability of approximately 0.67 of believing the election will be fair. At 25 percent, this probability falls to roughly 0.53, and at 50 percent it drops to approximately 0.38. The 29-percentage point decline underscores that the association between racial perceptions of election officials and fairness beliefs is statistically significant and substantively consequential.

These patterns are broadly consistent with Hypotheses 4 and 5, at least among white respondents. White respondents are significantly less trusting when they perceive there to be more Black and Latino election officials. The results for Black and Latino respondents are less conclusive due to small sample sizes, though the direction of effects for Latino respondents is generally consistent with a co-racial trust dynamic. The clearest takeaway is that among the largest racial group in the sample—white Americans—perceiving election officials as more non-white is associated with lower trust in elections. But, as we learned earlier, these perceptions are not necessarily grounded in reality.

Conclusion

Our findings extend the racial typification literature in an important direction. Udani, Manion and Kimball (2024) have shown that white Americans “other” illegal voters, overestimating the proportion who are racial minorities and favoring harsher punishments for voters of color. We show that this pattern of othering extends to the people who run elections. Just as Americans racially typify criminals, they racially typify election administrators—and the same political forces that drive the typification of lawbreakers appear to drive the typification of the bureaucrats charged with administering the law.

We also provide evidence that these racialized perceptions are associated with lower trust in elections among white Americans. While we are careful to present this trust analysis as associational rather than causal—the endogeneity of perceptions and trust is a genuine concern—the pattern is substantively important. White respondents who perceive election officials as more non-white express lower confidence in both local election officials and the broader fairness of the electoral process. This finding has troubling implications for a diversifying America. If confidence in elections erodes not because of performance failures but because of who voters perceive to be in charge, then election administration may be vulnerable to the same dynamics that have plagued other public goods in America. McGhee (2022) has documented how racial diversification eroded white support for schools, parks, swimming pools, and public infrastructure—the “drained pool” politics in which the majority group withdraws support from shared institutions as those institutions become more racially diverse. Our findings suggest elections may be next.

This makes low confidence in US elections a far more intractable problem than any story about voting machines, voter ID laws, or other electoral rules. Standard administrative reforms, such as improving procedures, increasing transparency, or auditing results, may be insufficient to address distrust rooted in racialized perceptions of institutional actors. Addressing this challenge will require both good election administration and a reckoning with the racial dynamics that shape how Americans perceive and evaluate the people who run their democracy. In this sense, our findings echo those of Brown et al. (2024), where the effectiveness of messages designed to restore trust in election officials on the basis of professional experience or shared community was conditional on shared racial identification with the messenger.

Several limitations of our analysis warrant discussion. The Black and Latino subsamples in our survey are small, limiting our ability to draw firm conclusions about minority respondents’ perceptions and trust. Our survey question about election officials does not specify a particular role, leaving ambiguity about whether respondents were thinking of poll workers, county clerks, or state election directors. We assess conspiratorial thinking with a single question rather than the full four-item American Conspiracy Thinking Scale index. Each of these limitations can be addressed by future work with a more robust sample and alternative question wording.

Most importantly, our cross-sectional data cannot establish causal ordering between perceptions and trust. While our theoretical framework posits that perceptions of descriptive representation shape trust, a substantial body of literature suggests the reverse pathway is plausible. We suspect the two operate as mutually reinforcing processes. Kunda (1990) establishes that directional goals bias cognitive processing, such that individuals arrive at conclusions they are motivated to reach while maintaining an “illusion of objectivity.” Applied to our context, respondents who already distrust election officials may be motivated to perceive those officials as demographically dissimilar from themselves. Taber and Lodge (2006) extend this logic to political attitudes and demonstrate that prior beliefs generate both confirmation bias (favorable treatment of attitude-consistent information) and disconfirmation bias (skeptical scrutiny of attitude-inconsistent information). A respondent who distrusts elections may thus selectively attend to or overestimate the presence of out-group officials while discounting evidence of in-group representation.

Future work using experimental or panel designs could help isolate whether racialized perceptions causally reduce trust, whether distrust generates racialized perceptions, or whether both processes operate simultaneously. For example, experimental vignettes might help to causally identify how implicit racial descriptions—like discussions of election fraud in urban areas—drive racial typifications of the election officials who work there (which would identify a key causal mechanism). Conjoint experiments could also be used to causally identify the effect of different racial compositions of elections administrators on trust in elections. Our results indicate that identifying both is essential to understanding questions of confidence and trust in elections. Despite these limitations, our fundamental finding remains that Americans’ perceptions of who runs elections are racially distorted, these distortions are politically structured, and they are associated with consequential differences in democratic trust.

Supplemental Material

Supplemental Material - The Racial Typification of Election Officials and Trust in Elections

Supplemental Material for The Racial Typification of Election Officials and Trust in Elections by Isaac Westfall, Soren Jordan, Spencer Goidel, Mitchell Brown in Political Research Quarterly.