The Southern Accent as a Heuristic in American Campaigns and Elections

Abstract

Are political candidates perceived differently based on the presence or absence of a southern accent? To address this question, we employ an experimental design that explores reactions to political candidates with a southern accent and a regionally neutral accent. We focus on perceptions of three general categories of candidate characteristics: candidate traits, candidate affect, and candidate issue positions. Overall, we discover that candidates with a southern accent are viewed more negatively, and they are thought to hold more conservative policy positions, than candidates with no discernible accent. Our findings suggest that the southern accent provides a heuristic that affects how voters perceive candidates.

Keywords

southern politics campaigns and elections political psychology voice

The American voter is awash with information. As a result, voters use information heuristics (also called candidate cues), to evaluate candidates for elected office (Miller & Shanks, 1996). Commonly explored heuristics range from the political, such as partisanship (Conover & Feldman, 1989) and incumbency status (Ansolabehere, Snyder, & Stewart, 2000); to the demographic, such as race (Matsubayashi & Ueda, 2011), gender (McDermott, 1997), and age (Webster & Pierce, 2018); to the trivial, such as appearance (Banducci, Karp, Thrasher, & Rallings, 2008; Todorov, Mandisodza, Goren, & Hall, 2005), ballot order (Krosnick, Miller, & Tichy, 2004), and voice pitch (Tigue, Borak, O’Connor, Schandl, & Feinberg, 2012).¹ Given the continued importance of region in voting and political behavior, regional accents may also provide a salient heuristic that provides information about a political candidate.

In this study, we focus on the southern accent, one of the most recognizable regional accents in the United States. Linguist Michael Montgomery (1993) argues that the southern accent, “is a kind of badge that signifies their self-identification as southern,” and that it reflects a person’s “upbringing, their loyalties, their education, and their roots” (p. 50). This badge does not come with uniformly positive connotations, however (Griffin, 2000). There is potential for voters to form opinions, particularly negative opinions, about political candidates with a southern accent. Even presidents are not immune from the negative perceptions that come with this type of voice. While giving a laudatory speech about President George W. Bush on the House floor, Representative Louis Gohmert (R-TX) noted,

He’s smarter than people give him credit, but as Jeff Foxworthy says, often when people who are not from the South hear a southern accent, they immediately deduct 50 IQ points from what they think the IQ of the speaker is. (155 Cong. Record, 2009)

Although southern accents can and do vary within the South (Reed, 2000; Wolfram & Reaser, 2014), people from a variety of regions are generally able to identify the southern accent as distinctive and discern the strength of a person’s southern accent (Boucher, Hammock, McLaughlin, & Henry, 2013; Fridland, 2008; Hartley, 2005). Furthermore, people with stronger southern accents are “more ‘southern’ in everything from their religious beliefs to their dietary preferences” (Reed, 2000, p. 102).

Despite the importance of candidate cues and the negative stereotypes that can be associated with the southern accent, we do not know how voters respond to political candidates who have one. In this article, we employ an experimental design to examine responses to male and female candidates with a regionally neutral accent and a southern accent. Discovering whether the accent provides a cue to voters will inform a growing literature about the vocal aspects of American campaigns and elections (Klofstad, 2015; Searles, Fowler, Ridout, Strach, & Zuber, 2017), and will contribute to the long-standing literature about the lingering importance of region in American politics (Black & Black, 2007). Given the role of a candidate’s voice in the modern campaign, our findings also suggest that candidates with a southern accent may need to be particularly attuned to what and how they communicate with voters.

Accents and Politics

Accents are frequently used to classify group membership and associate individuals with certain traits such as race, class, ethnicity, and education (Boucher et al., 2013; Giles & Billings, 2004; Rakic, Steffens, & Mummendey, 2011; Riches & Foddy, 1989). Although there are a variety of ways to categorize accents, the most common distinction is between “standard” and “nonstandard” groupings. “Standard” accents are associated with people in the mainstream, whereas “nonstandard” accents are generally associated with low socioeconomic status (Fishman, 1971).²

The southern accent has a long lineage and is perhaps the most recognizable nonstandard accent in the United States. The accent can be traced to at least the nation’s founding, but sociolinguists have discovered that differences between northern and southern speech increased rather substantially after the Civil War (Bailey, 2001). According to Walt Wolfram and Jeffrey Reaser (2014), “The post-Civil War cultural divide, the regional entrenchment, and the oppositional feelings toward the North offered a fertile environment for the independent development of southern speech” (p. 14). The continued resilience of “the South” as a social construct (Cooper & Knotts, 2017) suggests that the southern accent is likely to be relevant in the future. And, although there may be differences between the accents of Black and White southerners, “there can be no doubt that the vocabulary, pronunciation, and grammar of blacks and whites in the South are far more alike than different . . . ” (Montgomery, 1993, p. 56).

Individuals with a southern accent are perceived differently than those with a standard accent, and these differences are often negative and associated with stereotypical views of southerners (Fridland & Bartlett, 2006). Heaton and Nygaard (2011) found that standard-accented speakers were considered more intelligent, educated, and arrogant (indicators of higher status), whereas southern-accented speakers were considered more amusing, friendly, polite, nice, and cheerful (indicators of greater sociability). Wolfram and Reaser (2014) argue that “the collection of stereotypes results in an odd image of what a southerner must be: rustic, slow, and ignorant, yet genteel, kind, and pleasant” (p. 15). Others find that, on average, those with a southern accent are perceived as less competent and considered to speak with poorer “language correctness” and good grammar. Researchers, using a variety of samples, have replicated the findings about intelligence and friendliness (Boucher et al., 2013; Preston, 1999). Psychologists have shown that even young children classify individuals with northern accents as smart and southern accents as nice (Kinzler & DeJesus, 2012).

Because people often infer trait impressions from accents, it is not surprising that accents have consequences in a variety of social contexts. For example, people with nonstandard accents are more likely to be discriminated against in the job hiring process (Atkins, 1993; De La Zerda & Hopper, 1979; Giles, Wilson, & Conway, 1981; Hopper & Williams, 1973; Kalin & Rayko, 1978; Rey, 1977, but see Cargile, 2000). Southern accents can also have a negative effect in the marketplace. A recent study found that participants rated products advertised by a speaker with a southern accent less positively than those advertised by speakers with British accents (Morales, Scott, & Yorkston, 2012). This finding prompted the authors to note that “ . . . if Simon Cowell on American Idol spoke with a Southern accent rather than a British accent, perhaps his critiques would be discounted and appear less diagnostic” (Morales et al., 2012, p. 43). Although it is not much of a theoretical leap to suggest that the southern accent should affect politicians and political candidates in a similar manner, this topic remains unexplored.

Although we know very little about the political effects of the southern accent, a recent spate of work suggests that another vocal characteristic, voice pitch, affects perceptions of candidates and ultimately, vote choice. Experimental research suggests that voters express greater preference for presidents with lower voice pitch and perceive them to have greater integrity and physical prowess than their higher pitched counterparts (Tigue et al., 2012). Klofstad, Anderson, and Peters (2012) found that the preference for lower voice pitch applies to both male and female candidates. Female candidates with lower pitch were perceived by men and women as having attributes that indicate leadership capabilities (competence, strength, and trustworthiness). Male candidates with lower pitched voices were seen by women as more attractive (an indicator of electability) and by males as stronger and more competent. There is also evidence that candidates with lower pitched voices receive higher vote share in elections (Klofstad, 2015).

Hypotheses

With the knowledge that voice characteristics affect voter perception of candidates, and that the southern accents of ordinary citizens generate consistent stereotypes, we expect that listeners will infer personal characteristics about candidates based on whether or not they speak in this accent. We classify these characteristics into two categories—candidate traits and candidate affect. By candidate traits, we mean traits that participants or voters ascribe to candidates—such as intelligence, trustworthiness, and honesty. By candidate affect, we mean general positive or negative reactions to the candidate. For our first hypothesis, we predict that respondents will evaluate candidate traits of southern-accented candidates more negatively than candidates with a neutral accent. For our second hypothesis, we expect that respondents will express more negative affect towards southern-accented candidates, as compared to candidates with a neutral accent.

In addition, we expect that those who are exposed to a candidate with a southern accent also will use it to infer political information about the candidate. The role of inference for this type of candidate perception is well documented (Conover & Feldman, 1989; Feldman & Conover, 1983). When voters have limited information, or candidates are ambiguous about their political stances, individuals use the information they do have to fill in gaps in their mental portrait of a candidate (Conover & Feldman, 1984). Gender and race, for example, are often used to infer candidate ideology (Koch, 2000; McDermott, 1997; McDermott, 1998). Voters believe that women candidates, regardless of party, are more liberal than their male counterparts (Alexander & Andersen, 1993; Dolan, 2014; Koch, 2000) and that they have greater expertise and interest in certain policy areas such as reproductive rights and education (Dolan & Lynch, 2014; Fulton, 2012; Kahn, 1992; Koch, 1999).

Because accents are also a prominent, outward characteristic, we expect them to serve similar functions in the formulation of political judgments. Given the political transformation of the South toward a conservative, Republican-dominated region (Black & Black, 2002), we hypothesize that respondents will classify southern-accented candidates as more politically conservative (in general) than candidates without a southern accent, and more conservative across a range of specific policy positions such as gun rights, women’s rights, abortion, and defense spending.

Although the hypotheses above summarize our expectations for the main effects of our model, we do expect that certain factors may moderate the perceptions of the accent. For example, our predictions differ somewhat for male and female candidates. Women who run for office are thought to be more caring, ethical, honest, and trustworthy, and simultaneously perceived to be less assertive, less knowledgeable, and weaker than their male counterparts (Bauer, 2015; Dolan, 2014; Fox & Smith, 1998; Kahn, 1992; but see Brooks, 2011; Schneider & Bos, 2014). If a female candidate also has a southern accent, she may face voter stereotypes stemming from both her gender and her accent. In terms of candidate traits, we expect that female candidates with a southern accent will be perceived not only as more positively than male candidates on certain traits (honesty and trustworthiness) but also as less competent or intelligent. We also expect that southern-accented females will be perceived more negatively across all candidate traits than female candidates without the accent.

We also hypothesize that southern identity of the respondent will affect perceptions of the southern accent. Just as other shared group memberships such as race or gender can create a sense of linked fate in which an “individual’s life changes are inextricably tied to the group as a whole” (Simien, 2005, p. 529), we expect participants who identify as southern to express characteristics of linked fate that will mute the negative effects of the accent. There is evidence that those who speak with the accent of the in-group are trusted more than those who speak with the accent of an out-group (Birch & McPhail, 1999; DeShields & de los Santos, 2000; Fuertes, Gottdiener, Helena, Gilbert, & Giles, 2012), leading us to believe that accent may cue these feelings of linked fate. Similarly, linguists have argued that there may be a “hidden value associated with nonstandard speech” (Trudgill, 1972, p. 183), and speaking with a regional dialect can provide “covert prestige,” where the speaker is perceived by people who share the accent as having more positive attributes—precisely because they know that the speaker is robbed of “overt prestige” by society at large (Wolfram & Reaser, 2014). Therefore, we expect that southern participants will feel a sense of kinship with southern-accented candidates and, as a result, not levy a sanction against them.

Data and Method

To test our hypotheses, we used a 2 × 2 survey experiment, in which respondents listened to a 1-min campaign speech by a (fictitious) political candidate. We began with two speakers, one male and one female. Each speaker then produced two speeches—one with a southern accent and one with a regionally neutral accent.³ This yielded four conditions to which a respondent could be randomly assigned: a male candidate with a southern accent (male southern), a male candidate with a neutral accent (male neutral), a female candidate with a southern accent (female southern), or a female candidate with a neutral accent (female neutral). To record these speeches, we recruited two adults with acting experience to each produce a southern-accented version of the speech and a standard-accented version of the speech, resulting in four separate speeches. The speakers are both Caucasian, in their early 50s, born in the South, and currently living in a medium-sized southern city.

The speakers were given instructions to enunciate clearly and emphasize the same highlighted words in the script to keep intonation as uniform as possible. The audio files were inspected manually to ensure that the readings sounded similar. The gain on each file was then adjusted in Adobe Premier, an audio and video editing platform, to make the volume equivalent across recordings. The final four audio files (aiff format) were all between 74 and 81 s, and embedded into the survey. Participants had to click a “Play” button on the screen for the audio to begin. There were no accompanying visuals.

When creating the content of the script (see Appendix A), we purposefully avoided topics that were overtly partisan or ideological, reducing the probability that the participant would respond to the policy position rather than the candidate’s accent. For example, if the candidate discussed supporting prayer in schools, the respondent might perceive that candidate to be “extremely conservative” (see dependent variables below) regardless of the accent. Therefore, the speech focused on the candidate’s own life and on topics that are generally politically neutral such as job creation.

Participants were randomly assigned to one of the conditions and told that the recording came from a phone message that a political candidate used when calling potential voters. After listening to the speech, the dependent variables and controls were measured for all participants. The entire survey contained 31 questions and took the average participant slightly less than 9 min to complete.

We measured perceived candidate traits and candidate affect responses using 5-point Likert-type scales. The response categories were strongly disagree to strongly agree in response to the following statements: “The candidate is trustworthy,” “The candidate is honest,” “The candidate is intelligent,” “The candidate is competent,” “The candidate would look out for people like me,” and “I like this candidate.” In addition, respondents indicated whether they would consider voting for this candidate using a binary yes/no response.

Candidate issue positions responses were measured on a standard, 7-point ideology scale ranging from extremely liberal to extremely conservative. Participants were also asked to estimate the candidates’ political stances on 5-point scales ranging from strongly oppose to strongly support for the following four issues: the rights of gun owners, women’s rights, a woman’s right to an abortion, and increasing national defense spending. Full question wording can be found in Appendix A, and descriptive statistics for all of the variables in our primary models can be found in Table 1.

Table 1.

Descriptive Statistics.

Variable	N	M	SD	Minimum	Maximum
Trustworthy	757	3.303	0.797	1	5
Honest	757	3.378	0.820	1	5
Intelligent	757	3.588	0.795	1	5
Competent	757	3.598	0.773	1	5
Positive attributes	757	13.868	2.568	4	20
Would look out for me	757	3.370	0.869	1	5
Like candidate	757	3.314	0.918	1	5
Would vote for candidate	757	0.7107	0.454	0	1
Candidate conservatism	757	4.336	1.298	1	7
Candidate supports guns	757	3.371	0.960	1	5
Candidate supports women	757	3.654	0.967	1	5
Candidate supports abortion	757	2.812	0.982	1	5
Candidate supports defense spending	757	3.116	0.974	1	5
Southern identity	757	0.446	0.497	0	1

Source. Authors’ experiment.

Our sample consists of 757 participants from Alabama and Connecticut. We chose these states because both are locations where the southern accent has the potential to be influential: a southern state where the accent is prominent and familiar and a nonsouthern state where the southern accent is atypical. The sample was collected by Qualtrics, a survey and research consulting firm, in May 2017 and was matched to the U.S. Census for age, race, gender, and education level.⁴ As Searles et al. (2017) found, opt-in samples such as this one are useful for experimental work where voice manipulation is the primary variable of interest. To ensure high-quality responses, subjects had to commit to providing their best and most honest answers.⁵ They also had to spend a minimum of 3 min completing the survey.

Our first task was to ensure that the manipulation worked. Because any given voice is accented to some degree, the neutral conditions will never be perfectly “nonaccented.” What is critical, however, is that the southern treatments are perceived as southern and the neutral treatments are not perceived as southern or systematically regional. To determine whether the manipulation worked, we first asked participants, “Do you detect a regional accent?” The majority of participants in the southern conditions detected a regional accent and the majority of participants in the standard conditions did not. This check was placed directly after the dependent variables were measured to avoid making the accent unusually salient when respondents formulated perceptions. A Pearson’s chi-square test indicates that participants were indeed more likely to detect a regional accent in the southern condition, χ²(3, N = 757) = 142.47, p < .000. If a respondent did detect an accent, we then asked him or her to identify what type it was. Overall, 76% of subjects in the southern female condition (n = 190) identified a southern accent. Similarly, 77% of subjects in the southern male condition (n = 187) identified a southern accent. In the female neutral condition (n = 189), only 19% identified a southern accent and in the male neutral condition (n = 191), only 14% of subjects identified a southern accent. The relationship between the conditions and detection of a southern accent was significant, χ²(3, N = 482) = 148.81, p < .000. Although some participants perceived the voices to have alternative accents, such as northern or Midwestern, these responses were far fewer. These answers were also similar between the male and female versions of the readings within the southern and neutral conditions. For example, similar numbers of people (four and three, respectively) felt that the female southern voice and the male southern voice had a northern accent, providing reassurance that two speakers did not have major systematic differences.

Results

We begin our review of the results with a series of models predicting whether participants exposed to our treatment (a candidate with a southern accent) rated the candidate lower on a series of traits than participants who received the control (a candidate with a neutral accent). Each of the first four models examines a separate character trait—trustworthiness, honesty, intelligence, and competence. In the fifth model, we added responses to these questions together to create a scale of positive attributes, ranging from 4 to 20 (α = .82). We estimate all these models with bivariate ordinary least squares (OLS) regressions. Full model results are available in Appendices B and C; we present graphical evidence of the substantive results for the statistically significant relationships in Figure 1.

Figure 1.

Effect of candidate accent on perceptions of candidate traits.

Beginning with the top-left panel of Figure 1, we see that the candidate with a southern accent was considered less trustworthy (p < .05), less honest (p < .05), less intelligent (p < .001), and less competent (p < .001) than their counterparts with a neutral accent. Not surprisingly, the final model (an additive scale of each of the other four dependent variables) also demonstrated that the southern candidates were viewed as having fewer positive attributes than the neutral-accented candidates (p < .001). The differences in candidate assessments for the southern- and neutral-accented candidate are associated with a movement of 0.13 on the Trustworthiness scale, 0.12 on the Honesty scale, 0.27 on the Intelligence scale, and 0.20 on the Competence scale. Each of these scales ranges from 1 to 5. On the Positive Attributes scale (which ranges from 4 to 20), the difference between assessments of the southern- and neutral-accented candidates are associated with a difference of 0.73. These substantive effects could make the difference in a tight race.

These findings are reinforced in a separate investigation, which is available in the online appendix. Specifically, we follow the advice of Gerber and Green (2012) and also estimated models with covariates “to eliminate observed differences between treatment and control groups and to reduce the variability of outcomes” (p. 95). Our findings hold up in this multivariate context—controlling for other potential confounding factors, such as age, sex, race, southern identity, partisan identification, and political ideology.

Recall that while we expected participants to attribute more negative qualities to candidates with a southern accent on average, we expected this effect to be muted for southern participants. We predicted that the southern participants might feel some in-group kinship with the candidate with the southern accent that would cause them to sanction the candidate less than a nonsouthern participant.

To test this expectation, we ran a series of models interacting various measures of southern identity with the treatment group. Summaries of the statistical modeling portion of this investigation are included in Appendices B and C. In brief, there was no evidence of a positive in-group effect. On average, southern participants did not ascribe fewer negative attributes to southern-accented candidates than their nonsouthern counterparts. The same finding holds when we used region of residence, instead of southern identity, as the moderating variable. Having a southern accent, therefore, appears to be as much of a detriment when running in the South as it is outside of the South.

To determine whether the southern sanction was larger for female candidates, we estimated another series of OLS regressions, which included a variable interacting the gender of the candidate’s voice and the regional accent of the candidate’s voice (the results are also available in Appendices B and C). We found no evidence that the tendency for participants to rate the southern-accented candidate as having fewer positive attributes than the neutral-accented candidate varies by gender of the candidate. Separate analyses interacting party identification with the treatment condition also showed no effects. These are available in the online appendix.

Next, we move to a series of models that attempt to discern whether participants have more negative affect toward speakers with a southern accent. Similar to our hypotheses on candidate traits, we expected that, on average, participants would have a more negative affect toward the candidate with the southern accent, particularly if the candidate was female. We also expected that the influence of accent would be muted (or perhaps even reversed) among southern participants. Our results support the first of these predictions. Participants were less likely to indicate that the candidate with the southern accent would “look out for people like me,” less likely to indicate that they “like” the candidate, and less likely to indicate that they would “vote for the candidate.” Graphical results from these models are presented in Figure 2 and full model results are available in Appendix D.

Figure 2.

Effect of candidate accent on candidate affect.

As Figure 2 illustrates, on average, participants who were in the southern accent treatment placed the candidate 0.182 (on a 5-point scale) lower on the “would look out for me” scale than those who listened to the neutral speech. The difference for whether the person “liked” the candidate was 0.267 (also on a 5-point scale). Finally, the sine qua non of electoral politics (would the participant vote for the candidate?) showed substantively important effects as well (0.77 for the neutral-accented candidate vs. 0.65 for the candidate with the southern accent). Once again, these effects hold in the bivariate context and in separate multivariate analyses available in the online appendix. In sum, it is clear that candidates with a southern accent appear to face a sanction that is not faced by candidates with a neutral accent.

As reported in Appendix D, we find no evidence that the effect cited above is attenuated by southern identity. Participants who identify as southerners (either by state of residence or by indicating so on a questionnaire) apply the same sanction to candidates with a southern accent as their counterparts who do not live in the South or identify as a southerner. This runs counter to our initial hypothesis, and may suggest that, at least as it applies to political candidates, southerners do not display a positive in-group bias. Separate analyses interacting party identification with the treatment condition also showed no effects. These analyses are available in the online appendix.

Appendix D also indicates no difference between the southern sanction faced by female and male candidates in two of our three models. In one case, however, the gender of the speaker interacted in important ways with the accent of the speaker. The effect of the southern accent on likability is stronger among female than male candidates F(1, N = 753) = 6.012, p < .05. Figure 3 displays the effects of the southern accent treatment among the two groups and shows that the southern sanction on likability is greater for women than men. Results of our manipulation check appear in Table 2.

Figure 3.

Effect of candidate accent and gender on perceptions of candidate issue positions.

Table 2.

Full Manipulation Check Results.

	Female southern	Female neutral	Male southern	Male neutral
No accent	30 (16%)	107 (57%)	28 (15%)	110 (58%)
Southern	144 (76%)	37 (19%)	144 (77%)	26 (14%)
Northern	4 (2%)	15 (8%)	3 (2%)	20 (10%)
Midwestern	7 (4%)	22 (12%)	8 (4%)	27 (14%)
Western	2 (1%)	2 (1%)	1 (<1%)	2 (1%)
Other	3 (1%)	6 (3%)	3 (2%)	6 (3%)
Total	190	189	187	191

Source. Authors’ experiment.

Note. Southern and northern residents perceived the candidates similarly. Sixty-eight Alabama residents and 76 Connecticut residents perceived the female southern speaker as southern. Eighteen Alabama residents and 19 Connecticut residents perceived the female neutral speaker as southern. Seventy Alabama residents and 74 Connecticut residents perceived the male southern speaker as southern. Nineteen Alabama residents and seven Connecticut residents perceived the male neutral speaker as southern.

The final set of models (presented in Appendices E and F) examines whether candidates with a southern accent are perceived as having a distinct set of policy opinions and ideological beliefs. Candidates with a southern accent are thought to be more conservative than candidates with a neutral accent (p < .05). Similarly, they are perceived as being more likely to support gun rights (p < .001) and more likely to support restrictions on abortion (p < .001). We found no such effect as it relates to women’s rights or defense spending. It is unclear exactly why there was no effect in these two policy areas, but it may be that both are less closely associated with a particular ideology or region than are gun rights and abortion. Concerning women’s rights specifically, it may be that respondents presumed that, like themselves, any candidate would indicate support for women’s rights, if only because it was the socially acceptable position (Sigel, 1996).

Graphical representations of the statistically significant effects can be found in Figure 4. On average, hearing a southern accent (vs. a neutral accent) results in a participant rating a candidate as 0.23 more conservative (on a 7-point scale), 0.28 closer to the gun rights end of a Likert-type scale (with a range of 1-5), and 0.25 closer to the prolife end of the scale (on a 1-5 scale). We find identical results when running multivariate models. These are available in the online appendix.

Figure 4.

Effect of candidate accent on perceptions of candidate issue positions.

Although these policy and ideological positions do not necessarily come with the same stigma as the results reviewed previously, it is clear that participants in our experiment ascribe a set of beliefs to candidates with a southern accent that they do not ascribe to candidates without a southern accent. Once again, this finding is not dependent on any one model specification or operationalization of southern identity. Furthermore, it is not attenuated by southern identity or the gender of the candidate. There is one case, however, where the accent affects perceptions of the candidate’s opinion on women’s rights for Democratic participants, but not for Republican participants. Results of this are available in the online appendix. In sum, the southern accent operates as a consistent heuristic, providing voters with information shortcuts about who a candidate is and what she or he stands for.

Conclusion

A candidate’s voice is key to his or her political campaign. Whether communicating with voters through television advertisements, earned media, radio advertisements, or even YouTube clips, the candidate’s voice—and, therefore, accent—is a necessary and inevitable component of his or her communication style. Despite the ubiquity of regional accents in American political life, there is scant work on the effects of voices in political campaigning and no work that we can find that specifically addresses the effects of regional accent on candidate perceptions. This is a problematic oversight because of the frequency with which we hear candidate voices and the likelihood that people respond in systematically different ways to those with different accents. Indeed, commentary over Jeff Sessions’s June 2017 Senate Intelligence Committee testimony on Russia focused almost as much on his nasally southern drawl (and what that implied about this worldview) as it did on the content of his testimony (Lloyd, 2017). Politicians such as Bill Clinton have even been known to “turn on” a southern accent when they believe it will be beneficial—a phenomenon known as “style shifting” (Seabrook, 2006).

Using an experimental research design, we discover the presence of a southern sanction where participants view southern-accented politicians more negatively than neutral-accented politicians. Specifically, candidates with southern accents are viewed as less trustworthy, less honest, less intelligent, less competent, and less likable than nonaccented candidates. With the one exception, that female candidates with a southern accent are penalized more on likability than their male counterparts, the southern sanction also applies equally to both male and female candidates.

Perhaps, most important, we find that participants are less likely to say they would vote for the candidate with the southern accent. Indeed, employing eight different dependent variables with clear positive and negative poles, there was not one instance of the candidate with the southern accent being perceived more positively than the neutral-accented candidate.

Surprisingly, we found no evidence that any of these effects were muted by the southern identity of the participant. Southern participants (regardless of how “southerness” was measured) rated the candidate with the southern accent no more positively than did their nonsouthern counterparts. This suggests that the southern accent does not cue a positive in-group bias and may not generate covert prestige among southern voters. Consequently, politicians who are fond of “style shifting” may wish to reconsider when and why they move to the southern accent.

We also found that southern accents cause people to perceive candidates as more conservative in overall ideology and on certain issue positions. This suggests that if a southern-accented candidate holds moderate or liberal positions, he or she may need to work harder to communicate his or her agenda to voters to overcome the perceptions associated with this cue.

For scholars who wish to understand the modern campaign, our study adds to a growing literature (Klofstad et al., 2012; Klofstad, 2015; Searles et al., 2017), which finds that the audio component of campaigns and elections can be as important as their visual counterparts. Just as scholars have recently discovered that the mere presence of group images on a television advertisement provides signals to voters (Swigger, 2012), future work should continue to examine whether audio cues of group membership or affinity provide a similar heuristic.

For students of southern politics, our results also have important implications. First, our investigation reinforces the continued importance of region in American politics. Although qualitative, field-based work shows that southern politicians present a host of cues to signal their regional identity to their voters (Fenno, 2000; Glaser, 1996, 2005), our findings suggest that, at least as it relates to accent, those cues may not be received in the ways candidates intend. The question remains, of course, whether this is because regional identity and in-group attachment are fading or whether the southern accent itself simply is not an effective cue of regional rootedness. Future work should tease out this potentially important puzzle.

Although we believe this study has much to offer to our understanding of electoral and southern politics, it has several important limitations. Most obviously, our study is a single-shot experiment and has all the traditional challenges of external validity innate in this research design. We are particularly aware that we are asking voters to provide opinions about a candidate for whom they have no other information. This, of course, is not how modern campaigns and elections operate. In reality, voters develop complex cognitive schemas for politicians that include a host of other potentially competing cues about their policy positions, characteristics, and associations (Conover & Feldman, 1984; Feldman & Conover, 1983; Lodge & Taber, 2013). Although it is possible that the low-information environment we examine here mirrors low-salience elections at the local and state levels, it likely represents the upper bound of accent effects for elections at the federal level. Furthermore, all our speakers are White. As a result, our study uncovers how (presumed) White politicians with a southern accent are perceived, but provides no concrete information about how minority candidates with (or without) a southern accent are perceived. Clearly, this is a limitation, and one that future studies should ameliorate.

In addition, our decision to avoid overtly partisan or ideological speech content had drawbacks. For example, we were not able to say how much the southern accent influences perceptions of conservativeness, gun rights, or abortion policy. The nonpartisan nature of the speech we used also creates a more liberal test of the accent manipulation. Future studies should provide more stringent tests using partisan speech to see whether differences in perception still arise. Finally, our research design leaves us unable to determine whether the participants are responding to the accent as an independent heuristic, or merely responding to the accent as a cue to region of origin. Future studies should provide more complex designs that allow us to untangle this fundamental question.

Despite these limitations, this article has demonstrated that the southern accent can serve as a heuristic for voters. This cue affects voters’ opinions of the candidate, what policy positions they believe the candidate holds, and even whether they plan to vote for the candidate. Given that the southern accent seems to have a largely detrimental effect upon voter perceptions, an effect that persists among both southerners and nonsoutherners and for both male and female candidates, campaigns may wish to attend to the accent of their candidate and carefully consider how they want the candidate to sound to voters.

Footnotes

Appendix A

Appendix B

Candidate Traits I.

	Trustworthy			Honest			Intelligent
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9
Southern treatment	−.130*	−.097	−.122	−.124*	−.077	−.042	−.273***	−.259***	−.307***
Southern treatment	(0.057)	(0.077)	(0.081)	(0.059)	(0.079)	(0.083)	(0.056)	(0.076)	(0.080)
Southern identity		.161			.133			.119
Southern identity		(0.082)			(0.084)			(0.081)
Southern Treatment × Southern Identity		−.080			−.111			−.038
Southern Treatment × Southern Identity		(0.116)			(0.120)			(0.115)
Female treatment			.162*			.305***			.150
Female treatment			(0.081)			(0.083)			(0.079)
Southern Treatment × Female Treatment			−.017			−.166			.065
Southern Treatment × Female Treatment			(0.115)			(0.118)			(0.113)
Constant	3.368***	3.30***	3.28***	3.43***	3.38***	3.28***	3.72***	3.67***	3.64***
Constant	(0.040)	(0.053)	(0.057)	(0.042)	(0.055)	(0.058)	(0.040)	(0.053)	(0.056)
N	757	757	757	757	757	757	757	757	757
Adjusted R²	.005	.009	.012	.004	.005	.023	.028	.030	.039

Source. Authors’ experiment.

Note. Entries are ordinary least squares unstandardized regression coefficients. Standard errors in parentheses.

p < .05. **p < .01. ***p < .001.

Appendix C

Candidate Traits II.

	Competent			Positive attributes
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Southern treatment	−.199***	−.139	−.147	−.725***	−.573*	−.619*
Southern treatment	(0.055)	(0.074)	(0.078)	(0.185)	(0.248)	(0.260)
Southern identity		.175*			.587*
Southern identity		(0.079)			(0.263)
Southern Treatment × Southern Identity					−.371
Southern Treatment × Southern Identity					(0.371)
Female treatment			.181*			.797***
Female treatment			(0.078)			(0.259)
Southern Treatment × Female Treatment		−.141	−.104			−.222
Southern Treatment × Female Treatment		(0.112)	(0.111)			(0.367)
Constant		3.62***	3.60***	14.23***	13.98***	13.8***
Constant		(0.052)	(0.055)	(0.130)	(0.172)	(0.183)
N		757	757	757	757	757
Adjusted R²		.019	.021	.019	.023	.034

Source. Authors’ experiment.

Note. Entries are ordinary least squares unstandardized regression coefficients. Standard errors in parentheses.

*p < .05. **p < .01. ***p < .001.

Appendix D

Candidate Affect.

	Would look out for me			Like candidate			Would vote for candidate
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9
Southern treatment	−.182**	−.189*	−.184*	−.267***	−.215*	−.107	−.622***	−.654**	−.364
Southern treatment	(0.062)	(0.084)	(0.088)	(0.066)	(0.088)	(0.092)	(0.163)	(0.213)	(0.218)
Southern identity		.150			.162			.305
Southern identity		(0.089)			(0.094)			(0.252)
Southern Treatment × Southern Identity		.005			−.123			.043
Southern Treatment × Southern Identity		(0.126)			(0.133)			(0.333)
Female treatment			.147			.426***			.853***
Female treatment			(0.088)			(0.092)			(0.256)
Southern Treatment × Female Treatment			.002			−.323*			−.619
Southern Treatment × Female Treatment			(0.125)			(0.130)			(0.335)
Constant	3.461***	3.39***	3.38***	3.44***	3.37***	3.23***	1.22***	1.10***	0.855***
Constant	(0.044)	(0.058)	(0.062)	(0.046)	(0.061)	(0.064)	(0.123)	(0.157)	(0.158)
N	757	757	757	757	757	757	757	757	757
Adjusted R²/pseudo-R²	.010	.015	.014	.020	.021	.046	.02	.03	.02

Source. Authors’ experiment.

Note. Entries in Models 1 to 6 are ordinary least squares unstandardized regression coefficients. Entries in Models 7 to 9 are logistic regression coefficients. Standard errors in parentheses for all models.

*p < .05. **p < .01. ***p < .001.

Appendix E

Candidate Issue Positions 1.

	Candidate ideology			Candidate supports gun			Candidate supports defense spending
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9
Southern treatment	.233*	.228	.0942	.280***	.228	.375***	.090	.145	.302
Southern treatment	(0.094)	(0.126)	(0.133)	(0.069)	(0.126)	(0.097)	(0.070)	(0.095)	(0.100)
Southern identity		.129			.129			.101
Southern identity		(0.134)			(0.134)			(0.101)
Southern Treatment × Southern Identity		.000			.000			−.124
Southern Treatment × Southern Identity		(0.189)			(0.189)			(0.143)
Female treatment			−.366**			−.134			−.183
Female treatment			(0.132)			(0.096)			(0.099)
Southern Treatment × Female Treatment			.279			−.187			.123
Southern Treatment × Female Treatment			(0.187)			(0.137)			(0.141)
Constant	4.22***	4.16***	4.40***	3.23***	4.16***	3.29***	3.07***	3.02***	3.16***
Constant	(0.066)	(0.087)	(0.093)	(0.048)	(0.087)	(0.068)	(0.050)	(0.066)	(0.070)
N	757	757	757	757	757	757	757	757	757
Adjusted R²	.007	.007	.015	.020	.007	.034	.001	000	.003

Source. Authors’ experiment.

Note. Entries are ordinary least squares unstandardized regression coefficients. Standard errors in parentheses.

*p < .05. **p < .01. ***p < .001.

Appendix F

Candidate Issue Positions 2.

	Candidate supports women			Candidate supports abortion
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Southern treatment	−.097	−.119	−.215*	−.245***	−.154	−.345***
Southern treatment	(0.070)	(0.094)	(0.094)	(0.070)	(0.094)	(0.099)
Southern identity		.166			−.110
Southern identity		(0.099)			(0.101)
Southern Treatment × Southern Identity		.034			−.188
Southern Treatment × Southern Identity		(0.141)			(0.142)
Female treatment			.497***			.088
Female treatment			(0.094)			(0.099)
Southern Treatment × Female Treatment			.226			.197
Southern Treatment × Female Treatment			(0.133)			(0.141)
Constant	3.70***	3.63***	3.45***	2.93***	2.98***	2.89***
Constant	(0.049)	(0.065)	(0.066)	(0.050)	(0.065)	(0.070)
N	757	757	757	757	757	757
Adjusted R²	.001	.008	.102	.014	.025	.023

Source. Authors’ experiment.

Note. Entries are ordinary least squares unstandardized regression coefficients. Standard errors in parentheses.

*p < .05. **p < .01. ***p < .001.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors would like to acknowledge support from a Research and Development Grant and a Humanities and Social Sciences Dean’s Discretionary Award from the College of Charleston and a Summer Research Grant from the Graduate School and Research at Western Carolina University.

Supplemental Material

Supplemental material is available for this article online.

Notes

Author Biographies

Karyn Amira is an assistant professor in the political science department at the College of Charleston.

Christopher A. Cooper is a professor and department head in the department of political science and public affairs at Western Carolina University.

H. Gibbs Knotts is a professor and department chair in the political science department at the College of Charleston.

Claire Wofford is an assistant professor in the political science department at the College of Charleston.

References

Alexander

Andersen

(1993). Gender as a factor in the attribution of leadership traits. Political Research Quarterly, 46, 527-545.

Ansolabehere

Snyder

J. M.

Jr. Stewart

III . (2000). Old voters, new voters, and the personal vote: Using redistricting to measure the incumbency advantage. American Journal of Political Science, 44, 17-34.

Atkins

C. P.

(1993). Do employment recruiters discriminate on the basis of nonstandard dialect? Journal of Employment Counseling, 30, 108-118.

Atkinson

M. D.

Enos

R. D.

Hill

S. J.

(2009). Candidate faces and election outcomes: Is the face–vote correlation caused by candidate selection? Quarterly Journal of Political Science, 4, 229-249.

Bailey

(2001). The relationship between the AAVE and White vernaculars in the American South: Some phonological evidence. In Lanehart

(Ed.), Sociocultural and historical contexts of African American vernacular English (pp. 53-92). Philadelphia, PA: John Benjamins.

Banducci

S. A.

Karp

J. A.

Thrasher

Rallings

(2008). Ballot photographs as cues in low-information elections. Political Psychology, 29, 903-917.

Bauer

N. M.

(2015). Who stereotypes female candidates? Identifying individual differences in stereotype reliance. Politics, Groups, and Identities, 3, 94-110.

Birch

McPhail

(1999). Does accent matter in international television advertisements? International Journal of Advertising, 18, 251-268.

Black

(2002). The rise of southern Republicans. Cambridge, MA: Harvard University Press.

10.

Black

(2007). Divided America: The ferocious power struggle in American politics. New York, NY: Simon & Schuster.

11.

Boucher

C. J.

Hammock

G. S.

McLaughlin

S. D.

Henry

K. N.

(2013). Perceptions of competency as a function of accent. Psi Chi Journal of Psychological Research, 18(1), 27-32.

12.

Brooks

D. J.

(2011). He runs, she runs: Why gender stereotypes do not harm women candidates. Princeton, NJ: Princeton University Press.

13.

Cargile

A. C.

(2000). Evaluations of employment sustainability: Does accent always matter? Journal of Employment Counseling, 37, 165-177.

14.

Conover

P. J.

Feldman

(1984). How people organize the political world: A schematic model. American Journal of Political Science, 28, 95-126.

15.

Conover

P. J.

Feldman

(1989). Candidate perception in an ambiguous world: Campaigns, cues, and inference processes. American Journal of Political Science, 33, 912-940.

16.

Cooper

C. A.

Knotts

H. G.

(2017). The Resilience of southern identity: Why the South still matters in the minds of its people. Chapel Hill: University of North Carolina Press.

17.

De La Zerda

Hopper

(1979). Employment interviewers’ reactions to Mexican American speech. Communication Monographs, 46, 136-134.

18.

DeShields

O. W.

Jr. de los Santos

(2000). Salesperson’s accent as a globalization issue. Thunderbird International Business Review, 42, 29-46.

19.

Dolan

(2014). When does gender matter? Women candidates and gender stereotypes in American elections. New York, NY: Oxford University Press.

20.

Dolan

Lynch

(2014). It takes a survey: Understanding gender stereotypes, abstract attitudes, and voting for women candidates. American Politics Research, 42, 656-676.

21.

Feldman

Conover

P. J.

(1983). Candidates, issues and voters: The role of inference in political perception. The Journal of Politics, 45, 810-839.

22.

Fenno

R. F.

Jr. (2000). Congress at the grassroots: Representational change in the South, 1970-1998. Chapel Hill: University of North Carolina Press.

23.

Fishman

J. A.

(1971). Sociolinguistics: A brief introduction. Rowley, MA: Newbury House.

24.

Fox

R. L.

Smith

E. R.

(1998). The role of candidate sex in voter decision-making. Political Psychology, 19, 405-419.

25.

Fridland

(2008). Regional differences in perceiving vowel tokens on southerness, education, and pleasantness rating. Language Variation and Change, 20, 67-83.

26.

Fridland

Bartlett

(2006). Correctness, pleasantness, and degree of difference ratings across regions. American Speech, 81, 358-386.

27.

Fuertes

Gottdiener

Helena

Gilbert

Giles

(2012). A meta-analysis of the effects of speakers’ accents on interpersonal evaluations. European Journal of Social Psychology, 42, 120-133.

28.

Fulton

S. A.

(2012). Running backwards and in high heels: The gendered quality gap and incumbent electoral success. Political Research Quarterly, 65, 303-314.

29.

Gerber

A. S.

Green

D. P.

(2012). Field experiments: Design, analysis, and interpretation. New York, NY: W. W. Norton.

30.

Giles

Billings

(2004). Assessing Language Attitudes: Speaker Evaluation Studies. In Davies

Elder

(Eds.), Handbook of applied linguistics (pp. 187-209). Oxford, UK: Blackwell.

31.

Giles

Wilson

Conway

(1981). Accent and lexical diversity as determinants of impression formation and perceive employment suitability. Language Sciences, 3, 91-103.

32.

Glaser

(1996). Race, campaign politics and the realignment in the South. New Haven, CT: Yale University Press.

33.

Glaser

(2005). The hand of the past in contemporary southern politics. New Haven, CT: Yale University Press.

34.

Griffin

L. J.

(2000). Southern distinctiveness, yet again, or, why America still needs the South. Southern Cultures, 6(3), 47-72.

35.

Hartley

L. C.

(2005). The consequences of conflicting stereotypes: Bostonian perceptions of U.S. dialects. American Speech, 80, 388-405.

36.

Heaton

Nygaard

L. C.

(2011). Charm or harm: Effect of passage content on listener attitudes toward American English accents. Journal of Language and Social Psychology, 30, 202-211.

37.

D. E.

Imai

(2008). Estimating causal effects of ballot order from a randomized natural experiment: The California alphabet lottery, 1978-2002. Public Opinion Quarterly, 72, 216-240.

38.

Hopper

Williams

(1973). Speech characteristics and employability. Communications Monographs, 40, 296-302.

39.

Kahn

K. F.

(1992). Does being male help? An investigation of the effects of candidate gender and campaign coverage on evaluations of U.S. senate candidates. The Journal of Politics, 54, 497-517.

40.

Kalin

Rayko

D. S.

(1978). Discrimination in evaluative judgments against foreign-accented job candidates. Psychological Reports, 43(Suppl. 3), 1203-1209.

41.

Kinzler

K. D.

DeJesus

J. M.

(2012). Northern = smart and Southern = nice: The development of accent attitudes in the United States. The Quarterly Journal of Experimental Psychology, 66, 1146-1158.

42.

Klofstad

C. A.

(2016). Candidate voice pitch influences election outcomes. Political Psychology, 37, 725-738.

43.

Klofstad

C. A.

Anderson

R. C.

Peters

(2012). Sounds like a winner: Voice pitch influences perception of leadership capacity in both men and women. Proceedings: Biological Sciences, 279(1738), 2698-2704.

44.

Koch

J. W.

(1999). Candidate gender and assessments of senate candidates. Social Science Quarterly, 80, 84-96.

45.

Koch

J. W.

(2000). Do citizens apply gender stereotypes to infer candidates’ ideological operations? The Journal of Politics, 62, 414-429.

46.

Krosnick

J. A.

Miller

J. M.

Tichy

M. P.

(2004). An unrecognized need for ballot reform: The effects of candidate name order on election outcomes. In Criger

A. N.

Justin

M. R.

McCaffery

E. J.

(Eds.), Rethinking the vote: The politics and prospects of American election reform (pp. 51-73). New York, NY: Oxford University Press.

47.

Lloyd

(2017, June 14). Southern, small, and “looney”: Late-night TV takes on Jeff Sessions. Los Angeles Times. Retrieved from http://www.latimes.com/entertainment/la-et-entertainment-news-updates-june-late-night-does-jeff-sessions-1497457591-htmlstory.html

48.

Lodge

Taber

(2013). The rationalizing voter. New York, NY: Cambridge University Press.

49.

Matsubayashi

Ueda

(2011). Political knowledge and the use of candidate race as a voting cue. American Politics Research, 39, 380-413.

50.

McDermott

M. L.

(1997). Voting cues in low-information elections: Candidate gender as a social information variable in contemporary United States elections. American Journal of Political Science, 41, 270-283.

51.

McDermott

M. L.

(1998). Race and gender cues in low-information elections. Political Research Quarterly, 51, 895-918.

52.

Miller

W. E.

Shanks

J. M.

(1996). The new American voter. Cambridge, UK: Harvard University Press.

53.

Montgomery

(1993). The southern accent—Alive and well. Southern Cultures, 1(1), 47-64.

54.

Morales

A. C.

Scott

M. L.

Yorkston

E. A.

(2012). The role of accent standardness in message preference and recall. Journal of Advertising, 41, 33-46.

55.

155 Cong. Record 264 (2009) (statement of Rep. Louie Gohmert).

56.

Preston

D. R.

(1999). A language attitude approach to the perception of regional variety. In Preston

(Ed.), Handbook of perceptual dialectology (pp. 359-373). Amsterdam, The Netherlands: John Benjamins.

57.

Rakic

Steffens

M. C.

Mummendey

(2011). Blinded by the accent! The minor role of looks in ethnic categorization. Journal of Personality and Social Psychology, 100, 16-29.

58.

Reed

J. S.

(2000). The “southern accent.” Southern Cultures, 6, 102-103.

59.

Rey

(1977). Accent and employability: Language attitudes. Language Sciences, 47, 7-12.

60.

Riches

Foddy

(1989). Ethnic accent as a status cue. Social Psychology Quarterly, 52, 197-206.

61.

Schneider

M. C.

Bos

A. L.

(2014). Measuring stereotypes of female politicians. Political Psychology, 35, 245-266.

62.

Seabrook

(2006, October 1). Psychology behind the sudden southern drawl. National Public Radio. Retrieved from http://www.npr.org/templates/story/story.php?storyId=6175843

63.

Searles

Fowler

E. F.

Ridout

T. N.

Strach

Zuber

(2017). The effects of men’s and women’s voices in political advertising. Journal of Political Marketing. Advance online publication. Retrieved from https://www-tandfonline-com.web.bisu.edu.cn/doi/abs/10.1080/15377857.2017.1330723

64.

Sigel

(1996). Ambition and accommodation: How women view gender relations. Chicago, IL: University of Chicago Press.

65.

Simien

E. M.

(2005). Race, gender, and linked fate. Journal of Black Studies, 35, 529-550.

66.

Swigger

(2012). What you see is what you get: Drawing inferences from campaign imagery. Political Communication, 29, 367-386.

67.

Tigue

C. C.

Borak

D. J.

O’Connor

J. J. M.

Schandl

Feinberg

D. R.

(2012). Voice pitch influences voting behavior. Evolution & Human Behavior, 33, 210-216.

68.

Todorov

Mandisodza

A. N.

Goren

Hall

C. C.

(2005). Inferences of competence from faces predict election outcomes. Science, 308, 1623-1626.

69.

Trudgill

(1972). Sex, covert prestige and linguistic change in the urban British English of Norwich. Language in Society, 1, 179-195.

70.

Vannette

(2017, June 28). Using attention checks in your surveys may harm data quality. Qualtrics Blog. Retrieved from https://www.qualtrics.com/blog/using-attention-checks-in-your-surveys-may-harm-data-quality/

71.

Webster

S. W.

Pierce

A. W.

(2018). Older, younger, or more similar? The use of age as a voting heuristic. Social Science Quarterly. Retrieved from http://stevenwwebster.com/research/age_full_paper.pdf

72.

Wolfram

Reaser

(2014). Talkin’ tar heel: How our voices tell the story of North Carolina. Chapel Hill: University of North Carolina Press.

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.

0.00 MB

0.21 MB