The People’s Voice—The People’s Choice? How Vox Pop Exemplars Shape Audience Judgments as a Function of Populist Attitudes

Abstract

Exemplification research has consistently shown effects of vox pops’ exemplars on audience judgments, whereby people tend to follow the opinion of a few fellow citizens. In this study, we gain some insight into why—and especially for whom—ordinary citizens are such influential “opinion-givers.” Importantly, we look at populist attitudes as a potential moderator for exemplification effects by comparing news reports containing vox pops with purely journalistic news reports providing the same arguments. In a web-based experiment, we show that both perceptual and persuasive effects are moderated by participants’ populist attitudes, and thus, their resonance with the “voice of the people.”

Keywords

exemplification exemplars vox populi public opinion persuasion populism

Portraying the opinions of common people—or “ordinary citizens,” as they are often referred to in the mass communication literature (e.g., Hopmann & Shehata, 2011)—in the media has been thoroughly researched over the past three decades (for an overview, see Zerback & Peter, 2018). Although including citizens and their opinions seems desirable from a normative standpoint, the use of so-called exemplars as a stylistic device in news reports has been critically discussed among communication scholars, as journalists’ use of them is highly selective, yet highly influential in terms of people’s judgments (e.g., Beckers, Walgrave, & Van den Bulck, 2018; Daschmann & Brosius, 1999; Kleemans, Schaap, & Hermans, 2017). To date, there is still ambiguity about why exemplars, especially so-called vox populi, are so influential in the first place. One explanation that has been put forward, but not empirically tested, is that ordinary citizens as a source could be particularly trustworthy (especially, in comparison with other individual sources such as politicians). We argue that this should be especially true for people holding strong populist attitudes: As the concept of populism comprises a perceived dichotomy of the good and pure people and the evil and corrupt elite (Schulz et al., 2017), it could be that people holding populist beliefs are especially inclined to trust in the arguments of fellow citizens (and are more distrustful of arguments coming from elite sources).

Consequently, in this study, we are interested in whether including ordinary citizens as the source of arguments within a news report can increase the credibility of a news report and strengthen the effects on audience judgments, namely, on perceived public opinion and personal opinion. Most importantly, we look at whether populist attitudes moderate these effects, as people holding populist beliefs would seem to be more likely to favorably evaluate media content that gives ordinary citizens a voice.

Ordinary Citizens as a Stylistic Device in News Reports

The presence of ordinary citizens in media coverage has been studied in various fields, but especially in the context of exemplification (Zillmann & Brosius, 2000). Based on this research tradition, journalists integrate citizens’ stories or statements into news stories as a stylistic device to make the story appear more relatable and comprehensible for the audience (Beckers, 2017; Brosius & Bathelt, 1994; Daschmann, 2000). For example, a report on efforts to repeal Obamacare in the United States will—in addition to a description of the topic and statistics on how many people are affected—likely portray the plight of one Obamacare user in more detail. In this context, the specific person is not portrayed for his or her own sake but as an example of all people that would be affected by the repeal. Consequently, he or she is interchangeable with any other individual from that specific population, which is why these single cases are called exemplars (Zillmann & Brosius, 2000). This case-study exemplar type has often been investigated in the context of episodic framing (e.g., Gross, 2008; Iyengar, 1991).

Apart from the above-mentioned “case-study” exemplars, ordinary citizens can also function as “opinion-givers” in news reports. This type of exemplar is often referred to as a vox pop, short for vox populi (e.g., Daschmann, 2000; Lewis, Inthorn, & Wahl-Jorgensen, 2005). Instead of (or in addition to) featuring the plight or experience of a single citizen, journalists often present several citizens sharing their personal opinions on a topic such as whether repealing Obamacare is a good or a bad idea. Vox pops’ opinions are typically gathered in so-called man-on-the-street interviews, and they can be either affected or nonaffected individuals (Zerback & Peter, 2018). Analogous to case-study exemplars, vox pops are used to illustrate a larger population; their function is to represent public opinion. Bosch (2014) found that people tend to view this exemplar type as representative of typical ordinary citizens, as the “seeming randomness of these interviews . . . likely contribute[s] to the perception that these interviews represent the thoughts of everyday Americans” (p. 219).

Several content analyses have shown that exemplars are heavily present in media coverage, especially, not only in the U.S. and U.K. news (Lewis et al., 2005), but also in that of other European countries (Beckers et al., 2018; Hopmann & Shehata, 2011; Kleemans et al., 2017). Furthermore, news websites, in particular, seem to readily include social media comments as exemplars (Beckers & Harder, 2016). Consequently, in the era of social media, a good exemplar is only one click away. However, the use of exemplars has often been criticized due to their biased selection by journalists: Daschmann and Brosius (1999), for example, found a highly selective usage of exemplars, with almost all of them being in line with the general slant of the news story, thereby serving as opportune witnesses who support the media source’s viewpoint (Hagen, 1993). In line with this, Beckers et al. (2018) showed that a majority of news reports that included vox pops used them only to present one point of view.

Exemplification Effects: Impact on Perception and Persuasion

Media effect research on exemplification examines the impact of exemplars on audience judgments (for an overview, see Brosius & Peter, 2017). Most studies have compared exemplars with aggregated information about the population segment from which they were taken, so-called base-rate information (Bar-Hillel, 1980). For case-study exemplars, this would be the total number of people affected (e.g., the number of people who could lose health care). For vox pops, base-rate information would be opinion polls or other gauges of public opinion from which to draw conclusions (e.g., subjective statements such as “The majority of people think . . . ,” Lewis et al., 2005). Consequently, research has looked at both effects on the perceptual level (e.g., perception of public opinion, risk estimates) and persuasion level (e.g., personal opinion, behavioral intention) of judgments.

Studies on perceptual level effects have overwhelmingly shown that exemplars affect people’s perceptions about the respective base rate, even if more valid base-rate information was present in the report (e.g., Brosius, 1999; Brosius & Bathelt, 1994; Peter & Zerback, 2017; Zillmann, Perkins, & Sundar, 1992). Apart from that, there is some evidence that exemplars can also be powerful in influencing people’s personal opinions on a certain topic. In what is considered one of the seminal papers on exemplification, Brosius and Bathelt (1994) demonstrated such persuasive effects of exemplars on different issues. For example, an article with a majority of exemplars expressing a favorable opinion on cafeteria food led to more favorable personal opinions on this topic than a version of the article with a majority of exemplars being unfavorable. Although Brosius (2003) assumed that persuasive effects would only occur for issues of minor importance to the individual, some studies also found effects for political topics (e.g., Daschmann, 2000; Lefevere, De Swert, & Walgrave, 2012).

In a more recent study, Lefevere et al. (2012) showed stronger persuasive effects for arguments within a news report when they were put forward by ordinary citizens in contrast to other sources such as politicians or experts. Although they did not test it empirically, the authors argued that this might be due to the fact that ordinary citizens are a trustworthy source as they are similar to the audience, especially, when compared with other (elite) actors such as politicians.

The Role of Populist Attitudes Within Exemplification Theory

As described above, Lefevere and colleagues (2012) argue that the trustworthiness of exemplars can partly be attributed to the fact that the audience sees them as similar to themselves when compared with other sources. We argue that this should be especially true for people supporting populist ideas. Populism is defined as a political ideology (Mudde, 2004) as well as a communication style (Jagers & Walgrave, 2007) that, in its simplest form, consists of the construction of “the people” as an entity (Reinemann, Aalberg, Esser, Strömbäck, & de Vreese, 2017). In recent years, populist political actors have gained momentum in Europe as well as in the United States, leading researchers to concern themselves more deeply with this phenomenon (Aalberg, Esser, Reinemann, Strömbäck, & de Vreese, 2017). Although mostly discussed in the context of right-wing political attitudes, populism has been associated with both strong liberal and conservative views (Fawzi, 2018; Jagers & Walgrave, 2007; Schulz et al., 2017).

Thus, populist attitudes have been conceptualized as consisting of three dimensions (Schulz et al., 2017): antielitism (the belief that the elites are corrupt and betray the common people), a preference for popular sovereignty (the demand that the common people, and not politicians, should be in charge), and a belief in the homogeneity and virtuousness of the people. This results in a perceived dichotomy of “pure people” versus the “corrupt elite” (Akkerman, Mudde, & Zaslove, 2014; Mudde, 2004). Elite often refers not only to the political elite, but also to other elite segments such as the mainstream media that are seen as the henchmen of politics. Consequently, populist attitudes have been linked to low trust in the mainstream media (Fawzi, 2018).

Schulz et al. (2017) describe one core aspect of populist attitudes as “an understanding of a monolithic people that is altogether good, honest, and upright. In this vein, the people share the same values and interests” (p. 3). Consequently, holding strong populist attitudes could affect exemplification effects in several ways: (a) Regarding perceptual effects, inferences from a few single opinions in the form of vox pops on the broader public opinion could be stronger as the people’s opinions are seen as homogeneous and (b) Regarding persuasive effects, populist attitudes could result in more reliance on vox populi as the “voice of the people”: Vox pops—ordinary people voicing their opinion—might be perceived as even more similar by people with populist beliefs and thus as more trustworthy. At the same time, media content giving ordinary citizens a voice might weaken the perception that this media outlet is in league with the political elite, thus making it appear more trustworthy. In the following chapters, we will elaborate on how populist attitudes might moderate exemplification effects in terms of both perception and persuasion.

Populist Attitudes and Perceptual Effects

The main reason for exemplars’ strong impact on perceived public opinion is seen in heuristic processing, with people erroneously judging a few individuals to be representative of a larger population that shares the same features (representative heuristic; Kahneman & Tversky, 1972). In this sense, research has implicitly argued that people infer public opinion from media content containing vox pops because they see them as a reflection of what the rest of the public thinks, and thus they form the impression that said media content mirrors public opinion (Gunther, 1998; Zerback, 2016; Zerback, Koch, & Krämer, 2015). As argued above, we believe that those inference effects could be stronger for people holding populist beliefs as they believe in the homogeneity of the people and, thus, should assume that all people hold similar opinions. Consequently, even just a few vox pops should be perceived as representative of the broader public opinion by people holding populist beliefs, leading to stronger inference effects.

However, there is also reason to believe that a news report without vox pops will influence the audience’s perception of what the public thinks. Gunther (1998) has shown that people infer public opinion solely from the slant of the media content: In his seminal paper, he tested the influence of news reports with and without exemplars on perceived public opinion and found effects in both cases. For the version without exemplars, he assumed a mechanism that he labeled the “persuasive press inference” (see also Gunther, Christen, Liebhart, & Chia, 2001; Zerback, 2016). According to this, a two-step inference process is at work: First, people extrapolate the slant of a given news report to the overall media coverage on an issue (Gunther et al., 2001). Second, the slant of the media coverage is perceived to be influential on the audience—a phenomenon that has already been well established by research on the third-person effect and presumed influence (Gunther & Storey, 2003; Sun, Pan, & Shen, 2008; Tsfati, 2007). Overall, this leads people to assume that “what mass media are saying today must be what the public will be thinking tomorrow” (Gunther, 1998, S. 487). We argue that this persuasive press inference might also be stronger for people supporting populist ideas: They tend to mistrust the media as they are seen as part of the corrupt elite and are believed to report in a biased way (Schulz et al., 2017). As research on the hostile media phenomenon has shown, the perception of biased reporting and the associated mistrust in the media results in the perception of stronger effects on others (e.g., Barnidge & Rojas, 2014; Tsfati, 2007).

Following the above argumentation, we conclude two things. First, we assume that a news report containing vox pops will be more powerful in triggering the representative heuristic, resulting in a perceived reflection that shapes people’s perceptions about current public opinion. In addition, we assume that this effect will be more pronounced for people holding strong populist attitudes. Consequently, we assume the following:

H1a: A news report will affect people’s perceptions of current public opinion more strongly if the arguments within the article are put forward by vox pops rather than by the journalist.

H1b: The stronger the populist attitudes that people hold are, the more they will infer current public opinion from vox pops’ arguments and the less they will infer public opinion from the journalist’s arguments.

Second, news reports without vox pops will trigger perceived persuasion, which should result in stronger effects regarding the perception of public opinion in the future, meaning that people believe that public opinion will change in the direction of the news report. Consequently, although a news report containing vox pops should be more powerful in shaping current public opinion perceptions, reports both with and without vox pops should elicit a similar influence on the perception of future public opinion, as the overall slant of both versions (and thus potential persuasive press inferences) should be the same. Taken together, we hypothesize the following:

H2a: A news report will affect people’s perceptions of public opinion change, regardless of whether the arguments within the report are put forward by vox pops or by the journalist.

H2b: The stronger the populist attitudes that people hold are, the more they will infer public opinion change from both the vox pops’ and the journalist’s version of an article.

Populist Attitudes and Effects on Persuasion

People also tend to rely on exemplars when forming personal opinions on a specific topic (Allen & Preiss, 1997). Brosius and Bathelt’s (1994) study demonstrated the persuasive effects of exemplars for different issues. Daschmann (2000) replicated these findings for news coverage on a political topic. Within a news article on elections in a German federal state, five voters voiced their opinions on the state government (either in favor of or opposing the government). As a second factor, the poll results were manipulated (The state government will or will not receive the majority of votes in the next election). The results showed that after reading the version with vox pops who were progovernment, participants expressed a significantly more favorable opinion toward the state government than participants who read the version with antigovernment exemplars did. The poll results, however, had no significant effect on personal opinions.

One problem with the previous research on exemplification effects that we want to overcome in this study is the confounding of ordinary citizens as the source of arguments and the presence of the arguments themselves. As described above, most studies compared either the relative effects of exemplars (pro vs. contra exemplars on a topic) or exemplars with base-rate information. However, when the exemplars were absent (e.g., under base-rate conditions), so were the arguments they provided. For example, in the study by Daschmann (2000), one of the exemplars in favor of the state government argued that the government had introduced a law that ensured enough kindergartens for all children (p. 166). In the version with the poll results only, this argument (as well as the arguments advanced by the other four exemplars) was absent. Consequently, it cannot be determined whether exemplars are influential because ordinary citizens are a persuasive source or whether the strong effects of exemplification originate in the mere presence of (more) arguments provided within a news report.

A recent study tackled this problem by comparing exemplars with other sources (Lefevere et al., 2012). In this study, the exact same arguments were put forward by ordinary citizens, by a politician, or by an expert in a news report. The results showed stronger effects on personal opinions from the report featuring ordinary citizens as opinion-givers than from the report featuring an expert source did, and no effect was seen from the report featuring a politician as the source of arguments. In line with this, we assume that it is not the arguments themselves but the fact that they are put forward by ordinary citizens, which is responsible for the strong effects of exemplars. Consequently, we hypothesize the following:

H3a: A news report will affect people’s personal opinions more strongly if the arguments within the article are put forward by vox pops rather than by the journalist.

Again, we believe that populist attitudes may moderate these effects. Populist attitudes consist of a strong belief in “the people” as a unity that is contrasted against others, especially elites such as not only politicians, but also journalists (Fawzi, 2018; Reinemann et al., 2017). Consequently, we assume that people supporting populist ideas might be more inclined to believe in arguments that are put forward by other ordinary citizens rather than by a journalist:

H3b: The stronger the populist attitudes are that people hold, the more they will be influenced in their personal opinion by vox pops’ arguments and the less they will be influenced by the journalist’s arguments.

Several explanations have been put forth on the persuasive effects of exemplars. Lefevere and colleagues (2012) argue that ordinary citizens are a trustworthy source, especially, when compared with other actors such as politicians. Brosius (2003) already addressed the argument of trustworthiness when talking about case-study exemplars: Ordinary citizens provide their experience on a topic as firsthand witnesses, making their perspectives more credible than a retelling by a journalist. Lefevere et al. (2012) put forth two reasons why higher trustworthiness should also hold true for vox pop exemplars: First, vox pops are similar to the audience, as both the media consumers and the vox pops are, essentially, ordinary citizens. Source similarity, in turn, enhances source credibility (e.g., Pornpitakpan, 2004). Peter and Zerback (2017) have already shown that similarities between exemplars and the audience in terms of sociodemographic factors can increase persuasive effects.

The second argument advanced is that exemplars elicit less perceived intent to persuade than other sources do. People tend to trust others less when they feel that they are trying to persuade them, which is especially true of politicians (e.g., Koch & Zerback, 2013). In contrast, ordinary citizens (i.e., those “like you and me”) who are simply stating their opinions in a man-on-the-street interview are probably seen as less likely to be advancing any agenda. Consequently, scoring both high in similarity and low in persuasive intent makes exemplars highly trustworthy, which is one of the core aspects of credibility (Hovland, Janis, & Kelley, 1953). Source credibility, in turn, is associated with stronger persuasive effects (Hovland et al., 1953; for an overview, see Pornpitakpan, 2004; Wilson & Sherrell, 1993).

So, will people be more inclined to trust in media content if they see themselves represented in it? So far, several studies on credibility have shown spillover effects from the source onto its message (e.g., Hovland & Weiss, 1951; Koch & Zerback, 2013). In the case of exemplars embedded within a news report, we are actually looking at a spillover effect from a source presented within the message (the news report) onto the message itself. Consequently, we are interested in whether the inclusion of vox pop exemplars in a news report can make the news report itself seem more credible to the audience. Again, we argue that these effects should be more pronounced for people with populist beliefs: Vox pops might therefore be perceived as even more similar to those holding populist views and thus as more trustworthy. Consequently, an article that features ordinary citizens could be perceived as more trustworthy as it gives a voice to “the people.” In addition, other (elite) actors such as journalists might be perceived as particularly untrustworthy. Consequently, we assume the following:

H4a: A news report will be perceived as more credible if the arguments within the article are put forward by vox pops rather than by the journalist.

H4b: The stronger the populist attitudes that people hold are, the more credible they will find an article containing vox pops and the less credible they will find an article containing no vox pops.

Method

Participants

To test our hypotheses, we conducted a web-based experiment. The sample was recruited via the online access panel “SoSci Panel” (Leiner, 2016) to obtain a heterogeneous sample regarding gender (55.4% female), age (M = 41.68, SD = 16.34, range: 16-88 years), and formal education (68.9% held a higher formal educational qualification). This was necessary to guarantee that populist attitudes would be roughly equally distributed in the sample, which was the case (M = 2.76, SD = 0.76, range: 1.11-5.00). The panel is noncommercial and consists of members who have agreed to participate in scientific surveys and includes around 100,000 panelists from Germany, Austria, and the German-speaking part of Switzerland. Although this sample is not representative of the respective population—as panel members are generally younger and better educated—compared with traditional student samples, the SoSci Panel offers more heterogeneous samples in terms of age, education, geography, and personal interests (for more information on the composition and limitations of the SoSci Panel, see Leiner, 2016). A total of 2,100 invitations were sent to the panel, resulting in 353 completed questionnaires.

Experimental Manipulation

The experiment consisted of two experimental groups and one control group. As a stimulus for the experimental groups, we constructed a written news report about whether compulsory voting should be established in Germany. We chose compulsory voting as a topic for two reasons. First, it is a political issue that should be at least somewhat interesting and relevant for participants, which we deemed important, as much of the prior research on exemplification has been conducted using fictitious or at least nonpolitical topics (e.g., Brosius & Bathelt, 1994; Lefevere et al., 2012; Zillmann et al., 1992). Second, the topic had not been publicly discussed at the time of the data collection, so we expected participants not to have too much prior knowledge about it.

The article had the layout of a popular German online newspaper (thus representing the “mainstream” media) and clearly supported the idea of compulsory voting. This slant was created by three arguments in favor of compulsory voting within the article. The arguments were identified in a pretest (n = 29), in which we determined the strength of 10 arguments in favor of compulsory voting. Argument strength was measured following Zhao, Strasser, Cappella, Lerman, and Fishbein (2011). The three arguments with the highest ratings (between 3.08 and 3.77 on a 5-point Likert-type scale) were chosen for the article.

We created two versions of the news report. The headline, lead, and first paragraph were identical in both versions, and provided some general information about what compulsory voting entailed and which countries had already established the concept in their laws. The second paragraph was overwritten with “What are the advantages of compulsory voting?” Within this paragraph, the arguments in favor of compulsory voting were presented. In the first version (n = 131), the arguments were not put forward by an external source, but by the journalist as the author of the article. In the second version (n = 130), the three arguments were put forward by three ordinary citizens who were interviewed on the street. The arguments were presented as direct quotes, and the names of the citizens were provided (e.g., “‘If all persons entitled to vote were to do so, the result would be more representative of the German population’, said Daniela Huber”). Two exemplars were male, and one was female, with all of them having common German names.

Participants in the control group (n = 92) did not receive an article, but only a short informational paragraph on what compulsory voting entailed (the content was identical to the introductory paragraph of the news report).

Procedures and Measures

After some initial questions (e.g., an assessment of populist attitudes), both experimental groups read the corresponding news reports. The control group received no article, but instead was given a short informational paragraph about what compulsory voting entailed. The wording was identical to the information provided in the first paragraph of the article. After that, questions were posed about article quality (including credibility estimates) and about personal opinions on compulsory voting. Furthermore, we employed a treatment check: We asked participants whether people other than the journalist who wrote the article were given a voice within the article (answer categories were “yes,” “no,” or “I don’t remember”). In total, five participants who read the vox pop article said no other person than the author was given the chance to speak, and three readers of the journalist’s version erroneously recalled speakers besides the author. All other participants passed the treatment check. We decided to keep the eight participants for hypothesis testing for two reasons. First, removing participants based on the treatment check threatens the causal logic of the experiment. Second, even though they did not recall the source of the arguments, this does not mean that they did not base their judgments on the respective source. Especially, people forming and storing judgments during the reception phase are not always able to recall the specific information that their judgment was based on (e.g., Hastie & Park, 1986).

We measured perceptions of public opinion following Zerback (2016). Perceived current public opinion was assessed via two items (“A major part of the German population favors the implementation of compulsory voting” and “The majority of Germans oppose compulsory voting” [reverse coded], M = 2.17, SD = 0.87, α = .82). Perceived future public opinion change was also measured via two items (“In the future, more people will favor compulsory voting” and “More and more people will oppose compulsory voting in the future” [reverse coded], M = 3.07, SD = 1.01, α = .73). Answers were given based on a 5-point Likert-type scale.

Personal opinion toward compulsory voting was measured on a 5-point semantic differential (six items, for example, “unnecessary–necessary,” M = 3.07, SD = 1.37, α = .97). Perceived credibility of the news article was assessed via two items (“not credible–credible” and “inconclusive–conclusive,” M = 3.23, SD = 1.04, α = .78) within a 5-point semantic differential measuring several dimensions of article quality.

Populist attitudes as a moderator variable were assessed using the 9-item scale by Schulz et al. (2017), in which each of the three dimensions (antielitism, demand for the people’s sovereignty, and belief in the homogeneity of the people) was assessed via three items (overall populist attitudes: M = 2.76, SD = 0.76, α = .82). Answers were given based on a 5-point Likert-type scale. An overview of the mean values for the dependent variables between the experimental conditions can be found in Table 1.

Table 1.

Mean Values for Dependent Variables Between Experimental Groups and Control Group.

	Control(n = 92)		Vox pops(n = 130)		Journalist(n = 131)
	M	SD	M	SD	M	SD
Current public opinion	2.03	0.86	2.31	0.88	2.14	0.85
Public opinion change	2.57	1.04	3.05	1.01	2.99	1.15
Personal opinion	2.65	1.32	3.28	1.30	3.17	1.41
Article credibility	—	—	3.21	0.93	3.26	1.13

Results

To test our hypotheses, we conducted a series of moderation analyses using PROCESS (Hayes, 2018, Model 1) to account for the continuous scaling of our moderator populist attitudes. We introduced current public opinion, future public opinion change, or personal opinion as the dependent variable; experimental conditions (vox pop, journalist, no article) as the independent multicategorial variable; and populist attitudes as the moderator. To obtain values for the direct comparison between the journalistic article and the vox pop article (which was of interest in most of the hypotheses), we chose to compare the experimental groups via orthogonal contrasts (Helmert coding, see Hayes, 2018). To allow for an interpretation of conditional effects as the average effect (e.g., the effect of experimental conditions on the average value of populist attitudes), populist attitudes were mean centered before the analysis (see Hayes, Glynn, & Huge, 2012, for a discussion). Regarding perceived credibility as the dependent variable, we undertook a moderation analysis without the control group, as this group did not assess article quality.

Perceptual Effects: Current Public Opinion and Future Public Opinion Change

Regarding the current public opinion perception, we expected the vox pop article to exceed the journalist’s article in terms of influence (H1a) and that this would be more pronounced for people holding strong populist attitudes (H1b). The results showed neither a significant difference between the control and the experimental groups, nor between the journalist’s and the vox pops’ version (Table 2), so H1a must be rejected. Regarding populist attitudes as a moderator, the results reveal a pattern in line with H1a (Table 2, Figure 1): As populist attitudes became stronger, so did the difference between the journalist’s and the vox pops’ version. Consequently, people holding strong populist attitudes judged current public opinion to be more in line with the article after reading the vox pops’ version rather than the journalist’s version.

Table 2.

Moderation Analyses.

	Current public opinion			Public opinion change			Personal opinion			Credibility
	Coefficient	SE	p	Coefficient	SE	p	Coefficient	SE	p	Coefficient	SE	p
Conditional effects
Control vs. treaments	0.17	0.10	.092	0.43	0.13	<.001	0.53	0.16	<.001
Journalist vs. vox pops	0.18	0.11	.095	0.04	0.13	.742	0.08	0.16	.597	0.06	0.13	.657
Populist attitudes	0.24	0.06	<.001	0.31	0.08	<.001	0.53	0.09	<.001	0.07	0.08	.419
Moderations
Control vs. treaments × populist attitudes	0.04	0.14	.306	−0.24	0.18	.183	−0.02	0.22	.917
Journalist vs. vox pops × populist attitudes	0.33	0.14	.016	0.23	0.17	.165	0.44	0.21	.032	0.31	0.17	.061
Model summary	F(5, 346) = 5.67, p < .001, R² = .08			F(5, 345) = 6.26, p < .001, R² = .08			F(5, 346) = 10.36, p < .001, R² = .13			F(3, 256) = 1.38, p = .250, R² = .02

Note. All models were calculated using PROCESS (Hayes, 2018, Model 1). For analyses on current public opinion, future public opinion change, and personal opinion as the dependent variable, experimental conditions (vox pop, journalist, no article) have been entered as multicategorial independent variable (Helmert effect coding). For analysis on perceived article credibility, control group was omitted as this group did not assess this variable. To allow interpretation of conditional effects as the average, populist attitudes were mean centered before analysis. Unstandardized effects are displayed. Significant results are displayed in bold.

Figure 1.

Perceptual effects: Current public opinion and public opinion change.

With H2a, we expected both article versions to be equally influential in shaping perceptions of future public opinion change, meaning that we expected estimates to differ from the control group to the same extent. In addition, we assumed this effect to be stronger for people holding strong populist attitudes (H2b). As expected, the results revealed significant differences between both article versions and the control group (Table 2). To test for the equivalence of public opinion change estimates in the vox pops’ and the journalist’s versions, we conducted an equivalence test with r = .11 as the lowest effect size of interest (Weber & Popova, 2012). Thus, the results suggest that the two article versions are not equivalent, t(259) = −0.46, Δ = .11, p_eff = .167, as the mean value for the vox pop version was slightly higher than that for the journalistic condition (Table 1). Consequently, H2a can only be partly confirmed: Both article versions seem to elicit an effect (compared with the control group), but not to the same extent. For the interaction with populist attitudes, Figure 1 shows that the conditional effect between the control and the experimental groups can be largely attributed to people with rather low populist attitudes, whereas people scoring high seem not to infer public opinion from any of the articles (a conditional effect between the control and experimental groups for the 84th percentile level of populist attitudes: coefficient = 0.25, SE = 0.20, t = 1.25, p = .211). However, this interaction did not reach statistical significance, so H2b was also rejected.

Persuasive Effects: Personal Opinion and Article Credibility

We expected the vox pops’ version to be more influential than the journalist’s version in affecting personal opinion (H3a), with this effect being more pronounced for people holding strong populist attitudes (H3b). Again, we found both article versions to significantly differ from the control group, but not from each other (Table 2), so H3a was rejected. Yet, similar to effects on current public opinion, we found a significant interaction with populist attitudes in such a way that the stronger the populist attitudes of the participants were, the more they voiced a personal opinion in line with the vox pops’ version as compared with the journalistic version (Table 2, Figure 2). Consequently, H3b could be confirmed.

Figure 2.

Persuasive effects: Personal opinion and article credibility.

Finally, we expected an article containing vox pops to be more credible than the journalistic version (H4a), with this effect being more pronounced for people holding strong populist attitudes (H4b). The results in Table 2 show no difference between the experimental versions, and show only a borderline significant interaction effect with populist attitudes, so both hypotheses need to be rejected.

Noticeably, we found a conditional effect of populist attitudes on all dependent variables except for perceived article credibility, with stronger populist attitudes leading to more favorable opinions on compulsory voting and the public perception being more in favor of it.

Discussion

The present study investigated whether including ordinary citizens as the source of arguments within a news report (compared with the same arguments being presented by the journalist) could increase the perceptual and persuasive effects of the article, and whether these effects might be more pronounced for people holding strong populist attitudes, as people holding populist beliefs would seem to be more likely to favorably evaluate media content that gives ordinary citizens a voice. Consequently, to the best of our knowledge, this study is the first to explore the link between populist attitudes and exemplification effects.

As a first main finding, the results of this study add to the understanding as to why—and especially when—ordinary citizens as vox pops in news reports are powerful in influencing people’s personal opinions. The fact that overall, we did not find a stronger effect of vox pops on current public opinion (H1a) and personal opinion (H3a) suggests that the persuasive effects of exemplars in previous studies might at least partly be attributed to the arguments they provided: Most studies examined either relative effects (vox pops pro vs. contra the issue at hand) or compared news reports containing exemplars with those containing only base-rate information, with the latter lacking the specific arguments provided by the exemplars (Daschmann, 2000). Yet, if the same arguments are presented by the journalist as the author of the article, we find effects from this source, at least for part of the sample, as well.

This brings us to our second major finding: We found that exemplification effects seem to be dependent on people’s populist beliefs to some extent, and this seems to be true for both perceptual (Figure 1) and persuasive effects (Figure 2). Regarding current public opinion, our results show that although people scoring low on populist attitudes do not infer public opinion from any of the articles, people holding stronger populist beliefs judge current public opinion as being in line with an article only when the arguments are presented by fellow citizens but not by the journalist. This adds to the assumption that vox pops elicit an inference strategy that is based on perceived reflection: People view the opinions of a few ordinary citizens as reflective of the broader public opinion and thus use them to infer current public opinion, a heuristic judgment mechanism that is known as a representative heuristic (Kahneman & Tversky, 1972). It is also in line with the assumption that people holding populist beliefs, by definition, see the public as homogeneous, and that this perception would strengthen the inference from a few single citizen’s opinions to the broader public opinion. Interestingly, people scoring lowest on populist attitudes did not infer current public opinion from vox pops at all. For this group, we only found effects of the vox pop article on perceived public opinion change in the future, and the same held true for the journalistic version (Figure 1). The fact that this group infers public opinion change but not current public opinion from both report versions suggests that they may not see the media as a true reflection of public opinion, but rather as influential in changing it. What adds to this is that the vox pop version of the news report did not influence people with low populist attitudes in terms of changing their personal opinion on compulsory voting (Figure 2). Consequently, the results suggest that people holding minimal or no populist attitudes seem to be most susceptible to the persuasive press inference, and people scoring high on populist beliefs are not, which is in line with correlational evidence by Schulz et al. (2018).

The opposite seems to be true for people holding populist beliefs: Participants scoring at least moderately on populist attitudes inferred current public opinion only from the news report containing vox pops but not from the journalistic version (Figure 1). Regarding public opinion change, however, this group is not affected by either report version. Thus, it seems that people holding populist beliefs project vox pops onto current public opinion, whereas they do not perceive that there are persuasive effects from the media content on future public opinion. An explanation for this could be that people holding populist attitudes tend to mistrust the mainstream media, and thus might see themselves and others as part of the same homogeneous “people” and as rather immune to media influence. If this is true, people scoring high on populist attitudes should, in general, show lower third-person effects (which, to the best of our knowledge, has not yet been tested).

Regarding the persuasive effects, we found a similar pattern as for the effects on current public opinion: The results showed that when it came to personal opinions, the influence of the vox pops’ version increased as a function of populist attitudes, whereas, at the same time, the influence of the journalistic version decreased. Consequently, people scoring high on populist attitudes were only influenced in the direction of the article when arguments in favor of compulsory voting were voiced by fellow citizens, whereas for people scoring low on populist attitudes, the opposite seemed to be true, meaning that they were only influenced by the journalistic version (Figure 2). This is in line with the assumption that strong populist beliefs would be connected to more trust in fellow citizens, on the one hand, and less trust in other actors such as journalists, on the other. The results for the perceived credibility of the article point in a similar direction, but the effect was rather weak and not significant. Taken together, the results are remarkable insofar as vox pops are, in fact, not an independent source, but part of the journalist’s story, and thus deliberately chosen by him or her. As the content analyses have shown, the use of exemplars within news reports is highly selective, with them functioning as opportune witnesses through which to support the general slant of a news report (Beckers et al., 2018; Daschmann & Brosius, 1999). In addition, Beckers (in press) showed the effects of vox pops on both perceived public and personal opinion, even when people were informed that they were not representative of their population.

Limitations and Directions for Future Research

This study has several limitations that need to be addressed in further research on the topic. First, a potential confounding factor could be the number of sources presented within the article. In the vox pop version of the article, the arguments were put forward by three different ordinary citizens, and consequently by three different sources, whereas in the journalistic version, there was only one source. This might be a problem, as research on the truth effect has shown that people tend to have more trust in a message when they have obtained it from different sources (Unkelbach, 2006). However, if this were the case, then the vox pop version should have been perceived as more credible in general, yet, for people with low populist attitudes, we found the opposite tendency.

Second, to include vox pops in an external and valid way, we also included their names. These names provided additional information about the citizens (e.g., their gender). As research on similarity has shown, a match with sociodemographic variables can increase similarity (e.g., Goethals & Darley, 1977; Peter & Zerback, 2017). To account for this fact, we had two male exemplars and one female exemplar, and did not provide additional information such as the age or profession of the exemplars. However, future research should integrate such factors, especially, as they are apparent when we look at exemplification in audiovisual environments (e.g., TV coverage). This is linked to a third limitation of our study: The findings are limited to newspaper coverage. We chose a newspaper article as the stimulus because the manipulation of the sources for the arguments (vox pops vs. a journalist) could be realized with as little variance in additional information as possible, which increases the internal validity of the experiment. However, as the content analyses have shown that exemplars are often used in TV coverage, the effects should be further studied in this context, where additional factors such as perceived attractiveness and likability should be considered, as they influence source credibility (Pornpitakpan, 2004).

Fourth, regarding the perceptual effects, the inference mechanisms (reflection vs. persuasion) have been deduced from the results on different public opinion estimates, but they have not been measured directly to avoid priming effects. To further validate the assumptions underlying this study, the direct measurement of inference strategies as well as longitudinal studies is in order. With the latter, it would also be possible to establish causation between personal opinion and perceived public opinion and test for projection vs. adaption effects.

Fifth, the present study is the first to include populist attitudes as a moderator for exemplification effects. Populist attitudes are defined by a belief in the homogeneity of “the people” and a mistrust in elite sources (such as the mainstream media), which explain stronger effects for people with populist beliefs on perceived public opinion and personal opinion when the arguments are provided by fellow citizens, and less influence from an article with a journalist as the source of the arguments. As this study is the first to link exemplification and populism, further studies are to validate these results and test whether they hold true across different topics and media outlets. Certainly, in audiovisual environments, vox pops may add to the impression of a “firsthand testimony” (Brosius, 2003) that might bolster their effects in this context and also make their influence superior to the journalistic version for people who do not hold populist beliefs. Beckers (in press) showed vox pops’ effects on both personal and perceived public opinion with audiovisual news stimuli. Furthermore, it might also be important for future research to include political standpoints and distinguish between right-wing and left-wing populism and see how this interacts with vox pops and the respective topic under study.

Finally, a common problem of randomized experiments is forced exposure, which limits the external validity of the findings. In the present study, we found that the people holding strong populist attitudes were more persuaded by a news article from a mainstream media outlet when it contained vox pops; however, research has shown that people supporting populist attitudes tend to avoid the mainstream media (e.g., Fawzi, 2018; Schulz et al., 2018). In this context, it would be interesting to compare the mainstream media with alternative media outlets and populist communicators regarding their integration of “the people’s voice” within their communication strategies (Jagers & Walgrave, 2007).

An alternative explanation for the increased persuasiveness and also the slightly higher credibility of the vox pop version for people with populist beliefs could be that the vox pops are not the more trustworthy source, but the journalist behind the report could be, being rewarded with more trust for giving ordinary citizens a voice. In this context, it would be interesting to see what would happen if, prior to reception, awareness was raised regarding the fact that journalists, in many instances, use exemplars selectively and in a biased manner. Nevertheless, the results indicate that the impact of and the trust in journalistic content could—at least temporarily—be increased by giving ordinary citizens a voice. Further research needs to determine whether such effects hold true across different topics and outlets, how long-lasting these effects are, and what the cumulative effects of exemplification in this context are. Combining content analyses and a representative panel survey in three European countries, Jebril, Albæk, and de Vreese (2013) have shown that exposure to news including ordinary citizens can have beneficial effects on consumers as it decreases political cynicism. In times of populism and the associated mistrust of the mainstream media, the question arises as to whether giving ordinary citizens a voice could be a chance for journalism to win back the trust of the audience.

Footnotes

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) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Christina Peter

Author Biography

Christina Peter (PhD, 2014, LMU Munich) is a postdoctoral researcher at the Department of Media and Communication at the LMU Munich. She studied Communication Science, Political Science, and Psychology. Her research focuses on media usage and effects, political communication, and persuasion.

References

Aalberg

Esser

Reinemann

Strömbäck

de Vreese

C. H.

(Eds.). (2017). Populist political communication in Europe. New York, NY: Routledge.

Akkerman

Mudde

Zaslove

(2014). How populist are the people? Measuring populist attitudes in voters. Comparative Political Studies, 47, 1324-1353. doi:10.1177/0010414013512600

Allen

Preiss

R. W.

(1997). Comparing the persuasiveness of narrative and statistical evidence using meta-analysis. Communication Research Reports, 14, 125-131. doi:10.1080/08824099709388654

Bar-Hillel

(1980). The base-rate fallacy in probability judgments. Acta Psychologica, 44, 211-233. doi:10.1016/0001-6918(80)90046-3

Barnidge

Rojas

(2014). Hostile media perceptions, presumed media influence, and political talk: Expanding the corrective action hypothesis. International Journal of Public Opinion Research, 26, 135-156. doi:10.1093/ijpor/edt032

Beckers

(2019). What vox pops say and how that matters: Effects of vox pops in television news on perceived public opinion and personal opinion. Journalism & Mass Communication Quarterly, [Advance online publication]. doi:10.1177/1077699019843852

Beckers

(2017). Vox pops in the news: The journalists’ perspective. Communications: The European Journal of Communication Research. Advance online publication. doi:10.1515/commun-2017-0040

Beckers

Harder

R. A.

(2016). “Twitter just exploded”: Social media as alternative vox pop. Digital Journalism, 4, 910-920. doi:10.1080/21670811.2016.1161493

Beckers

Walgrave

Van den Bulck

(2018). Opinion balance in vox pop television news. Journalism Studies, 19, 284-296. doi:10.1080/1461670X.2016.1187576

10.

Bosch

(2014). Beyond vox pop: The role of news sourcing and political beliefs in exemplification effects. Mass Communication and Society, 17, 217-235. doi:10.1080/15205436.2013.779718

11.

Brosius

H.-B.

(1999). Research note: The influence of exemplars on recipients’ judgements: The part played by similarity between exemplar and recipient. European Journal of Communication, 14, 213-224. doi:10.1177/0267323199014002004

12.

Brosius

H.-B.

(2003). Exemplars in the news: A theory of the effects of political communication. In Bryant

Roskos-Ewoldsen

Cantor

(Eds.), Communication and emotion: Essays in honor of Dolf Zillmann (pp. 179-194). Mahwah, NJ: Lawrence Erlbaum.

13.

Brosius

H.-B.

Bathelt

(1994). The utility of exemplars in persuasive communication. Communication Research, 21, 48-78.

14.

Brosius

H.-B.

Peter

(2017). Exemplification theory. In Rössler

(Ed.), The international encyclopedia of media effects. Hoboken, NJ: Wiley-Blackwell. doi:10.1002/9781118783764

15.

Daschmann

(2000). Vox pop & polls: The impact of poll results and voter statements in the media on the perception of a climate of opinion. International Journal of Public Opinion research, 12, 160-181. doi:10.1093/ijpor/12.2.160

16.

Daschmann

Brosius

H. B.

(1999). Can a single incident create an issue? Exemplars in German television magazine shows. Journalism & Mass Communication Quarterly, 76, 35-51. doi:10.1177/107769909907600104

17.

Fawzi

(2018). Untrustworthy news and the media as “enemy of the people”? How a populist worldview shapes recipients’ attitudes toward the media. The International Journal of Press/Politics, 24, 146-164. doi:10.1177/1940161218811981

18.

Goethals

G. R.

Darley

J. M.

(1977). Social comparison theory. An attributional approach. In Suls

J. M.

Miller

R. L.

(Eds.), Social comparison processes: Theoretical and empirical perspectives (pp. 259-278). Washington, DC: Halsted-Wiley.

19.

Gross

(2008). Framing persuasive appeals: Episodic and thematic framing, emotional response, and policy opinion. Political Psychology, 29, 169-192. doi:10.1111/j.1467-9221.2008.00622.x

20.

Gunther

A. C.

(1998). The persuasive press inference: Effects of mass media on perceived public opinion. Communication Research, 25, 486-504. doi:10.1177/009365098025005002

21.

Gunther

A. C.

Christen

C. T.

Liebhart

J. L.

Chia

S. C.-Y.

(2001). Congenial public, contrary press, and biased estimates of the climate of opinion. Public Opinion Quarterly, 65, 295-320. doi:10.1086/322846

22.

Gunther

A. C.

Storey

J. D.

(2003). The influence of presumed influence. Journal of Communication, 53, 199-215. doi:10.1111/j.1460-2466.2003.tb02586.x

23.

Hagen

L. M.

(1993). Opportune witnesses: An analysis of balance in the selection of sources and arguments in the leading German newspapers’ coverage of the census issue. European Journal of Communication, 8, 317-343. doi:10.1177/0267323193008003004

24.

Hastie

Park

(1986). The relationship between memory and judgment depends on whether the judgment task is memory-based or on-line. Psychological Review, 93, 258-268.

25.

Hayes

A. F.

(2018). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach (2nd ed.). New York, NY: The Guilford Press.

26.

Hayes

A. F.

Glynn

C. J.

Huge

M. E.

(2012). Cautions regarding the interpretation of regression coefficients and hypothesis tests in linear models with interactions. Communication Methods and Measures, 6, 1-11. doi:10.1080/19312458.2012.651415

27.

Hopmann

D. N.

Shehata

(2011). The contingencies of ordinary citizen appearances in political television news. Journalism Practice, 5, 657-671. doi:10.1080/17512786.2011.603967

28.

Hovland

C. I.

Weiss

(1951). The influence of source credibility on communication effectiveness. Public Opinion Quarterly, 15, 635-650. doi:10.1086/266350

29.

Hovland

C. I.

Janis

I. L.

Kelley

H. H.

(1953). Communication and persuasion. New Haven, CT: Yale University Press.

30.

Iyengar

(1991). Is anyone responsible? How television frames political issues. Chicago, IL: The University of Chicago Press.

31.

Jagers

Walgrave

(2007). Populism as political communication style: An empirical study of political parties’ discourse in Belgium. European Journal of Political Research, 46, 319-345. doi:10.1111/j.1475-6765.2006.00690.x

32.

Jebril

Albæk

De Vreese

C. H.

(2013). Infotainment, cynicism and democracy: The effects of privatization vs personalization in the news. European Journal of Communication, 28, 105-121. doi:10.1177/0267323112468683

33.

Kahneman

Tversky

(1972). Subjective probability: A judgment of representativeness. Cognitive Psychology, 3, 430-454. doi:10.1016/0010-0285(72)90016-3

34.

Kleemans

Schaap

Hermans

(2017). Citizen sources in the news: Above and beyond the vox pop? Journalism, 18, 464-481. doi:10.1177/1464884915620206

35.

Koch

Zerback

(2013). Helpful or harmful? How frequent repetition affects perceived statement credibility. Journal of Communication, 63, 993-1010. doi:10.1111/jcom.12063

36.

Lefevere

de Swert

Walgrave

(2012). Effects of popular exemplars in television news. Communication Research, 39, 103-119. doi:10.1177/0093650210387124

37.

Leiner

D. J.

(2016). Our research’s breadth lives on convenience samples. A case study of the online respondent pool “SoSci Panel.” Studies in Communication and Media, 5, 367-396. doi:10.5771/2192-4007-2016-4-36769–134

38.

Lewis

Inthorn

Wahl-Jorgensen

(2005). Citizens or consumers? What the media tell us about political participation. Maidenhead, UK: McGraw-Hill Education.

39.

Mudde

(2004). The populist zeitgeist. Government & Opposition, 39, 542-563. doi:10.1111/j.1477-7053.2004.00135.x

40.

Peter

Zerback

(2017). The role of similarity for exemplification effects. Studies in Communication and Media, 6, 71-80. doi:10.5771/2192-4007-2017-1-71

41.

Pornpitakpan

(2004). The persuasiveness of source credibility: A critical review of five decades’ evidence. Journal of Applied Social Psychology, 34, 243-281. doi:10.1111/j.1559-1816.2004.tb02547.x

42.

Reinemann

Aalberg

Esser

Strömbäck

de Vreese

C. H.

(2017). Populist political communication: Towards a model of its causes, forms, and effects. In Aalberg

Esser

Reinemann

Strömbäck

de Vreese

C. H.

(Eds.), Populist communication in Europe (pp. 12-25). New York, NY: Routledge.

43.

Schulz

Müller

Schemer

Wirz

D. S.

Wettstein

Wirth

(2017). Measuring populist attitudes on three dimensions. International Journal of Public Opinion Research, 30, 316-326. doi:10.1093/ijpor/edw037

44.

Schulz

Wirth

Müller

(2018). We are the people and you are fake news: A social identity approach to populist citizens’ false consensus and hostile media perceptions. Communication Research. Advance online publication. doi:10.1177/0093650218794854

45.

Sun

Pan

Shen

(2008). Understanding the third-person perception: Evidence from a meta-analysis. Journal of Communication, 58, 280-300. doi:10.1111/j.1460-2466.2008.00385.x

46.

Tsfati

(2007). Hostile media perceptions, presumed media influence, and minority alienation: The case of Arabs in Israel. Journal of Communication, 57, 632-651. doi:10.1111/j.1460-2466.2007.00361.x

47.

Unkelbach

(2006). The learned interpretation of cognitive fluency. Psychological Science, 17, 339-345. doi:10.1111/j.1467-9280.2006.01708.x

48.

Weber

Popova

(2012). Testing equivalence in communication research: Theory and application. Communication Methods and Measures, 6, 190-213. doi:10.1080/19312458.2012.703834

49.

Wilson

E. J.

Sherrell

D. L.

(1993). Source effects in communication and persuasion research: A meta-analysis of effect size. Journal of the Academy of Marketing Science, 21, 101-112. doi:10.1007/BF02894421

50.

Zerback

(2016). Der Einfluss des Tenors von TV-Nachrichten auf die wahrgenommene öffentliche Meinung [The effect of television news slant on public opinion perception]. Publizistik, 61, 267-286. doi:10.1007/s11616-016-0266-8

51.

Zerback

Koch

Krämer

(2015). Thinking of others: Effects of implicit and explicit media cues on climate of opinion perceptions. Journalism & Mass Communication Quarterly, 92, 421-443. doi:10.1177/1077699015574481

52.

Zerback

Peter

(2018). Exemplar effects on public opinion perception and attitudes: The moderating role of exemplar involvement. Human Communication Research, 44, 176-196. doi:10.1093/hcr/hqx007

53.

Zhao

Strasser

Cappella

J. N.

Lerman

Fishbein

(2011). A measure of perceived argument strength: Reliability and validity. Communication Methods and Measures, 5, 48-75. doi:10.1080/19312458.2010.547822

54.

Zillmann

Brosius

H.-B.

(2000). Exemplification in communication. Mahwah, NJ: Lawrence Erlbaum.

55.

Zillmann

Perkins

J. W.

Sundar

S. S.

(1992). Impression-formation effects of printed news varying in descriptive precision and exemplifications. Zeitschrift für Medienpsychologie, 4, 168-185.