Testing the Attenuation of Protest Paradigm News Coverage Effects in the Presence of Prior Favorable Attitudes Toward a Cause

Abstract

This study explores the conditions under which the (adverse) effects of exposure to protest groups’ media portrayals are, in theory, attenuated. Two experiments manipulated the degree to which protest coverage adhered to the “protest paradigm.” One version presented the protesters as violent, unlawful, and disruptive; the other was more balanced. Group cause and news article’s source were also manipulated. Results from Study 1 showed that effects remained negative even when prior favorable attitudes toward the cause existed, with closely adherent coverage generally leading to more negative protester evaluations, lower levels of identification with protesters, and lower protest intention, often independent of participants’ prior attitudes. In contrast, Study 2 showed that the negative portrayals’ effects were rather inconsequential when participants’ ideology was considered.

“Protest paradigm” is the name given to a stable pattern of media coverage of protests, which arises from structural biases in the mainstream media, highlighting the nonnormative, violent, and deviant nature of protest groups (Chan & Lee, 1984; Gitlin, 1980; McLeod & Hertog, 1999), often in response to protest groups’ use of extreme tactics to gain media attention.

Research in the protest paradigm and its effects has traditionally focused on the impact of delegitimizing protest framings and devices in perceptions and attitudes toward relatively radical groups—with which the public has little to no direct contact—or otherwise counter-publics showcasing a conscious opposition to a dominant ideology. In such cases, the conclusions of a vast majority of empirical studies based in the United States align with traditional tenets of the protest paradigm that social movements demanding some structural social change (i.e., racial injustice), or movements employing more extreme or violent tactics, get more delegitimizing coverage than protests around policy issues having more nuanced and similar levels of support and opposition in the broader public (Boyle, McLeod, & Armstrong, 2012; Brown & Harlow, 2019; Gil-Lopez, 2021; Harlow & Brown, 2021). In the context of protest paradigm research, common framings and devices of representation considered delegitimizing are, among others, characterizations of protesters that make them appear as either eccentric or disruptive and violent (e.g., Gitlin, 1980; McLeod & Hertog, 1999), a focus on protest statistics (i.e., attendance or number of arrests) instead of its underlying causes, suggesting that there is a lack of focus or consensus within a movement and that it does not represent a societal majority (e.g., Dardis, 2006; McLeod & Hertog, 1999), or a strong reliance on official sources and definitions of protest above movement voices (e.g., Jha, 2007). In turn, the negative coverage that counter-public movements receive harms the legitimacy of the protest groups, worsening protester evaluations or identification with the protesters, among others (e.g., Gutting, 2019; McLeod & Detenber, 1999; Simpson, Willer, & Feinberg, 2018).

A different scenario is assumed regarding the coverage of protests around divisive policy issues, with equal support and opposition among the broad American public and often aligned with partisan affinities (Pew Research Center, 2023). By policy issues, we refer to issues involving conflict over the direction of government policy. Examples of issues addressed by public policy in the United States include public health, gun rights, immigration, and reproductive rights. Two main conditioning factors of protest paradigm mechanisms are critical to understanding how frames from the protest paradigm may be linked to such policy issues: First, commercial mainstream media are thought to be less capable of demonizing any given side of the issue. Since these issues are relevant to many people’s lives, protest coverage “may provide stories and images that appear as credible realizations of an issue’s importance to the news audience” (Smidt, 2012, p. 75). Accordingly, it is harder to communicate stereotypical ways of thinking about movements based on issues involving public policy than about movements in open opposition to dominant structures.

A critical justification for examining the attenuated effects of the protest paradigm in this study is the inherently divisive nature of certain political issues (e.g., immigration and gun control), which are deeply rooted in partisan ideological divides (Nayak, Fraser, Panagopoulos, Aldrich, & Kim, 2021; Westfall, Van Boven, Chambers, & Judd, 2015). Unlike traditional protest paradigm research, which predominantly examines protests associated with counter-publics or structurally oppositional movements (e.g., Brown & Harlow, 2019; Harlow & Johnson, 2011; McFarlane & Hay, 2003), this study focuses on causes that evoke strong, preexisting ideologies within the public. Our shift in attention is significant for two reasons: first, divisive policy issues (like immigration and gun control) provide a unique opportunity to examine how prior attitudes and ideologies (such as political ideology) influence protester evaluations, group identification, and protest intentions. Because these issues are closely tied to individual identity and belief systems, they allow for a more refined framework to examine whether exposure to delegitimizing coverage can override or reinforce these preexisting positions.

Furthermore, these issues highlight the interplay between identity, group dynamics, and media framing. Divisive policy issues are more likely to activate social identity mechanisms where individuals categorize themselves as members of in-groups or out-groups relative to the protest cause (Sanderson, Frederick, & Stocz, 2016). This context allows for an examination of whether delegitimizing media portrayals predominantly influences out-group evaluations or whether they also affect in-group perceptions. By exploring these mechanisms, the study sheds light on the limitations of the protest paradigm’s effects in an increasingly polarized public sphere, offering critical insights into the mechanisms that drive the attenuation of these effects in politically charged contexts.

Next, the public’s responses to media portrayals of protest are greatly influenced by preexisting attitudes (Brown & Mourão, 2022; Hsiao & Radnitz, 2020). Audiences often hold views on the protest issue, especially when a cause becomes public interest and can create opinion fault lines in societies. Yet so far, there has been relatively scant attention to the interplay between coverage characteristics and prior attitudes, particularly for those issues creating partisan-aligned opinion divides in the American public opinion. When readers hold firm opinions on an issue, their attitudes toward the protest cause and group (i.e., identifying with the movement, intending to participate in it) might be somewhat more resistant to coverage influence than when those opinions are not in place; therefore, an attenuation of protest paradigm effects seems feasible. Such two issues are immigration and gun control. In 2023, Republicans were “more than twice as likely as Democrats to say that illegal immigration is a very big problem” (70% vs. 25%, Pew Research Center, 2023, p. 5), while these differences remain just as extensive with regards to gun violence (81% of Democrats vs. 38% of Republicans, p. 4). Partisan divisions over specific gun policy proposals have grown wider over the past years (Pew Research Center, 2021). Both issues have generated a great deal of protest activity over the years in the United States in the recent past (e.g., Andone, 2018; Garsd, 2023; Roegiers & Almond, 2018; Suderman & Rankin, 2020).

In response to the above, we present two experimental studies examining the implications of protest paradigm-adherent coverage in news when individuals oppose or support a cause prior to exposure to protest coverage. Here, it is essential to ask if the presumably adverse effects of mainstream protest coverage (i.e., negative protester evaluations, low group identification, low protest intentions) are conditional upon citizens’ degree of sympathy for a group’s cause. Alternatively, discursive elements emphasizing deviant or nonnormative aspects of protests (i.e., group or event not conforming to a predominant view or to the norm/law), traditionally ascribed to the protest paradigm, may trigger perceptions that the protest group is an out-group; that is, unrepresentative of one’s own relevant social memberships and identity. If so, any coverage evoking images of violent and disruptive protesters, eccentric appearance/behavior, or arrests would still be sufficient to prevent readers from identifying with protesters and mobilizing, even for a cause they strongly support. The two studies presented here approach these questions from an intergroup perspective, proposing that the effects of protest paradigm framings differ from in-group to out-group portrayals. We consider individual differences in issue positions on two divisive issues in the current political opinion climate: immigration in Study 1, and gun control in Study 2. The study applies the protest paradigm to the opposite sides of each issue (i.e., pro-/anti- immigration/gun control).

Protest Coverage and Prevalence of the Protest Paradigm in the American Press

Prior literature has established that the patterns of the protest paradigm in the American press have been generally prevalent and stable over the past two decades (Gil-Lopez, 2021). These patterns showcase a general tendency to portray protesters and their causes negatively (Brown, 2021; Harlow & Brown, 2022; Kilgo, Harlow, García-Perdomo, & Salaverría, 2018; Richardson, 2020; Schmidt, 2024), although with some important nuances: for example, demonstrations against racial inequality have been reported to be among the most delegitimized in the U.S. mainstream media (Brown & Harlow, 2019), in contrast, among others, to immigration protests.

In turn, the extent to which news stories adhere to the standards of representation generally described by the protest paradigm has been shown to have a significant impact on perceptions of the legitimacy of protest groups or even social protest as a political tool (McLeod, 1995; McLeod & Detenber, 1999; Shoemaker, 1982). Ultimately, negative coverage leads to lower identification with the protest group (Simpson et al., 2018), perceptions of morality (Orazani & Leidner, 2019), group efficacy when protestors employ violent tactics (even among sympathizers, Thomas & Louis, 2014), and a lack of support and mobilization (i.e., protest intention) from the general public (Gutting, 2019). Some findings suggest that the latter (behavioral) outcomes could be mediated by identification (Wouters, 2019).

The link between media coverage of protests and individuals’ attitudinal responses is thus firmly established. We build on previous research, which has traditionally supported the idea that exposure to negative portrayals of protesters influences several outcomes, including evaluations of protesters, identification with them, and intentions to protest. This is because these outcomes are influenced by underlying principles of group legitimacy and support.

Case Studies: Protest Paradigm Adherence in Coverage of Immigration and Gun Policy Protests

As mentioned earlier, while movements demanding some structural social change tend to get more delegitimizing coverage than policy issues, both the coverage of immigration and gun policy protests are far from presenting homogeneous characteristics in the United States.

Immigration. The wave of pro-immigrants’ rights protests in 2006 attracted a great deal of media attention; the coverage was mixed despite the public’s massive support. McLeod (2007) underscored the positive tone of the coverage but noted that “no article provided a detailed explanation of the issues behind the protests” (p. 190). Other authors pointed to opposite patterns of coverage from Spanish and English-speaking press (Félix, González, & Ramírez, 2008; López-Sanders & Brown, 2020; cf. Baker-Cristales, 2009), with the latter contributing to solidifying an identification of the immigrants involved as “Mexicans” and a cultural threat (López-Sanders & Brown, 2020). Yet, when compared to movements fighting racial or ethnic injustice, coverage of pro-immigration protests was more legitimizing of the protesters’ demands and often employed the debate frame (Brown & Harlow, 2019).

Unexpectedly, there is a lack of comprehensive analysis on the coverage of anti-immigration protests. There is a notable scarcity of case studies, with the few available primarily focusing on European contexts. In the United States, some reports seem to suggest that certain nationwide anti-immigrant groups have successfully gathered sufficient media attention and encouraged the mainstream media to use them as legitimate sources (e.g., Rincon, 2021) . This is echoed by academic accounts collecting the opinions of pro-immigration organizers, who express frustration that the mass media “flock to cover the activities of tiny anti-immigrant groups” (Costanza-Chock, 2014, p. 29) while they face opposition from a media machine that is often resistant to engage with openly pro-immigration discourses. Suppose we use conservative movements as a broad proxy. In that case, we find that they are subject to marginalizing frames by ideologically opposed media channels (Weaver & Scacco, 2013), while conservative media coverage helps them foster community mobilization (Williamson, Skocpol, & Coggin, 2011). In a European context, anti-immigration protest groups were found to be reported extensively by mainstream media, but mainly in negative ways, featuring statements by critical political and pro-migrant observers (Lundstedt & Nissen, 2024).

Gun policy. Prior literature has accumulated enough evidence that gun control and gun rights groups have their unique framings—for example, public health and safety, rights, and democracy (Steidley & Colen, 2017); gun ownership as uniquely American, self-defense (Merry, 2018), respectively—and they seek to garner media attention to extend their frames to the public (Steidley & Colen, 2017). The conclusions of the reviewed studies all seem to suggest that gun control marches’ suggested frames are more effective in increasing news coverage than gun-rights protests (Steidley & Colen, 2017), pointing to more favorable coverage, a priori. The impact of gun control movements is also more significant in moving public support (Smidt, 2012) and policy changes (Sato & Haselswerdt, 2022), where pro-gun marches have no effect. Merry (2018) argues that protests being sparked by mass shootings (“focusing events”) may be a critical mediating factor in the relationship between mass shooting events and increases in both public support for gun control as well as gun policy. Yet, due to their financial advantages, the gun rights movement may be far less dependent on attention-grabbing moments such as protests as opportunities to grab media attention and define the debate (Laschever & Meyer, 2021). This could explain the overall high rates of media attention to gun rights organizations (e.g., NRA), whose share rarely dropped below 80% of the gun rights movement’s coverage in the New York Times between 1945 and 2015 (Laschever & Meyer, 2021). These studies were all focused on case studies, and the sample of media outlets was reduced; the specific strategies commercial mainstream media used to frame protests from either side remain under-investigated. More recent developments have warranted associations of gun rights with broader White nationalist and supremacist movements (Bogel-Burroughs & Oppel, 2020), and studies have reported high adherence to protest paradigm standards, for instance, in coverage of Charlottesville’s Unite the Right protests in 2017 (Ismail, Torosyan, & Tully, 2019), appealing to public order framings and political responsibility of the Trump government as the protests progressed (Ophir, Forde, Neurohr, Walter, & Massignan, 2023).

These findings suggest that the analyzed portrayals could refine existing theories on protest paradigm framings, particularly concerning policy issues. In today’s media landscape, the variety of responses to protests has increased, and mainstream media’s portrayals may not be the only sources shaping the public’s understanding of protest issues. Several studies have noted that partisan news media, which increasingly educate the citizenry, provide different coverage patterns of the social movements behind policy issues (Brown & Harlow, 2019, 2021, 2022; Jackson, 2021; Oliver & Meyer, 1999; Weaver & Scacco, 2013). These ideological media actively promote images of opposite ideological groups and causes that hurt their agenda. In the end, this creates a scenario where strong divisions over such partisan or ideologically rooted issues as immigration and gun rights in the United States are sufficient conditions for partisan media to produce delegitimizing portrayals of antagonist protest movements to sway public opinion in favor of their ideology. Thus, we argue that in today’s polarized political and media ecology, it is possible for protest groups favoring one side of an issue to become de facto “counter-publics” for the opposite side.

This section has reviewed how the media have traditionally covered the opposite sides of the case studies just presented (pro-immigration and gun control, anti-immigration and gun-rights) because the present study will apply the protest paradigm to both sides of each issue in a systematic manner, which may or may not coincide to any degree with how mainstream media usually portrays them.

Prior Attitudes and Group Membership Cues

A majority of protest paradigm research has traditionally focused on portrayals of relatively radical collectives or counter-publics showcasing a conscious opposition to the dominant ideology (see Brown & Harlow, 2019; Kilgo et al., 2018 for protests relating to racism and colonialism). In these cases, individual attitudes around the protest group and their cause are often strongly shaped by hegemonic media representations by the U.S. media, prioritizing privileged opinion groups (e.g., Herman & Chomsky, 1988; Richardson, 2020). The consequences of being exposed to portrayals highlighting negative group characteristics can potentially be harmful. The cognitions these portrayals elicit about the group are a priori susceptible to the systematic biases and distortions reproduced by mediated communications (Bodenhausen, Kang, & Peery, 2012). Mainstream portrayals describing movements as a deviant “other,” unrepresentative of the public opinion, predispose the public to attribute negative characteristics to these groups, which they perceive to be outside the boundaries of their in-groups. Social Identity Theory (Tajfel & Turner, 1986) helps explain why individuals are unlikely to challenge negative out-group representations as long as such representations do not negatively impact their self-concept (Tajfel, Billig, Bundy, & Flament, 1971). Any given group presented by the news media as radical, or simply uncivil and disruptive, may easily be detached from normative ideas of being a law-abiding, good citizen. As long as media portrayals underscore positive definitions of oneself concerning the radical other, they may be accepted and helpful in informing one’s attitudes toward the out-group.

A different picture is presented when media coverage portrays opinion groups formed around highly divisive public opinion issues, such as immigration (Pew Research Center, 2023) and gun policy (Pew Research Center, 2021), both based on broad partisan divides in the United States. Here, the mainstream public is likely to be familiar with and even share the positions advocated by protest groups. As the saliency of such policy issues grows, certain predispositions may be developed toward groups based on personal views. The perceptions that mainstream coverage elicits are then likely to be complemented by such existing attitudes toward movements and groups (Brown & Mourão, 2022; Hsiao & Radnitz, 2020; Masullo, Brown, & Harlow, 2023; McLeod & Detenber, 1999; Wouters, 2019).

Further, strong positions on an issue may trigger perceptions of shared group membership with the members of a protest group when people are exposed to protest portrayals. Relevant group membership cues might gain relevance in shaping evaluations of the portrayed group. On policy issues such as the ones under study, these group memberships are often assumed to be mostly partisan or ideological. However, other aspects of identity—race, citizenship—are also deeply entrenched into immigration and gun policy positions in the American public. These additional group membership cues also manifest in media coverage and often coincide with the most negative media framings: for instance, immigration in the United States is usually thought of as a “Latino” or even “Mexican” issue (Zepeda-Millán, 2014), and the issue has frequently been racialized and gendered in the press, with Latino men becoming the quintessential “illegals” (“bad hombres,” see Silber Mohamed & Farris, 2020) and “offenders” of U.S. immigration law (Rim, 2009).

Notions of non-citizenship and illegality, in essence, group nonmembership, underscore the immigrant-native boundary and translate into exclusionary narratives of criminal threat (e.g., López-Sanders & Brown, 2020; Silber Mohamed & Farris, 2020; Zepeda-Millán, 2014), potentially shaping migrant protest participation (Zepeda-Millán, 2014), public opinion, and pushing conservative groups into advocating for harsher legislation. The fact that it is mostly pro-immigration and barely any anti-immigration group portrayals are being discussed in the literature suggests that group categorizations reflect power imbalances in representation. Silber Mohamed and Farris (2020) argue that “the media helps to develop social constructions of ‘target groups,’ referring to the population most affected by a particular policy debate, which in turn shapes popular opinion about policy issues […] and towards immigration specifically” (p. 2). Thus, protest-paradigm framings can alter the group imagery associated with an issue, leading to evaluations of issues based on attitudes toward the group in question rather than the issue at hand (Gilens, 1999; Nelson & Kinder, 1996). Importantly, in the long run, these changes in individual issue positions could turn into shifts of core political and partisan identities (Abrajano, Hajnal, & Hassell, 2017), one of the primary membership cues articulating our studies.

Likewise, the gun policy debate appeals to culturally rooted identities, from what it means to be an American citizen (i.e., gun ownership and use as normal and uniquely American; Steidley & Colen, 2017) and associations to the Second Amendment to racial legacies on who is the legitimate subject of gun rights (Driscoll, 2023). Merry (2018) points to implicit biases in gun control groups’ strategic emphasis on keeping guns out of the hands of “dangerous criminals.” Competing frames around the gun debate are actively pushed by advocate groups on both sides, which strongly impact individuals’ issue interpretation and policy preferences, particularly now that groups have direct communication with the public (Merry, 2020).

We, therefore, assume that reactions to protest-paradigm frames will be driven, at least in part, by a set of relevant symbolic political and social identities (e.g., Democrat, liberal, or Hispanic) that can affect the relevance of the information to the individual. Accordingly, when strongly favorable predispositions toward the protest group exist before exposure, readers may be less susceptible to the effects of negative story frames (Hsiao & Radnitz, 2020). This is consistent with the theoretical proposition, from persuasion research, that individuals rely on prior evaluations rather than considering new information to evaluate once their attitudes are formed (Roskos-Ewoldsen, 1997; Sanbonmatsu & Fazio, 1990). As prior research suggests (Hsiao & Radnitz, 2020; Power, Murphy, & Coover, 1996), in-group evaluations can remain favorable independent of whether portrayals favor the group or not. In contrast, evaluations of a group not perceived as one’s own could vary accordingly (Arpan et al., 2006). As an illustration, Brown and Mourão (2021) considered how prior attitudes toward movements influenced how protest paradigmatic coverage affected individuals, finding that they influenced exposure outcomes more than the media frames. Protest-paradigm mechanisms may be subject to audience expectations after all: when individuals’ stereotyped beliefs about a protest group were not reflected in protest coverage, they deemed more legitimizing coverage less credible (Masullo et al., 2023). Ultimately, then, studying the interaction between framings and prior attitudes or symbolic identities is relevant to extending the explication of protest-paradigm mechanisms, especially considering that coverage may solely influence mobilizing behavior (i.e., protest participation) when paired with meaningful symbolic identities (Merolla, Pantoja, Cargile, & Mora, 2013).

In what follows, we introduce predictions based on experimental manipulations of media framings (low vs. high adherence to protest paradigm) and protest group cues (pro- or anti-immigration/gun policy) and test the moderation effect of prior attitudes toward a cause (strongly favoring—strongly opposing). First we hypothesize that:

H1: The more favorable individuals’ attitudes are to the protest cause (pro-/anti-immigration), a) the less negative their evaluations of protesters will be, b) the greater their identification with the protesters, and c) the greater their protest intention.

Next, to test attenuation effects, we propose:

H2: Coverage that adheres more closely to the protest paradigm will lead to a) more negative protester evaluations, b) lower identification with the protesters, and c) lower protest intention than coverage that adheres less closely to the paradigm, but this pattern will be less pronounced among individuals who strongly favor the protest cause (pro-/anti-immigration) than among those who strongly oppose it.

Method

Study 1

Participants. A total of 609 respondents participated in this study. The sample was drawn from Amazon Mechanical Turk, a web-based platform for recruiting workers to complete tasks. Participants received a monetary payment of $1 to complete a 20-min online survey. To increase sample quality, we only recruited workers from the United States who had at least 500 jobs or HITs (Human Intelligence Tasks) approved and a HIT approval rate greater than 98%. We also deleted 64 cases (10%) due to the extremely short/long completion time (M ± 3 SD), and 16 respondents (3%) failed to pass data quality checks. The final sample was composed of 529 cases (M age = 37.2 years; 54.6% men). Complete sample descriptive statistics are reported in Table 1.

Materials and procedures. Four different versions of a newspaper article covering a protest event were prepared (see Figure 1 and Appendix  A, Figures A1–A4). First, one experimental manipulation was that the protest group portrayed in the stories was either in favor of or opposed to immigration in the United States. Stories were pretested to ensure that they were about protest and that protesters were for or against immigration. The two conditions differed significantly from one another (on a 5-point scale, in favor of immigration, M = 3.65, SD = 1.508; against it, M = 2.19, SD = 1.460, t(176) = −6.537, p < .001), and from a neutral position (pro, t(92) = 4.125, p < .001; against, t(84) = −5.127, p < .001). Two versions were created for each position (pro/against), each considered a different type of coverage in relation to its degree of adherence to the protest paradigm: one utilized some of the most common devices of representation traditionally ascribed to the protest paradigm (Figure 1), the other did not. All versions ranged from 235 to 275 words. The protest paradigm framing was the experimental version, and the low adherence version was the control.

Therefore, the experiment uses a 2 (adherence to protest paradigm: low or high) × 2 (advocacy group: pro- or anti-immigration)¹ posttest only, between-subjects design with random assignment to condition. Data² were collected between 19th and 29th March 2019. Upon agreeing to participate in the study on Mechanical Turk, participants were redirected to the online survey hosted by Qualtrics. The participants provided their consent and then answered a set of survey items measuring the valence and strength of their position on several issues, including immigration. Respondents indicated their agreement with the statement “U.S. restrictions on immigration have gone too far” (−3 = strongly disagree, 3 = strongly agree). They also rated the personal importance of each issue (0 = not at all important, 4 = very important). Participants were then randomly assigned to one of the four conditions. First, they read a fictitious news article about a demonstration organized by the group “Immigration Reform Now!.” The measurement of the dependent variables and manipulation checks followed the experimental exposure.

Measurement. After reading the manipulated story, respondents answered questions about their evaluations of protesters, identification with the group, and perception of protesters’ intentions. Negative protester evaluations measured agreement with various statements about the protesters participating in the demonstration, such as “The protesters were disrespectful” or “The protesters were out of line,” on a 7-point scale. A composite scale was created by averaging scores, with higher numbers denoting more unfavorable attitudes (M = 3.74, SD = 1.606). Identification with protesters was also measured by four items ranging from 1 to 7 (e.g., “I am like the members of [group]”; M = 3.57, SD = 1.922). These two scales were adapted from McLeod (1995). Finally, protest intention was assessed following Gutting’s (2019) operationalization, with an item asking participants how likely they were to attend a future protest event if “Immigration Reform Now!” were to organize it in their area (M = 2.40, SD = 1.854). The independent and control variables were the exposure group, political ideology, political engagement, trust in government, and past protest participation³ (item and scale statistics are reported in Appendix  B, Table A1). These were assessed using existing measures from the ANES (n.d.).

Analytic strategy. To reflect participants’ stance on immigration, we multiplied their positions’ valence and strength (e.g., neutral, somewhat opposed) by their interest in the issue, following Gunther and Christen (2002). The variable ranged from −12 (strongly opposes immigration; personally very important) to 12 (strongly favors; very important; M = 1.16, SD = 6.456). Scores closer to 0 reflected less extreme stances, less interest in the issue, or both. The measure was z-normalized for analyses. All hypotheses were tested using linear regression models.

Results. To ensure that random assignment of participants to conditions was successful, one-way ANOVAs were conducted on independent measures. No significant differences were observed in immigration stance or personal importance of immigration (see Appendix  C). Manipulation checks were also successful (see Appendix  D).

The first set of hypotheses predicted that outcomes of exposure would be less negative the more favorable individuals’ attitudes were to the group’s stand on immigration. Supporting immigration was related to less negative evaluations of pro-immigration protesters (β_{stance × group} = −0.38, p < .001), but immigration stance was unrelated to evaluations of anti-immigration protesters (Table 2). Thus, H1a was only partially supported. Identification with protesters was negatively correlated with more favorable opinions on immigration in the anti-immigration group (β = −0.46, p < .001), and positively correlated in the pro-immigration condition (β_{stance × group} = 0.88, p < .001). Likewise, the more favorable participants were to immigration in America, the more likely they were to attend a protest organized by the portrayed group when such group defended immigration (β_{stance × group} = 0.43, p < .001), and the opposite was true for the anti-immigration condition (β = −0.10, p < .001). H1b and c were supported.

The second set of hypotheses proposed that favorable prior attitudes would attenuate the effects of protest paradigm framings. We found that highly adherent coverage (i.e., violent, disruptive protest) led to significantly more negative protester evaluations than the control condition, independent of how favorable individuals were to the causes portrayed (Table 2). High-adherence portrayals also lowered identification with protesters (H2b) (β_{stance × pp} = 0.16, p <.05), although these effects were relatively weaker for participants who supported immigration. For protest intention, there was a significant 3-way interaction between paradigm adherence, exposure group, and immigration stance (β = −0.21, p <.01). It indicated that adherent portrayals consistently decreased participants’ protest intentions in the pro-immigration manipulation, and that this effect was not as consistent in the anti-immigration manipulation (i.e., applied mostly to sympathetic readers, that is, to participants who supported an anti-immigration stance, see Figure 2). Thus, none of the three hypotheses were supported.

Conclusion. Study 1 showed that prior attitudes toward immigration impacted how adverse the outcomes of exposure were: for pro-immigration portrayals, all three outcomes of exposure were less negative when individuals favored immigration, while for anti-immigration portrayals, identification was greater among opponents than among supporters of immigration.

Contrary to our expectations, when coverage type and participants’ immigration stance were jointly tested, our results showed that closely adherent coverage overall led to more negative protester evaluations, lower levels of identification with protesters, and lower protest intention, often independent of participants’ prior attitudes about the policy issue.

Study 2

Overview. To ascertain the robustness of the findings of Study 1, we conducted a second survey experiment in March 2021. Some of the asymmetries noted above between supporters and opponents of immigration (i.e., the relationship between favorable attitudes and more positive group perceptions was almost exclusively observed among supporters) led us to consider a second polarizing topic that often causes greater fervor in conservative circles: gun control. Also, since Study 1’s findings seemed to suggest that attenuation of adverse effects did not take place, the question remained whether such lack of attenuation was because issue positions are somewhat dynamic, perhaps situationally dependent, and whether evaluative judgments of the protest group are relatively independent of the issue position—that is, that protest issue and group can be independently evaluated, to some degree. Thus, Study 2 looked at more stable predispositions (i.e., ideology as moderator) to test if they are more resistant to the impacts attributed to delegitimizing, high-adherence portrayals. While our hypotheses remain the same, a revised version is provided for clarification, now in terms of ideology as a relevant symbolic identity informing attitudes toward a cause:

H1: The more favorable individuals’ ideology is to the protest cause (gun control/gun rights), a) the less negative their evaluations of protesters will be, b) the greater their identification with the protesters, and c) the greater their protest intention.

H2: Coverage that adheres more closely to the protest paradigm will lead to a) more negative protester evaluations, b) lower identification with the protesters, and c) lower protest intention than coverage that adheres less closely to the paradigm, but this pattern will be less pronounced among individuals whose ideology is favorable to the protest cause (gun control/gun rights) than among those whose ideology is opposed to it.

Political ideology was measured on a 7-point scale ranging from very conservative to very liberal (M = 3.81, SD = 1.753), like in Study 1. All other aspects relative to design and operationalization were identical to Study 1. Once again, item and scale statistics are reported in Appendix  B.

Procedure. The data for this study were collected by the market research firm Qualtrics. Adopting an online panel-based approach, a sample was drawn from a group of U.S. nationals who closely matched the population parameters focusing on age and gender. Respondents who finished the survey too soon (one third of the median time) were excluded. In addition, 69 participants failed to pass data quality checks and were discarded from the final sample (N = 1,812). Finally, random assignment and manipulation checks were successful; they are reported in Appendices  C and  D.

Results. Following the same analysis as in Study 1, we found that more favorable ideologies toward the protest cause were consistently related to less negative protester evaluations (β_{ideology × group} = −0.44, p < .001), greater identification with protesters (β_{ideology × group} = 0.99, p < .001), and greater protest intention (β_{ideology × group} = 0.49, p < .001). Thus, all three hypotheses were supported (see Table 3).

Concerning the interaction between protest paradigm adherence and ideology, high-adherence coverage increased negative protester evaluations relative to low adherence for both group portrayals (β_pp = 0.35, p <.001), independent of ideology (β_{ideology × group × pp} = 0.1, ns). Identification with protesters and protest intention were not affected by the degree of adherence to the protest paradigm but rather entirely defined by participants’ ideology. Thus, the second set of hypotheses was rejected (Table 3, see also Figure 3).

Conclusion. Contrary to Study 1, identification with protesters and protest intention were unaffected by the portrayals’ degree of adherence to the protest paradigm. While hypothesis H1 assumed that prior favorable ideologies would cause attenuation of protest paradigm effects, it still predicted some negative effect of protest paradigm framings. This was the case in Study 1, but in Study 2, identification and protest intention related solely to ideology.

General Discussion

This study aimed to test the effects of protest paradigm framings on perceptions and attitudes toward protest groups in the presence of prior favorable attitudes that could potentially attenuate their influence.

The findings have important implications for protest paradigm literature. First, the results confirm that negative coverage of protest groups can have harmful effects, even beyond favorable preexisting attitudes. Before considering the effect of these attitudes, we ran a preliminary analysis testing the effects of protest paradigm framings alone (see Appendix  E). We found that focusing on elements that highlight the violence and deviance of protest movements directly affected people’s immediate judgments of the subjects in the portrayals. Recent studies have shown that aspects such as contentiousness and disrespect for authorities reduce support and intention to mobilize (Gutting, 2019), and this was the case in Study 1. Extreme protest tactics also reduced support for a protest group as participants’ social identification decreased (Feinberg, Willer, & Kovacheff, 2017). These patterns are in line with the findings of Study 1, where overall, protest paradigm-adherent portrayals lowered in-group identification even when prior attitudes toward a cause were strongly positive. One possible explanation is that such portrayals threaten a positive self-concept associated with the specific social identity being portrayed. Thus, individuals fail to identify with the group (Ethier & Deaux, 1994).

However, Study 2’s results challenged Study 1’s findings by showing that participants’ identification with protesters and protest intentions were independent of the nature of portrayals and related solely to ideology. One explanation is that the strength of individuals’ positions could depend on the issue. For instance, views on gun rights may be relatively stable along the ideological spectrum and less likely to be impacted by specific media framings. The findings could also entail that identifying with a particular ideology (i.e., strongly conservative), more so than holding a position on a specific issue (i.e., demanding restrictions on immigration), relates to evaluative judgments about social groups in rather fundamental ways. If that is the case, negative portrayals are unlikely to impact how much one identifies or is willing to support a group they sympathize with. Taken together, the results of Studies 1 and 2, although they may seem contradictory at first sight, align with previous protest paradigm studies, which found that some predispositions (i.e., sympathy toward protesters, support for the specific movement of Black Lives Matter) significantly influenced several outcomes of exposure while others (i.e., attitudes towards Black people) did not (Brown & Mourão, 2021). We argue that this is because some preexisting attitudes inform relevant symbolic identities associated with concrete societal groups (i.e., self-identifying as a conservative—Study 2—, supporting BLM’s fight for racial justice) while others (i.e., supporting a specific direction of government migration policy—Study 1) may be relatively inconsequential and, ultimately, insufficient to threaten a positive self-definition associated with the specific social identity symbolized by the portrayed group (Tajfel et al., 1971). We consider that this process holds important implications for an explication of protest paradigm mechanisms, namely attenuation of effects that seems to apply across protest causes, be it around policy issues or fighting hegemonic structures. While immediate protester evaluations may be hurt in response to negative media coverage, more enduring attitudes remain positive despite it; in other words, effects are attenuated, though only so long as relevant social identities are made accessible to readers and engaged. Political ideology may bear greater weight in shaping perceptions of social membership and identity around highly polarized issues than mere policy stances.

It is pertinent to point out that the findings could also inform strategies for media messaging. The insight that ideological alignment outweighs the effects of negative media portrayals could lead to a strategic shift toward strengthening ideological connections with potential supporters, where links to relevant self- and group identities would act as a powerful mobilizer of protest support. While our experimental manipulations featured relatively common standards and representation devices of the protest paradigm, they did not explicitly appeal to such intrinsic group membership cues as race/ethnicity or citizenship as much as the reviewed mainstream media portrayals have traditionally tended to do. These are likely to further shape the direction of protest paradigm effects in ways that could strengthen in-group/out-group sentiments. This calls for an urgent exploration of these portrayals, given their potential implications for (de)polarization processes.

Second, our findings also hint at a nuanced range of effects of protest coverage that may be contingent on individual differences, one of which is ideology. In study 1, fundamental ideological differences may help explain why the adverse effects of protest paradigm standards remain unattenuated, particularly for immigration opponents. Prior research suggests that conservatives, more likely to hold such a position on immigration (Carney, Jost, Gosling, & Potter, 2008), may be more susceptible to resisting certain protest definitions than liberals. There is also ample evidence of an asymmetry in protest participation between liberals and conservatives (Schussman & Soule, 2005; Van der Meer, Van Deth, & Scheepers, 2009). Study 2 addressed the asymmetry found in Study 1 by directly testing the role of ideology as a moderator and using a second topic, gun control, about which conservatives are presumed to feel strongly. We observed that their support for and identification with gun-rights protesters was consistently strong and unaffected by negative media portrayals. This is consistent with prior studies, which found significant effects of prior attitudes on protest coverage evaluations for conservatives and liberals (e.g., Masullo et al., 2023).

The contrasting findings between issue stance and ideology as moderators may also point to more fundamental interconnections among the multiple relevant symbolic identities that may inform individuals’ relation to a protest cause. We have argued that media images are populated with motifs that tie both immigration and gun policy to race, legality, citizenship, and otherness; in essence, culturally ingrained notions of group (non-)membership. While our experimental manipulations did not include direct references to other group membership cues, we can speculate that a long tradition of partisan and ideologically based debates surrounding these two divisive policy issues has possibly strengthened ties between latent political values and such symbolic group identities as being a legitimate citizen or bearer of rights (e.g., Berg, 2015).

This study has several limitations. First, while using a different issue in Studies 1 and 2 enabled us to observe a broader range of participant responses to different protest coverage types along an ample continuum of prior attitudes, contextual factors could have also impacted participant responses in ways that our experimental design is unable to discern, or formally test. The recent assault on the U.S. Capitol only 2 months prior to Study 2, and a year of nationwide protests against police brutality amidst a global pandemic may have contributed to significant breakups in public opinion concerning protests. While we mostly attribute the effects to the experimental manipulations, we know that attitudes on the issues at hand could have changed more fundamentally as situated responses to societal phenomena. A more robust design would have included both topics in the initial experiment so that any differences could have been tested.

Second, our findings were centered on the impact of media portrayals within particular scenarios; it is likely that additional influential factors—such as the political climate or individual personality traits—could affect public support for protests. Future research could build upon these results by exploring these other dimensions.

Finally, our assumption of a direct correlation between public perceptions of protest tactics and support for the cause simplifies the potential complexity of public responses to these tactics. This limitation suggests an opportunity for researchers to identify and investigate more intricate mechanisms that could influence these dynamics.

References

Biographical Notes

Teresa Gil-Lopez (PhD Communication, University of California-Davis, 2019) is a Postdoctoral researcher in Political Communication and Media Effects at the University Carlos III of Madrid. She investigates the ways in which digital technologies may have altered the relationships between social movements, the media, and the citizen discourse.

Saifuddin Ahmed (PhD, University of California-Davis) is an Assistant Professor at the Wee Kim Wee School of Communication and Information at Nanyang Technological University, Singapore. His research interests lie in new and emerging media, political communication, election studies and public opinion. He is particularly interested in the transformative role of new media in civic and political engagement in technologically emerging societies.

Laramie D Taylor has expertise in the uses and effects of traditional and new media, including effects related to sex and sexuality, gender, violence, and idealized bodies. Taylor also studies media involvement and fans.

Funding

This research was supported by the Research Initiative Young Talent Fund, University of Koblenz-Landau.

Footnotes

Appendix A. Experimental Manipulation Examples⁴

Appendix B. Descriptive Statistics for Variables in the Analysis

Appendix C. Random Assignment Checks

Study 1

First, no significant differences were found across conditions in mean levels of immigration issue position [F(7, 538) = 0.696, p >.05] and personal importance of immigration [F(7, 538) = 0.934, p >.05], nor in the combined measure on the nature of participants’ opinion [F(7, 538) = 0.386, p > .05]. There were also no significant differences across conditions for various demographic measures: age [F(7, 538) = 1.562, p > .05], gender [χ² (7, N = 546) = 12.517, p > .05], ethnicity [χ² (28, N = 546) = 14.954, p > .05], education [F(7, 538) = 0.415, p > .05], or income [F(7, 538) = 1.300, p > .05]. Finally, no significant differences were found across conditions for a set of key variables of interest, among which are political ideology [F(7, 538) = 0.262, p > .05], political engagement [F(7, 538) = 0.289, p > .05], protest participation [χ² (7, N = 546) = 5.023, p > .05], trust in government [F(7, 538) = 0.534, p > .05]. Overall, it can be concluded that random assignment was generally adequate and that the different experimental conditions were equivalent concerning individual-level factors.

Study 2

No significant differences were found across conditions in mean levels of issue position [F(7, 759.899) = 0.973, p > .05] and personal importance of gun control [F(7, 761.087) = 0.899, p > .05], nor in political ideology [F(7, 761.146) = 0.302, p > .05]. There were also no significant differences across conditions for age [F(7, 761.025) = 1.866, p > .05], gender [χ² (14, N = 1,812) = 5.165, p > .05], ethnicity [χ² (42, N = 1,812) = 43.307, p > .05], education [F(7, 761.343) = 0.719, p > .05], and income [F(7, 760.845) = 0.592, p > .05]. Furthermore, no significant differences were found across conditions for any of the control variables, political engagement [F(7, 761.718) = 0.643, p > .05] and past protest participation [χ² (7, N = 1,812) = 9.193, p > .05]. Overall, it can be concluded that random assignment was generally adequate also in Study 2 and that the different experimental conditions were equivalent concerning individual-level factors. However, there were significant differences in trust in government across the eight experimental conditions [F(7, 762.405) = 2.711, p < .01]. Thus, trust in government was added as a control variable to the regression models.

Appendix D. Manipulation Checks

Study 1

Participants correctly identified whether the group was in favor of or against immigration depending on which group they were exposed to [pro: true; against: false, χ² (1, N = 529) = 301.837, p < .001]. On the aggregate, they also correctly recalled that the protest paradigm high-adherence article featured an estimate of protest attendance (“Hundreds march…,” “About 800 activists…”) and that the low-adherence article did not state the number of protesters [χ² (1, N = 529) = 149.439, p < .001]. Overall, these results suggests that the experimental manipulations were appropriately perceived by participants.

Study 2

Study 2 improved the manipulation checks from Study 1. Participants correctly identified whether the group called for stricter gun laws or not depending on which group they were exposed to [gun control: true; gun rights: false, χ² (1, N = 1,812) = 272.848, p < .001]. There were also significant differences between the high- (M = 2.98, SD = 0.995) and low- (M = 2.37, SD = 0.927) protest paradigm adherence conditions on the average level of violence featured on the events portrayed by the news article [t(1810) = −13.546, p < .001]. Together, these tests suggest that the experimental manipulations worked appropriately.

Appendix E. Preliminary analyses on the impact of protest paradigm framings on outcomes of exposure (no prior attitudes considered)

Study 1

We found that readers exposed to stories that closely adhered to the paradigm displayed more negative evaluations of protesters (M = 4.61, SD = 1.47) than those exposed to less adherent coverage (M = 2.89, SD = 1.23), F(1,523) = 209.13, p < .001. Protest intention was also significantly lower in the high adherence (M = 2.23, SD = 1.76) than in the low adherence condition (M = 2.57, SD = 1.93), F(1,523) = 6.233, p < .05. Identification with protesters was not significantly different across conditions⁵.

Study 2

Following the same analysis as in Study 1, we found that greater adherence to the protest paradigm led to more negative protester perceptions (M = 4.48 for high adherence vs. M = 3.66 for low, F(1, 1806) = 224.945, p < .001), but did not have a direct effect on identification with protesters nor protest intention.