From Hostile Media Perceptions to Action: The Role of Racial Empathy in the Black Lives Matter Movement

Empathy is a core psychological orientation that can help communities come together disentangling social and political turmoil, as empathy incentivizes people to assist others in situations of distress (McDonald & Messinger, Citation2011). In political campaigns, electoral processes, policy-related rallies, and street protests, to the health and sanitary measures that a pandemic or natural disaster may trigger, empathy can become a clear counterpoint to social division and ultimately political polarization.

Bearing in mind that many images of what happens in the world are experienced indirectly, through mediated accounts, it stands to reason that the empathy, or lack thereof, that we develop for others, is in part contingent on media representations of those others, as well as our perceptions of the truthfulness of those representations. In this study, we argue that perceptions of media hostility toward a social group can lead to feelings of empathy toward that group, which in turn may result in increased likelihood to engage in actions that group takes as part of its struggles for social acceptance.

The demonstrations against police brutality sparked by the killing of George Floyd quickly grew into a historically large nationwide movement calling for racial justice. The fact that BLM in 2020 amassed support across ethnic boundaries (Buchanan et al., Citation2020) raises important questions for communication scholars as to what factors might have galvanized citizens to further the movement with actions, in a context of extensive media coverage and social media chatter about anti-racism and against violence toward the Black community. To contribute to the growing trend of investigating motivational forces for action amidst political polarization (Carnahan et al., Citation2023), it becomes important to delve deeper into the underlying mechanisms driving individuals, particularly across diverse ethnic backgrounds, to actively participate in such movements.

We employ the Orientation-Stimuli-Orientation-Response (O-S-O-R) framework, derived from the communication mediation model in political communication effects research (Cho et al., Citation2009; McLeod et al., Citation2001). The aim is to offer a structured approach that elucidates the relationship between party identification (O₁ – an individual’s predispositions prior to encountering a stimulus) and perceptions of bias in mainstream and social media (S – media messages and interactions that shape individuals’ attitudes), which subsequently fosters empathy toward racial minorities (O₂ – shifts in one’s motivational and cognitive states), motivating engagement in BLM actions (R – ensuing behaviors).

From O₁ to S: Mainstream Media Coverage and Social Media Discourse Surrounding BLM

Social movements seeking to challenge the existing power structure are often perceived as politically liberal and Democratic as they clash with the conservative ideal of preserving the status quo (Kostelka & Rovny, Citation2019). Though their main purpose is generally grounded in humanitarian values such as “equality,” the politicization of social movements might account for the reason why mainstream media coverage of protests generally fall under the two broad categories of “legitimizing” and “delegitimizing” protest frames.

Legitimizing frames are typically seen as a journalistic effort to increase public support for collective action against perceived inequality. They usually emphasize a movement’s social critique and “treat protesters as a legitimate entity” (Brown & Mourão, Citation2021, p. 5), presenting a sophisticated discussion of protesters’ grievances and goals (Hertog & McLeod, Citation2001). A BLM news story that employs a legitimizing frame may voice the protesters’ viewpoint, for instance, putting emphasis on the signs that the protesters are holding or on protesters’ demands for police accountability (Mourão et al., Citation2021). In a different vein, delegitimizing protest frames aim to negatively affect public opinion concerning the movement to advance/protect a particular political agenda. Some of the most common types of delegitimizing protest frames include “confrontational frame” and “riot frame,” both of which revolve around physical actions of protesters, which often lack explanations into the reasons behind the movement. In stressing protester-opposition (e.g., police) clashes, confrontational frames may depict protesters as combatants rather than social critics (Hertog & McLeod, Citation2001). Existing works note that frames that deliberately or inadvertently delegitimize protesters and their collective efforts are more common than legitimizing frames (Hertog & McLeod, Citation2001; Mourão et al., Citation2021).

The riot frame, in particular, is known to exaggerate chaos created by protesters and highlight criminal activity that may occur during a protest, such as looting. Such coverage tends to characterize protesters as menaces and a danger to society while portraying the bystanders/public as being hostile toward them (Hertog & McLeod, Citation2001). A recent analysis of newspaper frames revealed that journalists were more likely to employ a riot frame to cover the anti-Black racism protests than they did to report protests about health and environment (Brown & Harlow, Citation2019). Similarly, Mourão et al. (Citation2021) found that the initial newspaper coverage of the 2014 BLM movement sparked by the shooting death of Michael Brown tended to stress the episodes of violence and riot over protesters’ grievances. The adoption of delegitimizing frames to cover a social movement seems to be present in most media regardless of their ideological leaning (Leopold & Bell, Citation2017).

The pervasiveness of delegitimizing frames in mainstream media coverage of protests calls for scholarly inquiry into its influence on citizen support for antiracism movements. A national survey conducted after the 2014 BLM protests showed that there were more Americans who supported the movement than those who opposed it – the level of support was especially high among Blacks, but the findings were consistent across different racial groups including Whites (Horowitz & Livingston, Citation2016) – raising an interesting question about how news audience with preexisting ideological preferences would perceive the contemporary trend in mainstream media coverage of antiracism protests.

Here we draw attention to the hostile media perception (HMP), a robust communication phenomenon that has been somewhat rarely explored in relation to BLM. Previous studies have shown strong evidence for people’s tendency to see news coverage as biased against their own ideological viewpoint, and such tendency is known to be particularly pronounced in the context of a politically polarized issue such as immigration (Watson & Riffe, Citation2013) and climate change (Feldman et al., Citation2017). In the context of BLM, the fact that those that are highly involved in politics tend to have greater HMP than nonpartisans (Gunther et al., Citation2009) should lead one to assume a significant positive link between party ID strength and the likelihood of perceiving bias in media coverage of the BLM movement. And considering that supporters of the Democratic Party are significantly more likely to support the BLM movement than their ideological opponents (Horowitz & Livingston, Citation2016), greater support for the Democratic Party would be associated with the likelihood of perceiving that the protesters were being unfairly depicted by the mainstream media. The original concept of HMP states that one’s political leaning affects her/him to perceive hostility in neutral media reports (Vallone et al., Citation1985), but given that much of BLM coverage in recent years were found as having either a legitimizing or delegitimizing frame, we expect the presence of “relative HMP” among those who follow mainstream news on BLM, perceiving protest coverage to be “more hostile to [or less agreeable with] their own point of view than the opposing group sees it” (Gunther & Chia, Citation2001, p. 690).

When it comes to the contents circulated on social media platforms such as Facebook and Twitter regarding BLM, they tend to be largely supportive of social movements, challenging the narrative formulated by the government and the mainstream media regarding protesters (McGarry et al., Citation2019). Research, however, notes the consistent presence of information that expresses counter-movement sentiments (Ince et al., Citation2017). A recent analysis of tweets by the U.S. Congress members revealed that Republicans’ posts with BLM hashtags largely focused on criticizing perceived rioting and looting that took place during the protests (Panda et al., Citation2020).

This suggests more similarities than differences between mainstream media coverage of BLM and information on BLM circulated on social media platforms in terms of the frames they adopt to express support or disapprove the movement. Applying the notion of relative HMP, one could expect Democrats to perceive bias even in the information they consume on social media, a setting in which the discourse around BLM is still largely supportive of the protesters. We therefore expect Democratic Party supporters to perceive that the BLM information they encounter on social media are relatively hostile to their expectations – that is, focusing too much on the violent aspects (e.g., rioting and looting) of the protests. The present study considers social media as interactive communication channels have the potential to expose users to both ideologically congruent and incongruent opinions rather than those that isolate users in partisan echo chambers, a view that has increasingly been questioned by communication scholars (e.g., Dubois & Blank, Citation2018).

Perceiving bias in mainstream media is likely associated with perceiving bias in social media due to the flow of information between these platforms. Recent studies highlight that information often flows between mainstream media and social media platforms (Valenzuela et al., Citation2017; Zuckerman, Citation2021). Biased narratives or content perceived in mainstream media can be shared and amplified on social media platforms, potentially reinforcing perceptions of bias across both mediums. We thus predict the following:

H1:

Party identification is positively related to perceiving hostile bias in BLM information, such that greater Democratic Party support positively relates with perceiving that (a) mainstream media coverage and (b) social media information focus too much on the violent aspects of the protests.

H2:

Perceiving bias in mainstream media coverage of BLM is positively related to perception of social media bias regarding BLM.

From S to O₂: Racial Empathy as a Potential Consequence of Hostile Media Perceptions

Conceptualizing empathy as the imagining of the feelings, thoughts, and actions that others are experiencing (Mehrabian & Epstein, Citation1972), scholars argue that the ways in which media discourse frames events and those involved may elicit empathetic responses such as sympathy and compassion among the observers. For instance, Gross (Citation2008) found that frames that provide a narrative on the plight of a particular individual to illustrate a broader social issue increased empathetic reactions, conforming to the notion that human interest details are persuasive and emotionally engaging to the audience.

Although such findings may accordingly imply that delegitimizing protest frames devoid of detailed descriptions of individual protesters would not foster favorable feelings among viewers, several studies suggest that perceiving hostile media bias could lead individuals to react against the message rather than the subjects of the message, potentially eliciting empathetic attitudes toward the group members portrayed in the news. Individuals may perceive injustice/unfairness when they perceive news coverage as biased (e.g., Hwang et al., Citation2008), which could consequently trigger empathetic attitudes toward the target of the unjust treatment, given that empathetic, “other-oriented” responses arise when one witnesses others’ suffering (Cikara et al., Citation2011). This might be particularly the case when the ingroup rather than outgroup members are perceived as oppressed or in need – McKeever (Citation2015) found that people are more likely to empathize with those that are socially similar to themselves. Haider‐Markel et al. (Citation2007) showed that perceiving negative media portrayal of Black victims following Hurricane Katrina led African Americans to become more empathetic toward storm victims.

In this view, perceiving BLM coverage to be unfairly slanted against one’s ideological position could generate empathy in pro-movement citizens for the victims of the hostile media environment, that is in this case, protesters being depicted by news media as violent rioters. More specifically, HMP in the context of a social movement fighting for racial justice may evoke racial empathy, which in this study is conceptualized as empathic thoughts toward members of racial minority groups (Wang et al., Citation2003).

Based on previous works that shows the role of empathy in motivating one to engage in action to help victims of injustice (Batson, Citation1998; Cartabuke et al., Citation2019), the present research predicts a significant positive association between racial empathy derived from perceptions of bias in BLM coverage and citizens’ participation in the protest movement. Classified under the category of moral emotions that motivate prosocial tendencies (Pagano & Huo, Citation2007), empathy has often been identified as an important predictor of ethical decision-making and altruistic behavior in a political situation (Batson, Citation1998). In particular, Cartabuke et al. (Citation2019) indicated that those higher on empathy may be more likely to hold positive attitudes toward social justice issues, hinting at empathy’s potential to generate socially responsible behavior. Drawing on the work by Saab et al. (Citation2015) that illustrates a serial path model in which perceived injustice significantly relates to collective action tendencies through empathetic attitude, the following hypotheses are set forth:

H3:

Perceiving bias in (a) mainstream media and (b) social media coverage of BLM is positively related to racial empathy.

H4:

Racial empathy is positively related to participating in (a) peaceful BLM action and (b) confrontational BLM action.

Despite the lack of extensive evidence on the direct link between party identification and racial empathy, some have argued that Democrats would be more likely than Republicans to empathize with racial and ethnic minorities, if empathy in the context of misfortune were to involve understanding others’ emotional states and feeling compassionate for those in need (Hasson et al., Citation2018). Previous studies found a significant positive association between support for liberal policies and self-reported empathetic feeling toward others in general (Iyer et al., Citation2012) and suggest that those high in social dominance orientation, notably conservatives, tend to score relatively low on empathic concern (Hasson et al., Citation2018). Moreover, liberals are more likely than conservatives to attribute external rather than internal causes for others’ misfortune (e.g., Cozzarelli et al., Citation2001) – Democratic Party supporters in the context of BLM would thus be more empathetic toward those that are subject to racial profiling than Republicans, perceiving systemic racism as the cause for the way Black people in America are treated by police.

Given the significant role mass media outlets and social media platforms played in spreading information and discourse around racism after the killing of George Floyd, one should note the potential difference between the political left and right in terms of the targets of their empathetic concern. Waytz et al. (Citation2016) argue that liberals are inclined to expend empathy toward “larger, farther, less structured, and more encompassing social circles” (p. 64) compared to conservatives who tend to prioritize smaller/closer circles over larger ones (e.g., family over humanity). In this sense, Democrats’ centrifugal concern for the targets of police brutality – about whom they have largely learned through mediated experience – would have increased their support for the social justice movement.

Additionally, it is worth noting African Americans’ allegiance to the Democratic Party, whose supporters by and large have endorsed policies to reduce racial discrimination and hold more favorable images of Blacks than Independents and Republicans (Welch & Sigelman, Citation2011). Based on the notion of intergroup empathy bias, which posits a difference in empathy for in- vs. out-group members (Cikara et al., Citation2014), it is plausible to expect Democrats to share and be affected by the emotion that their group members feel. We therefore propose:

H5:

Party identification is positively related to racial empathy, such that greater Democratic Party support positively relates to higher levels of empathy.

From O₂ to R: BLM Protests as “Corrective” Actions

According to earlier works on HMP, perceiving media to be biased against one’s view may directly promote political participation. The argument rests on the following premises: (1) the hostile media phenomenon tends to occur when an information channel is perceived as having a broad reach (Gunther & Liebhart, Citation2006), (2) believing that large audiences would be affected by hostile media motivates one to take political action in an attempt to sway public opinion in favor of one’s own views, that is, to “correct” public opinion (Barnidge & Rojas, Citation2014; Rojas, Citation2010). Strong partisans are known to be more prone to bias perceptions than weak partisans or independents and thus are more likely to engage in corrective action or to speak out against the perceived majority (Feldman et al., Citation2017). Communication research has repeatedly demonstrated a significant link between HMP and expressive forms of corrective behaviors that occur in an offline setting (e.g., participating in a protest; Rojas, Citation2010). HMP may stimulate a sense of “media indignation,” or perception of injustice and unfairness in news media (Hwang et al., Citation2008; see also Vallone et al., Citation1985), which could eventually lead to activities aimed at correcting the injustice.

Prior studies indicate that HMP regarding a specific political issue would result in activism related to the very issue one thinks is unfairly treated by the news media. Feldman et al. (Citation2017) found a significant association between HMP in media coverage of climate change and the likelihood of engaging in climate change activism such as donating money to an organization to reduce global warming. Similarly, Ho et al. (Citation2011) suggest that HMP among issue publics (i.e., those especially invested in a particular issue) may directly contribute to issue-specific (e.g., stem cell research) participation. From the context of our study, perceiving mainstream and social media information on BLM to be biased against one’ views may have a significant positive relationship with the odds of participating in BLM actions. More specifically, considering that audiences would be willing to correct negative media portrayal of protests (e.g., rioting and looting) and its possible influence on public opinion, bias perceptions would be associated with peaceful rather than confrontational activities.

H6:

Perceiving bias in (a) mainstream media coverage and (b) social media information is positively related to participating in peaceful BLM action.

We additionally ask whether party identification directly relates to the likelihood of engaging in BLM activities. A recent national survey found that nearly 80% of those who attended a BLM protest after the death of George Floyd identified with the Democratic Party (Barroso & Minkin, Citation2020), alluding to a direct positive association between Democratic Party support and BLM activities. Here we assume BLM participation to be of a non-confrontational type, acknowledging the lack of evidence for (and thus the need to explore) the link between party support and confrontational actions.

H7:

Democratic Party support is positively related to participating in peaceful BLM action.

Additionally, although the 2020 BLM movement was found to have been overwhelmingly peaceful (Chenoweth & Pressman, Citation2020), the current study investigates into the possibility that confrontational types of action may arise from peaceful actions. Recent findings from multinational data suggest that nonviolent protests may escalate into violence (1) when the protests are unorganized (i.e., lacking identifiable hierarchical structures), which would allow violence-oriented participants to easily break group norms, and (2) when peaceful forms of participation are repressed by authorities, thus making nonviolent demonstrations as costly as the violent ones (e.g., similar risk of imprisonment) (Ives & Lewis, Citation2020). These may explain instances of escalated tensions at the George Floyd protests (Taylor, Citation2021), a decentralized, nondestructive movement that repeatedly clashed with police throughout its duration (Barker et al., Citation2021).

At an individual level, a peaceful action may lead to a deeper, more extreme form of engagement through the process of “self-persuasion.” One of the basic rules of persuasion, is that commitment to an action is very persuasive (Cialdini, Citation2008). It has been argued that even an action taken without strong conviction has the potential to initiate a sequence of self-justification to eventually strengthen previously held attitudes and mount to committed action (De Meyer et al., Citation2020). Participating in a protest might not persuade those watching the protest about its legitimacy, but it has a powerful impact on the participants themselves. Therefore, through action the persuader is persuading him or herself of the correctness of his or her actions and thus if nothing else, protesting should persuade the “protester” that further action is required thus engaging in a cycle of behavior.

It is plausible that protests participants increasing sense of the righteousness of their cause coupled with the negativity bias of media coverage, that is focusing for example on violence, would lead participants to legitimize and engage in increasingly confrontational forms of protesting. It is also plausible that selective exposure, that focuses the attention of engaged participants on the police repression of the protest can legitimize more confrontational forms of protest. Literature on self-persuasion posits that individuals are naturally inclined to reduce cognitive dissonance, or the perception of inconsistency between their actions and originally held beliefs (Aronson, Citation1999), and that they would often do so through the use of external self-justification strategies that diminish personal responsibility for their behaviors (Holland et al., Citation2002). Perceived violence by police may serve as a cognitive means through which protesters justify, therefore self-persuade, their intention to engage in physical aggression. In view of negativity bias as a widespread phenomenon that occurs among traditional and social media users (Park, Citation2015), we propose:

H8:

Participating in peaceful BLM action is positively related to participating in confrontational BLM action.

Despite limited evidence demonstrating the effect of party preference on the tendency to participate in physically aggressive political action, considering Republicans’ support for stricter policies toward crime (Fine et al., Citation2019), this study additionally asks whether meaningful association exists between party identification and the likelihood of engaging confrontational form of BLM activities.

RQ:

Does party identification significantly relate to participating in confrontational BLM action?

Our expectations and research question are summarized and presented as a sequential path model in Figure 1.

Figure 1. Proposed path model predicting BLM actions.

Display full size

Method

Results

To test our hypotheses and research questions, we analyzed our theoretical path model using Stata 16.0 software with maximum likelihood estimates with missing values. The proposed model indicated good fit to the data based on the following fit indices: chi-square goodness-of-fit index (χ²₍₅₎ = 5.263, p = .385), comparative fit index (CFI = 1.000), root mean square error of approximation (RMSEA = .010; p = .861) and standardized root mean square residual (SRMR = .005). The results of our path analysis are summarized in Figure 2 – standardized coefficients are reported.

Figure 2. Results of path analysis (N = 489).

Coefficients estimated by maximum likelihood (ML). χ²₍₅₎ = 5.263, p = .385; CFI = 1.000; RMSEA = .010, p = .861; SRMR = .005.

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

Display full size

Greater support for the Democratic Party was positively related with the likelihood of perceiving mainstream and social media information as biased against one’s views (H1a: γ = .15, p < .01; H1b: γ = .10; p < .05). Additionally, perception of mainstream media bias was significantly associated with that of social media bias (H2: β = .62, p < .001). Our results indicate that perceiving hostile bias in BLM information fully mediates the link between Democratic Party support and empathetic attitudes (H3a: β = .16, p < .01; H3b: β = .11; p < .05) – the direct path from party ID to racial empathy (H5) was not significant.

Our analysis revealed a statistically significant relationship between racial empathy and peaceful BLM action (H4a: β = .12, p < .01), but empathy was not directly associated with confrontational action (H4b) in the proposed model. We report a direct path from perception of mainstream media bias to peaceful action (H6a: β = .11, p < .01), which was not significantly correlated with the likelihood of perceiving bias in BLM information on social media (H6b). The direct relationship between party ID and peaceful action was not confirmed (H7). As expected, we found a strong positive association between the two types of BLM actions (H8: β = .55; p < .001). Although Democratic Party support was negatively related with the odds of engaging in physical confrontations in our model (RQ: β = −.12; p < .01), the total (direct and indirect) effect of party ID on confrontational action was not significant (θ = −.08; p > .05). Standardized direct, indirect and total effects are presented in Table 2. Overall, the model explained about 18% of the total variance (R²) in perceived mainstream media bias, 19% for social media bias, 31% of variance for racial empathy, 41% for peaceful action and about 40% of total variance in confrontational BLM action.

We performed a series of post-hoc analyses to (1) account for the possibility that physical confrontations could precede peaceful actions and (2) test whether the path coefficients would significantly differ between the White and non-White samples. The alternative model in which confrontational action was treated as a predictor of peaceful participation yielded a bad model fit: χ²₍₅₎ = 7.704, p = .006; CFI = .994; RMSEA = .115, p = .050; SRMR = .006. Our multigroup path analysis assessed the differences between White (n = 371) and non-White participants (n = 118) by comparing the unconstrained model (free structural parameter coefficients) to the constrained model (parameters equally constrained across two groups). The Chi-square difference test suggested that our theoretical model was variant between Whites and non-Whites: χ²_diff (9) = 17.66, p < .05. Based on the critical ratio criteria of ± 1.96 at a significance level of .05, our subsequent path-by-path comparison revealed significant racial group differences in two paths in the model: Democratic Party support was more strongly associated with perception of mainstream media bias among Whites than non-Whites (Z = −3.04, β_White = .41, p < .001; β_non-White = −.09, p > .05) whereas the relationship between party ID and perception of social media bias was stronger among non-Whites than Whites (Z = 2.45, β_White = −.08, p > .05; β_non-White = .21, p < .05).

Discussion

Taken together, the present findings propose a theoretical model grounded in the O-S-O-R framework, elucidating how partisan support (O₁) positively relates to the odds of participating in peaceful and confrontational forms of BLM actions (R) via HMP (S) and empathetic attitudes toward racial minorities (O₂). First, we report a significant association between Democratic Party ID strength and the likelihood of perceiving both mainstream and social media information on BLM to be biased – that is, regardless of the actual ideological bias presented in the coverage of the protests – suggesting the presence of relative HMP in the context of the most recent BLM campaign.

Perhaps a more compelling set of findings for this work pertains to the positive direct relationship between HMP and racial empathy, which in our sequential model was treated as a consequence of perceived media hostility. Contrary to our expectation, support for the Democratic Party was not directly associated with empathetic attitudes, as the link between two constructs was fully mediated by the perception that mainstream and social media information on BLM focused too much on the violent instances of the anti-racism protests.

Our findings imply that this may in turn trigger protest actions among politically liberal citizens in support of the perceived victims of the hostile media environment, rather than urging them to succumb to the unfavorable media climate for whom they identify as ingroup members. Here, we highlight racial empathy as a key mediator in the link between HMP and political participation, providing additional insights into the role of bias perceptions in the media in stimulating “corrective” actions (Rojas, Citation2010).

Also worth noting is that while racial empathy was directly and positively related to the likelihood of engaging in peaceful actions, empathetic responses to HMP were not shown to have a direct relationship with physical confrontations. Though the results are congruent with the notion that more intensified forms of participation may arise from peaceful actions (Ives & Lewis, Citation2020), we nonetheless show that the feelings of empathy, which in the present context would motivate one to help the perceived target of injustice, directly elicit nonviolent types of protest behaviors rather than physical forms of action. Yet, we acknowledge the limitation of our cross-sectional study design to decisively claim the causal relationships between racial empathy and the two forms of participation, as well as the effect of HMP on empathetic attitudes, which underscore the need for further longitudinal, experimental evidence to confirm such findings.

Likewise, some of the measures employed in this study may be expanded or additionally narrowed down to provide a more comprehensive/detailed assessment of the variables explored. For instance, though our single-item measures that assessed perceived prevalence of hostile information on mainstream and social media might have been sufficient to support the previously reported link between party ID strength and perceiving hostility in the media environment, future research may focus on particular media channels and online platforms (e.g., Fox News and Instagram). Our BLM action items may also be developed into a more exhaustive list of protest-related activities, although one should note that empirical evidence on the different types of confrontational protest actions remains sparse. And finally, it is worth noting that the empathy measure used in the study was not specifically designed to capture reactions to perceived media bias in BLM information. Future research may create and utilize measures designed for this specific purpose to allow investigations into whether such adjustments yield strengthened findings and offer deeper insights into the state of racial empathy and its manifestations in protest actions.

Our results also provide some unanticipated findings that are worthy of future investigation. Though our model hints at the direct effect of Republican Party support on the odds of engaging in confrontational protest actions, the absence of a significant total effect suggests future research needs to continue to explore the association between party ID and more confrontational forms of protest. The estimates from the proposed path model also indicate possible differences between bias perceptions in mainstream media coverage and those in social media contents on BLM protests in terms of their potential to elicit peaceful actions, as only the former was to shown to be directly related to protest activities. The results from our subsequent multigroup analysis further stress the medium differences in HMP, showing that certain racial groups could be more prone than others to perceive information on a particular medium as hostile toward their views. These should motivate future research to test whether “perceived reach” of a news source plays a moderating role between HMP and actions, and to look into possible racial differences in the patterns of media usage, as well as preferences for specific information channels, all of which were not directly measured in this study.

Nevertheless, the present findings provide important insights concerning the psychological underpinnings of political engagement in a contemporary social justice movement. Advancing the existing literature, which suggests that individuals may prioritize news that reinforces their own racial in-group preference (Holt & Carnahan, Citation2020), we propose that among Democrats, perception of unfair news media coverage toward racial minorities may extend beyond ethnic boundaries, demonstrating the potential for cross-ethnic solidarity and collective action. We suggest racial empathy as a direct outcome of perceiving media hostility against one’s ideological allies, as well as a driver of action for change, drawing attention to the previously unexplored mediating role of empathy for the subjects of biased information in the link between HMP and corrective action. The present work additionally provides a conceptual model through which confrontational forms of protest are understood as a possible consequence of peaceful actions, which may be used as an important reference for further analysis on the emergence of protest violence. While our findings at large may confirm the uncompromising nature of political beliefs, we believe they furthermore illustrate the altruistic side of political belongingness. In serving as a potential agenda for political communication scholars to expand the discussion on news media use and prosocial actions, as well as a direction for further works on the motivation potential of racial empathy, this study may help illuminate BLM actions as a humanitarian response to perceived injustice.