Black Lives Matter and Partisan Media

Abstract

Analyzing data from a survey of U.S. adults (N = 509), our study examines the role of partisan media in (a) shaping people’s perceptions of the Black Lives Matter (BLM) movement, (b) producing an echo-chamber effect, influencing users to believe that other people have an opinion similar to their own, and (c) creating a polarized public opinion environment, where people hold more extreme perspectives on BLM. Left-wing media consumption was positively correlated with favorable perceptions of BLM, while right-wing media consumption had a negative correlation, suggesting that partisan media could influence users’ own perceptions of BLM. Also, left-wing media consumption was positively correlated with the belief that others held favorable views of BLM, while right-wing media consumption was negatively correlated with such a belief, suggesting that partisan media could affect one’s beliefs about others’ perceptions. Supporting its role in producing an echo-chamber effect, frequent right-wing media usage was associated with a smaller difference between one’s own views and their assessment of others’ views regarding BLM. Finally, there was a positive correlation between left-wing media consumption and having extreme perceptions of BLM, suggesting that the use of left-wing media could play a role in creating a polarized public opinion climate.

1. Introduction

With the advances in digital technology, there has been a large increase in the number of news sources from which audiences form their opinions on social and political issues. Many news sources today tend to represent highly partisan perspectives rather than offering balanced interpretations of important issues, leading to a highly fragmented media environment (Guess et al. 2021). The way audiences consume news seems to be biased as well, selectively exposing themselves to like-minded messages and news sources. Some of the consequences of this selective exposure may involve an echo-chamber effect, where partisan media users overestimate the number of other people sharing opinions similar to their own because they are constantly exposed to like-minded messages and perspectives (Guess et al. 2018). Also, partisan media users may hold more extreme and polarized opinions on controversial issues because they may have limited opportunities to be exposed to or think about dissenting viewpoints. Their own opinions are rarely challenged, so they become increasingly confident about the validity of their perspectives, leading to holding more extreme perceptions and opinions (Levendusky 2013). Using the Black Lives Matter (BLM) movement as a case, our study examines the role of partisan media usage in (a) shaping people’s perceptions of BLM, (b) producing an echo-chamber effect, influencing their users to believe that other people have an opinion similar to their own, and (c) creating a polarized public opinion environment, where people tend to hold more extreme perspectives on BLM.

2. Literature Review

2.1. Increase in Partisan Media and Fragmented Media Environment

The media landscape in the United States has gone through a major transformation over the last decades, primarily characterized by a notable increase in partisan media outlets. This transformation is rooted in multiple factors, including advancements in digital communication technology, strategic choices made by media organizations, and a range of consequences that reverberate throughout American society. Wide adoptions of the Internet and social media platforms substantially expanded the media ecosystem, allowing for the emergence of news sources that explicitly cater to specific ideological niches and partisan leanings (Jurkowitz et al. 2020; Stocking et al. 2022). Multiple news outlets operate on different logics, competing as well as collaborating for audience attention (Klinger and Svensson 2015).

Today’s media system is also marked by partisan division among the audience. News composition and audiences on the political right are quite distinct from those on the political left (Baum and Groeling 2008; Faris et al. 2017; Stroud 2011), and liberals and conservatives largely, though by no means exclusively, gravitate toward their ideological news channels (Colleoni et al. 2014; Yarchi et al. 2021). A study by the Pew Research Center in 2020 (Jurkowitz et al. 2020) found that Fox News, a prominent conservative-leaning news outlet, has become a trusted source for a significant portion of Republicans, with approximately 65% of Republicans and Republican-leaning audiences expressing trust in the network. Similarly, CNN has garnered the trust of 67% of Democrats and Democratic-leaning audiences, demonstrating highly divided partisan preferences in media consumption. Also, the same study found evidence of growing distrust in news sources between 2014 and 2019 on both sides of the aisle. For example, distrust in CNN grew from 33% in 2014 to 58% in 2019 among Republicans, The Washington Post from 22% to 39%, and The New York Times from 29% to 42% during the same five years. Among Democrats, distrust in the Sean Hannity radio show grew from 32% to 38% and in Breitbart News from 7% to 36%. These statistics highlight the segmentation of the media landscape, with none of the 30 sources surveyed in the study showing a level of trust above 50% for all U.S. adults and instead merely having the trust of people with certain political leanings.

The digital revolution in the form of Internet access and widespread use of social media platforms seem to play a role in reshaping the media landscape. The ease of access to information through online platforms intensified competition among media outlets, and the highly competitive nature of the media industry drives media organizations to make strategic decisions to ensure their viability. In this context, news organizations have increasingly gravitated towards extreme and sensational content that resonates with specific ideological segments of the population, as it is believed to be more financially advantageous to cater to niche audiences than to attempt to appeal to a broader, more general audience (Finkel et al. 2020). A similar trend is visible on social media platforms, where user-generated content is the primary source of audience engagement. For instance, Hong and Kim (2016) found that politicians with extreme ideological positions tended to have more Twitter followers; that is, both left- and right-wing extremists were found to have a larger Twitter readership than their moderate peers.

2.2. Black Lives Matter and Partisan Media

It has been about a decade since #BlackLivesMatter first appeared on social media to spark conversation on racism, violence, and criminal justice (Bestvater et al. 2023). In July 2013, the movement emerged as a response to the acquittal of a neighborhood watch volunteer who fatally shot 17-year-old Trayvon Martin in Sanford, Florida, in 2012 (Morrison 2023). After the 2014 incident of the shooting death of 18-year-old Michael Brown at the hands of police in Ferguson, Missouri, the phrase “Black Lives Matter” became a rally cry for racial and judicial justice, and BLM has played an important role in organizing online and offline protests against police violence and racism (Associated Press 2019). BLM aims to eradicate white supremacy, empower local communities to intervene in violence against Black communities, and promote the overall improvement of African Americans’ lives and experiences (Black Lives Matter n.d.).

In the summer of 2020, BLM reached the peak of its global impact after George Floyd, a 46-year-old unarmed African American man, died after a white Minneapolis police officer, Derek Chauvin, knelt on his neck for more than nine minutes in May 2020 (Taylor 2021). Unprecedented waves of social turbulence since the Civil Rights Movement ignited massive protests in cities throughout the United States and across the globe (Deliso 2021). According to a report from the Pew Research Center, a slight majority of Americans seem to have positive perceptions of BLM, with 51% of adults expressing support for the movement in 2023, though it was a decline from 56% in 2022 (Horowitz et al. 2023). However, there are questions about the integrity of the movement as well, and a lawsuit, for example, was filed in 2022 against an executive of BLM for allegedly stealing millions of dollars in donations (Bellamy-Walker 2022), which was widely reported by media outlets of all political leanings.

BLM has been in the news headlines for years when there was an incident involving police violence against African Americans. Journalism research reports that news media in general tend to downplay social protests and often employ anti-protest frames to delegitimize race-related movements, such as BLM (Kilgo and Harlow 2019; Lane et al. 2020). Nevertheless, there seem to be noticeable differences between right- and left-wing media in their presentations of such movements, with conservative right-wing media being less likely to portray the movements in a positive way than liberal left-wing media (Mehta 2020). For example, a recent study (Cox 2022) of news stories posted on Facebook by major U.S. news outlets provided evidence for the ideology-based media bias in presenting BLM. According to the study, during the 90 days immediately following George Floyd’s death, about 35.8% of right-wing Fox News stories portrayed BLM and its protests in a positive frame, while as much as 82.0% and 92.1% of stories used a positive frame in CNN and MSNBC, which are left-wing news channels. Also, Fox News was more likely than CNN and MSNBC to use terms that could carry negative connotations when describing the actions of BLM proponents and referring to protest participants. On average, the term “riot” was used about 1.15 times in Fox News stories, while the same term was used 0.46 times in CNN stories and 0.24 times in MSNBC stories. BLM protesters were referred to as “rioters” about 0.34 times on average in Fox News stories, while the same term was used less frequently in CNN (0.04 times) and MSNBC stories (0.08 times).

The political and ideological leaning of each news organization seems to affect the way BLM, or a social movement, is presented in news stories. Research shows that the right- or left-wing orientation of the publisher is often mirrored in the editorial tone or certain managerial practices of news organizations, as it may function as a form of organizational constraint (Kim et al. 2010; Reese and Shoemaker 2016). Also, by offering content that aligns with the specific beliefs and desires of primary audiences, news organizations can secure a loyal viewership base, leading to higher ratings, increased advertising revenue, and greater profitability (Jurkowitz et al. 2020). As Entman (2007) points out, the publishers’ deeper financial and organizational resources for skilled management seem to put a limit on how far right or left their editors can go on the ideological spectrum.

Previous research has examined whether the use of partisan media could affect audiences’ perceptions of and attitudes toward BLM. From a U.S. national panel survey, Kilgo and Mourão (2019) found that greater uses of conservative media (Fox News and Breitbart) were associated with more negative attitudes toward BLM, suggesting that unfavorable presentations of BLM in conservative media might have affected audience attitudes towards the movement. However, the uses of mainstream and liberal media were largely unrelated to audience attitudes. Ilchi and Frank (2021) reported from their survey data that exposure to news coverage of police misconduct was not significantly correlated with negative attitudes toward BLM. Instead, such attitudes were most strongly and significantly associated with being white and holding symbolic racism, which referred to a belief in negative stereotypes about racial minorities. In a more recent study, Brown and Mourão (2022) used a combination of content analysis of news about BLM protests and an online survey of a national panel to investigate whether news consumption and individual predispositions could affect public support for BLM. Their findings revealed that while consuming Fox News was negatively and significantly correlated with support for the movement, respondents’ political affiliation and ideology were more closely corelated with the support, suggesting that individual predispositions might be much stronger determinants of support for BLM than media consumption. Furthermore, despite a heavy emphasis on the violence of BLM protests in CNN and AP, consuming those two news sources was positively, rather than negatively, associated with support for the movement, suggesting again that individual predispositions and preexisting attitudes might be more robust predictors of support for BLM than news coverage.

While previous studies have linked news consumption to perceptions of BLM, our study attempts to investigate further implications of such perceptions by analyzing the role of partisan media usage in creating an opinion echo chamber and political polarization, which may hinder improving the current state of race relations and promoting constructive consensus building. The proliferation of partisan media has resulted in a greater variety of news channels, each catering to specific ideological inclinations, making it easier for audiences to selectively expose themselves to like-minded channels and messages, creating a self-reinforcing cycle where individuals are constantly validated in their beliefs (Finkel et al. 2020). The increase in selective exposure contributes to the formation of echo chambers, which, by design, shield individuals from dissenting viewpoints, creating a perception that most other people share opinions similar to their own and fostering an environment where their beliefs are rarely challenged. This insulation may lead to increased confidence in the validity of their perspectives and contribute to the entrenchment of more extreme political perceptions and opinions (Hetherington and Weiler 2018). The insulation may also lead to forming a world view where opposing partisans become more socially distant, ideologically extreme, politically engaged, contemptuous, and uncooperative than is actually the case, thereby exacerbating political polarization (Mason 2018). For example, Republicans estimate that 32% of Democrats are LGBT when in reality it is 6%; Democrats estimate that 38% of Republicans earn over USD 250,000 per year when in reality it is 2% (Ahler and Sood 2018). Also, studies show that exposure to partisan media, particularly on social media platforms, can intensify existing partisan divisions (Zhang et al. 2022).

3. Research Hypotheses

Our first set of hypotheses examines the role of partisan media usage in shaping audiences’ perceptions of BLM. There is evidence to suggest that left-wing media are generally more supportive of civil rights movements, such as BLM, while right-wing media often provide critical perspectives on such movements (Cox 2022; Mehta 2020). Therefore, it can be hypothesized that the use of left-wing media will be positively correlated with holding favorable perceptions of BLM, while the use of right-wing media will show a negative correlation. Also, selective exposure can reinforce such correlations. Left-wing media’s positive presentations of BLM can resonate with their viewers who already lean left politically, and these left-leaning viewers’ use of left-wing media may reinforce their existing favorable perceptions of the movement. That is, the alignment between media content and viewer predispositions can strengthen the positive correlation between left-wing media consumption and favorable perceptions of BLM. Right-leaning viewers’ use of right-wing media will have a similar reinforcing effect, strengthening the negative correlation between the use of right-wing media and favorable perceptions of the movement. We test the following hypotheses:

H1a. 

The use of left-wing media will be positively associated with having favorable perceptions of BLM.

H1b. 

The use of right-wing media will be negatively associated with having favorable perceptions of BLM.

Our second set of hypotheses investigates whether the use of partisan media can affect audiences’ assessments of other people’s perceptions of BLM. Research has demonstrated that people tend to infer public opinion from the overall tone or impression of news coverage, believing either that such news coverage has a substantial influence on other people’s opinions (Davison 1983) or that the news coverage reflects the general sentiment of the public (Gunther 1998). That is, perceived news slant, or the degree to which news appears to be favorable or unfavorable toward a topic, can influence individuals’ perceptions of public opinion in a corresponding favorable or unfavorable direction (Kim et al. 2004). As such, left-wing media presenting BLM as a widely supported movement may influence their users to perceive greater public acceptance of the movement. That is, individuals who regularly consume left-wing media may become more likely to believe that others hold favorable perceptions of BLM. Conversely, right-wing media portraying the BLM movement in a less favorable light can lead their users to believe that the movement is not widely supported. Therefore, those who predominantly consume right-wing media may become less likely to believe that other people have favorable perceptions of BLM. We test the following hypotheses:

H2a. 

The use of left-wing media will be positively associated with believing other people to have favorable perceptions of BLM.

H2b. 

The use of right-wing media will be negatively associated with believing other people to have favorable perceptions of BLM.

Our H3 examines the role of partisan media in producing an echo-chamber effect, influencing the audience to believe that other people have an opinion similar to their own. In highly fragmented and polarized media environments, it becomes a lot easier for the audience to selectively expose themselves to news stories that align with their own beliefs (Flaxman et al. 2016). For example, those who have favorable perceptions of BLM may selectively use left-wing media, which tend to present BLM in a positive way, and this will in turn lead to the belief that the majority of other people have favorable perceptions of BLM as well. Individuals who frequently consume left-wing or right-wing media are more likely to interact with like-minded messages and individuals (Steppat et al. 2021), and as a consequence of these interactions, there will be a smaller difference between their own perceptions of BLM and what they believe others think about the movement. We propose the following hypothesis:

H3. 

There will be smaller differences between own and other people’s perceptions of BLM among those who use partisan media (either left- or right-wing) more frequently.

Our last hypothesis (H4) investigates the role of partisan media usage in creating polarized public opinion environments, where people hold extreme, rather than moderate, perspectives on BLM. By selectively exposing themselves to news channels that align with their own ideological and partisan leanings, individuals may constantly interact with like-minded messages, and such interactions may allow them to have their own opinions validated and reinforced rather than questioned or challenged (Steppat et al. 2021). As a result, those who use partisan media sources frequently may become more confident about their own beliefs and opinions, which in turn can lead to holding more extreme, rather than moderate, perspectives. As Levendusky and Malhotra (2016) observe, partisan media consumption may reinforce and amplify pre-existing political beliefs, making individuals more polarized and divided in their views. We hypothesize that those who frequently consume left- or right-wing media are more likely to develop extreme and polarized perceptions of the BLM movement.

H4. 

More frequent users of partisan media (either left- or right-wing) will have more extreme (extremely favorable or unfavorable) perceptions of BLM.

4. Materials and Methods

4.1. Sample

Data were collected using an online survey distributed by Qualtrics in March 2021. Using non-probability quota sampling, members of the Qualtrics national panel were recruited based on age, gender, and race quotas—ages 18–34 (30.4%), 35–54 (34.4%), 55+ (35.2%); male (49.0%), female (51.0%); non-Hispanic White (62.0%), non-Hispanic Black (12.0%), Hispanic (17.0%), other (9.0%)—designed to reflect the demographic makeup of the U.S. population. Respondents were not allowed to participate in the survey once their subgroup (age, gender, race) reached its quota. Invitation emails served as a consent form, and IRB approval was obtained in March of 2021. Several data quality checks were used to make sure the Qualtrics respondents paid close attention to the questions and their responses when taking the survey. First, the survey involved several attention check questions, and the respondents who failed to pass the attention questions were immediately moved to the end of the survey. Also, a speeder check was employed, using the median completion time (10 min). The respondents who spent less than five minutes on the whole survey were eliminated from the final sample. Straight-liners were eliminated from the final sample as well. Finally, we employed the forced response option to make certain the respondents would not skip over any questions. Respondents were allowed to move on to the next survey page only if they answered—including “don’t know” and “not applicable”—entire questions on a previous page. Only fully completed surveys were included in our data, and the final sample size was 509.

4.2. Measurements

Our first dependent variable, having favorable perceptions of BLM, was measured by asking respondents to rate their level of agreement (1 = strongly disagree; 5 = strongly agree) with six positive or negative statements about the movement (e.g., “BLM is a peaceful civil rights movement.”). Actual question wordings and descriptive statistics are reported in

Table 1. Question Wording and Descriptive Statistics.

Question Wording	M	SD
Having Favorable Perceptions of BLM (Self) (α = 0.77)
“The following are statements that some people made about BLM. Please indicate how strongly YOU personally agree or disagree with each:” (1 = strongly disagree; 5 = strongly agree)	19.23	5.58
BLM is a militant movement. (reversed)	2.87	1.45
BLM opposes White Americans. (reversed)	2.81	1.40
BLM makes Black Americans look bad. (reversed)	2.88	1.43
BLM is a peaceful civil rights movement.	3.40	1.35
BLM wants to prevent the unjust deaths of Black people at the hands of the   police.	3.74	1.22
BLM supports the racial equality of all Americans	3.52	1.33
Believing Other People to Have Favorable Perceptions of BLM (α = 0.69)
“Now, let’s talk about other Americans. How strongly do you think average Americans would agree or disagree with the following statements?” (1 = strongly disagree; 5 = strongly agree)	18.06	4.71
BLM is a militant movement. (reversed)	2.60	1.26
BLM opposes White Americans. (reversed)	2.51	1.21
BLM makes Black Americans look bad. (reversed)	2.62	1.26
BLM is a peaceful civil rights movement.	3.26	1.29
BLM wants to prevent the unjust deaths of Black people at the hands of the   police.	3.65	1.13
BLM supports the racial equality of all Americans	3.42	1.25
Perceived Racial Inequality in the U.S. (α = 0.73)
“This section is about your views on race relations in the United States. Please indicate how strongly you agree or disagree with the following statements:”	19.92	4.90
The United States has made all of the required changes to make Black people   equal. (reversed)	2.84	1.34
Black people in the United States are held back by racial discrimination.	3.50	1.23
If Black people can’t succeed, it is mostly because of their own actions and   decisions. (reversed)	2.72	1.31
The police are more likely to arrest a Black person than a White person   Simply because of their race.	3.58	1.23
In general, the police are more likely to use violence against peaceful Black   protesters than peaceful White protesters.	3.57	1.22
It is much more difficult to be a Black person in the United States than a   White person	3.70	1.19

. These six items were combined into a single index (α = 0.77) to represent respondents’ overall perceptions about BLM. To measure believing other people to have favorable perceptions of BLM, respondents were asked to assess how strongly average Americans would agree or disagree (1 = strongly disagree; 5 = strongly agree) with the same six statements about BLM (see Table 1). Again, the six measures were combined into a single index (α = 0.69). The difference between one’s own and other people’s perceptions of BLM was an absolute value of the difference obtained from subtracting one’s assessment of other people’s perceptions from his or her own perceptions of BLM. Higher scores indicated a greater difference (M = 2.77; SD = 3.87). To measure having extreme perceptions of BLM, the 5-point scale (1 = strongly disagree; 5 = strongly agree) measurements of own perceptions of BLM were folded over into a 3-point scale (0 = neither agree or disagree; 1 = somewhat agree or disagree; 2 = strongly agree or disagree) and then combined into a single index (α = 0.89; M = 13.05; SD = 3.92). Higher scores indicated having more extreme (either favorable or unfavorable) perceptions.

Our primary independent variable, partisan media usage, was measured with the use of left-wing media and the use of right-wing media separately. On a 5-point scale (1 = never; 5 = very often), respondents were asked about the frequency of using four highly liberal media outlets (Huffington Post, Buzzfeed, the Late Show with Stephen Colbert, and MSNBC), and these four measures were combined into an additive index to represent the use of left-wing media (α = 0.86; M = 10.13; SD = 5.15). Similarly, the use of right-wing media was measured by the frequency of using four highly conservative media outlets (the Rush Limbaugh Show, Fox News, One American News, and News Max). Again, these four measures were combined into a single index (α = 0.86; M = 9.86; SD = 5.35).

Our demographic controls included: Age (median = 42 years), gender (50.5% female), income (median = USD 50,000 to USD 74,000), race (White 62.9%, Black or African American 11.8%, Hispanic 16.5%, Asian 5.9%, Other 2.9%), and education (median = associate degree in college). Respondents’ ideology (1 = very liberal; 5 = very conservative, M = 3.10; SD = 1.33) and political party (Democrats 55.2%, Independents 25.3%, Republicans 19.4%) were also measured as control variables. Despite the carefully constructed quota sampling, our sample was on average slightly more educated (an associate degree in college as the median) and included more Democrats (55.2% of the sample) than the nation as a whole. Therefore, those who tend to have more favorable views on BLM—the highly educated and Democrats (Horowitz et al. 2023)—were over-represented in the sample, limiting the generalizability of our findings. Nevertheless, inasmuch as our focus here is on exploring the hypothesized patterns of relationships between measurements rather than providing a description of specific views of the population, it is unlikely that this over-representation of certain demographics will significantly bias the outcomes of our hypothesis testing.

Finally, perceived racial inequality in the U.S. was measured as another control variable. On a 5-point scale (1 = strongly disagree; 5 = strongly agree), respondents were asked to indicate their agreement with six statements regarding overall discrimination, as well as discriminatory police practices, against Black people in the United States (e.g., “Black people in the United States are held back by racial discrimination”; see Table 1 for question wording and descriptive statistics). These six measures were then combined into an additive index (α = 0.73).

The use of a standard agree-disagree scale poses the issue of acquiescence bias, in which respondents tend to select positive, rather than negative, response options (e.g., “agree” rather than “disagree”) disproportionately more frequently without consciously considering the content of the question or their actual preferences (Baron-Epel et al. 2010). To minimize the acquiescence bias, we employed a balanced scales approach in designing our measurement items. For example, the measurement of perceived racial inequality (see Table 1) involved a combination of positively and negatively framed items (e.g., “Black people in the United States are held back by racial discrimination” vs. “If Black people can’t succeed, it is mostly because of their own actions and decisions”). There were no positive and statistically significant correlations between the positively and negatively framed items, which would indicate the presence of substantial acquiescence bias. However, the positively and negatively framed items were not negatively or significantly correlated either, suggesting that acquiescence bias was present but not very substantial.

5. Findings

Our H1a and H1b test whether one’s partisan media usage is correlated with his or her own perceptions of BLM. The hypotheses were tested using a regression model, where the level of favorable perceptions of BLM was regressed onto the frequencies of left-wing and right-wing media usages (see

Table 2. Regressions (OLS) Predicting Perceptions of BLM (N = 509).

	Having Favorable Perceptions of BLM (Self)	Believing Other People to Have Favorable Perceptions of BLM	Difference between Own and Other People’s Perceptions of BLM	Having Extreme Perceptions of BLM
Age	0.074 *	0.071	−0.007	0.124 *
Gender (female)	0.061	0.038	−0.059	−0.112 *
Household Income	0.024	0.022	−0.026	0.071
Education	−0.057	−0.053	0.049	−0.028
Race (base: White)
African American	0.080 *	0.110 *	−0.000	0.078
Hispanic	0.059	0.079	0.018	−0.045
Asian	0.060 *	0.088 *	−0.046	−0.028
Other	0.058	0.034	−0.038	−0.013
Ideology (conservative)	−0.133 ***	−0.012	−0.132 **	0.069
Political Party (Republican)	−0.104 **	−0.130 **	0.015	−0.046
Perceived Racial Inequality	0.496 ***	0.263 ***	0.238 ***	0.141 **
The Use of Left-Wing Media	0.296 ***	0.200 **	−0.104	0.203 *
The Use of Right-Wing Media	−0.321 ***	−0.166 *	−0.164 *	−0.054
R-Square (%)	58.6 ***	21.8 ***	15.9 ***	13.3 ***

Note: Entries are standardized regression coefficients (β) after all variables entered the model.* p < 0.05; ** p < 0.01; *** p < 0.001.

, the “Having Favorable Perceptions of BLM (self)” column). Demographic controls (age, gender, income, education, race) entered the regression first. Because people’s partisan media usage could be related to their political and ideological orientations, which could also be related to their perceptions of BLM, respondents’ ideology and political party were controlled for in the regression as well. Similarly, because a person’s overall perceptions of racial inequality could be related to his or her usage of partisan media as well as perceptions of BLM, a measure of perceived racial inequality in the U.S. entered the regression as a control variable. Finally, two measures of partisan media usage (the use of left-wing media, the use of right-wing media) entered the regression. Multicollinearity diagnostics were carried out to check the potential severity of collinearity between the independent variables. The highest VIF in our regression was 4.02, suggesting that multicollinearity remained within the acceptable range (1.0 < VIF < 5.0, James et al. (2013)).

Supporting H1a, the frequency of using left-wing media was positively and significantly (β = 0.296; p < 0.001) associated with favorable perceptions of BLM even after controlling for one’s demographics, ideology, political party, and perceived racial inequality. H1b was supported as well. As shown in Table 2, the use of right-wing media was negatively and significantly (β = −0.321; p < 0.001) correlated with favorable perceptions. These correlations suggest that the use of partisan media may affect audience perceptions of BLM.

H2a and H2b examine whether partisan media usage is related to audiences’ assessments of other people’s perceptions of BLM. As shown in Table 2 (see the “Believing Other People to Have Favorable Perceptions of BLM” column), the use of left-wing media was positively associated with believing other people to have favorable perceptions (β = 0.200; p < 0.01), while the use of right-wing media was negatively correlated with the same belief (β = −0.166; p < 0.05). Therefore, both H2a and H2b were supported in our study. More positive presentations of BLM in left-wing media might have affected the audience to believe that other people have positive perceptions of the movement, while negative presentations in right-wing media might have led to an opposite belief.

H3 examines an echo-chamber effect and predicts that there will be a smaller difference between own and others’ perceptions of BLM among those who use partisan media—either left- or right-wing—more often. Supporting the hypothesis, the use of right-wing media was negatively and significantly (β = −0.164; p < 0.05) associated with the difference (see Table 2, the “Difference between Own and Other People’s Perceptions of BLM” column). However, the use of left-wing media was not significantly correlated with the difference. Therefore, our H3 was supported only for the use of right-wing media.

Our H4 examines the polarization of perceptions as a consequence of using partisan media. We hypothesized that frequent users of partisan media would have more extreme—either extremely favorable or unfavorable—views about BLM. Table 2 (see the “Having Extreme Perceptions of BLM” column) shows that the use of left-wing media, as hypothesized, was associated positively and significantly (β = 0.203; p < 0.05) with having extreme perceptions of BLM. However, the use of right-wing media was not significantly correlated with extreme perceptions. Therefore, our H4 was supported only for the use of left-wing media.

6. Discussion

Our study provides important insights into the role of partisan media in shaping the nature of public opinion environments regarding a civil rights movement. In the context of the Black Lives Matter (BLM) movement, we examined whether the use of partisan media was correlated with public perceptions of BLM, as well as other potential consequences, such as forming an echo-chamber effect and creating a polarized public opinion environment.

Our findings support the idea that partisan media can play a role in shaping users’ own perceptions of BLM. As hypothesized, left-wing media consumption was positively correlated with favorable perceptions of BLM, while right-wing media consumption showed a negative correlation. Studies have revealed that left-wing media tend to present civil rights movements in a positive light, while right-wing media’s portrayals are relatively less positive (Cox 2022; Mehta 2020), and such differences in the presentation of the movement might produce opposite directions of correlations. Even after controlling for political ideology, party affiliation, and perceptions of racial inequality, these correlations remained statistically significant, suggesting that partisan media may play a role in affecting individual perceptions of BLM.

Also, our findings showed that partisan media could affect not only one’s own perceptions but also his or her beliefs about other people’s perceptions of BLM. Left-wing media consumption was positively correlated with the belief that others held favorable views of BLM, while right-wing media consumption was negatively correlated with such a belief. As demonstrated in previous studies (Gunther 1998; Kim et al. 2004), people seem to infer public opinion, or the opinion of others, from the overall impression of news coverage. Positive portrayals of BLM in left-wing media seem to produce perceptions of a public opinion climate favorable toward the movement, while somewhat negative presentations in right-wing media seem to create perceptions of an unfavorable public opinion climate.

We found evidence to suggest that partisan media may function to produce echo chambers, where the users are constantly exposed to like-minded messages and individuals, coming to believe that other people share opinions similar to their own. Frequent right-wing media usage was significantly correlated with a smaller difference between one’s own views and his or her assessment of others’ views on BLM, suggesting that the use of right-wing media may produce an echo-chamber effect. However, the use of left-wing media did not show such a significant correlation. One possible explanation is that frequent users of left-wing media may include not only strong supporters of BLM but also some individuals who have skeptical views about the movement, so that there is a significant difference between these individuals’ own and their assessment of others’ views. On the contrary, the large majority of right-wing media users may have critical, rather than highly favorable, views about BLM, so that most of them perceive a small difference between their own and others’ views on BLM. Certainly, it is beyond the scope of our study to test the validity of this explanation, and future research may further investigate whether and why the echo-camber effect can be more pronounced among consumers of right-wing media.

Finally, we examined the role of partisan media in creating a highly polarized public opinion climate regarding the BLM movement. We found a positive and statistically significant correlation between left-wing media consumption and having extreme perceptions of BLM, suggesting that frequent usage of left-wing media can lead to polarization, affecting the users to have more extreme, rather than moderate, views. However, the use of right-wing media was not significantly correlated with extreme perceptions. Taken together, these findings suggest that both left-wing and right-wing media usage may potentially have an effect that is not constructive for a democratic debate. Echo-chamber effects were pronounced among right-wing users; heavy users of right-wing media demonstrated smaller differences between their own and their perceptions of other people’s opinions on BLM. When it comes to left-wing media, heavy users of left-wing media indicated having more extreme views on BLM, suggesting that the use of left-wing media may function to produce greater opinion polarization.

Before further discussing our findings, it is necessary to mention some of the limitations of this study. First, the cross-sectional nature of our survey data limits our ability to make a strong inference about causality. For example, regarding the correlation between the use of left-wing media and favorable perceptions of BLM, it is somewhat unclear whether left-wing media, as we hypothesized, do in fact affect audience perceptions, or whether most users of left-wing media simply happen to be strong supporters of BLM in the first place. Certainly, this ambiguity of causal interpretation applies to the correlation between right-wing media usage and perceptions of BLM as well. When it comes to the correlation between partisan media usage and audience assessments of other people’s perceptions of BLM, it is also unclear whether the media, as we hypothesized, affect such an assessment that may create an echo chamber, or whether frequent users of left-wing media, who tend to have favorable perceptions of BLM, simply project their own perceptions onto other people, and frequent users of right-wing media do the same, creating a correlation between the use of partisan media and audience assessments of public opinion. To better demonstrate the causal directions of our hypothesized relationships, future research can employ a longitudinal survey, where the respondents’ partisan media usage and their perceptions of public opinion are measured at multiple time points.

Another limitation is that we simply “assumed” that left-wing media would be supportive of BLM, while right-wing media would present the movement in a negative light. Although this was an assumption based upon previous content analyses of news coverage of BLM and other social movements (e.g., Cox 2022; Mehta 2020), we did not actually analyze news coverage of BLM in the four specific left-wing news outlets and the four specific right-wing media used in our study. Without carrying out an extensive content analysis, it is unclear whether news coverage of BLM was in fact favorable in those four left-wing media and unfavorable in the four right-wing news outlets. Future research may combine survey data with a content analysis to adequately validate the link between partisan media usage and its presumed effects. Researchers can draw hypotheses based on the findings of their own content analysis and use a survey to test the hypotheses.

With these and other limitations in mind, our findings can offer important implications for understanding how the use of partisan media can affect the nature of the public opinion environment. Certainly, the potential role of partisan media usage in promoting a highly polarized political climate deserves scholarly attention, as the United States has experienced an increasingly stark divide between Democrats and Republicans on a long list of social and political issues (Dimock and Wike 2020). Our study provides evidence for the idea that partisan media in the highly fragmented media environment today could be at least partially responsible for the polarization of American politics. We found that frequent uses of right-wing media were correlated with an opinion echo-chamber effect, and frequent uses of left-wing media were associated with extreme views. The polarizing role of both left- and right-wing media is concerning, given the seemingly deteriorating state of race relations in the U.S. today. In a 2019 Pew Research Center survey (Horowitz et al. 2019), about 58% of Americans believed that race relations in the country were in a negative state. Among these respondents, as many as 70% felt that the situation was getting even worse. The potential connection between partisan media and political polarization can provide new opportunities to generate more sophisticated hypotheses than those advanced in the previous literature, and our study may offer useful insights for theory building that may make a significant contribution to the literature.

Though not formally hypothesized, perceived racial inequality was one of the strongest predictors of all four dependent measures in our study (see Table 2). The significant role of racial attitudes in shaping support for BLM does align with past studies. For example, Ilchi and Frank (2021) found that negative attitudes toward BML were strongly associated with holding symbolically racist attitudes, which captured such a belief that racism was no longer a serious problem in the U.S that minorities’ situation in society tended to be the result of their lack of hard work, and that they should be respectful of the police, which are comparable to our measures of perceived racial inequality (see Table 1). As Ilchi and Frank (2021) explain, negative perceptions of BLM may be driven at least partly by resentment and bias against racial minorities, and it is not surprising that racial bias is significantly correlated with negative attitudes toward a civil rights movement, like it has been related to opposition to other efforts aimed to address racial inequality, such as affirmative action and public assistance (Tuch and Hughes 2011). Also, as shown in Table 2 (see the “Having Favorable Perceptions of BLM (self)” column), our findings suggest that preexisting attitudes, such as racial attitudes, may play a larger role than news consumption in shaping perceptions of a civil rights movement (Brown and Mourão 2022). It is possible that those who perceive great racial inequality may still have a favorable attitude toward BLM despite their exposure to negative news coverage of the movement. As Brown and Mourão (2022) explain, exposure to violent BLM protests in the news can signal necessary moral pursuits among these individuals that grant legitimacy. Violence is justified as necessary and may even boost their support for the movement. On the contrary, for those who perceive little racial inequality, exposure to the violent protests in the news may primarily reinforce their preexisting racial attitudes, creating more unfavorable perceptions of BLM.

As Bellovary et al. (2021) point out, partisan media, regardless of their political orientation, tend to often employ negative affect to elicit emotional responses and engagement from their audiences, suggesting that they can promote emotionally charged beliefs and attitudes. Political echo chambers and polarization, combined with emotionally charged political discourse, can manifest most starkly in what has been characterized as hate-based politics (Finkel et al. 2020). Hate-based politics extends beyond mere policy differences; it encompasses the realm of identity and values, creating a stark “us versus them” mentality (Hetherington and Weiler 2018). As Mason (2018) witnesses, polarization in U.S. politics may have moved beyond disagreements about policy matters and seems to be rooted in social identity. Both parties tend to view each other with distrust, holding highly negative views of their opponents. The proliferation of partisan media and the creation of echo chambers may have made the “us versus them” rhetoric a powerful force in shaping the nature of political debates and electoral choices in American politics. It is also likely that the presentations of civil rights movements and race relations in partisan media may have contributed to forming the “status threat” among traditionally high-status Americans (i.e., whites, men, Christians), a perception that their status as the dominant social group is threatened by growing domestic racial diversity, leading to an increased support for a candidate who promises to reestablish status hierarchies in the past (Mutz 2018). Therefore, one of the important contributions of our study was to demonstrate that frequent uses of partisan media might lead to an opinion echo chamber and polarization, which in turn might contribute to the developments of the “us versus them” mentality, hate-based political discourse, and status threat, some of the important features of American politics today that may characterize the current public opinion environment. The polarized environment created by partisan media makes it increasingly difficult for political actors to reach agreements and compromises. The political discourse becomes dominated by impassioned rhetoric and partisan grandstanding, with little room for meaningful dialogue (Strickler 2018). Constructive dialogue becomes hindered in an atmosphere where individuals rarely encounter opposing viewpoints (Morris and Morris 2022). The culmination of hate-based politics and heightened emotions can lead to eruptions of violence (Finkel et al. 2020), as exemplified by the 2021 Capitol invasion. The echo chamber effect, with its role in reinforcing extreme opinions and limiting exposure to alternative perspectives, might have played a contributing role in this disturbing event.