Selective Exposure During Uprisings: Examining the Public’s News Consumption and Sharing Tendencies During the 2019 Lebanon Protests

Abstract

This study examines selective exposure and trust during uprisings. It studies major uses of traditional and social media and assesses the public’s trust in these media and their engagement in sharing news. Focusing on the October 2019 Lebanon uprising, the research surveys thousand Lebanese individuals using a nationally representative probability sample. The findings provide support for selective exposure theory beyond the Western context, particularly for people undergoing stressful political change. The findings also provide support to the robustness of issue publics as a theoretical construct that predicts selective exposure to pro-attitudinal news content, as well as support for attitude strength and attitude importance as two separate predictors.

Keywords

Lebanon selective exposure Arab media conflict confirmation bias protest issue publics

Research on the effects of conflict on media uses and media-related needs is scarce, with only a number of studies assessing media uses during political turmoil (cf. Dotan and Cohen 1976; Georgiou 2012; Kozman and Melki 2018; Lev-On 2011; Melki and Kozman 2019). As a region plagued by continuous conflict and civil strife, the Middle East is ripe for media and conflict research, offering scholars the opportunity to expand knowledge of how audiences manipulate the information overload by partaking in the pool of news formation through their own content creation. Central to this idea is the concept of selective exposure that parts from the strong media effects paradigms to extend agency to members of the public.

A decade of uprisings throughout the Arab region has ushered a new era of conflict, uncertainty, and radical change. Mass demonstrations that spread across the region in 2010 lead to the fall of governments in Egypt and Tunisia; civil wars in Libya, Yemen, and Syria; and political unrest in other countries. Having experienced one of the most recent wave of protests that debuted in late 2019, Lebanon presents a unique case, where largely peaceful demonstrations paralyzed the country and the ruling elite, bringing down the government (Council of Ministers) within a few days (“A Surge of Public Anger Sends” 2019). In such fluid situations, both media coverage and audience attention to the news increase (Melki and Kozman 2019).

Focusing on the Lebanon protests, this survey builds on the theoretical framework of selective exposure to examine public opinion and reported behavior of a nationally representative sample of thousand individuals living in Lebanon during the height of the protests. The study explores news consumption habits, comparing audiences’ attitudes toward the protests to their news selection patterns. In addition, it assesses the criteria people follow in posting and sharing information on social media.

This study is particularly significant in advancing extant literature on selective exposure beyond the Western hemisphere and in examining the relationship between attitudes and behaviors of people undergoing stressful political changes. The theoretical contribution of this study is additionally evident in testing the robustness of issue publics as a construct that predicts selective exposure to pro-attitudinal news content, alongside measuring attitude strength and attitude importance as two separate predictors. This study is also timely given that selective exposure has rarely been applied to conflict and protest situations—a matter surprisingly understudied despite the record number of protests and conflicts globally and the significant role of news media in instigating, sustaining, and propelling them.

By employing the proposed theoretical framework to examine the attitudes of people during protests, the study contributes to the already considerable and worthy body of work on selective exposure. The objective is to demonstrate that selective exposure can inform media uses and trust not just among general populations in everyday situations but also among persons whose everyday lives are marked by uncertainty and who may nevertheless turn to media for a multiplicity of needs.

Background: The WhatsApp Revolution

On October 17, 2019, small groups of young Lebanese gathered in various areas around the country to protest the newly announced decision to tax WhatsApp calls. What was the topic of memes and jokes during the first few days quickly became a symbol of the public’s power to demand and successfully force change (“Lebanon: WhatsApp Tax” 2019).

Although the protests were sparked by the WhatsApp tax, the issue masked the public’s growing resentment toward the corrupt sectarian ruling class that abused the country’s resources for decades and failed to provide even the most basic public services, including electricity, water, and garbage disposal. The subsequent dwindling economy, low trust in government institutions and political parties, mass economic migration, and skyrocketing unemployment rates provided a diverse public with the impetus to revolt in such massive numbers for the first time in Lebanon’s history (Yee and Saad 2019). Guided by a nationwide anger against the government and other nonstate actors, the demonstrations paralyzed the country, closing schools, banks, and various other public and private institutions (see also Chehayeb and Sewell 2019; Sullivan 2019).

Amid these tumultuous times, most Lebanese media took a stance, largely polarizing themselves along pro- and anti-protest dimensions, as observed by the authors and other analysts (Lteif 2019; “Spinning, Downplaying, Ignoring” 2019). For instance, the three pro-protest Television channels (al-Jadeed, LBCI, and MTV) carried slogans and hashtags of the uprising, whereas the three anti-protest channels (OTV, al-Manar, and NBN) discredited the protestors and accused them of receiving foreign funding and threatening national security (for details about each station, see Bou-Hamad and Yehya 2020).

Whereas traditional media went about their partisan tendencies, parallel discussions were being held on social media. Aware of its ability to drive civic conversations, the public took part in content creation, engaging in selection and dissemination of information it deemed important. Acting in an emotionally and politically charged environment, Lebanese people produced and shared content, commented on news, and followed specific media sources.

Measuring public opinion during protests is valuable for both research and practical reasons. The public’s engagement with traditional and social media could inform us about selection patterns, whether they are exposed to diverse media content that may challenge their views or follow content reflective of their biases, and whether they share news on social networks or elect to retreat from engaging with the networked public sphere. These activities in turn could reveal behavioral measures people resort to during turmoil.

Selective Exposure and Attitudes

The high-choice media environment that characterizes twenty-first century society warrants selectivity on the part of the public to make sense of the continuous flow of information. This process, while seemingly simple, is a complex set of elements that interact in individuals’ brains to prompt them to select one type of content while ignoring another. At the heart of this selection process lies the concept of selective exposure that is rooted in psychology. Based on Festinger’s (1957) theory of cognitive dissonance, selective exposure contends people seek information they agree with to avoid dissonance that could cause psychological discomfort (Knobloch-Westerwick et al. 2020). In the media context, selective exposure suggests individuals seek media content that confirms their attitudes while avoiding content that challenges these attitudes (Westerwick et al. 2013). A meta-analysis of 91 studies of selective exposure using 8,000 participants revealed people are twice as likely to select pro-attitudinal than counterattitudinal content (Hart et al. 2009).

Selective exposure, however, is neither straightforward nor simple. There exist various elements that motivate individuals to select attitude-consistent media, some of which are not for the purpose of reducing dissonance (Metzger et al. 2020). In his revision of cognitive dissonance, Festinger (1964) contended that individuals’ commitment to a position is crucial for selective exposure to attitude-supportive content. This suggests that attitude strength drives selective exposure (Thomas et al. 2018). Indeed, a plethora of studies have found that beyond the simple existence of attitudes, their strength increases exposure to attitude-consistent media (Brannon et al. 2007; Garrett 2009; Westerwick et al. 2013; Wojcieszak 2019). In certain cases, however, strong attitudes do not predict selecting pro-attitudinal content (Peralta et al. 2017), where participants who deem an issue important and hold strong attitudes regarding that issue might choose to consume balanced content instead of pro-attitudinal material only (Peralta et al. 2017). The reasons behind both confirmations and dismissals of attitude strength as a predictor of selective exposure are a mix of motivations and psychographics: People could choose either to reach validated conclusions about a personally important issue through selecting balanced content that refutes counterarguments or to resist persuasion from the opposite side (Peralta et al. 2017). At times, individuals with strong attitudes engage with selective exposure more so to reinforce their opinion than avoid challenging opinions (Garrett 2009).

Regardless of the motivations behind these selections, participants’ backgrounds and interest in specific issues have been able to predict exposure. Working in tandem with attitudes, individuals’ political preferences could also predict selective exposure. Whereas some have found evidence for political beliefs to drive exposure to attitude-consistent content (Dahlgren et al. 2019; Dorison et al. 2019; Feldman and Hart 2018; Hameleers & van der Meer 2019; Stroud 2008), others have found that political interest and party affiliation drive exposure to attitude-discrepant content (Knobloch-Westerwick and Meng 2009). The latter finding is perhaps due to the notion that those who have stronger stances are more confident in their opinions and are thus less likely to be shaken by counterarguments (Westerwick et al. 2013). In the case of the former, one reason for partisan selective exposure driven by political belief is people’s inflation of the potential impact of their exposure to counterattitudinal content (Dorison et al. 2019). The partisanship of the media themselves could also act as a driving force for selective exposure. A study comparing selective exposure to television in Japan and Germany, which are nonpartisan by law, to exposure to television in the United States, whose television coverage has become increasingly partisan, revealed confirmation bias to be stronger for the latter (Knobloch-Westerwick et al. 2019). In Lebanon, the media are predominantly partisan with political affiliations that render some virtual mouthpieces of the party they support (Melki 2014). This partisan coverage became even starker during the 2019 uprising that split the public and the media into two distinct groups marked by pro- and anti-government sentiments (“Spinning, Downplaying, Ignoring” 2019). This makes a strong case for politically motivated selectivity to occur in this study, reinforced further by the political nature of the anti-governmental protests in Lebanon. Although research has shown confirmation bias to be a universal phenomenon that occurs across countries and issues (e.g., Jahng 2018; Knobloch-Westerwick et al. 2019; Wojcieszak 2019), this effect is greater for political issues, as Hart et al.’s (2009) meta-analysis of selective exposure research revealed.

The Lebanese protests present a unique case that allows us to link strong attitudes toward the protests to preference for media that supported the protests, and vice versa.

With strong and polarized attitudes, the protestors’ rhetoric on the street was unyielding and centered on a hashtag that targeted the entire political class and that translates into “all of them means all of them” (كلن_يعني_كلن#) (“A Moment for Change” 2019). Such a strong, uncompromising attitude is expected to surpass mere opinion and be related to behavior, measured through media selectivity. Extant research on attitudes has conceptualized attitude strength as how strongly participants felt about an issue, such as gun ownership/control, abortion, minimum wage, health care, and climate change (Knobloch-Westerwick and Meng 2009; Peralta et al. 2017; Westerwick et al. 2013; Wojcieszak 2019). Consistent with these and similar studies on selective exposure and attitudes, we hypothesize people’s strong attitudes about the protests to be related to exposure to media that supported the protests, and vice versa:

Hypothesis 1 (H1): Attitude strength will be positively related to preference for pro-attitudinal news content and negatively related to preference for anti-attitudinal news content.

Besides strength, several characteristics of attitudes have been able to predict selective exposure. Attitude certainty, for instance, is positively associated with exposure to pro-attitudinal media and negatively associated with exposure to counterattitudinal media (Knobloch-Westerwick and Meng 2009). Research has revealed that attitude importance increases exposure to like-minded media (Westerwick et al. 2013; Wojcieszak 2019). One main reason for this is that greater dissonance would occur regarding topics of personal importance, which will in turn heighten confirmation bias and exposure to pro-attitudinal content (Knobloch-Westerwick et al. 2005). Attitude importance, however, could also act in opposite ways and drive exposure to counterattitudinal content (Knobloch-Westerwick and Meng 2009). Due to differences in research findings about attitude importance, we were motivated to seek out how this concept is related to selective exposure in the context of protests. Whereas previous studies have tested attitude importance in general settings, we believe the situation changes when the setting is based on politically charged anti-governmental protests. Similar to existing research (Knobloch-Westerwick and Meng 2009; Peralta et al. 2017; Westerwick et al. 2013; Wojcieszak 2019), we conceptualized attitude importance to be how important participants believed the issue, that is, the protests, to be. We thus propose the following hypothesis:

Hypothesis 2 (H2): Attitude importance will be positively related to preference for pro-attitudinal news content and negatively related to preference for anti-attitudinal news content.

Selective Exposure and Issue Publics

Research on selective exposure typically uses attitude strength and attitude importance as two separate predictors of selectivity (Knobloch-Westerwick and Meng 2009; Peralta et al. 2017). Rarely ever did a study combine these two measures into a construct, although they are related (Peralta et al. 2017). In this study, we attempt to combine these measures through the theoretical construct of issue publics. Extant research defines issue publics as groups of people who are highly knowledgeable about an issue that interests them (Converse 1964; Iyengar et al. 2008). Issue publics comprise people with high attitude importance and high attitude strength (Peralta et al. 2017). Since the two variables are highly related, we test the robustness of issue publics as a construct in predicting selective exposure:

Research Question 1 (RQ1): Issue publics will be (a) more likely to prefer pro-attitudinal news content and (b) less likely to prefer anti-attitudinal news content than non-issue publics.

A motive that drives scholars’ interest in opinions and attitudes is their presumed effect on behavior. In the context of protests, civic participation is one form of political behavior, whether online or offline. Given the intensity of the Lebanese protests, we were driven to explore the relationship between attitudes and reported behavior, conceptualized as participation in street protests. In this study, it is probable for those with strong attitudes about the protests and who deemed them important to show their civic agency by joining street protests, especially in light of the massive social media recruitment efforts that took place. In a high-tension environment, social media could encourage political participation as activists use them as a tool to spur people into action (P. S. Lee et al. 2015). Lebanese who felt strongly about the protests could have been additionally motivated to participate by an internal feeling of being able to effect political change, which serves as a “psychological incentive for participation” (Valenzuela et al. 2012: 164). Studies have found political behavior to be partly driven by collective efficacy—people’s belief that their group is able to achieve its goals (Lee 2010a)—even in the online environment (Velasquez and LaRose 2015). To examine the link between attitudes and political behavior in the Lebanese protests, we draw on the construct of issue publics to test whether people with high attitude strength and importance toward the protests tended to show their passion by taking part in street protests:

Research Question 2 (RQ2): Are issue publics more likely than non-issue publics to participate in street protests?

Selective Exposure and Trust

Participants’ attitudes and beliefs are not the sole predictors of behavior. Individuals engage with selectivity due to a myriad of reasons that lie outside internal psychological mechanisms and could be related to context and message characteristics (Knobloch-Westerwick et al. 2019). One such motivation is information utility (Sears and Freedman 1967), which argues that individuals select information from sources they perceive as useful, overriding confirmation bias (Knobloch-Westerwick et al. 2005). Alongside information utility, trust in the media is yet another element that drives people’s media selection (Tsfati and Cappella 2003). As it moderates the relationship between users and media content (Jackob 2010), trust becomes a significant factor in media effects (Tsfati 2003). It is also plausible that trusting sources means people find them important for their information needs, especially during times of turmoil and uncertainty (Melki and Kozman 2019). In research, trust has been used as one component of credibility, along with professionalism and bias (Metzger et al. 2020). As far as it concerns selective exposure, source credibility is deemed important to online users who spent more time on messages from high-credibility sources than on messages from low-credibility sources (Knobloch-Westerwick et al. 2015). In another instance, preference for counterattitudinal content was stronger for low-credibility messages, whereas preference for pro-attitudinal content was stronger for messages from experts (Wojcieszak 2019). Some have even used source credibility as a predictor of selective exposure to challenge the concept of dissonance (Metzger et al. 2020) or as a mediator between political party identification and exposure to partisan news (Stroud and Lee 2013). Research also reveals source credibility can override attitudes when it comes to selecting information sources that both support and challenge people’s viewpoints (Johnson and Kaye 2013). Trust in the media was found to be inversely related to confirmation bias (Knobloch-Westerwick et al. 2019), with low media trust leading to higher likelihood of exposure to attitude-consistent content (Tsfati and Ariely 2014). A comparative analysis of media across 44 countries also revealed trust to be positively correlated with political interest (Tsfati and Ariely 2014). Although studies have examined trust in media as a whole, research on trust in a partisan media environment has yielded similar results. Jones (2004) found low public trust in the media to be related to low trust in government, especially among self-identified American conservatives who consume partisan news. He further contended that conservatives seem to prefer exposure to conservative political talk radio, seeing it as an avenue that aligns with their ideological views (Jones 2004). Partisanship and ideology were also shown to be associated with trust in the media in the United States during a time when the conservative Fox News was gaining popularity (Lee 2010b). What these studies suggest, at times indirectly, is that media trust in a partisan media environment is connected to people’s own partisanship. So far, only one study has used trust to examine selective exposure in the Arab world. The survey of Syrian citizens during the Syrian civil war revealed trusting television news channels was positively related to following these channels (Melki and Kozman 2019). For Lebanese, media trust is often synonymous with trust in the media channels that reflect their sectarian ideologies (Melki 2014).

The above studies indicate that trusting a source could influence individuals’ decisions of whether they would select that source or not. In the context of this study and based on Melki and Kozman’s (2019) findings, we can expect trust to play a role in selective exposure during political turmoil, decreasing trust in the government, and a partisan Lebanese media environment. We test this relationship regarding people’s trust and subsequent exposure to local television channels:

Hypothesis 3a (H3a): Trusting pro-protest television channels is positively related to following pro-protest channels.

Hypothesis 3b (H3b): Trusting anti-protest television channels is positively related to following anti-protest channels.

Social Media and Protests

Communication technologies have added another layer of meaning to selective exposure. Due to their customizability and algorithm-driven content, social networking sites offer users the chance to engage in selectivity, resulting in a decrease in exposure to politically challenging content (Dylko et al. 2018). Research has also highlighted the potential of social networking sites to foster heterogeneity based on the premise that users are exposed, perhaps unintentionally, to various opinions about various issues (Jahng 2018), promoting selection of attitude-discrepant content (Beam et al. 2018). Not all people, however, harness the power of social media to diversify their diet of news. Whereas we expect the public to make use of the unlimited sources of information that the online world offers, people’s selections do not always reflect this variety. Sometimes users are consistent in their selection of attitude-supporting messages, raising concerns about social isolation and polarization (Song et al. 2020). Studies on online networks have revealed partisan selective exposure leads to polarization (Lu and Lee 2018; Stroud 2010), increasing the risk of creating filter bubbles and echo chambers that produce a highly polarized and fragmented society (Sunstein 2001). This could be due to the limited variety of content available to them or the selective sharing that increases the likelihood of exposure to attitude-consistent content (Liang 2018). Another reason is attributed to the collective frames people in the same network produce through sharing posts in social networks and thus altering the frequency of content on their social media walls (Aruguete and Calvo 2018). These effects, however, could subside over time, reducing the degree of selective sharing and thus increasing the diversity of sharers (Liang 2018). In recent years, social networks have been steadily moving to the online environment, with social media becoming an agitating force behind participation in the streets. Studies on protests in several countries have found social media, especially Facebook, to be directly related to protest activity as was the case in Chile (Valenzuela et al. 2012) and indirectly to protest activity in Hong Kong (Zhu et al. 2017) and Egypt (Tufekci and Wilson 2012). During the demonstrations in Lebanon, WhatsApp and Facebook groups and other social networks were widely used to signal support for the protests, organize marches, and disseminate news about widespread corruption (Lteif 2019). Based on the literature on online civic activism, we test the relationship between support and sharing news through the following hypothesis:

Hypothesis 4 (H4): Supporting the protests is related to sharing news about the protests on social media.

Method

The study used a cross-sectional researcher-administered survey of Lebanese individuals aged eighteen years and older, conducted face-to-face during the height of the protests (December 5–12, 2019).

Sample

An overall sample size of thousand participants was calculated based on a population of six million, a 95 percent confidence level, and a ±3.1 percent sampling error. Questionnaires were distributed proportionally to the number of residents in each governorate. Moreover, each governorate was stratified into districts, and questionnaires were distributed in each capital city and main town to ensure a proper representation by religious sect (confession) in each district. Since Lebanon has no reliable census records, election records were used to ensure proper distribution by sect. The number of questionnaires assigned to a district was proportionally distributed based on the number of voters at the town level. The study adopted a multistage probability sampling technique to ensure a random representative sample. The first stage consisted of selecting neighborhoods inside each selected city or town to represent the confessional diversity of the areas. The second stage consisted of selecting households based on a systematic random sample in each neighborhood. Finally, the third stage sampled a primary respondent within each household based on the most recent birthday. If the selected person was not at home, a revisit up to one time was conducted before declaring a nonresponse. This method ensured that everyone has an equal chance of inclusion, with no one allowed to self-select into the sample.

In addition, to ensure sufficient representation from protestors in the streets, 30 percent of the surveys were conducted with respondents present at the main protest squares. Researchers used a systematic random sample in this case, depending on the size of the protest, picking every third person for small and medium-sized gatherings and every fifth person for large gatherings.

Interviewers explained the objectives of the study and informed the selected participants about their rights, in line with the institutional review board (IRB)-approved procedure. Once informed consent was received, face-to-face surveys were conducted using tablets.

Instrument and Measures

The questionnaire comprised twenty-four close-ended questions, required approximately ten minutes to complete, and generated eighty-four variables. All media uses variables followed a four-point ordered response scale measured at the ordinal and interval levels. In addition to demographics, the questionnaire assessed attitudes and behavior regarding the protests. The internal reliability of all composite scales was measured using Cronbach’s alpha.

Demographics and psychographics

These measures included gender, education, employment status, income, party affiliation, allegiance (national/sectarian), and interest in Lebanese politics (Table 1).

Table 1.

Sociodemographic Profile of the Survey Participants (N = 1,000).

Sociodemographic Variable	% (n)
Gender
Male	51.4 (514)
Female	48.6 (486)
Age
18–22	18.1 (181)
23–30	26.5 (265)
31–45	26.7 (267)
46–65	24.3 (243)
Older than 65	4.4 (44)
Education
Elementary school or less	7.7 (77)
Middle school	22.3 (223)
High school	51.7 (517)
Bachelor’s or higher	18.3 (183)
Employment
Unemployed	46.2 (462)
Part-time	16.1 (161)
Full-time	37.7 (377)
Income (monthly in USD) ^a
No income	14.0 (140)
Less than $500 (minimum wage)	41.3 (413)
$500–$1,000	33.7 (337)
$1,001–$2,000	9.7 (97)
More than $2,000	1.3 (13)
Party affiliation
Party member	3.0 (30)
Party supporter	25.6 (256)
Not a member/supporter	71.4 (714)
Allegiance
To religion	3.4 (34)
To country	69.8 (698)
To country and religion equally	22.9 (229)
Neither	3.9 (39)

Income was based on the official conversion rate of 1 USD = 1,515 Lebanese Pounds (LBP; rounded to 1,500), although the unofficial exchange rate at the time had already exceeded 2,000 LBP per 1 USD.

Attitudes and behavior

As per Hart et al. (2009), we define attitude as an individual’s evaluation of an event: the protests. And we define behavior “as an overt action performed in relation to an entity”: participation in street protests (p. 556). Similar to previous studies (Knobloch-Westerwick and Meng 2009; Peralta et al. 2017; Westerwick et al. 2013; Wojcieszak 2019), we measured attitude strength by asking respondents whether they opposed or supported the protests on a four-point scale (19.2 percent strongly opposed, 15.5 percent somewhat opposed, 27.9 percent somewhat supported, and 37.4 percent strongly supported; M = 2.84, SD = 1.13), and we measured attitude importance by asking respondents how important they deemed the protests a four-point scale (20.6 percent said not at all important, 9.7 percent not that important, 27.5 percent somewhat important, and 42.2 percent very important; M = 2.91, SD = 1.16). For participation in protests, participants were asked how often they joined the protests on the streets (70 percent never joined, 2.1 percent rarely, 9.3 percent sometimes, and 18.6 percent often; M = 1.77, SD = 1.22).

Issue publics

Relying on Peralta et al.’s (2017) conceptualization of issue publics as groups with high attitude strength and importance, we created a scale for issue publics by averaging answers for two variables: attitude strength and attitude importance about the protests (M = 2.87, SD = 1.11, Cronbach’s α = .96), where higher numbers reflected stronger issue publics. In other words, the higher the value, the stronger the support for the protests, and vice versa.

Selective exposure

We operationalized selective exposure as exposure to pro-attitudinal content. This was measured along two dimensions: pro-protest local television channels and anti-protest local television stations. Participants were asked how often they followed each of the six local television stations on a four-point scale (1 = never, 4 = often). To measure how often participants followed pro-protest television stations, we computed a scale by averaging answers for Al-Jadeed, LBC, and Murr TV (M = 2.59, SD = 0.88, α = .72). For following anti-protest television stations, the scale averaged answers for Al-Manar, NBN, and OTV (M = 1.63, SD = 0.76, α = .76).

Trust

Respondents were asked how much they trust each of the six local television stations for news about the protests. Following the same procedure as above, we created two scales: one for trusting pro-protest television stations Al-Jadeed, LBC, and Murr TV (M = 2.41, SD = 0.85, = 0.72) and one for trusting anti-protest television stations Al-Manar, NBN, and OTV (M = 1.62, SD = 0.79, α = .78).

Sharing or posting news

Participants were asked how often they shared/posted news about the protests on Facebook, Twitter, Instagram, and WhatsApp (1 = never, 4 = often). The following reflects the breakdown: 37.6 percent of participants sometimes or often shared or posted news about the protests via WhatsApp, whereas 26.3 percent did so on Facebook, 10.6 percent on Instagram, and 9.3 percent on Twitter. To get a composite measure of sharing or posting news about the protests, we created a scale by averaging the answers for each of these social media platforms (M = 1.6, SD = 0.76, α = .79).

Limitations

Interviewers faced the following limitations during the field work: 6.1 percent of cases required a follow-up interview because the selected respondent was not at home, 15.3 percent of approached respondents refused to participate, and Lebanese who made an income higher than $1,500 were underrepresented, whereas those who made less than the minimum wage or no income were overrepresented. Although the study was conducted during a tumultuous period, field researchers reported that respondents were mostly welcoming and cooperative.

Results

H1 hypothesized that attitude strength will be positively related to preference for pro-attitudinal political content and negatively related to preference for anti-attitudinal political content. It was supported for the overall model. A multivariate analysis of covariance (MANCOVA) was used to assess the effect of attitude strength on preference for pro-attitudinal and anti-attitudinal media content while controlling for gender, education, party affiliation, political interest, and interest in knowing about the protests (when applicable, all were recoded into binary measures). To avoid the interpretation complexity that all covariates are inclined to add to multivariate tests, we only considered the variables that were found to have an effect on the two dependent variables. Among all five covariates, only party affiliation, political interest, and interest in the protests were significant. The results of the MANCOVA revealed the main effect variable, attitude strength, was significant: Pillai’s trace F(6, 1978) = 5.36, p < .001, partial η² = .016. Pillai’s trace is reported instead of Wilks’ lambda since it is the most conservative and is robust against violations of homogeneity of variances and covariances. While the covariates alone were not individually significant, they all interacted with the main variable to produce a significant result: F(8, 1978) = 3.7, p < .001, partial η² = .015. Evaluating the effect of attitude strength on each of the two dependent variables, factoring in for the presence of controls, the univariate tests indicated that it is only significant on exposure to pro-attitudinal media—F(3, 989) = 8.9, p < .001, partial η² = .027, whereas political interest was significant for anti-attitudinal media—F(1, 989) = 4.74, p < .05, partial η² = .005. The results support H1 in that attitude strength was significant in the overall model, affecting participant’s choice to follow pro-attitudinal or anti-attitudinal media at the same time, whereas it interacts with party affiliation, political interest, and interest in the protests to produce an effect on exposure to anti-attitudinal media. In particular, low attitudes (those who strongly opposed the protests) were related to a 0.87-point decrease (SE = 0.18, p < .001) in preference for pro-content compared to a 1-unit increase in high attitudes (those who strongly supported the protests), whereas those who somewhat opposed the protests were related to a 0.36-point decrease (SE = 0.18, p = .052) in their preference for pro-content for every 1-unit increase in those who strongly supported the protests. This could be interpreted as people who opposed the protests were 87 percent less likely to expose themselves to pro-protest content compared to those who strongly supported the protests, and those who somewhat opposed the protests were 36 percent less likely to expose themselves to pro-protest content. As a covariate, political interest was positively related to preference for anti-attitudinal content (SE = 0.14, p < .05).

H2, which predicted attitude importance will be positively related to preference for pro-attitudinal political content and negatively related to preference for anti-attitudinal political content, was supported for the overall model. Following the same statistical steps as H1, the MANCOVA revealed the main effect variable, attitude importance, was significant, Pillai’s trace F(6, 1978) = 3.64, p < .001, partial η² = .011, alongside covariates political interest and interest in knowing about the protests, which also had an interaction effect with the main independent variable. Univariate tests showed attitude importance was significant only for preference for pro-attitudinal content, F(3, 989) = 5.92, p < .005, partial η² = .018, whereas political interest, F(1, 989) = 5.9, p < .05, partial η² = .006, and interest in protests were significant for anti-attitudinal media only, F(1, 989) = 9.43, p < .005, partial η² = .009. Specifically, the only significant difference in attitude importance was found between the two extremes, that is, those who said the protests were very important and those who considered them not at all important, where the latter were less likely than the former to choose pro-protest content (B = −0.78, SE = 0.21, p < .001). As for the other two significant variables, political interest was positively related to preference for anti-attitudinal content (B = 0.29, SE = 0.09, p < .005), whereas interest in knowing about the protests was negatively related to anti-protest content (B = −0.31, SE = 0.12, p < .05).

We also wanted to test issue publics as a construct versus using attitude strength and attitude importance separately through two multiple regressions on following pro-protest television (RQ1a) and following anti-protest television (RQ1b) as dependent variables. In the two tests, we used gender, education, party affiliation, allegiance to country/sect, interest in knowing about the protests, and political interest as predictors in the first block and issue publics in the second block.

For RQ1a, the regression model was significant, F(7, 953) = 32.012, p < .001, improving the model by 6.3 percent when issue publics was entered (adjusted R² = 18.4 percent, p < .001). Among the predictors, gender (B = 0.121, SE = 0.053, p < .05), education (B = 0.277, SE = 0.067, p < .001), and political interest (B = −0.168, SE = 0.06, p < .01) were significant, alongside the main effect variable issue publics (B = 0.328, SE = 0.038, p < .001). These results indicate women are more likely than men, and the more educated are more likely than the less educated, to prefer exposure to pro-attitudinal media, whereas the more political interest meant less likelihood of choosing pro-protest media. Beyond these characteristics, the stronger the support for the protest meant stronger preference for pro-protest content.

For RQ1b, the regression model was significant, F(7, 953) = 36.39, p < .001, improving the model by 3.9 percent when issue publics was entered (adjusted R² = 20.5 percent, p < .001). Among the predictors, political interest (B = 0.388, SE = 0.051, p < .001), party affiliation (B = 0.439, SE = 0.05, p < .001), and issue publics (B = −0.221, SE = 0.032, p < .001) were significant. This means participants who are interested in Lebanese politics, and those who are either members of a party or affiliated to it, were more likely than the uninterested and unaffiliated to choose anti-attitudinal media. Beyond these controls, those who identified as pro-protest (higher issue publics values) were less likely to choose anti-protest content compared to the rest in their groups.

RQ2 asked whether issue publics will participate in street protests. A binary logistic regression was used to measure the relationship while controlling for political interest. The dependent variable was recoded into a binary joined and not joined. The predictors were entered in two blocks, using the Enter method. In the first block, political interest and party membership were entered, and in the second block, issue publics were entered. The model was significant, χ²(3) = 459.28, p < .001, Nagelkerke’s R² = .52, beyond political interest as a covariate (B = 0.56, SE = 0.11, Wald = 27.13, p < .001). The stronger the issue publics, the more likely they were to join the protests (B = 2.11, SE = 0.17, Wald = 151.83, p < .001). Interestingly, party affiliation was significant in the first block, where party-affiliated individuals were less likely to join the protests (B = −0.99, SE = 0.18, Wald = 31.18, p < .001). However, when issue publics was added as the main predictor, party affiliation became nonsignificant.

H3a, which suggested trusting pro-protest television stations is positively related to following pro-protest stations, was supported. This hypothesis predicted that trusting pro-protest television stations is positively related to following pro-protest stations, regardless of how important the protests are. The linear regression model was significant, F(2, 997) = 1,612.42, p < .001, with 76.3 percent of the variance in the dependent variable accounted for by trust alone (B = 0.9, SE = 0.02, p < .001), where a 1-unit increase in trust was related to a 0.9-unit increase in following pro-protest stations.

H3b, which predicted trusting anti-protest television stations is positively related to following anti-protest stations, was supported. The linear regression model was significant, with 80.2 percent of the variance in the dependent variable accounted for by predictors, F(2, 997) = 2,021.87, p < .001. Trust was a significant predictor (B = 0.87, SE = 0.01, p < .001) even after controlling for importance of protests (B = 0.02, SE = 0.01, p < .05), where a 1-unit increase in trust was related to a 0.87-unit increase in following anti-protest stations.

H4, which predicted supporting the protests is related to sharing news about the protests on social media, was supported. A t-test revealed significant effect for attitude on sharing news about the protests, t(995.83) = −13.53, p < .001. Those who supported the protests were more likely (M = 1.77, SD = 0.82) than those who opposed them (M = 1.22, SD = 0.46) to share news.

Discussion

This study examined selective exposure and trust in media during uprisings. The evidence strongly supported selective exposure tendencies, especially for legacy media and particularly television. First, both attitude strength and attitude importance were positively related to preference for pro-attitudinal political content on television and negatively related to preference for anti-attitudinal political content, regardless of gender, education, partisanship/affiliation, political interest, and importance of allegiance to country or sect. The results are consistent with previous studies on selective exposure, particularly affirming the concept of confirmation bias among the public (Hart et al. 2009; Knobloch-Westerwick et al. 2020). Similar to extant research, both attitude importance and attitude strength increased exposure to attitude-consistent media (Brannon et al. 2007; Garrett 2009; Knobloch-Westerwick et al. 2005; Westerwick et al. 2013; Wojcieszak 2019). The role of education, partisanship, and political interest in predicting selective exposure warrants a closer look. As the results revealed, individuals’ attitude strength and importance regarding the protests determined selectivity above and beyond their education, their party affiliations, and political interest. Similar to previous studies, the findings here indicate political interest drives exposure to both pro-attitudinal content (Dahlgren et al. 2019; Dorison et al. 2019; Feldman and Hart 2018; Hameleers & van der Meer 2019) and anti-attitudinal content (Knobloch-Westerwick and Meng 2009). In the case of the latter, only partisanship predicted selectivity.

Second, the stronger the issue publics, the more likely individuals were to join the protests in the streets. Third, trusting pro-protest television stations was positively related to following pro-protest stations, and trusting anti-protest television stations was positively related to following anti-protest stations. This confirms the presence of trust as a significant factor in media effects (Tsfati 2003) and as a predictor of following particular media outlets (Melki and Kozman 2019). When it came to social media, evidence also supported selective exposure tendencies, where supporting the protests was related to sharing news about the protests on social media.

The findings show that, overall, selective exposure is present during times of turmoil. One would speculate, though, that during such political upheavals and major changes in society, people may try to seek the most holistic picture and follow multiple news sources—especially in an overtly partisan media system and a precarious political context, but the evidence demonstrates the opposite. People who face uncertainty and tense situations mainly trust and follow the media that reconfirm their attitudes. This is likely due to intolerance to any information that might cause cognitive dissonance. Indeed, the main postulate of selective exposure rests on people’s tendency to reduce dissonance that could lead to psychological discomfort (Festinger 1957; Knobloch-Westerwick et al. 2020).

This explanation is also supported by the facts on the ground. For those who supported the protests, taking to the streets after many years of inaction and tolerance of the status quo meant that they were both emotionally exasperated and mentally prepared to fight the system all the way. After the gradual collapse of the economy, it was only expected for those supporting the protests to blame the ruling elites for their misfortunes. In such cases, emotional involvement is expected to lower the public’s tolerance of opposing attitudes, and thus decrease the capacity to handle news that defends the same government which the protestors believe has led to the present bankruptcy. On the contrary, for those who opposed the protests, it was a momentous existential crisis, especially given the protestors’ call for the total elimination of the current political class and the eradication of the sectarian state: “all of them means all of them.” For these die-hard partisans, it became a zero-sum game that threatened the raison d’être of the ruling elite—their benefactors. Despite the economic collapse that erased most of their fortunes, those who still benefited from the clientelist political system and believed in its geopolitical worldview feared that losing the status quo meant losing everything. For some, this loss also meant local and regional geopolitical defeat, especially after some leaders hinted at secret foreign intervention and funding from embassies. Therefore, the same emotional involvement for this group is expected to decrease the capacity to handle news that praises the protests and attacks their leaders.

People’s strong attitudes toward their beliefs and the media are evident in the entire sample and not only for those who support the protests. Similar to exposure to pro-attitudinal media, individuals who reported to have lower attitude strength and importance (oppose the protests and deem them unimportant) were more likely to expose themselves to anti-attitudinal media.

Concerning the Lebanese context, our data show an overwhelming majority of the public supported the mass protests. This diverse body of individuals represented a group of people whose attitudes toward the government were recently formed, as evidenced by the fact that many of its members were former supporters of political parties and factions that continued to control the government during the onset of the uprising. This conclusion is also based on the high proportion of Lebanese who said they support the uprising (around two-thirds)—despite having one-third of the respondents sampled from the protests, and the unusually high percentage of those who said they do not support political parties and declared primary allegiance is to their country and not their religious sect. It is also evident in the backlash faced by some media outlets who had previously supported smaller protests in recent years and turned against the 2019 uprising. Al-Akhbar newspaper, for example, faced harsh criticism by the protestors and several of its journalists quit, whereas television stations that fully supported the uprising witnessed a surge in their audiences.

As for gender, the data show that women were more likely than men to prefer exposure to pro-attitudinal media, but only within the context of issues public and not in the general model. Further analysis shows that although there were no significant gender differences for support or opposition of the protests overall, significantly less women (24.9 percent) than men (34.8 percent) joined the protests (p = 0.001), and less women (25.5 percent) than men (31.5 percent) supported or held membership in political parties (p = 0.036). In Lebanon’s patriarchal society, where leaders of political parties and major politicians are mainly men, women may be more driven by political issues rather than ideologies and political affiliations, especially when many protest slogans called for women’s rights and adopted feminist narratives. Furthermore, in a society where some women’s bodies and freedom of movement may be restricted and controlled by men, particularly the freedom to physically participate in the protests, many women may only find agency in preferences that are invisible to their oppressive environments, such as preference for pro-attitudinal media.

In a highly partisan and deeply sectarian country like Lebanon, the most plausible conclusion is that at least a sizable proportion of this pro-uprising group had recently shifted their attitude from pro-government (anti-protests) to anti-government (pro-protests). This interpretation is particularly bolstered by the uprising’s popular slogan “all of them means all of them” and the very low percentage of respondents who supported political parties and declared primary allegiance to their sect. Therefore, these people who underwent a change of heart (attitude) during the protests shifted their attention to the media that supported their new attitudes instead of balancing their media diet. Perhaps the change in attitude was so powerful to effect such change in media selection. In other words, people may be more inclined to change attitudes and shift to the new attitude-confirming content rather than balancing political content that confirms their new attitude with content that had confirmed their previous attitude.

Conclusion

The current survey explored the theoretical framework of selective exposure during social uprisings by relying on the opinions and attitudes of a representative sample of Lebanese people at the height of the 2019 anti-government protests. Examining the strength and importance of the public’s attitudes toward the protests in relation to its media preferences found support for the concept of confirmation bias in exposure to pro-attitude and anti-attitude media. In line with most research about selective exposure (Hart et al. 2009), this study thus extends the utility of confirmation bias as a predictor of selective exposure to a situation rarely studied before. Although individuals’ education levels, partisanship, and interest in politics were shown to be positively related to selective exposure, they could not undo people’s tendency to expose themselves to media sources that confirm their biases.

This study also reinforces the ability of confirmation bias to predict selectivity in a small country whose media system is characterized by high partisanship and diversity. In addition to extending the theory beyond Western contexts, this study corroborates the role of media partisanship in influencing selectivity (Bou-Hamad and Yehya 2020). Although we did not compare the Lebanese media system (as highly partisan) to other media systems, the findings nevertheless support Knobloch-Westerwick et al.’s (2019) study that found confirmation bias to be stronger for U.S. media compared to Germany and Japan whose media are nonpartisan by law.

Our findings further suggest a tendency toward people to be more extreme in their views, choosing to trust and follow media sources that reinforce their preconceived attitudes. As Sunstein (2001) previously warned, echo chambers are breeding grounds for polarization that pull people further away from acceptance and tolerance toward the other. Without such tolerance, democratic deliberation is set to fail. Furthermore, people’s selective practices are multiplied through their sharing habits on social media, potentially amplifying particular messages and thus nurturing extreme views in echo chambers. At the same time, our finding that pro-protest participants were more likely than others to post or share content on social media means the public was proactive in its online usage, taking advantage of the wide reach of social media to spread its message and recruit more protesters.

The implications of the findings—particularly how both attitude importance and strength were positively related to selecting pro-attitudinal media and negatively related to anti-attitudinal media—are grave for Lebanon and for what they tell us about public opinion. In a high-choice media environment where people are free to choose from an infinite number of sources, the tendency remains to coalesce and huddle in a filter bubble that excludes opposing views. To make the situation worse, media partisanship could exacerbate these attitudes. As media outlets depart from their role as objective instruments for relaying facts on the ground and as they become increasingly more involved in propaganda for their parties, they play a significant role in driving the country into extremes.

This study exhibited several limitations. Survey methodology, which relies on self-reporting, often suffers from inaccurate responses. This study was particularly influenced by prevalent patriotic fervor that peaked during the field work (December 5–12). Participants may have underreported their allegiance to their religious sect in this chronically sectarian country. As an expected social desirability bias, they may have also underreported their support to political parties, given the widespread public anger against them at the time. Nevertheless, our study could mimic field experiments since the data were collected at the height of the protests (Stroud 2010) and are therefore as accurate as possible in these circumstances. Still, as a snapshot in time, the survey only captured the attitudes and media preferences of the participants during the height of the uprising. This may have significantly changed after the appointment of new Prime Minister (December 19); even before the new government took power, political realignments and public sentiments started to change, and the sectarian rhetoric slowly returned to the screens as some political parties attempted to jump on the uprising’s bandwagon. This does not change the findings and theoretical implications of this study, which targeted a specific time. However, future research that utilizes longitudinal data can help further confirm this phenomenon and track changes during and after any uprising to discern whether people normally return their old habits and selective biases after the dust settles or whether new habits and biases emerge. Furthermore, given the increasing number of uprisings around the world, studies like these that compare several countries may shed light on the role of the government and media systems in exacerbating or ameliorating confirmation bias.

Footnotes

Declaration of Conflicting Interests

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

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded and supported by the Lebanese American University.

ORCID iD

Jad Melki

Author Biographies

Jad Melki, PhD, is associate professor of journalism and media studies, chairperson of the Communication Arts Department, and director of the Institute of Media Research and Training (IMRT) at the Lebanese American University (LAU). His research is at the intersection of media literacy, conflict journalism, media and war, and gender studies in the Arab region.

Claudia Kozman, PhD, is assistant professor of multimedia journalism in the Department of Communication Arts and Research Director of the Institute of Media Research and Training (IMRT) at the Lebanese American University. Her research focuses on news content, with particular attention to news values, sourcing, framing, and intermedia issue transfer in Arab media. She has conducted studies on the Syrian conflict, including the media uses of displaced Syrian nationals, as well as the Lebanese and American news media coverage of the war.

References

A Moment for Change: The Lebanese Uprisings of 2019. 2019. Democracy Reporting International, December 6. https://democracy-reporting.org/dri_publications/a-moment-for-change-the-lebanese-uprisings-of-2019.

Aruguete

Calvo

2018. “Time to #Protest: Selective Exposure, Cascading Activation, and Framing in Social Media.” Journal of Communication 68 (3): 480–502. doi:10.1093/joc/jqy007.

A Surge of Public Anger Sends Lebanon’s Politicians Reeling. 2019. The Economist, October 24. https://www.economist.com/middle-east-and-africa/2019/10/24/a-surge-of-public-anger-sends-lebanons-politicians-reeling.

Beam

Hutchens

Hmielowski

2018. “Facebook News and (De)polarization: Reinforcing Spirals in the 2016 US Election.” Information, Communication & Society 21 (7): 940–58. doi:10.1080/1369118x.2018.1444783.

Bou-Hamad

Yehya

2020. “Partisan Selective Exposure in TV Consumption Patterns: A Polarized Developing Country Context.” Communication Research 47 (1): 55–81. doi:10.1177/0093650216681896.

Brannon

Tagler

Eagly

2007. “The Moderating Role of Attitude Strength in Selective Exposure to Information.” Journal of Experimental Social Psychology 43 (4): 611–17. doi:10.1016/j.jesp.2006.05.001.

Chehayeb

Sewell

2019. “Why Protesters in Lebanon are Taking to the Streets.” Foreign Policy, November 2. https://foreignpolicy.com/2019/11/02/lebanon-protesters-movement-streets-explainer/.

Converse

P. E.

1964. “The Nature of Belief Systems in Mass Publics.” In Ideology and Discontent, ed. DA Apter, 206–261. New York, NY: Free Press.

Dahlgren

Shehata

Strömbäck

2019. “Reinforcing Spirals at Work? Mutual Influences Between Selective News Exposure and Ideological Leaning.” European Journal of Communication 34 (2): 159–74. doi:10.1177/0267323119830056.

10.

Dorison

Minson

Rogers

2019. “Selective Exposure Partly Relies on Faulty Affective Forecasts.” Cognition 188:98–107. doi:10.1016/j.cognition.2019.02.010.

11.

Dotan

Cohen

1976. “Mass Media Use in the Family during War and Peace.” Communication Research 3 (4): 393–402. doi:10.1177/009365027600300403.

12.

Dylko

Dolgov

Hoffman

Eckhart

Molina

Aaziz

2018. “Impact of Customizability Technology on Political Polarization.” Journal of Information Technology & Politics 15 (1): 19–33. doi:10.1080/19331681.2017.1354243.

13.

Feldman

Hart

2018. “Broadening Exposure to Climate Change News? How Framing and Political Orientation Interact to Influence Selective Exposure.” Journal of Communication 68 (3): 503–24. doi:10.1093/joc/jqy011.

14.

Festinger

1957. A Theory of Cognitive Dissonance. Stanford: Stanford University Press. doi:10.2307/1420234.

15.

Festinger

1964. Conflict, Decision, and Dissonance. 3rd Edition. Stanford: Stanford University Press.

16.

Garrett

2009. “Politically Motivated Reinforcement Seeking: Reframing the Selective Exposure Debate.” Journal of Communication 59 (4): 676–99. doi:10.1111/j.1460-2466.2009.01452.x.

17.

Georgiou

2012. “Seeking Ontological Security beyond the Nation.” Television & New Media 14 (4): 304–21. doi:10.1177/1527476412463448.

18.

Hameleers

van der Meer

2019. “Fight or Flight? Attributing Responsibility in Response to Mixed Congruent and Incongruent Partisan News in Selective Exposure Media Environments.” Information, Communication & Society 23:1327–1352. doi:10.1080/1369118x.2019.1566394.

19.

Hart

Albarracín

Eagly

Brechan

Lindberg

Merrill

2009. “Feeling Validated Versus Being Correct: A Meta-Analysis of Selective Exposure to Information.” Psychological Bulletin 135 (4): 555–88. doi:10.1037/a0015701.

20.

Iyengar

Hahn

K. S.

Krosnick

J. A.

Walker

2008. “Selective Exposure to Campaign Communication: The Role of Anticipated Agreement and Issue Public Membership.” Journal of Politics 70:186–200. doi:10.1017/S0022381607080139.

21.

Jackob

2010. “No Alternatives? The Relationship between Perceived Media Dependency, Use of Alternative Information Sources, and General Trust in Mass Media.” International Journal of Communication 4:589–606. doi:1932–8036/20100589.

22.

Jahng

2018. “From Reading Comments to Seeking News: Exposure to Disagreements From Online Comments and the Need For Opinion-Challenging News.” Journal of Information Technology & Politics 15 (2): 142–54. doi:10.1080/19331681.2018.1449702.

23.

Johnson

Kaye

2013. “The Dark Side of the Boon? Credibility, Selective Exposure and the Proliferation of Online Sources of Political Information.” Computers in Human Behavior 29 (4): 1862–71. doi:10.1016/j.chb.2013.02.011.

24.

Jones

D. A.

2004. “Why Americans Don’t Trust the Media: A Preliminary Analysis.” Harvard International Journal of Press/Politics 9 (2): 60–75.

25.

Knobloch-Westerwick

Carpentier

Blumhoff

Nickel

2005. “Selective Exposure Effects for Positive and Negative News: Testing the Robustness of the Informational Utility Model.” Journalism & Mass Communication Quarterly 82 (1): 181–95. doi:10.1177/107769900508200112.

26.

Knobloch-Westerwick

Johnson

Westerwick

2015. “Confirmation Bias in Online Searches: Impacts of Selective Exposure before an Election on Political Attitude Strength and Shifts.” Journal of Computer-Mediated Communication 20 (2): 171–87. doi:10.1111/jcc4.12105.

27.

Knobloch-Westerwick

Liu

Hino

Westerwick

Johnson

2019. “Context Impacts on Confirmation Bias: Evidence from the 2017 Japanese Snap Election Compared with American and German Findings.” Human Communication Research 45:427–49. doi:10.1093/hcr/hqz005.

28.

Knobloch-Westerwick

Meng

2009. “Looking the Other Way: Selective Exposure to Attitude-Consistent and Counter-Attitudinal Political Information.” Communication Research 36 (2): 426–48. doi:10.1177/0093650209333030.

29.

Knobloch-Westerwick

Mothes

Polavin

2020. “Confirmation Bias, Ingroup Bias, and Negativity Bias in Selective Exposure to Political Information.” Communication Research 47 (1): 104–24. doi:10.1177/0093650217719596.

30.

Kozman

Melki

2018. “News Media Uses during War: The Case of the Syrian Conflict.” Journalism Studies 19 (10): 1466–88. doi:10.1080/1461670x.2017.1279564.

31.

Lebanon: WhatsApp Tax Sparks Mass Protests. 2019. Deutsche Welle, October 18. https://www.dw.com/en/lebanon-whatsapp-tax-sparks-mass-protests/a-50880357.

32.

Lee

F. L. F.

2010a. “The Perceptual Bases of Collective Efficacy and Protest Participation: The Case of Pro-Democracy Protests in Hong Kong.” International Journal of Public Opinion Research 22 (3): 392–411.10.1093/ijpor/edq023

33.

Lee

T. T.

2010b. “Why They Don’t Trust the Media: An Examination of Factors Predicting Trust.” American Behavioral Scientist 54 (1): 8–21.10.1177/0002764210376308

34.

Lee

P. S. N.

C. Y. K.

Leung

2015. “Social Media and Umbrella Movement: Insurgent Public Sphere in Formation.” Chinese Journal of Communication 8 (4): 356–75.

35.

Lev-On

2011. “Communication, Community, Crisis: Mapping Uses and Gratifications in the Contemporary Media Environment.” New Media & Society 14 (1): 98–116. doi:10.1177/1461444811410401.

36.

Liang

2018. “broadcast versus Viral Spreading: The Structure of Diffusion Cascades and Selective Sharing on Social Media.” Journal of Communication 68 (3): 525–46. doi:10.1093/joc/jqy006.

37.

Lteif

2019. “The Lebanon Revolution Takes on the Media: A Resource on Alternative News Outlets.” Jadaliyya, December 23. https://www.jadaliyya.com/Details/40379.

38.

Lee

2018. “Partisan Information Sources and Affective Polarization: Panel Analysis of the Mediating Role of Anger and Fear.” Journalism & Mass Communication Quarterly 96 (3): 767–83. doi:10.1177/1077699018811295.

39.

Melki

2014. “The Interplay of Politics, Economics and Culture in News Framing of Middle East Wars.” Media, War & Conflict 7 (2): 165–86. doi:10.1177/1750635214537405.

40.

Melki

Kozman

2019. “Media Dependency, Selective Exposure and Trust during War: Media Sources and Information Needs of Displaced and Non-Displaced Syrians.” Media, War & Conflict. Published electronically July 10. doi:10.1177/1750635219861907.

41.

Metzger

Hartsell

Flanagin

2020. “Cognitive Dissonance or Credibility? A Comparison of Two Theoretical Explanations for Selective Exposure to Partisan News.” Communication Research 47 (1): 3–28. doi:10.1177/0093650215613136.

42.

Peralta

Wojcieszak

Lelkes

Vreese

2017. “Selective Exposure to Balanced Content and Evidence Type: The Case of Issue and Non-Issue Publics about Climate Change and Health Care.” Journalism & Mass Communication Quarterly 94 (3): 833–61. doi:10.1177/1077699016654681.

43.

Sears

Freedman

1967. “Selective Exposure to Information: A Critical Review.” Public Opinion Quarterly 31 (2): 194–213. doi:10.1086/267513.

44.

Song

Cho

Benefield

2020. “The Dynamics of Message Selection in Online Political Discussion Forums: Self-Segregation or Diverse Exposure?” Communication Research 47 (1): 125–52. doi:10.1177/0093650218790144.

45.

Spinning, Downplaying, Ignoring: Lebanon Protests and the Media. 2019. Al-Jazeera, October 28. https://www.aljazeera.com/programmes/listeningpost/2019/10/spinning-downplaying-ignoring-lebanon-protests-media-191027033939860.html.

46.

Stroud

2008. “Media Use and Political Predispositions: Revisiting the Concept of Selective Exposure.” Political Behavior 30 (3): 341–66. doi:10.1007/s11109-007-9050-9.

47.

Stroud

2010. “Polarization and Partisan Selective Exposure.” Journal of Communication 60 (3): 556–76. doi:10.1111/j.1460-2466.2010.01497.x.

48.

Stroud

Lee

2013. “Perceptions of Cable News Credibility.” Mass Communication and Society 16 (1): 67–88. doi:10.1080/15205436.2011.646449.

49.

Sullivan

2019. “The Making of Lebanon’s October Revolution.” The New Yorker, October 29. https://www.newyorker.com/news/dispatch/the-making-of-lebanons-october-revolution.

50.

Sunstein

2001. Republic.com. Princeton: Princeton University.

51.

Thomas

Bourdeau

Tagler

2018. “Interhemispheric Communication and the Preference for Attitude Consistent Information.” Laterality: Asymmetries of Body, Brain and Cognition 24 (3): 342–54. doi:10.1080/1357650x.2018.1520860.

52.

Tsfati

2003. “Media Skepticism and Climate of Opinion Perception.” International Journal of Public Opinion Research 15 (1): 65–82. doi:10.1093/ijpor/15.1.65.

53.

Tsfati

Ariely

2014. “Individual and Contextual Correlates of Trust in Media across 44 Countries.” Communication Research 41 (6): 760–82. doi:10.1177/0093650213485972.

54.

Tsfati

Cappella

2003. “Do People Watch What They Do Not Trust?.” Communication Research 30 (5): 504–29. doi:10.1177/0093650203253371.

55.

Tufekci

Wilson

2012. “Social Media and the Decision to Participate in Political Protest: Observations from Tahrir Square.” Journal of Communication 62 (2): 363–79. doi:10.1111/j.1460-2466.2012.01629.x.

56.

Valenzuela

Arriagada

Scherman

2012. “The Social Media Basis of Youth Protest Behavior: The Case of Chile.” Journal of Communication 62 (2): 299–314. doi:10.1111/j.1460-2466.2012.01635.x.

57.

Velasquez

LaRose

2015. “Youth Collective Activism through Social Media: The Role of Collective Efficacy.” New Media & Society 17 (6): 899–918.10.1177/1461444813518391

58.

Westerwick

Kleinman

Knobloch-Westerwick

2013. “Turn a Blind Eye If You Care: Impacts of Attitude Consistency, Importance, and Credibility on Seeking of Political Information and Implications for Attitudes.” Journal of Communication 63 (3): 432–53. doi:10.1111/jcom.12028.

59.

Wojcieszak

2019. “What Predicts Selective Exposure Online: Testing Political Attitudes, Credibility, and Social Identity.” Communication Research. Published electronically May 20. doi:10.1177/0093650219844868.

60.

Yee

Saad

2019. “To Make Sense of Lebanon’s Protests, Follow the Garbage.” The New York Times, December 3. doi:nyti.ms/2DCMtZD.

61.

Zhu

Skoric

Shen

2017. “I Shield Myself from Thee: Selective Avoidance on Social Media during Political Protests.” Political Communication 34 (1): 112–31. doi:10.1080/10584609.2016.1222471.