Barriers to Participation in Polarized Online Discussions About Covid‐19 and the Russo‐Ukrainian War

Even though social networking sites create a unique online public space for the exchange of opinions, only a small share of citizens participate in online discussions. Moreover, research has depicted current online discussions as highly uncivil, hostile, and polarized, and the number of heated discussions has escalated in the last two years because of health, social, and security crises. This study investigates the perceived barriers to participation in Facebook discussions, focusing on two topics: the Covid‐19 pandemic and the Russo‐Ukrainian War. It explores the role that the negativity of these online discussions has on participation. To investigate the perspectives of users and their personal experiences with online dis‐ cussions in times of crisis, we apply a qualitative research method and interviews with participants. We collected and analyzed 50 semi‐structured interviews with Czech Facebook users who participated in discussions during the spring of 2021 (i.e., Covid‐19) and the spring of 2022 (i.e., Russo‐Ukrainian War). The results show that, after initial mobilization at the beginning of the pandemic, the crisis reinforced several crucial barriers to participation in discussions due to the perceived persistence of polarization (e.g., the spread of disinformation, the bipolar character of discussions, negative perception of opponents), which subsequently spread to other areas and issues. The data also implies that these barriers tend to demobilize less active participants, those who do not have strong opinions, and participants who think the subject matter is not worth the heated exchange of opinions.


Introduction
In recent years, political discussions have increasingly taken place online, which has inspired prolific research in the fields of political science and communication.This attention is not unjustified, because informal political talk is considered an essential pillar of a healthy democracy, allowing citizens to learn about and interact with matters of public concern, form and articulate opinions, and have more opportunities for political engagement (Conover & Searing, 2005;Dryzek, 2000;Gutmann & Thompson, 2004).Informal political talk is seen as important across different models of democratic citizenship, and it is particularly important from the standpoint of deliberative democracy because it may contribute to the enhancement of the public sphere.From this standpoint, political discussion is often judged based on its deliberative potential, which lies in the exchange of diverse opinions, critical thinking, re-evaluation of opinions, and orientation toward the public good (Mutz, 2006;Rossini & Stromer-Galley, 2019).Through informal political discussions, citizens are exposed to and contribute to raising new topics and perspectives into the public spherebeing a precursor to more sophisticated forms of political engagement (Habermas, 1989).Even though online political discussions are perceived as a valuable form of political participation (Ohme, 2019), scholars have raised concerns about access and new barriers to the online public sphere, which might exclude some voices from the discussions (Habermas, 2022;Kennedy et al., 2021;Vochocová et al., 2016).
Much of the research on online political talk has been oriented by the normative principles of deliberation, such as reflexivity, openness for dialogue, reasongiving, and publicly oriented citizens (see Dahlberg, 2001;Habermas, 1984).But the reality of online discussion differs, and citizens do not strive for the fulfilment of the quality criteria of the public sphere (Rossini & Stromer-Galley, 2019).The lack of reflexive conversation between those who hold different opinions (Štětka & Vochocová, 2014) and the incivility characterized by disrespect is a relatively common part of the online political talk (Kim et al., 2021;Rossini, 2022).As such, scholars have argued that the value of online political talk should not be restricted to the elusive normative expectations of discursive quality because they may often be detached from the reality of counter-attitudinal opinion exchange (Rossini & Stromer-Galley, 2019).
Research suggests that people have different perceptions and reactions to uncivil and hostile political talk.Some people withdraw from such debates.Others find them engaging and entertaining (see Sydnor, 2019).While scholars have paid considerable attention to online political talk since the popularization of social media, the dynamics that underlie these conversations may have changed in recent years because heated discussions have escalated in light of health and security crises (Jiang et al., 2020;Zhu et al., 2022).In this context, instead of focusing on the perceived quality of online discussions from a normative standpoint, this study focuses on the discussion dynamics that may have a demobilizing effect on political participation.We examined Czechia-based online political discussions during two recent global crises: the Covid-19 pandemic and the Russo-Ukrainian War.Considering the polarizing nature of these two topics, we focus on people's experiences with engagement in contentious and heated discussions, that is, being exposed to disagreement and incivility.We focus primarily on Facebook discussions because Facebook is the most popular social networking site in Czechia.
Prior research on online discussions has mainly focused on quantitative approaches, such as content analysis to capture the content and character of interactions (Andersson, 2022;Numerato et al., 2019;Rossini, 2022), survey or experimental research to investigate the behavior of discussants (Kenski et al., 2017;Rösner et al., 2016), or research on the effect of discussions (Hwang et al., 2014).Less attention has been given to the qualitative methods that address the meanings and perspectives of users who engage in online discussion spaces.To provide a more nuanced account of people's experiences in online debates in times of crisis, we implemented a qualitative research design with semi-structured interviews.Qualitative interviews allow us to explore the experiences of the participants of online discussions and engage them more deeply in their reflections on their and others' past and current behavior and to access the changes within their behavior.Whereas attention is often paid to the drivers and factors that influence political participation (Ohme, 2019;Vochocová et al., 2016), we would like to shed light on the perception of the dynamics (e.g., the tone, content, heterogeneity of opinion expression) that discourage citizens from participation in the debate.
This research focuses on perceived discussion dynamics that have a demobilizing effect on participation in political discussions (i.e., barriers) in the context of two crises: the Covid-19 pandemic and the Russo-Ukrainian War.The uniqueness of these two highly disruptive and polarizing events enables us to examine the challenging dynamics of online political discussions during contentious times that may deter participation.Our study shows that, while participants seem to share some consensus when describing discursive dynamics that disincentivize them from participation (e.g., the conflict between the normative expectations regarding the quality of discussions, the reality of crosscutting exchange), the context for these two crises has contributed to deepen divisions and further demobilize participation (e.g., the spread of disinformation, the divisive character of discussions, negative perception of opponents).This was particularly true during the pandemic because of the perceived ongoing polarization in the discussions, which subsequently spread to other areas and issues (including discussions about the Russo-Ukrainian War).After initial mobilization at the beginning of the pandemic, the growing polarization in public attitudes about the government response to the pandemic led participants to withdraw from debating these issues online-and that is particularly true for the less active discussants.This is concerning because the demobilizing effect may persist despite the eventual ends of these specific crises and their potential to increase opinion polarization.Ultimately, if citizens perceive cross-cutting discussions to be hostile and if it will lead them to refrain from participation, this could lead to the prevalence of more extreme-and potentially homogeneous-opinions online and contribute to the increased perceptions of polarization.Insofar as online discussions have the power to form opinions about current issues and insofar as social media are essential sources of information in crisis time (Van Aelst et al., 2021), it is crucial to understand how distinct discussion dynamics may demobilize citizens' participation.Further implications in the context of the online public sphere and deliberative democracy are discussed.

Theoretical Background
Scholarship in political communication has scrutinized how important events, such as the Covid-19 pandemic, may influence how citizens consume and engage with digital media and news.Research conducted at the beginning of the pandemic shows key changes in media practices and news consumption because of the initial need for information.Several research studies indicate that the overall consumption of news increased (Mihelj et al., 2021;Van Aelst et al., 2021), including online news and social media usage (Van Aelst et al., 2021).However, it seems that these changes were rather short-term, and the audience practices quickly returned to their previous states (Kormelink & Gunnewiek, 2022).Furthermore, research also revealed subsequent avoidance of media content about Covid-19, which proved to be a stressful and overwhelming topic (Mihelj et al., 2021).
Another factor that could reinforce the decrease in interest in the pandemic is the high level of polarization that developed around Covid-19, including divisive views related to vaccination and governmental measures (Jiang et al., 2020).Since the beginning of 2022, another global crisis has affected public debate in Czechia: the full-scale invasion of Ukraine by Russia.This conflict is particularly salient in Czechia because the country was significantly engaged in military aid and help to protect refugees, despite some opposition.The perception of the Russo-Ukrainian conflict might also be influenced by the long-lasting Russian propaganda within post-communist Czechia and the geographical and cultural closeness between Czechia and Ukraine.However, the first wave of solidarity was later affected by other concerns, like the economic difficulties faced by Czech citizens (Münich & Protivínský, 2023).
Both of the crises fueled divides in Czech society at different levels and are arguably the most controversial topics of the last two years.Unprecedented events and crises are likely to lead to polarized debates (see Hiaeshutter-Rice & Hawkins, 2022;Lee & Nerghes, 2018), mainly because people hold different opinions and may express themselves more harshly in light of heightened emotions by resorting to hate speech, threats, and attacks (Schudson, 1997).Hostile and uncivil communication and heterogeneous opinion exchange were previously recognized as patterns for online discussions that might discourage citizens from participation (see Sydnor, 2019;Vraga et al., 2015).As the importance of social media during crises is widely recognized (Malova, 2021), we focus on how citizens articulate and perceive the dynamics of online discussions that may disengage them from participating.We ask: RQ1: What discussion dynamics can demobilize participation in online debates during global crises?
Besides the role of crises, we also specifically address the barriers associated with cross-cutting discussions and emphasize the conflicts within opinion exchange.Cross-cutting discussions are based on various diverse opinion exchanges, where people are likely to experience exposure to disagreement (Lu & Lee, 2021).Participating in cross-cutting discussions is an important element of active democratic citizenship because it might help to develop critical thinking and raise awareness about diverging opinions, potentially leading to increased respect towards the other side (Mutz, 2006).However, there are concerns that cross-cutting conversations might result in uncertainties about political opinion and further demobilization (Chen & Lin, 2021).It might also lead to cutting ties due to disagreement (Choi, 2021) or an increase in polarization (Hwang et al., 2014), which is supported by perceived social distance and the prevalence of the feeling they have less in common with people who hold opposing views (Duggan & Smith, 2016).Moreover, the negative character of online debates based on disagreement leads to the discouragement of future conversations, which strengthens the polarizing effect of cross-cutting conversations (Marchal, 2022).
Willingness to participate in cross-cutting discussions -discussions in which participants are exposed to counter-attitudinal viewpoints-is shaped by individuallevel characteristics, such as political interest (Lu & Lee, 2021), general active engagement in online political discussions (Heatherly et al., 2017), and conflict avoidance (Sydnor, 2019;Vraga et al., 2015).We do not know much about the aspects of heterogeneous conversations that may deter people from engaging in them.More generally, we do not know the extent to which such discussion dynamics may undermine people's perceptions of the value of those conversations.The role of disagreement in online political talk has been primarily examined through survey-based research (Choi, 2021) and little is known about how participants experience conversations where they are faced with counter-attitudinal opinions online.To better understand the extent to which cross-cutting discussions-and the associated polarizing dynamicsmay pose barriers for people to participate in online political talk, we ask: RQ2: What role do cross-cutting discussions and their characteristics play in the unwillingness to participate in discussions on Facebook?RQ3: How are online cross-cutting discussions related to the perception of polarization among the public?Online discussions are often described as problematic due to their negative attributes, such as incivility, which potentially trigger negative emotions (e.g., anger, frustration) that make it hard for some to participate (Chen, 2017;Sydnor, 2019).Incivility is commonly operationalized as expressions that violate social norms.It refers to rude or harsh opinion expressions (Rossini, 2022), like name-calling, aspersions, lying, vulgarity, and pejorative speech (Coe et al., 2014).However, people's perceptions of incivility may differ for several reasons, such as personal characteristics (Bormann, 2022;Kenski et al., 2017), conflict orientation (Sydnor, 2019), and experiences with online discussions (Coe et al., 2014;Hmielowski et al., 2014).Moreover, incivility from like-minded groups tends to be evaluated as less uncivil than when it comes from "the other side."People are more tolerant when incivility is targeted at arguments instead of personal characteristics (Muddiman, 2017).Some warned that incivility may deepen the divide between people who hold different opinions and increase polarization (Anderson et al., 2014;Hwang et al., 2014).However, reactions to incivility also vary.For some, incivility might fuel negative feelings (see Rösner et al., 2016) and lead to unfriending (Goyanes et al., 2021).Others might find it to be an acceptable way to communicate (Sydnor, 2019).Much of this research has leveraged quantitative approaches, such as surveys and experiments, to investigate perceptions and effects.As such, we lack a more nuanced explanation for why incivility seems to come with the territory for some but is perceived as unacceptable by others.We are also interested in understanding how people experience and cope with incivility in online discussions.Concerning the possible effects of incivility on participants in discussions, we ask: RQ4: How does incivility affect active participation in discussions on Facebook?

Methods
We use qualitative semi-structured interviews and thematic analysis (Braun & Clarke, 2006;Brett & Wheeler, 2022), with a focus on identifying the individual experiences of users and their meanings.

Data Collection and Sample
Semi-structured interviews were collected in two crises.The first period (March-April 2021) covers a hard Covid-19 lockdown in Czechia.The interviews focused on online discussions about Covid-19 and were conducted mostly online (n = 20).The interviews in the second period (March-April 2022) focused on both the online discussions about the Russo-Ukrainian War, which started with an invasion on 24 February, and Covid-19 (n = 30).These were conducted both online and face-to-face (based on the preferences of participants).We used purposive sampling.The trained interviewers looked for research participants who were active in any online discussions on Facebook (including Facebook groups and private/public pages or personal/friends' Facebook wall) and had specific experience with active engagement (e.g., writing comments, reacting with the like/emoji button) in discussions about the two crisis topics.Participants were recruited via the snowball method by contacts retrieved by interviewers who verified selec-tion criteria with potential participants.Although we did not provide any financial compensation, the response rate was relatively high (only five contacted participants refused to participate).This might be related to the character of the topic and the shared interest: active discussion engagement.The final sample included participants with various socio-demographic characteristics and levels of engagement.We aimed to intentionally involve people with various socio-demographic characteristics, because these may shape online political participation (see Kennedy et al., 2021;Vochocová et al., 2016).This allows us to capture different experiences with online discussions.Despite our focus on Facebook, many participants also had experience with discussions on other social network sites, especially Twitter.Interviews were conducted after informed consent and lasted approximately 60 minutes.Then they were transcribed and anonymized for analysis.The interview guide covers three sections: general use of social network sites, especially Facebook; engagement in online discussions and specific experiences with the discussions of crises topics; and perceptions of Facebook as a discussion environment and selective activities (e.g., unfriending, blocking, homogeneity of the network, negative/positive experiences with cross-cutting discussions).
The final sample (N = 50) varies with regard to the age of the participants from 21 to 74 (Mean = 35, Median = 29), gender (32% female), education level (prevalence of participants with higher education), residence (dominance of bigger cities with more than 100,000 inhabitants), marital status, and having children (32% declared to have at least one child), which seem to have an impact on political participation (Norris et al., 2004; see a detailed description of the data sample in the Supplementary Material).Participants also differed with respect to their communication strategies and their roles in the online discussions (e.g., correcting false information, enjoying conflict, conflict avoidance), their previous experiences with online discussions and history capturing the unique dynamic of various debates, and their attitudes about both crisis topics (e.g., provaccine and anti-vaccine and pro-Russian/pro-Ukrainian).Although we did not ask participants explicitly about their attitudes, the sample varied in this regard.The subsequent analysis shows satisfactory theoretical saturation in the sample for different strategies for engaging in debates and the perceptions of discursive dynamics in online discussions.

Analysis
Anonymized data from the interviews were coded by four trained coders (including the two authors).ATLAS.ti was used for coding and data analysis.Intercoder reliability was ensured through weekly training sessions during the ongoing coding process.The codebook was built through careful review of the coded interviews and the repeated reading of each other's coded interviews.
Differences were discussed and solved within the team.Data were then inductively analyzed by implementing a process of thematic analysis (Braun & Clarke, 2006;Brett & Wheeler, 2022).We started with an initial reading of all of the interviews and recorded the emerging themes.Then, we developed coding frames based on 10 interviews and elaborated the frames by merging some codes and adding new sub-codes.We first focused on experiences with active participation in online discussions (e.g., perception of other discussants, emotions connected to discussions, characteristics of opinion exchange), challenges to participation, and the implications of the experiences for further participation (or attitudes).Subsequently, while analyzing the data, we developed themes that targeted the role of cross-cutting discussions and the perceived and experienced incivility in the participants' willingness to engage in discussions.All relevant segments related to those themes were captured via more nuanced sub-codes.We tracked the new codes and their descriptions, including systematic, repeated interview reading and re-coding.In the final analysis stage, we generated all of the information segments related to the chosen codes (or group codes) and focused primarily on the differences similarities among participants and key themes.
Interviews were conducted according to ethical standards for qualitative interview research (Brett & Wheeler, 2022).Voice records were deleted and transcripts were stored with password protection, in accordance with the ethical code at Masaryk University.We avoided asking specific questions about sensitive political opinions and attitudes unless the participants wanted to share their views.Additionally, the protection of the participants and interviewers was guaranteed by the possibility to end the interview at any time and without any particular reason.

The Main Barriers to Participation in Discussions on Facebook in Times of Crisis
To answer RQ1, we found that participants clearly noted the impact of both crises in how they perceive and the extent to which they are willing to participate in online discussions.Participants felt that the already fragmented social media environment became more polarized during the pandemic.Alžběta (female, 45) and Karel (male, 29) observed that the pandemic was capable of splitting groups that had had similar political opinions beforehand.The perceived opinion polarization experienced during (and after) the first year of the pandemic quickly spread to other political content.This led participants to avoid not only discussions regarding Covid-19 but also other topics (described as "polarizing" or "controversial").This included the Russo-Ukrainian War, which had two extreme sides and the same dynamics (e.g., the people who were denying Covid-19 and, later the vac-cines, were, based on participants' observations, most likely supporters of Russia): "Covid [discussions] have polarized society terribly.There's always been some consensus in those political discussions, but here there's no in-between.One is either a fanatical supporter of regulation or a fanatical opponent.There's rarely half-and-half" (Jaroslav, male, 40); "But the fact is that a lot of people, or a lot of people who were against the measures and were swearing at Covid fascism and so on, today are swearing at the Ukrainian fascist" (Adéla, female, 61).
The frustration with the dynamic of the discussions on these topics reflects that participants are not willing to listen to the other side.Some participants say they are exhausted by discussing these topics and unwilling to engage in discussions they describe as "pointless."Moreover, participants perceive these debates as extremely divided.Especially in discussions about the pandemic, participants perceived no room for a middle ground, with balanced opinions being pushed aside.Because extreme opinions were predominant, participants felt that balanced opinions were not endorsed (Jonáš, male, 30).These dynamics reinforce a withdrawal from discussions (Askay, 2015)-Cross-cutting discussions in times of crisis are perceived as aggressive due to extreme opinions, driving those in the middle to disengage: There are only opinions that I'm either extremely against or extremely for.There's no compromise, no middle ground anywhere.People aren't able to accept the other side's arguments at all.(Václav, male, 28) Because the loudest voices from the extremes are heard on social media.It often seems to me that even though 99% of the issues are some kind of spectrum and nothing is black and white, it's the black or the white that's being addressed on those social networks and there's nothing in between.(Oliver, male, 34) Participants from the second round of interviews in the spring of 2022 admitted that in the case of Covid-19, especially when the pandemic started, there was considerably more motivation to get involved in the discussion, and they were quite mobilized.This is explained by the initial need to make sense of what was going on, which led people to spread the information they perceived as correct to help others and also to stop the spread of disinformation (e.g., Šimon, male, 22).But as the situation progressed, mobilization decreased due to growing negative experiences, a perceived decrease in the meaningfulness of the efforts, and the perceived value of the overall discussions-which was significantly affected by conspiratorial sources and disinformation.The avoidance of negative experiences in these debates was also explained in light of the stressful pandemic.Evženie (female, 52) tried to avoid conversations that could make her angry because she felt frustrated that she was stuck at home for a long time.These negative discussions about the pandemic led many participants to intentionally avoid discussions about the war.Many of the participants expressed frustration with polarized discussions built on low-quality and questionable sources, fake news, and propaganda, which could lead to unfriending or blocking certain people and content: There's an awful lot of overlap between these groups, it seems to me, which I think is logical because it's going to be similar people who are susceptible to the propaganda that's just coming from the same (disinformation) channels as the Covid one before and the anti-Ukrainian one today.(Daniel, male, 22) I have to say, since the war started, I unfortunately removed about six people from my friends because I couldn't take their covert aggression, ridicule, and contempt for people anymore.(Šárka, female, 38)

The Polarizing Role of Cross-Cutting Discussions in Unwillingness to Participate in Discussions on Facebook
With declining mobilization after the initial phase of the pandemic, we observed, regarding RQ2, that participants became increasingly unwilling to engage with opposing opinions and to participate in cross-cutting discussions, especially about controversial issues (e.g., #MeToo, migration).Most participants reported increasingly avoiding cross-cutting discussions about such topics, which often featured aggressive and emotional responses.Participants believe it is not a good idea to contribute to these conversations, either because they hold strong opinions (Jonáš,male,30;Ondřej,male,46), or because they wanted to avoid extreme and unpleasant discussions due to previous negative experiences: So those types of topics [e.g., a story about a brutally raped woman], I know that the majority of discussants in the Czech Republic will focus on the fact that those women are responsible for what happened to them, so I refuse to participate.This simply does not make any sense.(Lada,female,23) Besides the polarizing character of the discussed issues, barriers to participation in cross-cutting discussions are mainly rooted in participants' expectations of how counter-attitudinal opinion exchanges should look.Put simply, some participants have higher expectations to engage in polite opinion exchanges where the "best arguments" should prevail, with participants willing to change their views.For many, the inability to change the opinions of those on the other side demotivates them from engagement because they do not see the benefits of investing time and energy in discussions that are not productive.Besides changing others' opinions, some people mentioned that these discussions often lack rational opinion exchange and constructive dialogue.Thus, participants perceive cross-cutting discussions as unproductive because people talk across one another instead of engaging with divergent views: I won't expect this person to say, "Oh, Jesus, yes now I see the point.It is absolutely like you say."I know it won't happen, but from this discussion, you feel that the other person is unwilling to think and keeps telling his own story.(Filip, male, 30) A critical factor is the perceived level of homogeneity or heterogeneity for the opinion in the discussions.Many participants (especially those less active in writing comments in online discussions and those less assertive in pushing their arguments forward and defending their opinions) feel discouraged from participating in discussions where their own opinions diverge from the predominant views.On the other hand, perceived alignment with majority opinions may encourage participants to share their views.However, some participants feel like homogeneous discussions are not productive and believe that heterogenous discussions are more meaningful for understanding others and seeing their point of view (Luboš, male, 30): When someone throws an opinion out there in their bubble, whether strongly held or as part of a trend, they have people who agree with it.There are maybe 60 responses in agreement.And you can just write something completely opposite.Well, they'll come down on you and put you down! (Čeněk, male, 56) There are two poles, but there are certainly also many people who just move between them or have some completely alternative, slightly different opinion that doesn't fit even on that one continuous scale.And it's actually interesting to read how those people think about it.(Jáchym, male, 24) The perception of polarization (RQ3) is largely explained by an overall negative evaluation of those on the other side.This is especially true with respect to communicative skills, intellect, and media literacy.The negative perception of opponents is likely exacerbated by the abovementioned ongoing polarization, which seems to be both opinion-and ideology-driven, and more affectively based (Iyengar et al., 2012).Participants who apparently support governmental restrictions are labelled as "sheep" who follow rules without thinking or described as radicals who take the rules too seriously (Čeněk, male, 56; Jitka, female, 29).Also, the lack of lived experience with Ukrainian immigrants, for instance, tends to be used to undermine the opposing view.The communication strategies of those "on the other side" were described as not worthy of a conversation because they are unable to listen or lack critical thinking.Overall, their argumentative skills and mental capacity are degraded: "Those people, you won't convince them.They just have their own perception of the world.I think that some of them are unable to absorb arguments that you try to explain them.It is beyond their mental capacity" (Jan, male, 60).
Participants also perceived those on the "other side" as being more vulgar (Jan, male, 60).This is noticeable when participants refer to their side as polite commenters and "they" as toxic and more aggressive participants (David, male, 25).Moreover, the negative perception of the "others" is often associated with the perceived inability of the opponents to use relevant or factual information (or even the intentional use of disinformation).The participants tend to describe their opponents as being less educated and unable to have critical discussions.Participants also refrain from participating in conversations when they feel other people are spreading disinformation.However, several participants made a clear distinction between false information, which they felt made sense to correct, and disinformation, which is too radical to engage.If there is a feeling that the other person's opinions may be changed, there is more motivation to get involved than to leave the discussion.Mutual antipathy based on negative prescribed characteristics between two opposing groups seemed to deepen the divide between people who hold different opinions: Even though the comment is civil, it's completely based on total bullshit because someone writes, "I'm sorry, but you want to support Bill Gates, who's here…" or [then they write] "Nazi Zelensky."And stuff like that.The ones that are completely confused by disinformation, those strike me as being over the edge.(Luboš, male, 30)

Incivility and Its Effects on Active Participation in Discussions on Facebook
Uncivil opinion expression represents a very significant barrier to participation for participants who expect the discussions to be polite (RQ4).For these participants, verbal attacks and antinormative intensity are seen as problematic in cross-cutting discussions, and that may lead participants to perceive other groups as being hostile or irrational.However, we note that participants who avoid uncivil discussions were typically less active in online discussions, which corroborates the argument that incivility may become normalized for those who more frequently participate (Hmielowski et al., 2014), but which may also deter others from engagement.Those who avoid uncivil discussions feel that it is pointless to be part of irrational discussions where people just shout at each other.
Participants shared various coping strategies when exposed to incivility.While some are discouraged from participating in discussions altogether, others responded by leaving the discussion, and some continue to read the comment thread without participating.Others admit to punching back by being hostile or aggressive in response.Most participants described a mix of these reactions.Additionally, different types of incivility matter according to participants' responses.Personal attacks are seen as the most harmful, especially ad hominem attacks that comment on other online discussions participants' hair style, body, or age, in alignment with prior findings (Muddiman, 2017).In the case of more serious attacks, such as racism, threats of violence, or attacks against human rights, many participants reacted by reporting the behavior.Whereas participants had different sensitivity toward various levels of incivility, racist, and serious violent threads were unanimously condemned: Instead of making an effort to foster argumentation, it will turn into personal attacks.So simply I wrote to this person that we will just stop, that I am not interested when he scolds me.So, bye!I finish it because it does not make any sense to continue in such a discussion.(Jonáš,male,30) When he downloads pictures from a profile of kids of the other discussants and writes down threats, it is something that I really do not like.I always report it.(Bára,female,25) Incivility is not perceived as a universal incentive to demobilize.Another reaction to incivility is resilience, which refers to the ability to counter the presence of incivility in online discussions by ignoring it (see Humprecht et al., 2020), which is contrary to other strategies that include avoidance.Perceived resilience is-besides the character of incivility and the impact on readers' perception-also related to participants' experiences with online discussions.Some participants feel like they became less affected by personal attacks over time (Alžběta,female,45) and that this resilience is justified by the importance of "not giving up" cross-cutting discussions.As noted by prior research, frequent discussants tend to be less affected by anti-normative discourse (Hmielowski et al., 2014).Patrik (male, 28) described how he persevered and strove for mutual understanding in debates.But when people use arguments without thinking (i.e., "verbal vomit"), he kept his distance.Others went a bit further and started to use uncivil attacks to react to previous vulgar comments addressed to them or other discussants, to defend their points of view.Only very few participants sometimes participated in heated discussions.There is a conclusive connection between the willingness to fight for other people's rights to protect them against others and sharing their point of view.Courage to enter the environment and be part of discussions that they considered hostile was supported by fuelling angry reactions and a tendency to not overlook it or leave it as it was (Jarmila,female,45).Those who find it easier to cope with incivility tend to enjoy participating in heated conversations and are themselves uncivil.
An interesting perspective is captured by those who say that incivility is how some people communicate, which is quite rare and apparently helps overcome a decline in motivation to participate in uncivil conversations, and some participants feel resistant to vulgar attacks related to their personalities (Patrik, male, 28), but this is a minority perspective in our sample.For most participants, uncivil opinion expressions negatively affect the perception of those on the other side.Discussants who resort to vulgar and aggressive language are mostly described as less intelligent or educated ("It says something about their intellect"; Evženie, female, 52), seen as extremists who do not follow the informal norms of opinion exchange.The perception of lower education is justified by grammatical mistakes, the spread of disinformation, and argumentative errors.The perceived characteristics of those who resort to incivility strengthen the unwillingness of participants to engage with them.

Conclusions
The aim of this article was to examine and explain the perceived barriers to participation in online discussions on Facebook in Czechia during two major global crises.Our findings are related to the barriers perceived by people who are (or were) typically engaged in this type of discussion.As such, we are unable to speak about more general reasons for avoiding discussions.
We a rapid decrease in the willingness to participate in political discussions on Facebook as crises unfolded.After a boost in participation at the beginning of the pandemic, perhaps for similar reasons that led to an increased interest in news consumption (Van Aelst et al., 2021), the mobilization sharply declined over the course of the pandemic and was much lower in the Russo-Ukraine war.It is explained by participants as the reinforcement of the negative aspects of online discussions.While the first year of the pandemic in Czechia was characterized by an intense wave of solidarity and support (Buštíková & Baboš, 2020), the discussions on Facebook became increasingly perceived as more aggressive, uncivil, divided, and significantly affected by disinformation.These perceptions appear to be connected to several factors, such as the long-lasting crisis time and repeated negative experiences with opinion exchange in an online environment.Moreover, according to the study participants, these crucial barriers were first attributed to polarized discussions about Covid-19, and then the second crisis in 2022, the Russo-Ukrainian War, which only deepened the trends.The geographical closeness of the conflict to Czechia and its leading position within Europe with regards to Ukrainian refugees per capita (Münich & Protivínský, 2023) contribute to the sensitivity of the topic and the frustration about counter-attitudinal opinions, which were often contrasted to the lived experiences with refugees.
Perceived polarization experienced during the pandemic and the subsequent conflicts that penetrated other areas of society was apparent via the reinforced uncivil character of communication, criticism about the low quality of the discourse, and the negative evaluation of their opponents.Similar to the research of Hwang et al. (2014), we found a link between incivility and polarization based on the negative perception of "the others" (Iyengar et al., 2012).We identified a clear tendency for participants to delineate "us" versus "them" in several ways: their style of communication (e.g., vulgar), their communicative skills, intellect (such as following government measurements without critical reflection), education, lifestyle (e.g., profile pictures), and media literacy (e.g., quality of sources and disinformation).Remarkably, according to the study of Numerato et al. (2019), harsh criticism and the denunciation of opponents were already present in the vaccination debate on Facebook in 2016 (i.e., in the United States), several years before the pandemic.Importantly, the dynamics of villainizing the opposition could contribute to further dividing those who are on opposing sides of an argument (Hwang et al., 2014), which may lead to further withdrawal from controversial and polarized debates (Marchal, 2022).
Although incivility is considered to be problematic during heterogeneous debates, it does not represent a universal incentive to demobilize because a considerable group of users is resilient toward it.Several participants find uncivil discussions to be normal (see Sydnor, 2019), although this was a minority perspective in our sample.For many participants, incivility and perceived opinion polarization are seen as significant barriers to engaging in discussions, which possibly contribute to the demobilization of certain groups.This is particularly true for those who are less active in online discussions, less uncivil, less extreme, and have a more strict attitude toward normative ideals for discussions and their requirements for the participants and their behavior.
To sum up, the aforementioned barriers to participation became more apparent and problematic as both crises unfolded, and this was particularly consequential to discourage less active commenters from engaging in online debates.Subsequently, these contribute to further polarization in discussions (and society) because some voices might be systematically excluded from the online public sphere (Askay, 2015).An imbalance of shared opinions might have serious consequences, because opinions expressed online may impact others' perceptions of relevant issues.This is particularly relevant during crises characterized by uncertainties, fear, and a demand for relevant information.Our research also shows that, while the described barriers to participation seem to currently be more urgent for many participants, their effects can vary slightly for the different groups of users based on their personal characteristics.Results also need to be interpreted in the context of a particular platform, like Facebook, where users usually use their own identity and connect with people they know from real life.People's experiences may differ on other social media platforms that have distinct affordances, such as a level of anonymity or social cues (see Rossini, 2022), and that might also shape the dynamic of opinion exchanges.
This study and its findings have limitations.Our inferences refer to perceived experiences reported by a limited sample of participants, and self-reports of prior experiences may also differ from actual behavior.Future work could combine qualitative interviews with a diary that could provide useful support for tracking political participation online (see Mihelj et al., 2021).Although we strived for variability in our sample, our findings are not representative of the population.The sample also lacks diversity in terms of representation of minority groups.Future research could address this gap and focus on participants who are part of racial minorities and who have personal experiences with more harmful comments, such as racism, which might bring a different perspective to the barriers to active engagement in online discussions.Lastly, disinformation appears to significantly affect participants' willingness to engage in political discussion, but prior research has not shed light on this relationship.Future work needs to further explore the relationship between the role of disinformation and political expression online.