Political conversations and regret: A qualitative evaluation on the aftermath of political discussions on social media

The fundamental role of political discussions in democracy has been frequently highlighted by quantitative and qualitative literature at the intersection of political communication and media effects. Most research has revolved around whether, and if so under what conditions, social media platforms constitute public spaces where democracy can be nurtured and promoted. Building on this literature and underscoring the importance of individuals’ self-effects theories, this qualitative study, based on 42 in-depth interviews, clarifies how social media users navigate political discussions and their ulterior affective and cognitive processes, introducing the notion of political discussion regret. Specifically, this concept fundamentally emphasizes the sterility of partaking in political discussions as the main motivation for users’ cognitive lamentation, which indirectly cancels the presumed muscle of social media as the sphere of public and private political discussion and deliberations. Implications of the study’s findings and main theoretical consequences for the political discussion literature are also provided.


Introduction
In recent years the gateways for news use have mushroomed on social media, substantially altering citizens' consumption patterns and challenging users' objective and subjective political knowledge (Lee;Diehl;Valenzuela, 2022).In the case of Spain, the share of people who used Facebook for news neared 44% in 2020, a slight decline from the previous year -47%- (Statista, 2021).Beyond these practices of news distribution and consumption (Guallar et al., 2016), research has also suggested that social media enable users to partake in political discussions when exposed to news posted, linked, or shared across these platforms (Valenzuela;Kim;Gil de Zúñiga, 2012); regardless of whether it takes place in public spaces -as the case of Twitter and Facebook-, or in more private, backstage environments -such as WhatsApp (Treré, 2020).
However, not all users, who consume news, discuss politics actively, as some users may choose either to ignore the news, contribute to the discussions, or attend discussions as mere spectators (Jordá;Goyanes, 2022).Within these practices, media effects literature has documented that cognitive elaboration is an important indirect mechanism to be politically informed when exposed to news or engage in political discussions (Eveland, 2001) and, arguably, one of the most examined self-effects (Shah, 2016;Valkenburg, 2017).In short, people use social media for an array of motivations (Rojas, 2015;Gil de Zúñiga;Valenzuela;Weeks, 2016), triggering behavioral and cognitive consequences, such as political discussion, cognitive elaboration, and learning outcomes.
Quantitative research examining the antecedents and consequences of political discussion have been prolific in the last decade (Conover;Searing;Crewe, 2002;Eveland, 2004;Valenzuela;Kim;Gil de Zúñiga, 2012;Rojas, 2015).Despite this vast knowledge, what remains unclear is to what extent and under what conditions social media political discussions lay fertile paths for engendering negative self-effects, such as regret.As Valkenburg, points out, "few communication theories have conceptualized how creating or sending messages for the purpose of communicating to others may affect oneself" (Valkenburg, 2017, p. 478).
This study takes on this call and focuses on how users navigate political discussions and how the dynamic interplay of practices of persuasion, tolerance, and incivility, articulates an interesting message release effect (Pingree, 2007): social media political discussion regret.
Ultimately, feelings of regret provide an indirect mechanism to calibrate citizens' perceptions on the intrinsic value of holding political discussions on social media, enabling us to better theorize about users' cognitive and emotional sensemaking in the aftermath of such interactions, whether they were civil or uncivil.Based on a set of 42 of qualitative interviews, our findings first illustrate the main rationales behind the initiation of political exchanges on social media and how the political thematic patterns set in the agenda are the most common triggers, typically with strong ties in WhatsApp, and to a lesser extent with weak ties on Twitter.
Our study also shows that when individuals participate in political discussions, some of them activate affective and cognitive mechanisms to reflect upon such interactions.Due to this reflection on political discussions, sometimes individuals may realize that they could have handled certain situations differently, which we broadly define as social media political discussion regret.With this concept we add nuance to prior findings by showing that most users find political discussions on social media sterile, typically lamenting the participation in the proper political discussion itself rather than about its terms, origin and nature.This study contributes to the budding literature on self-effects (Pingree, 2007;Shah, 2016;Valkenburg, 2017), proposing a new concept to examine the potential muscle political exchanges have in activating users' feelings of regret.

Literature review 2.1. The dynamics of social media discussions
Defined as the "episodes of political conversation and discussion that take place between the non-elite members of political community" (Schmitt-Beck, 2008, p. 341), political discussion is considered as a major component of the political process (Valenzuela; Kim; Gil de Zúñiga, 2012).
Scholars have consistently argued that in addition to enabling knowledge exchanges, political discussions also involve interpretive frameworks helping to process novel information (Eveland, 2004 Due to the growth of users on social media platforms, an increasing number of people discuss politics in these ecologies, which allow for "creation and exchange of user-generated content" (Hampton; Shin; Lu, 2017, p. 1092).
Scholars believed that social media platforms have constituted online public spheres in which a deliberative democracy can be nurtured and promoted (Hampton;Shin;Lu, 2017).A defining characteristic of a deliberative democracy speaks to the free flow of information, which provides accountability and justification for the political order (Chambers, 2003).Therefore, free and public exchange of political information, ideas, and ideals (i.e., political discussions) on social media may provide an arena for the sustainability of democratic societies (Halpern;Gibbs, 2013).
A plethora of studies has probed the motivations, antecedents, and consequences of political discussions on social media platforms.Many argue that motivations and antecedents of political discussions include information seeking, opinion formation, self-presentation, persuading others, as well as entertainment and relaxation (e.g.Prior research argues that frequent political discussions on social media provide individuals exposure to political viewpoint diversity, which is beneficial for the development of discussion network heterogeneity (Kim;Chen, 2015).A heterogeneous discussion network, in turn, can enhance digital citizens' political tolerance and "awareness of rationales for their own political opinions," providing possibility of "reasoned public opinion and deliberative democracy" (Kim;Chen, 2015Chen, , p. 2347)).Moreover, during these processes of information exchanges, individuals' political knowledge can also increase (Park, 2017).Political discussion online can enhance coordinated political efforts, such as civic engagement and political participation (Su;Xiao, 2022).This is due to not only the functioning of the social media affordances in channeling social pressure and reinforcement, which effectively persuade participants to engage in activism (Halpern; Valenzuela; Katz, 2017; Gil de Zúñiga; Ardèvol-Abreu; Casero-Ripollés, 2021), but also the fact that frequent and consistent political discussion can help construct shared identity among digital citizens (Halpern;Valenzuela;Katz, 2017).
However, political discussion can also lead to negative outcomes.For instance, as political discussions are not always heterogeneous, many people may tend to discuss with those who share similar backgrounds, potentially reducing their cognitive dissonance (Boulianne; Koc-Michalska; Bimber, 2020).These homogeneous discussions would facilitate the formation of echo chambers online, strengthening ideological polarization (Taylor; Mantzaris; Garibay, 2018), escalating populism (Boulianne; Koc-Michalska; Bimber, 2020), and intensifying the viral dissemination of misinformation (Wang;Song, 2020).
Moreover, even when discussions may fall within the more positive context of heterogeneous networks, not all discussions on social media platforms are polite.As many social media platforms afford anonymity, their users may ease in expressing themselves online in terms of politeness and interpersonal interaction norms and moral code (Hwang;Kim;Kim, 2018).These uncivil online discussions may trigger moral judgment of others, strong model condemnation, interpersonal hostility, and moral indignation (Ng;Detenber, 2005), all of which further erode justice and legitimacy of discussions, hinder the expression of disagreements, and jeopardize a more deliberative democracy (Hwang;Kim;Kim, 2018).
In sum, what we know thus far about the effects of social media political discussions and the main dynamics of discussing politics in such ecologies revolves around the main, general antecedents and effects, addressing both positive and negative outcomes, while little is known as to the self-effects of social media political discussions, such as political regret.

Regret as self-effect
In general lines, regret has been observed as a corollarial aspect within the choice-making literature, especially as part of emotional experiences while making decisions (Connolly;Zeelenberg, 2002).Although it is commonly believed that people usually make behavioral decisions to maximize expected utility such as economic profit, emotional pleasure, or political benefit, many asserted that emotions such as impulse and whim can also lead individuals to some decisions, which further elicits regret (Landman, 1987).Among several of these emotional dimensions in decision making (i.e., relief, anxiety) the one that arguably has attracted most research attention is regret (Connolly;Zeelenberg, 2002).
Regret has been studied from various perspectives.For instance, marketing science has framed regret as post-choice evaluation when considering forgone alternatives in the comparison between products (Inman; Dyer; Jia, 1997).Psychologists defined regret as a cognitively based emotion that helps people realize that the negative situation is due to their behavior and that they would have been better off if they chose a different action (Zeelenberg, 1999).These conceptualizations collectively entail a general defining character of regret: unlike other emotions such as anger or fear, regret requires more cognitive elaboration (Connolly;Zeelenberg, 2002).In other words, regret might be viewed as a normal, Not all users who consume news discuss politics actively, as some users may choose either to ignore the news, contribute to the discussions, or attend discussions as mere spectators inevitable, and direct consequence of rationality, which more likely occurs when an individual realizes that a situation is not ideal or that his/her behaviors were inappropriate (Connolly;Zeelenberg, 2002)).Therefore, regret is considered a matter of "intertwined reason and emotion" (Wang et al., 2020, p. 5).
In a nutshell, people who experienced regret would have a higher chance to alter their future behaviors to avoid similar experiences of regret.
Although research on political regret on social media is scanty, the concept is relevant for an established area of research on self-effects (Gil de Zúñiga; Molyneux; Zheng, 2014;Valkenburg, 2017).Specifically, self-effects are "the effects of messages on the cognitions (knowledge or beliefs), emotions, attitudes, and behavior of the message creators/senders themselves" (Valkenburg, 2017, p. 487).
Within this literature several studies have shown how expressing online can indeed affect oneself.For example, self-expression via blogging can increase the perception of support (Baker;Moore, 2008), or a higher sense of self-esteem (Schmitt-Beck, 2008).Although the extant research primarily supports the positive impact of self-effects, online political expression can have certain negative consequences such as a feeling of regret.Political regret can be thought of as release effects (Pingree, 2007), which refers to the "cognitive, attitudinal, emotional and behavioral side effects" (Valkenburg, 2017, pg.482) that can occur after someone has sent a message or posted on social media.Combining both the literature on regret and self-effects in communication, this study aims to inductively understand the cognitive aftermath of political discussion.Accordingly, this study poses the following research questions: RQ1.Under what circumstances do social media users regret their political discussions?RQ2.Beyond regret, which other self-effects are salient in the aftermath of holding social media political discussions?

Method
Data for this study comes from a corpus of in-depth, semi-structured interviews with Spanish social media users (n = 42).Spain is one of the liberal democracies with the greatest affective polarization in the world (Miller, 2020), a fact that may have a significant impact on the political discussion emerged in social media.In addition, more than 40 million citizens use social media (Statista, 2023), turning Spain in a fertile soil for partaking in political discussion that many may end regretting.As the main objective of this study is to navigate the complexity of social media political discussions from citizens' perspectives, we rely on rich data from interviews rather than quantitative observations from survey data, as most prior scholarship has done (Kim; Chen, 2015; Rojas, 2015; Gil de Zúñiga; Valenzuela; Weeks, 2016).Accordingly, our methodology enables us to capture and distill users' emotional reactions stemming from political discussions on social media, "across a whole range of situations and contexts" (Couldry, 2004, p. 110).
Drawing on the work of Boczkowski and colleagues, for data collection we relied on two different, but complemental techniques: snowball sampling, and random interviews with potential respondents in public places, including libraries, bookstores, or coffee shops (Boczkowski;Mitchelstein;Matassi, 2018).Regarding snowball sampling, this method was implemented due to the lockdown restrictions implemented in Spain during the COVID-19 outbreak (April, 2020).Although snowball sampling is not often used to achieve this type of representation, a growing number of qualitative studies have implemented it to gather qualitative evidence online (Goyanes; Demeter, 2022) due to its effectiveness, cost, and external validity.Face-to-face interviews in public spaces were carried out in June 2020, and participants were randomly selected, to maximize their representation and diversity in accordance with the Spanish population.
All interviews were carried out and transcribed by a team of research assistants and eventually coded and analyzed by the first author.Interviews were carried out until saturation of ideas was achieved, following three main ethical principles: all participants were granted anonymity, were fully informed about the intended nature of the study, and permission was requested to use verbatims.Accordingly, names reported in the results are pseudonyms.Interviewees from the snowball sampling were conducted in two different digital platforms: Google Meets and Skype.The length of interviews was similar in the snowball sampling, public interviews, and across the two different digital platforms, and lasted between 25-40 minutes each.By complementing face-to-face and online interviews a balanced diversity of participants was satisfactorily achieved.
As previously stated, interviews followed a semi-structured protocol in which respondents were asked about their expe-What remains unclear is to what extent and under what conditions social media political discussions lay fertile paths for engendering negative self-effects, such as regret riences with political discussions on social media, their cognitive effects (regret), and the general emotional consequences they have generated (see appendix for the interview guide in Spanish).All interviews were carried out in Spanish and all verbatims were team traduced from this language.For the selection of potential participants, we rely on Patton's, (2002) recommendations to design a diverse sample.The diversity of participants, based on key demographic variables, such as gender, age, education, region, income, and work positions, enabled us to create a rich and diverse sample, ensuring quality data to inform our research questions (see Table 1).
After data transcription, a thematic analysis (Braun;Clarke, 2006) was implemented.Thematic analysis typically encompasses six phases, which were followed to identify, analyze, and report patterns (i.e., themes) within data.These six phases were developed in subsequent steps to first familiarize oneself with the data for generating the initial codes.Then, after inductively generating themes, we interactively reviewed their congruence, defined and named the main ones.
After and before finishing the data report, we discussed the main findings with two independent researchers to clarify potential inconsistencies and clarify conceptual issues.Throughout the interviews, different thematic patterns emerged: the motivations to consume news on social media, the motivations to regret political discussions on social media, and other complemental emotional effects rooting from the political discussions.We structured the results accordingly.

Motivations and platforms for partaking in political discussions
After inquiring about respondents' political discussions on social media, our interview guide first focused on clarifying the main patterns of social media use.Not surprisingly, all our respondents declared using different social media platforms hinging on the intrinsic gratifications they cover.Twitter, Facebook, WhatsApp, Instagram or YouTube were, above all, the most referenced, but other more recent ones such as TikTok or Twitch were also mentioned by many of our participants.Beyond the common patterns of social media use addressed by our respondents-and systematically portrayed by extant research-, such as for entertainment or for being in touch with "friend and relatives," most of our respondents used Twitter and Facebook as the most salient platforms for being informed about public affairs and politics.
However, there are some important nuances among our interviewees when it comes to assessing the appropriate context of social media for political discussion purposes.According to our evidence, some respondents raised critical concerns regarding the misleading nature of the news circulating massively on social media which, according to them, might spark an uninformed public opinion.Sebastián, an unemployed mechanic from Valencia, acknowledged that he "does not trust the news shared on social media," and consequently, he claimed not to follow current news on social media, but he typically watched the news on television instead.Ana, a university sophomore from the Canary Islands, said that social media was not a suitable place for getting news and said that when she wanted to be informed about a particular (political) issue, she typically sought more actively that content on the Internet.
When respondents were asked about the frequency and nature of the political discussion they had on social media, our evidence shows an incredible diversity of opinions.Indeed, most of our respondents acknowledged their involvement in political discussions, providing three main rationales: -because they enjoy political deliberations and think they can deliver informed opinions on certain topics, -to challenge deceptive arguments or persuasive discourses that may mislead public opinion (i.e., misinformation) and -because they are open to learning from others' perspectives (i.e., being persuaded), to reinforce their ideas, and to attempt to persuade others.
Susana, a schoolteacher from Santiago de Compostela, perfectly portraits her main trigger, when she blatantly explains that many people share misinformation about political leaders because they dislike them.Francisco, a carpenter from Zamora, said that the COVID-19 health crisis had activated an array of perspectives on how to manage the pandemic and, interestingly, he thought that such diversity of perspectives "provides him a more ample and complex vision about the current situation".Accordingly, he valued others' thoughts, and, on many occasions, he directly acknowledged changing his opinion about different political issues, such as the restrictions of the Spanish government during the lockdown.
The most typical platform for discussing politics is, by far, WhatsApp.Most respondents, who initially declared having political discussions, brought up this platform, arguing that their affordances permeated the close discussion with "people I care," as Alba, a senior architect from Segovia said.The connection with strong ties in this platform furnishes and an environment "of trust and security in which you can show your political ideals with any fear of retaliation" as Fernando, a car dealer from Zaragoza, mentioned.Despite the secure context in which those discussions emerge, Paula, a bartender from Madrid, believed that such interactions were "exhausting and pointless" as she considered that "it is very difficult to learn something".Besides using WhatsApp, some respondents also used Twitter and Facebook.Two participants, who mentioned Twitter as the most appropriate platform, considered that the constant flow of news and the availability of information threads motivated their willingness to share their political opinion.
Another participant, Manuela, a tourist guide from Toledo, said that Facebook was also a potential platform for political expression, explaining how she elaborated a text to persuade people to cast their vote: "When I discussed politics, I used Facebook, since it allows you to write long texts.For example, when we were in a situation of political blockade and people said they would not go to vote, I wrote a text talking about the rights and duties of citizens and social responsibility to encourage people to vote."

Political discussion, outcomes, and regret
When we inquired about the main topics of respondents' discussions, most of them related the thematic patterns to the political issues set in the agenda.Current issues about Spanish democratic institutions, such as laws discussed in Parliament, government performance, political party leaders' opinions, and policy agendas were commonly brought up by many of our interviewees.In addition, many other respondents also referred to news events related to immigration, economics, gender, racism, education, or social rights.While there is an ample spectrum of thematic patterns, those above mentioned were arguably the most common topics that our participants discuss, and also represent the most salient topics our interviews eventually end up regretting discussing.
David, an industrial worker from A Coruña, pointed out that the topics he discusses the most were typically those revolving around education, health (mostly Covid-19) immigration and the related "agendas of Vox 1 and Podemos 2 " Fernando shared a similar thought, saying that many people and party politics prioritized the "extermination" of the universal health care system, especially, "for those most disadvantaged who do not hold a work permit [i.e., immigrants]."Efrén, a waiter from Barcelona, referred, as an example, to the coalition agreement between the Socialist Party and Podemos and the great number of ministries that the new government has.While for some respondents, political discussions on those topics were sporadic, for other participants-mostly depending on the political diversity of their network-were recurrent, even repetitive.All in all, our findings illustrate that the agenda set by the media, and the political differences that such content activate, are typically the main factors that facilitate our subjects' political discussions.
Our evidence seems to suggest that those who frequently discussed politics typically did so with people "they care," such as family or friends."I have discussed with both direct family members and lifelong friends," Sonia, a nurse from Albacete, said.Manuela also indicated that she had discussed political issues with close friends who had polarized opinions "and always want to be right," while Paula revealed she typically discussed such issues with her father "because he has a different ideology."These discussions among strong ties, are fundamentally taken place on instant message applications, mainly, WhatsApp, and the main rationale referred by our participants is because this platform enables privacy.
However, other participants said that they did not only discuss political issues with friends or family, but also with strangers, mostly on Twitter."I mostly discuss about politics with unknown users," said José, an unemployed TV producer from Cáceres.Similarly, Sergio, an industry worker from Murcia, said that he usually engages in political discussions with unknown users, because the people he knows on Twitter are politically aligned.Finally, there is a notable portion of respondents who clearly acknowledged that they had never participated in political discussion on social media, arguing that they perceived such interactions as "pointless" and "uncivil."Different people make different evaluations of social media political discussions they get engaged in and they end up regretting.To evaluate the usefulness of these discussions for the participants themselves we need to consider that different participants might hold different ideas about the purpose of this type of discussion.First, some respondents found political discussions on social media platforms useless for changing others' minds (i.e., persuade).However, some interviewees acknowledged that such discussions had moderated their viewpoints about a topic (for instance, abortion).Fernando said: "Indeed, they have been very useful to me [the discussions], since I have learned other points of view and have been enriched by the opinions of those who think differently.Somehow, I have moderated my perspective," Other participants, Fátima, a senior nurse from Barcelona, believed that even if none of the participants had changed their mind, "discussions are good for better understanding others' viewpoints." Learning new information is also a crucial outcome of engaging in political discussions, although a significant number of participants blatantly state the contrary.Sebastián said: "In my opinion, discussions on social media are typically not very useful.They serve to test the waters, but they do not generally generate value in establishing rational debates and sharing different and moderate points of view".
The new information acquired may be about the political facts or news participants discuss, other people's views, or different arguments that help participants to reinforce their previous position.Sofía, a taxi driver from Castellón, said: "As I have little political knowledge, the few times that I have discussed politics, I have obtained new information that I previously lacked, for instance about the measures implemented during the pandemic".
Many participants felt that sharing information with others was a common effect of engaging on social media political discussions, while for many others developing civic skills was the ultimate outcome.
Interestingly, civic skills such as tolerance, participation in reasonable debates and civil exchange of opinions, are common elements brought by many of our interviewees.For instance, Mariña, an unemployed schoolteacher from Vigo, believed that she had learned from engaging in political discussions about how to better communicate, especially "by being calm and more resilient."While some interviewees reported that they had taught or showed others how to be more tolerant, some others said that it was impossible to have civic discussions with some individuals because of the dynamics social media ecologies foster (i.e., misunderstandings, incentives for vehement and passionate arguments, not a good space for rational debate), as the following quote illustrates: "What I have learned is that you have to know with whom and where to discuss politics.Ultimately, I think I have learned that on social media people tend to be more passionate and it is difficult to establish a rational dialogue." When individuals participate in political discussions on social media, some of them activate effective and cognitive mechanisms to reflect upon such interactions and is what we broadly define as social media political discussion regret.Specifically, and firstly, social media political discussion regret entails a common response to political discussions that aim to facilitate users' sensemaking of the perceivable role of social media in democratic processes.While many internet users may regret a specific behavior, for example, being more moderate during a discussion, social media political regret fundamentally encapsulates challenging the fact itself: it entails the cognitive and effective elaboration about the paucity of the value of discussing on social media.There are many reasons for regretting a political discussion on social media, but beyond the assumed importance of uncivil interactions (i.e., behavior), most of our respondents clarified that what they regretted the most was not the terms and nature of the conversation, but the conversation itself.This illustrative finding suggests that social media are no longer, if they have ever been, the deliberative spaces to foster pro-democratic values, as reflected in the following responses: "I do not regret the terms of the discussion, but the discussions themselves because generally, they do not have any value" [Sebastián, unemployed mechanic] "Yes, sometimes I regret having discussions about politics because generally it just creates tension and they don't bring me much new, let's say they don't evolve" [Susana, schoolteacher] "I regret it because the political discussion has wasted my time.They don't have value" [Alba, senior architect] Being this the most salient rationale for activating social media political discussion regret, many other testimonies also mentioned some other motivations.Alberto, a lawyer from Sevilla, regretted the vocabulary he used when discussing politics with a friend.Specifically, Alberto has labeled his friend "fascist", something Alberto now believes should not have done, introducing an example of social costs.
What we know thus far about the effects of social media political discussions and the main dynamics of discussing politics in such ecologies revolves around the main, general antecedents and effects, addressing both positive and negative outcomes, while little is known as to the self-effects of social media political discussions, such as political regret e320302 Profesional de la información, 2023, v. 32, n. 3. e-ISSN: 1699-2407 8 "The discussion was about the Parental PIN Law 3 .My friend mentioned that on certain issues parents should have a say over their children.Although this person did not agree with the ideals behind that Law, I labeled him fascist." Similarly, Alejandra regretted a conversation because it had "transpired to real-life" and affected her friendship, as she preferred to refrain from discussing politics with her friend anymore.Some other participants reflected upon other political discussions they ended up regretting.Sergio, for instance, regretted discussing political issues shared by his colleagues on WhatsApp about Vox or Podemos."It is common for some of my contacts to share opinions or proposals of these parties and use them as a throwing weapon, as if one or the other were 'my party,' 'my team,' 'my life'." Finally, some interviewees pointed out how social media political discussions were different from offline discussions.On social media, there are more instances of misunderstandings, which might lead to one of the discussants being particularly offended.As Manuela said, "We are confined and if I discuss with someone in my family, I feel bad, I cannot see him in person and speak and reason in person, it would be all very different.By WhatsApp, the words are misunderstood many times." Other participants emphasized the fact that on social media, people were more likely to misbehave or get more "passionate" or be more "intransigent."For most "regretters," political discussion with strong ties starts after some friend or relative shares some news links or audio-message, usually on WhatsApp, and people getting involved in weak tie discussion starts on Twitter because someone shares a news link or there is already a discussion going on there.
Table 2. Motivations for activating social media political discussion regret

Rationale Explanation Example quote
The discussion itself Lack of value of the political discussions "I do not regret the terms of the discussion, but the discussions themselves because generally they do not have any value" Waste of time "I think I've wasted my time during the discussion" Nothing new to learn "I did not learn nothing new… The uncivil nature of the discussion Incivility of the conversation "The discussions I had where not very civil, I mean, people on social media do not behave as they should, in my opinion" Insulting others "…suddenly, he started to insult me" The fact that the discussion happened on social media Misunderstandings "We are confined and if I discuss with someone in my family, I feel bad, I cannot see him in person and speak and reason in person, it would be all very different.By WhatsApp, the words are misunderstood many times" Social media exacerbate passion among participants "Some people when they engage in discussions on social media become more passionate about their perspectives and sometimes this ends badly"

Effects Explanation Example quote
In-person Social media discussion transpires into real life interactions "We keep discussing [in-person] and get angry to each other during the day" Online Break of a dyad relation in-person and online (i.e., unfriending, blocking, unfollowing) "The great thing about social media is that you can block people, and that is actually what I did" Altogether, users participate in political discussions on social media for many reasons, both with strong and weak ties.Some respondents received positive outcomes from such interactions, while the vast majority fully negated their value.However, when it comes to political discussion regret, all our evidence points to a clear direction: what respondents regretted the most is to have engaged in these discussions themselves, rather than the nature and terms of the conversations, especially with strong ties or people they care.
Despite the complexity of univocally addressing common features between participants who have acknowledged regretting about social media discussions, we found certain commonalities that are interesting to highlight.First the most obvious is that they have all engaged in political discussions on social media in the past and have somewhat strong political convictions, regardless of their political ideology inclination (right or left leaning).In addition, our evidence also suggests that most of these "regretters" are young adults who report heavy social media use, and who typically adopt a passive attitude towards social media discussions, but who are particularly inclined to partake in such interactions when certain levels of intolerance or lies are significantly breached.

Complemental self-effects of social media political discussion
The cognitive elaboration of social media "regretters" also fosters other emotional self-effects that are important to understand the aftermath of political discussions on social media.Despite some of our respondents' statements that such discussions do not typically generate any kind of emotional response in their real lives, and thus do not affect them to a great extent, others acknowledge that other affective emotions are saliently connected to the relationship established between the peers involved (strong vs. weak ties) and the tension generated in the discussion.Indeed, those respondents who strongly regret past political discussions are more prone to neglect future political interactions, fundamentally due to their lack of perceived value, while those less apprehensive do not mention their future intentions or simply say that they will discuss politics if they think they should.
Most self-effects that our respondents agreed upon revolve around "emotional ones," as Sebastián mentioned, and fundamentally entail annoyance, anger, tension, anxiety, and sadness.Accordingly, if the discussion materializes with strong ties, the reported self-effects are typically annoyance, anger, and tension, while if the discussion emerges with weak ties, the reported effects revolve around anxiety and sadness.Here, the interviewees directly connected those effects with being attacked or insulted on social media.
Regarding strong ties emotional self-effects, when José discussed with a friend about the political measures during the lockdown on Facebook he felt "momentarily" angry (i.e., only that day), and was anxious about the radical position of his friend.Likewise, Susana, who supports liberal politics, acknowledged that she discussed political issues with her father, who strongly supports Vox, on social media, and when they got together, they "keep discussing and get angry at each other during the day." Finally, when it comes to weak ties, sadness and anxiety are the most common and complemental responses to social media political discussions.For instance, Sergio said that during "heated discussion, even though they were unknown people," he sometimes felt anxious for being blatantly exposed and attacked.He also provided an example about how usually the conversation evolves on Twitter: "On Twitter, other people can join the thread and talk.Sometimes it happens that other users enter the conversation and start to discuss [a topic] with you, let's say from the side of the person with whom you are discussing, and they come to attack instead of debate.Such situations have caused me a bit of anxiety." Similarly, Raquel, an unemployed waitress from Alicante, got sad when she realized that there were "bad people, who spend most of the times on social media insulting people they do not know".
She went on to introduce one of the most common tactics to curtail uncivil behavior, above all, with weak ties: blocking and unfriending, saying that "The great thing about social media is that you can block people, and that is actually what I did."

Discussion
Political discussions on social media are fairly widespread and yet the examination of individuals' cognitive self-effects is relatively scarce.Although previous studies have examined the main antecedents and political outcomes of engaging in political interactions on social media (Conover;Searing;Crewe, 2002;Rojas, 2015), in this study we put the spotlight on a presumably important issue, but relatively neglected in the literature as an emotional and cognitive factor: regret (Wang et al., 2020).Specifically, drawing upon qualitative data from in-depth interviews with adults from diverse backgrounds and geographical areas of Spain, this study aims to better understand under what circumstances social media users hold political discussions that they end up regretting, providing some insightful empirical findings and one tentative construct explication.
At the empirical level, our findings first capture and describe the main rationale behind the initiation of political discussions on social media.Developing civic skills, learning, enjoying public or private deliberations, curtailing misinformation, and persuading and being persuaded, are the most salient motivations.Prior scholarship has indeed suggested the relevance of these motivations in triggering users' discussions (Taylor; Mantzaris; Garibay, 2018; Wang; Song, 2020) but our study clearly theorizes about their constructive nature: all of them are positive for a healthy democracy.Beyond the display of uncivil behaviour, also very common in these virtual spaces (Goyanes; Borah; Gil de Zúñiga; 2021), attitudes towards political interactions are fairly positive, as most citizens who engage in discussions believe they can enrich civic life, whether by means of providing their point of view or by implementing corrective actions to reduce the circulation of fabricated content.Our findings thus provide a fresh and optimistic breath to the typically apocalyptic discourse around social media, highlighting users' proactive behavior in their attempt to contribute to the deliberative process of democracy.
Similarly, as also suggested by extant research (Goyanes; Demeter, 2020), political discussions on social media emerge from an amalgam of news and political issues trending in a particular moment.These thematic patterns tend to speak to the diverse political exchanges in these ecologies and their connection with news consumption (Hampton;Shin;Lu, 2017).In addition, our findings also suggest that such political discussions are Regret might be viewed as a normal, inevitable, and direct consequence of rationality, which more likely occurs when an individual realizes that a situation is not ideal or that his/her behaviors were inappropriate linked to the established connection between the peers involved, which directly speak to the backstage (in WhatsApp with strong ties) and frontstage (mostly in Twitter with weak ties) of political discussions on social media (Pont-Sorribes; Besalú; Codina, 2020; Treré, 2020).
Our findings illustrate that the affordances of social media enable diverse discussion dynamics which trigger different consequences.For instance, users are more prone to discuss political issues in private environments, especially with people they know or care about, and less inclined to commence political discussions in public forums with strangers, in which civic terms are not fully warranted.Despite this general inclination, the social costs associated to this generalized pattern is paradoxical: while people are more inclined to discuss politics with strong ties in private environments such as WhatsApp (because participants know each other and feel safe), the personal costs in terms of unfriending and flaming are undoubtedly higher than engaging in conversations with the general public, for instance on Twitter.
At a theoretical level, our study fleshes out the concept of social media political discussion regret to examine the aftermath of political exchanges.Specifically, this conceptual contribution aims to examine the nature, nuances, and characteristics of political conversations that users end up regretting.While beyond cognitive elaboration (e.g., Eveland, 2001) extant research has limitedly examined self-effects (Pingree, 2007;Shah, 2016;Valkenburg, 2017), our study seeks to capture under what circumstances users that held political discussions ultimately reflect upon them, causing the feeling of regret.As reported in our findings, users found most political discussions sterile, despite the pro-democratic rationales that motivated their engagement.Accordingly, there was a surprising gap between users' intended contribution to the public (Facebook and Twitter) or private (WhatsApp) sphere and the outcome of such political exchanges: people's prior motivations are promising, yet aftermath outcomes are fairly disappointing.
The concept of 'social media political discussion regret' also hones a strand of public and academic discourse revolving its potential effects.At times, this discourse essentially takes for granted that social media are fully responsible for diverse deleterious effects in the democratic process, as if they were fully and irresponsibly managed by artificial intelligence and biased algorithms.Contrary to these assumptions, which are indeed critical, social media political discussion regret primarily emphasizes the human nature of social media conversations, and the fact that beyond the presumed uncivil nature of many discussions (Goyanes; Borah; Gil de Zúñiga; 2021), what citizens end-up lamenting is not the terms and conditions of the political exchanges, but what is even more worrying, the conversation itself.By focusing on users' cognitions and emotions, our study clearly portrays a widespread rejection of the democratic value of political exchanges and deliberation opportunities on social media ecologies.For many, political discussions on social media become essentially pointless, and that resonates as one of the main motivations as for why users end up regretting.
However, as suggested in our results, many respondents acknowledged that they have never participated in political discussions (N = 8), and thus the theorization on regret is drawn from the rest of participants (N = 34).Accordingly, future research on social media discussion and regret should focus specifically on participants that manifest some level of regret to capture further nuances and motivations of such behavior.In addition, since the interview guide addressed questions that may resonate in participants' minds as normative, their responses may include, in some instances, desirability bias.However, as extant research has largely shown, people have different motivations to discuss politics, and those motivations, in fact, have different effects (Gil de Zúñiga; Valenzuela; Weeks, 2016).Accordingly, our findings should be read in terms of expectations: while many participants seem naïve in their motivations to discuss politics on social media, once they engage in such discussions the outcome is rather disappointing.
Our findings may also point to a double mechanism of political discussions regret, which directly speaks to the irrelevance some discussions trigger.On the one hand, some respondents may not even reflect upon the political discussions they participated in, should they not be directly inquired about their nature.Once such interrogation is raised, participants may connect the absurdity and lack of value of their past discussions in the present time, reporting feelings of regret.This cognitive mechanism could be associated with low levels of regret, thus triggering limited emotional and cognitive effects, and presumably not preventing future political exchanges.A large portion of respondents may fit well in this category, and future studies may focus on how different levels of discussion engagement may intervene to activate different levels of regret.
On the other hand, other respondents may display stronger political regret, which may be directly associated with the cognitive burden political discussions have triggered and the impact they had in participants' everyday lives.Some of these consequences may include, for instance, online and offline unfriending and strong discussions translated to the real world, especially with strong ties.Such a strong cognitive burden may have a subsequent impact on users' behavior, limiting or fully inhibiting future political discussions on social media.Future quantitative studies may be better equipped to dig into these nuances and provide empirical evidence regarding the potential mechanisms.
There was a surprising gap between users' intended contribution to the public (Facebook and Twitter) or private (WhatsApp) sphere and the outcome of such political exchanges: people's prior motivations are promising, yet aftermath outcomes are fairly disappointing In conclusion, the findings presented in this study provide a much-needed nuance over some of the main effects of engaging in political discussions on social media, as well all as important motivations that trigger people's future participation in this type of political discussion.In doing so, social media political discussion regret emerges as a vital variable that future research should consider when casting a more nuanced light on the effects of social media deliberative and discussion processes in democracy.

Table 1 .
Sociodemographic of the sample vs Spanish Census