Political participation in the digital technology era: a symposium on Outside the Bubble: Social Media and Political Participation in Western Democracies By Cristian Vaccari and Augusto Valeriani

Rosa Borge: the contribution of social media to the quality of democratic participation

Outside the Bubble: Social Media and Political Participation in Western Democracies by professors Cristian Vaccari and Augusto Valeriani is a key theoretical and empirical contribution to the study of the role of social media in enabling the dissemination of diverse political content, widening citizens' exposure to politics and promoting political participation. Through a fine-grained empirical analysis comparing nine Western countries (Denmark, France, Germany, Greece, Italy, Poland, Spain, UK, USA) and based on a survey of 16,500 citizens, it questions many of the common beliefs about the role of social media in producing echo chambers and enveloping citizens in filter bubbles. It is also a well-written book, enjoyable, showing coherence between conceptualization, hypotheses, and empirical testing. The authors are honest in their conclusions and the book is entertaining featuring interesting vignettes at the beginning of each chapter.

The authors have a remarkable history of research, first on the uses of ICT in electoral campaigns (Vaccari, 2008) and the participatory and communicative features of political parties' websites (Vaccari, 2013), and more recently on the effects of social media in different aspects of political communication and action (Vaccari et al., 2016; Valeriani and Vaccari, 2018). They stand out among other scholars because of their continuous comparative studies of Western countries, which mainly focus on individual behaviour and attitudes, but always take into account the evolution of technological affordances and the political context. This new book is a peak in their empirical effort to understand the role of evolving digitalization in political communication, participation, and democracy. Moreover, this new push is directed at answering relevant theoretical and normative questions related to the quality of democracy, such as to what extent social media reinforce echo chambers and leave out less politically engaged citizens, and how do social media affect equality and diversity in participation.

I am delighted to have finally read a book that analyses the contribution of social media to the public sphere and their impact on political participation in a nuanced, comprehensive, and comparative manner. In my opinion, one of its main virtues lies in focusing on specific and relevant political outcomes or experiences of citizens' use of social media, that is, levels of political agreement encountered on social media, accidental exposure to political news, and being targeted by electoral mobilization, rather than on general assumptions of their effects that disregard specific impacts on political participation and the diversity among different social groups. Due to such generic explanations, research narratives often fluctuate between pessimism and optimism towards the role of social media in public opinion and democracy, when what is needed is systematic and nuanced empirical research (pp. 218–219). The almost universal adoption of the Internet and the many different ways and purposes for using social media make generic questions such as ‘does social media affect participation’ theoretically unclear and empirically unfruitful (p. 63). I think that in this way, the authors are able to draw useful conclusions that can enlighten and frame future work by other researchers.

The critical book's addition to theory is rooted in the two first chapters following the introduction, where the authors set out their position, hypotheses, and expectations regarding the role of social media in widening the political exposure of the less politically engaged and in strengthening different forms of political participation, and in the conclusion, where the findings are put in relation to the quality and equality in participation that democracy should ensure. In these chapters, the authors unfold a complete review and dialogue with the literature that not only helps to understand their specific choices but can also serve as a wonderful guide for graduate and undergraduate students, other researchers in the fields of communication and political science, journalists, and even politicians and regulators interested in social media.

It is relevant to highlight that this literature review led them to conclude that three types of fallacies have limited previous studies: (1) the affordances-as-destiny fallacy; (2) the one-effect-fits-all fallacy; and (3) the contextual vacuum fallacy. I will structure my review following these useful reflections and the book's empirical results related to them.

The affordances-as-destiny fallacy refers to the assumption that the technical features of social media tend to produce specific political outcomes automatically. The main example is related to the fact that social media users can choose whom they connect with and what to read and comment on. Given this feature, people who are uninterested in politics are expected to continue isolating themselves from political content, while people who are interested and have defined views are expected to avoid disagreeing viewpoints and engage mainly in opinion-reinforcing interactions. However, numerous studies have already shown that these assumptions are doubtful. Thus, the solution given by the authors to overcome this fallacy is to empirically study concrete and theoretically relevant outcomes of social media's affordances and measure them as variables, not as constants. In the book, they focus on three relevant political outcomes or experiences that they then operationalize with survey questions and are considered as both dependent or independent variables in their regression models: (1) levels of respondents' encounters with agreeing or disagreeing political messages; (2) respondents' accidental exposure to political news while on social media for other reasons; and (3) whether they were targeted by electoral mobilization messages to vote for a party or candidate.

In the empirical chapters, the authors check how widespread these political experiences are among social media users, which type of users undergo them more or less frequently (chapter 3), and what relationship there is between these experiences and different forms of political participation aimed at influencing policies, the selection of public officials, and other citizens (chapters 4, 5, and 6). First, in chapter 3, the three political experiences are treated as dependent variables and analysed through logistic regression models to predict the likelihood of encountering them depending on a large number of variables related to sociodemographic characteristics, political attitudes, social media use, and country of residence. Second, in the following chapters, the direct effects of the three political experiences on levels of political participation (dependent variable) are calculated through Poisson regression models that include – as controls – the same large number of variables used before.

Just the simple analysis of the frequencies of the three political experiences on social media shows – as the authors correctly argue – that social media should not be blamed for increasing levels of polarization in Western democracies. Exposure to disagreeing political opinions is more common on social media (28%) than in face-to-face conversations (20%). Most social media users report that they predominantly encounter information and opinions they do not agree with (30.6%) or are balanced content messages (51%). Forty-six per cent of social media users frequently or always encounter unsearched political news on social media when they are going on social media for other purposes. Moreover, platforms matter: frequent use of Facebook and Youtube is positively associated with inadvertent exposure to politics, while the frequency of use of Twitter and Instagram is not. Furthermore, interest in politics, political efficacy, or satisfaction with democracy are not significant predictors of receiving electoral mobilization messages. In addition, younger voters are more likely to be targeted. Frequent usage of Facebook raises the probability of experiencing electoral mobilization much more than the other platforms. Finally, the explanatory analyses reveal that the three political experiences contribute to citizens' repertoires of political participation, being the impact of electoral mobilization the strongest one, the second one the exposure to supportive electoral content and the third one the accidental exposure to political content.

Second, the one-effect-fits-all fallacy implies that social media usage's direct and generalized effect on political participation is similar across different population subgroups. However, the differential effects between and within different subgroups can vary greatly. Previous research has mainly focused on how sociodemographic characteristics influence political participation, but without analysing the relationship between specific political experiences on social media and participation. The authors consider that, because of the extension and pervasiveness of digital technology, the divides related to political attitudes (political interest, ideology, extremist positions) are more persistent. The authors recommend abandoning the belief that only the politically interested citizens access political content on social media and are aware of the various political opportunities that different groups may encounter online (p. 29). The logistic regressions show that citizens who are more likely to come across political information accidentally on social media are not the ‘usual suspects’ but rather the non-interested in politics or political campaigns, young people, women, and the unemployed.

In addition, the increase in the number of participatory repertoires undertaken by the less politically involved (i.e. citizens who describe themselves as less interested and attentive to political campaigns) when affected by the three political experiences on social media is much greater than for the already politically involved. The less involved benefit more from these experiences and thus we can consider that some of the pervasive inequalities in participation related to political motivation could disappear. The concern here is to ascertain if the levelling of the participatory gaps encourages the participation of specific ideological groups. However, this is not the case. Participatory gains produced by the three political experiences on social media are rather evenly distributed across voters of different ideologies and do not benefit the extreme positions or those who vote for populist parties or leaders. Thus, again, political experiences on social media are not contributing to political polarization.

The contextual vacuum fallacy means that in most of the studies systemic and institutional dimensions have been forgotten when analysing the relationship between social media and participation. Moreover, the research has been focused on the Unites States or on single-country case studies. The political impact of social media has been understood as being similar across countries because the same platforms are worldwide and being used in different countries. However, the three political experiences on social media mentioned above can vary depending on country-specific political institutions and media systems that may shape political communication and action. Thus, the authors theorize that three institutional characteristics moderate the relationship between political experiences on social media and repertoires of political participation: the patterns of electoral competition (majoritarian or proportional), whether the political system is party-centric or candidate-centric, and the type of media systems (liberal, polarized pluralist, or democratic corporatist). The hypotheses are: (1) citizens in majoritarian democracies who encounter agreeing viewpoints on social media will increase their participation compared with proportional systems; (2) electoral mobilization on social media will have stronger effects on participation in countries where political organizations are party-centric; and (3) accidental exposure to political news on social media will raise participation in countries with liberal mass media systems. In chapter 6, the authors confirm that the electoral and party systems moderate the relationships between the political experiences on social media and participation, but mass media systems are not affecting these relationships. However, overall, the system-level relationships are weaker than the individual-level factors included in the models. The authors caution us about taking these weak influences as a static phenomenon because social media can alter the impact of those systemic structures.

Other findings confirm American exceptionalism and therefore the authors argue that this country should not be considered a standard. US citizens are more likely to experience one-sided supportive content and be targeted by electoral mobilization messages on social media than the rest of the selected countries. Other countries also differentiate from the rest of the sample: German citizens are the least likely to receive electoral mobilization messages and have the strongest probability of being exposed to messages they agree with.

The theoretical reflections and empirical analyses on political participation are especially valuable because they offer an updated perspective on what political participation means today in a context where the relationship between citizens and political institutions has dramatically changed. Active citizenship has evolved in the past decades into distrust towards institutional politics and more personalized and autonomous forms of political action that are often promoted by event-based political organizations and loose networks of community action. Vaccari and Valeriani are very conscious of the role of digital technologies here: social media platforms have accelerated and enabled these trends in ways that have further expanded modes of political participation that now involve online networking and discursive and political expression behaviours (pp. 60, 61). Finally, the authors define political participation as political actions, whether online or offline, aimed at influencing specific policies, the selection of public officials and other citizens' political preferences and behaviours. The activities selected in their study are contacting a politician to support a cause; signing petitions and subscribing to referenda; financing a party, candidate, or electoral campaign; taking part in public meetings and electoral rallies; distributing leaflets to support a political or social cause; and trying to convince someone to vote for a party or candidate.

Most of these behaviours can take place on digital media or in face-to-face environments. In this regard, the authors adapt the conceptualization of political participation to the new times, where citizens and political actors continuously combine online and in-person political activities. Only attending rallies and leafleting can occur solely in physical spaces. The authors consider that the refusal to include digital behaviours as forms of political participation because they are easier and less costly is too simplistic. Taking political stances or trying to convince others to vote for a party via social media can be very costly with regards to reputation, keeping personal relationships and possible harassment.

The most popular forms in all of the nine democracies included are signing petitions and trying to convince other people to vote for someone, which are undertaken by more than 40% of respondents. The authors suggest that the reason for their popularity lies not only in that they can be undertaken online and offline, but in that they are not so heavily monitored by political organizations and involve acting on citizens' personal terms, a fact that is congruent with the new style of active citizenship explained above. Concerning the effect of the three political experiences on social media, they do not disproportionally encourage participation in easier and digital activities, but they foster hybrid repertoires occurring online and face to face.

Throughout some chapters and in the conclusions the authors recognize several limitations of their work. The authors highlight two of them: the fact that they do not analyse the type, quality, and sources of the content that users are exposed to, and other relevant political experiences on social media that might also influence political participation such as exposure to hate speech, fake news, or disinformation. They call for future research to study these critical aspects. I think that Vaccari and Valeriani's work is deep and wide enough and it is impossible to cover all the relevant factors related to the role of social media on political participation. Precisely, that will be contrary to the successful research design they have unfolded. They focused on three politically relevant outcomes of social media (encountering political agreement, accidental exposure to political content, and being targeted by electoral mobilization) that contribute to citizens' participatory repertoires. And that is a precious piece of advice to us as researchers: when arguing that social media play any (positive or negative) democratic role, please specify, focus, and unravel which type of political experience on social media you are going to study (p. 153).

In my opinion, some more reflection is needed about the limitations of surveys to ascertain whether social media induce echo chambers and filter bubbles. A systematic review of the literature shows that self-reported data and studies focused on content exposure underestimate polarization and hinder the detection of echo chambers on social media (Terren and Borge, 2021). Nevertheless, in the online Appendix, the authors correctly offer several reasons why surveys are better suited than digital trace data to study the relationship between social media and participation, that is, broader coverage of different types of actions and environments being online or offline; data uniformity, standardization, and comparability among countries; the possibility of measuring users' attitudes, motivations, and activities that occur outside the social media and collecting data about the political and social context; detection of not so frequent users; and assessment of relevant listening activities that do not leave visible traces on social media. Thus, the benefits of surveys for analysing the political impact of social media could surmount their limitations.

To conclude, this book is an important achievement in the empirical study of the relationship between social media and political participation in Western countries. Nowadays, political information, expression, and action are being channelled and transformed by social media, making this contribution timely and necessary. Their findings show that social media can contribute to the quality of democratic life through fostering the participation of non-politically involved citizens and young people, and incrementing the participatory repertoires and the volume of participation. As the authors warn, these findings do not fit well the current pessimist climate of opinion around social media, but maybe it is time to take a more balanced view and follow the example of this book.

Shelley Boulianne: the role of echo chambers and incidental news exposure for political participation

This book examines the relationship between social media use and political participation in nine Western democracies (Denmark, France, Germany, Greece, Italy, Poland, Spain, the UK, and the US) from 2015 to 2018. My review describes the contribution of this book to research on social media and political participation, scholarship about comparative politics, and setting a research agenda for the field of political communication over the next few years.

Social media and political participation

In terms of research on social media and political participation, this book is an ambitious undertaking. The book seeks to contribute to a field of research that has expanded exponentially. In 2014, less than 40 papers had been published that used survey data to assess the role of social media use in civic and political participation (Boulianne, 2015). Within two years, the number of studies on this topic had tripled (Boulianne, 2019). With hundreds of studies now published on this topic, it is ambitious to make a distinct contribution and to try to change the course of research on this topic. The book focuses on the role of social media in the creation of echo chambers, incidental news exposure, and electoral recruitment attempts. I particularly like the connection between echo chambers and incidental news exposure, as these two concepts should be understood together. Both activities raise questions about the extent to which we control what we see on social media platforms (user agency). Echo chambers and incidental exposure have been focal points of this subfield for the past 10 years and the book makes a significant contribution to this scholarship. While these topics are well studied, this book contributes a cross-national perspective to these topics and offers a robust way to measure echo chambers.

In particular, the book seeks to shape the discussion of echo chambers. The authors offer an original way to measure echo chambers: one-sided supportive, two-sided, and one-sided oppositional. They find that ‘across the Western democracies we study, most social media users encounter similar amounts of agreeable and disagreeable political content, and it is much more common for users to predominately see political messages they disagree than agree with’ (Vaccari and Valeriani, 2021: 7).

The field has focused on the possibility of echo chambers, but a significant group of people do not see any political information or engage in discussion of public affairs. Vaccari and Valeriani (2021) report that 38.8% of respondents did not post about politics and 24.4% of respondents did not read posts about politics (Figure 1.1, p. 18). We do not know if this differs by platform or by country, which is an unanswered question with major implications. The Digital News Report lists platform adoption by country as well as the use of platforms for news (Newman et al., 2018). There is a large gap between the percentage of people who use a platform and the percentage who consume news on the platform. Understanding this gap is important for understanding variations in experiences of politics on social media.

Political communication has moved towards assessing the role of specific platforms. This book presents four platforms in their analysis, but these findings are sidelined in favour of other theories and findings. In light of current debates about algorithms and affordances, I believe the book makes some important contributions to this discussion (pp. 24–27). I summarize the platform-specific findings below:

• • Facebook use is the only platform that significantly predicts exposure to one-sided oppositional viewpoints (Table 3.1, Model 3, pp. 91–93). The relationship is negative, meaning that greater Facebook use decreases exposure to one-sided oppositional viewpoints. Twitter and YouTube usage increases exposure to one-sided supportive viewpoints (Table 3.1, Model 3, pp. 91–93).

• • Facebook and YouTube usage is positively correlated with accidental exposure to political news on social media (Table 3.2, pp. 105–106), whereas Instagram and Twitter are not correlated with accidental exposure.

• • Use of each of these social media platforms increases exposure to electoral mobilization messages on social media, but Facebook seems to have a slightly larger role than the other platforms (Table 3.3, pp. 114–5).

• • In terms of predicting political participation, Facebook use slightly decreases participation, Twitter does not have a significant relationship with participation, and YouTube and Instagram use increases political participation. However, for each platform, the relationships are weak and just make the P < 0.05 threshold (Table 4.1, pp. 130–131).

Studying specific platforms is a challenge because the popularity of platforms varies by country (Newman et al., 2018). I summarize these adoption patterns in the table below. Studying the same platform in different countries helps to (indirectly) reveal digital affordances. If we see similar outcomes across multiple countries, then we might attribute the outcome to the platform and its affordances. If we do not see similar outcomes, then the results are not easy to interpret; we do not know if the political context is what shapes social media use and its outcomes. The book seems to favour political context as an explanation of differing effects. However, if we think about a country's platform adoption as a key contextual variable, then we might point to Germany's low use of Facebook (52%) (see Table 1).

Table 1. Cross-national differences in social media adoption

Source: Newman et al., 2018.

Comparative politics

The book uses a large sample survey, pooling results from Denmark, France, Germany, Greece, Italy, Poland, Spain, the UK, and the US. I agree with the authors' claim that the distinctiveness of their project lies in its cross-national comparison. A handful of cross-national studies have been conducted in recent years. Dutton et al. completed a seven-country survey (France, Germany, Italy, Poland, Spain, the UK, and the US) in January 2017 and find that the filter bubble argument and echo chambers are overstated ‘as Internet users expose themselves to a variety of opinions and viewpoints online and through a diversity of media’ (Dutton et al., 2017: 5). They report side-by-side statistics on key variables for each country, but do not test for cross-national differences or outline a theory of these differences. Vaccari and Valeriani (2021) test for cross-national differences and offer a theory about why differences may be present.

Gil de Zuniga and colleagues also conducted a cross-national survey (22 countries) on digital media and participation (Strauß et al., 2020). This work is published as a series of articles, which limits the opportunities to elaborate on why any differences could be present. However, in one article, they consider how structural factors such as Internet connectivity, gross domestic product (GDP), press freedom, and literacy rates predict incidental news exposure. The descriptive statistics suggest that the UK stands out as having low scores on incidental news exposure, whereas Brazil has a high score. They did not find that press freedom, GDP, and Internet connectivity explain incidental news exposure, but did find that literacy rate has a significant negative relationship with incidental news exposure. Most of this cross-national work appears in journal articles, allowing maybe 500–1000 words in the literature review to cover why there might be cross-national differences.

Finally, the Digital News Report is important for offering a big picture view of platform use and news exposure on platforms across 46 markets. Using a subset of seven countries (Austria, Denmark, Germany, Norway, Spain, the UK, and the US) from the 2020 Reuters Institute Digital News Report, Fletcher et al. (2021: 21) find that ‘5% of the population are in either a left or right echo chamber in most countries. The one clear exception to this the US, where an average of 10% are in the left echo chamber and 3% are in the right echo chamber’. This work contributes to the US exceptionalism discourse related to digital media effects.

In sum, this other comparative work is presented as a series of articles or reports, rather than offering a comprehensive account in a single monograph. As such, the existing scholarship is not as fully theorized nor as comprehensively assessed as Vaccari and Valeriani's (2021) book. As such, they make an important contribution. The authors argue that political context is important (media system, electoral competition, political organizations) for understanding cross-national differences (Table 2.2, p. 80). They present many interesting ideas about why differences are evident. Some examples of their findings are as follows:

• • In relation to echo chambers, Greek respondents are significantly different from US respondents (Table 3.1, Model 4, pp. 91–93). Greek respondents are more likely to report exposure to one-sided oppositional views, compared to US respondents.

• • Respondents from Greece have higher levels of accidental exposure on social media compared to respondents from the US (Table 3.2, Model 4, pp. 105–6). Respondents from Germany, the UK, and France have much lower levels of accidental exposure on social media compared to US respondents.

• • Electoral mobilization on social media is lower in all countries when compared to the US (Table 3.3, Model 4, pp. 114–115). Germany is the most different from the US in terms of being the least likely to receive electoral recruitment messages (see discussion p. 119).

• • US respondents have higher levels of political participation when compared to respondents from the other countries (Table 4.1, pp. 130–1). The difference between Poland and the US is not significant at the P < 0.05 level.

Overall, I do not see how the factors identified in Table 2.2 (p. 80) explain these patterns. The authors attempt to explain the Greek findings about accidental exposure (pp. 108–109) but point to topics not covered by the theoretical model in Table 2.2. I believe this discussion implies that we need to consider more dynamic factors to explain country differences (referendums as part of the election, protests, a second election), rather than focus on stable institutional factors.

Returning to my summary table (Table 1), another possible explanation of country differences relates to the population's adoption of a platform. Greece is distinctive in relation to low Internet penetration (69%, Table 1). As mentioned, Germany's low use of Facebook (52%) might explain the different results on echo chambers, accidental exposure, and electoral mobilization on social media. German respondents are the least likely to report exposure to electoral mobilization messages and Facebook use increases exposure to these types of messages (Table 3.3, Model 4, pp. 114–115). Platform adoption at both the individual and aggregate levels shapes political experiences on social media.

In Chapter 6, the cross-national comparison is exchanged for a comparison of majoritarian vs proportional systems (Table 6.1, pp. 194–195), type of media system (Table 6.2, pp. 202–203), and type of political systems in terms of candidate-centric, party-centric, or legacy party-centric (Table 6.3, pp. 211–212). This is certainly an original approach that I have not seen in cross-national comparative works related to social media and political participation (Boulianne, 2019). However, because I was primed to think about Germany, Greece, and the US in Chapter 3, I looked for explanations of these exceptional cases and could not find one. As part of setting the agenda for future research, this book suggests that we need to keep looking for contextual variables that might explain patterns of differences. In particular, we might look for why Germany, Greece, and the US are distinctive.

Setting a research agenda

In addition to the suggestions made in the prior sections, the book sets up various other topics for future research. In particular, the book focuses on electoral participation. The measures of political participation are still largely focused on elections and data collected during election campaigns: donating to campaigns, participating in electoral rallies, distributing leaflets, and trying to convince others to vote. The index also includes ‘contacting a politician to support a cause’ and ‘signing petitions and subscribing to a referenda’ (p. 69). However, elections only happen once every 4–5 years and represent unique opportunities for citizens to participate by persuading each other to vote, donating to campaigns, going to campaign rallies, and distributing leaflets. What about the political participation that happens every day of the year, such as signing petitions? In the pooled sample, 46% of respondents had signed a petition (p. 70). In this case, should we expect social media to play a different role in more popular activities and activities that are possible to do on a daily basis instead of every 4–5 years? What about the variety of other online political activities that are enabled through social media (Theocharis, 2015)?

In addition, the book advances scholarship about social media and elections. The authors' findings can be used in a number of ways to launch campaign-specific research. For example, YouTube and Instagram increase citizens' electoral participation, as mentioned. How might political groups and campaigns optimize the use of these platforms for campaign communication? Facebook does not increase electoral participation; therefore, will we see campaigns move away from this platform? In particular, what do the trend lines look like for citizens' participation or for parties using social media to mobilize citizens? Are parties going to invest more heavily in these social media platforms or move to other platforms (TikTok, Twitch, Reddit, WhatsApp, Telegram, etc.)? Will the next set of elections in these countries reveal a different platform to communicate and mobilize citizens? What comes next in digital campaigning and political uses of social media?

In sum, the book offers some conclusive evidence related to incidental news exposure and echo chambers, as well as an agenda for future scholarship.

James Dennis: not all participation is created equal: Outside the Bubble and the impact of social media on marginalized citizens

The publication of Outside the Bubble represents a turning point within the scholarly work exploring the impact of social media on political participation. Perhaps most surprising to readers is its optimism, which is particularly noteworthy given the current popular discourse about the role digital tools play in politics. I write this essay following the recent appointment of Nadine Dorries MP as Culture Secretary of the UK government. In her first interview after taking the post (BBC News, 2021), she reflected on the corrosive impact that social media is having on how citizens encounter politics in everyday life, pointing to the prevalence of filter bubbles – the claim that personalized algorithms show social media users like-minded content separating them from perspectives they disagree with – and how left-wing activists had ‘hijacked’ political debate online. While these comments are politically loaded, they indicate a broader trend in attitudes towards technology companies like Google, Meta, and Twitter. In the aftermath of the controversial results of the 2016 referendum on membership of the European Union in the United Kingdom and the 2016 United States presidential election, social media has been seen to be part of the ‘problem’. As the comments from Dorries illustrate, the parameters of precisely what this problem is can be difficult to define, but scholarly work has provided significant evidence of serious issues, such as the spread of harmful misinformation and disinformation (Freelon and Wells, 2020), evidence of misogynistic and racist communication aimed at elected representatives (Southern and Harmer, 2021), and the lack of transparency in the collection, storage, and use of personal data by political organizations (Dommett, 2020).

In this context, this book's robust comparative analysis of political participation on social media across nine Western democracies is notable, as Vaccari and Valeriani evidence how social platforms expose citizens to relevant political information and provide access to a range of participatory opportunities. Outside the Bubble details the marginal benefits citizens can accrue when using social media in everyday life, including (1) users who see more supportive political content participate more; (2) the more citizens accidentally encounter news, the more they participate; and (3) users targeted by messages to vote for a political party or candidate tend to engage more (Vaccari and Valeriani, 2021: 222). Importantly, they also illustrate that these associations do not benefit populist parties more than supporters of non-populist parties.

But beyond these findings and the comprehensive rebuttal they offer to the filter bubble critique, which they claim is ‘at best exaggerated and at worst unfounded’ (Vaccari and Valeriani, 2021: 7), Vaccari and Valeriani also offer a profound theoretical framework that will influence future research on the relationship between social media and political engagement; how do digital technologies impact different citizens differently? To explain its significance, I will first look backwards, briefly summarising the history of research in this area, before considering how Outside the Bubble can inspire a new research agenda.

The three waves of research on the impact of social media use on political participation

In reflecting on the history of scholarly work on the relationship between social media and political participation, I suggest that three waves of research can be identified.

Firstly, the early adoption of social media in political contexts was characterized by a utopian-dystopian dichotomy, where commentators narrowly focused on the efficacy of these tools in achieving systemic change. Such analysis was typically found in popular discourse, where optimists lauded the revolutionary potential of social platforms to drastically minimize the organizational costs associated with collective action (Shirky, 2008), while pessimists lamented how these technologies could distract citizens from political activity or, worse still, be used by authoritarian regimes as a means of control (Morozov, 2011). This dichotomy was evident in news commentary of several significant events around the turn of the decade, such as the Arab Spring, the election of US President Barack Obama in 2008, and the 2011 riots in England. This reliance on anecdotal, context-dependent case studies to formulate sweeping generalizations on the dynamics of political behaviour failed to comprehensively examine the impact of social media (O'Loughlin, 2011; Wright, 2012).

Secondly, as Boulianne's (2015) extensive meta-analysis shows, a rich body of work has since illustrated the positive relationship between social media use and political participation. This impact is evident across both traditional forms of engagement, most typically associated with elections and the lobbying of elected representatives, and those activities once deemed unconventional, such as demonstrations, political consumerism, and civic participation. Kreiss (2012), Stromer-Galley (2014), and Vaccari (2013) demonstrate how political parties have used social media during elections to involve citizens in campaigning activities, such as amplifying campaign communication, organising rallies, and driving crowdfunding initiatives. While this does not necessarily democratize decision-making within political parties, with elites still making critical strategic decisions, this work demonstrates the significance of citizen participation on social media to a successful campaign. Collective organising has also benefited. Bennett and Segerberg (2013) point to los indignados and Occupy Wall Street to show how social movements can use platforms like Facebook and Twitter to form, grow, and mobilize wide networks without organizational support. Even within more formal advocacy groups, Dennis (2018), Karpf (2012), and Vromen (2017) demonstrate how the feedback loops built into the design of social platforms can provide opportunities for grassroots members to have more influence over decision-making processes. By moving beyond the dichotomy of revolutionary change, this second wave of research suggests that social platforms can make engagement more accessible, providing access to information, networks, and requests for involvement in a time-efficient manner.

But is increasing the efficiency of participatory opportunities necessarily always good for democracy? As Vaccari and Valeriani (2021: 218–219) identify, the aforementioned electoral shocks of 2016 – and the role of social media in facilitating these – represented a ‘tidal shift’ in the focus of academic research and the start of the third wave of research. Described by Chadwick (2019: 8) as the ‘engagement gaze’, the aftermath of these events saw scholars question the assumption that ‘more engagement unproblematically creates more democratic goods for the media system and the polity’; participation is not a universal good. Instead, scholars have questioned who actually benefits from well-intended citizen engagement, with Zuboff (2019) identifying how platforms profit from the commodification of such data, while others have become increasingly concerned with forms of participation that harm democracy. Defined by Quandt (2018: 40) as ‘dark participation’, these acts are ‘characterised by negative, selfish or even deeply sinister contributions’. A vast body of research has since emerged examining the production, consumption, and effects of such communication on democracy (see Freelon and Wells, 2020 for example).

While much of this literature focuses on state actors and political organizations as producers of deceptive content and the role that platforms play in amplifying such messages, the behaviour of citizens has also come under scrutiny. Several studies have documented the abusive messages that elected representatives receive on social media and the devastating personal and professional consequences this can have (McLoughlin and Ward, 2020; Southern and Harmer, 2021). Similarly, when discussing politics with fellow citizens, some users engage in intolerant discourse ‘that threatens, harasses, or silences people or groups’ (Rossini, 2022: 417). News sharing has also been examined, with Chadwick et al. (2018) finding that Twitter users in the UK who share tabloid journalism are more likely to amplify misleading content. Collectively, this third wave has provided rich empirical evidence of how citizen participation on social media can destabilize democratic processes.

However, while scholarly critiques of platform power, toxic online abuse, and propaganda are urgently needed, across much of the popular discourse on how social media impacts politics, broadly conceived, there still remains a desire to speak in generalities. When the parameters of debate focus on the universal effects of digital tools on society, the tone can resemble the utopian/dystopian divide that characterized the first wave. While the depth and sophistication of research has developed leaps and bounds from these anecdotal generalizations, the discussions on this topic in news coverage, in elected chambers, and even occasionally in academic circles can still retain this dichotomous framing. In doing so, I believe we overlook the importance of specificity.

This can come in several forms. Temporal specificity is important. As Karpf (2020) observes in his critique of terms such as the ‘digital age’ or ‘digital era’, used to differentiate new platforms, tools, and practices from the broadcast media era, such framing overlooks the differences that exist between different periods of Internet design and usage. Platform specificity also matters, with the need to recognize the unique affordances between and within social platforms when we reflect on their participatory potential (see ‘The Affordances-as-Destiny Fallacy’, Vaccari and Valeriani, 2021: 24–27). But a significant omission within popular discourse is the awareness of how these tools do not impact all citizens in the same way. In the next section, I suggest that by theorising and empirically supporting this idea of differential effects, Vaccari and Valeriani offer a new pathway for research to understand the impact of social media on marginalized groups.

The differential effects of social media use on participation

The concept of differential effects represents a move away from considering the impact of social media on political participation in generalities. This is what Vaccari and Valeriani (2021: 27–30) describe as the ‘one-size-fits-all’ fallacy, ‘the assumption that the relationship between political interactions on social media and political participation can be expected to be roughly similar, if not the same, across different subgroups within the population’. Instead, the authors seek to understand how the strength of effect varies across different groups and identify which groups benefit and which lose out, as ‘a relationship may be positive for one group and negative for another’ (Vaccari and Valeriani, 2021: 27). This is particularly significant when we consider those marginalized by or disinterested in politics, who are typically not included in research designs that analyse evidence of public-political expression on social media.

By using a measure of political interest to compare citizen groups, Outside the Bubble illustrates how political experiences on social media have stronger, positive relationships with future participatory acts for those users who report low levels of interest. This mobilising potential includes exposure to supportive material on social media and accidental encountering political news, which is particularly significant when considering how younger citizens and those from low-income backgrounds typically rely on algorithmically curated news on these platforms more than average (Kalogeropoulos and Nielsen, 2018). This relationship is partially explained by the fact that these citizens are less likely to be exposed to such content elsewhere, unlike their compatriots with higher interest levels (Vaccari and Valeriani, 2021: 223).

This does not represent the first call to study the differing effects of social media, with a vast body of research on the digital divide and participatory inequalities, including studies showing how class (Schradie, 2018) and gender (Bode, 2017) can negatively shape political experiences online. However, the timing of this intervention is significant. Just as contributions from Chadwick (2019) and Quandt (2018) problematized the notion that all participation is normatively good in the aftermath of the seismic events of 2016, Vaccari and Valeriani (2021: 27–30) push back on the one-size-fits-all fallacy in a time where gender, ethnicity, and age are the driving forces behind urgent social movements, such as #MeToo, #BlackLivesMatter, and Fridays for Future. Therefore, I argue that Outside the Bubble is a necessary call to action for scholars to study how existing theories in political communication impact different social groups in distinct ways.

Understanding the quality of participation for marginalized groups: a new wave of research

In light of the findings of the third wave of research and the ill-effects of some citizen engagement on social media, one might conceivably ask what this contribution means for democracy more broadly. Outside the Bubble is explicit in stating that more participation is not a benefit in and of itself, but that increasing the diversity of those who engage within it can be. In what I suggest represents a fourth wave of research, Vaccari and Valeriani (2021: 225) observe how academic work should turn to the quality of participation for these citizens mobilized on social media, ‘disentangling democratically desirable and undesirable relationships’. To flesh out what this new research agenda might look like, both positive and negative political experiences must be observed. Along with the aforementioned movements that have emerged from different manifestations of inequality, some of the most pressing concerns within the field of political communication, such as the impact of harmful content, media literacy, and radicalization, also disproportionately concern marginalized people in society. I argue that using the theoretical lens of differential effects will provide a more nuanced understanding of how these issues impact dissimilar groups.

Future research should also explore those omitted from the book's sampling frame. This includes those citizens who ‘“cannot” participate because they lack the time, knowledge, and resources that are necessary to bear the individual costs of participating’ (Vaccari and Valeriani, 2021: 223). Consider, for instance, the 14% of respondents who are never accidentally exposed to political news on social media. While defining inequality is notoriously tricky, especially when moving beyond measures of income, the differential effects model is a necessary call to action to understand the impact of social media on those who experience a complex array of disadvantages in everyday life. Morrison (2022) provides a useful starting point for such research by offering a concept of ‘left behindness’, which encompasses dimensions of economic, cultural, political, and geographic disconnection that underpin marginalization in society.

Scholars can also build upon the book by diversifying the repertoire of political acts studied, exploring if the same findings hold true for other forms of engagement. Vaccari and Valeriani (2021: 226) focus on six specific behaviours. While these include a mixture of conventional and unconventional forms of participation, the authors also recognize that this list is not exhaustive. As Theocharis and van Deth (2018) observe, what constitutes political engagement is evolving and includes many more expressive and passive digitally-mediated acts.

In my ongoing research examining the participation of young people from England who experience inequality, I was struck by this; citizens who feel alienated by political parties and legacy news media, who consider themselves to be ‘politically disengaged’ but are actually consumers and producers of rich forms of political communication within networks of private Instagram accounts. They are mobilized by representations of lived experience and contribute to democracy through creative and expressive acts. Outside the Bubble, and the concept of differential effects, has profoundly impacted how I have approached this project and, more significantly, how I understand the relationship between political participation and social media. It has encouraged me to go beyond the general impact of new technologies and instead consider the meaning and significance of these behaviours for the specific circumstances that marginalized groups face. I feel it will do the same for many others.

Cristian Vaccari and Augusto Valeriani: a rejoinder from the authors: the who, the what, the where, and the why of social media and political participation

We worked on the book project that became Outside the Bubble for almost four years. We spent a lot of that time writing and revising drafts, but we also devoted countless hours to passionate discussions on theories, interpretations of our findings, and implications for future directions of research. When we eventually submitted the manuscript, after incorporating very helpful feedback by anonymous reviewers and the series editor, our main hope was that the book would stimulate others in joining such discussions. We are thankful to the IPSR/RISP editors who have provided an exceptionally valuable platform for this discussion, and even more grateful to our colleagues Rosa Borge, Shelley Boulianne, and James Dennis for engaging with our work in such an insightful and thought-provoking way.

In this brief rejoinder we will offer some initial responses to the acute observations of our three reviewers and, on this basis, present what we envision as the most important challenges that research on digital media and participation should address in the future.

Borge, Boulianne, and Dennis all define Outside the Bubble as proposing an optimistic reading of the implications of the diffusion of social media for democratic governance. However, more than bringing in a new wave of optimism in a debate that has been dominated by pessimism in the aftermath of the 2016 shocks, our main goal was actually to get out from what we consider the ritualistic, unproductive, and somewhat artificial juxtaposition of two contrasting worldviews on the relationship between digital media and politics. We sought to contribute to moving this debate forward by abandoning what we term the ‘one-effect-fits-all fallacy’ and focus on the differential relationship between social media and politics. In other words, instead of asking ‘whether’ social media may enhance participation, we asked ‘for whom’ it might do so more and for whom it might do so less, or perhaps not at all. As Rosa Borge points out in her review of our book, generic questions (and generic answers) about political experiences on social media, and on their relationship with participation, are unavoidably misleading in a context where these technologies are almost universally adopted, at least in the Western world, and thus impact very different life experiences, social backgrounds, and worldviews among different groups. Our findings show that different citizens not only experience politics on social media in different ways, but are also affected by these experiences in different ways.

Based on these premises, James Dennis suggests that Outside the Bubble may inaugurate a fourth wave in the study of social media and political participation. The first age, as Dennis reconstructs it, was characterized by techno-determinist prophecies of an imminent revolution in political participation (for good or for bad, depending on the analyst's perspective). The second age developed cumulative evidence that digital platforms can – in general – facilitate political engagement. And the third age, according to Dennis, has suggested and provided evidence for a more cautious approach that, rather than adopting an unconditional ‘engagement gaze’ (Chadwick, 2020) insensitive to the different, and not always democracy-enhancing, goals and outcomes of participation, more clearly recognizes that some forms of participation feature a democratic ‘dark side’ (Fiorina, 1999; Quandt, 2018).

It is clearly too soon to tell whether Outside the Bubble will contribute to an emerging fourth wave of theory and research on social media and participation. One might even argue that our conclusion that social media, by closing the participatory gap between citizens with high and low levels of political engagement, contribute to democratic equality, still adopts the ‘engagement gaze’ of the third wave. At the same time, we hope that, by combining some of the key points we make in the conclusions of our book with some of the insightful comments of our reviewers, we will be able to highlight some of the crucial issues that the fourth wave of research envisioned by Dennis might address. These issues are related to the Who, the What, the Where, and the Why of social media and political participation.

The who (citizens)

In our book we took a first step towards assessing whether and how political experiences on social media may affect different citizens in different ways. We hope that future research starts from the premise that there are substantial differences between those who experience political content on social media (and our book shows that robust percentages of social media users do) and those who do not (and we show that these users also do exist). Investigating the drivers of this gap would help more robustly test our theory that social media may reduce democratic inequalities. If, as we contend in our book, political experiences on social media spur participation, then it becomes even more crucial to study those who are at risk of being ‘left behind’ (to use Morrison's (2022) term as proposed by Dennis), as they are relatively isolated from politics online, and to understand what the democratic consequences of these inequalities are.

For this approach to be productive, however, it must be as granular and nuanced as possible, avoiding black-and-white simplifications and straw-man arguments. Ideally, studies should account for a combination of different factors, such as platforms' affordances, national contextual specificities (as Boulianne suggests in her review), structural forces that explain traditional inequalities (as Dennis proposes), but also individual choices and attitudes. From this standpoint, it may be productive to differentiate between those who are actually ‘left behind’, that is, who would not be opposed to engaging with politics online to some degree but end up not being exposed to political content and opportunities due to factors beyond their control, and those who deliberately avoid political experiences on social media – whom we might say are ‘leaving behind’ politics online.

Even if we specifically consider deliberate avoidance of news and politics online, recent studies suggest approaches sensitive to differential effects might be the most appropriate. First of all, avoidance might not necessarily mean lack of exposure, but also an unconscious or deliberate lack of attention, that is, ‘information skipping’ (Bode et al., 2017). Other studies differentiate between simple ‘scanning’ and substantial ‘processing’ of unsearched political news (Matthes et al., 2020). In turn, the different reasons behind avoidance could be associated to different outcomes that might result in conflicting contributions toward democratic quality and equality. For example, avoidance could be an adaptive response to users' perception of overload (Schmitt et al., 2018), or a protective strategy when news is overly negative and ultimately depressing (de Bruin et al., 2021), or the reaction to an increased lack of trust towards an environment that one perceives to be ‘poisoned’ by false or exaggerated information and inauthentic or malicious actors. Finally, in contemporary environments where users' experience is saturated by multiple social media, individual and collective decisions of temporarily or permanently disconnecting from specific platforms are increasingly frequent and may be based on multiple reasons – from ‘digital detox’ to political resistance to ‘platform capitalism’ (Kaun and Treré, 2020).

In Outside the Bubble, we argue that political experiences on social media do not make the rich richer but reduce gaps between the already engaged ‘usual suspects’ and traditionally less active citizens. Broadening the scope of our models to include the social media left-behind and those who deliberately opt out of digital politics would enhance our understanding of how social media contribute to the health of contemporary democracies.

The what (modes)

Borge, Boulianne, and Dennis all point out that future studies of social media and political participation should take into account the ongoing expansion of political participation repertoires, not least thanks to the opportunities provided by digital platforms. According to our reviewers, Outside the Bubble only partially addresses these issues. As Boulianne highlights, in developing our participation index we mainly focus on forms of participation that occur in and around elections, and only in a very limited way (i.e. in considering signing petitions) on modes of political action that ‘happens every day of the year’. Dennis underlines how some of the citizens who feel politically disengaged and have lost trust in political elites ‘are actually consumers and producers of rich forms of political communication’ on digital platforms. Then there are the increasingly diverse repertoires and opportunities for collective and connective action in the realm of contentious politics (Bennett and Segerberg, 2013) that we deliberately left out from our research to avoid overstretching our concepts, theories, and empirical tools. We agree with our reviewers that testing hypotheses similar to ours by focusing on different realms and modalities of participation will be relevant for future research. Defining what can be considered as political participation in a given context represents a never-ending challenge for scholars, and one that digital media have further complicated.

Some scholars have already helped solve these puzzles with insightful theoretical and conceptual proposals that were of great inspiration to our work. We are indebted to the concept of ‘digitally networked participation’ proposed by Yannis Theocharis (2015; see also Theocharis and Van Deth, 2018), which provides a convincing path to establishing a dialogue between classical theories on participation (e.g. Verba and Nie 1972; Verba et al., 1995) and later theorizations of digital platforms as spaces for political action (e.g. Papacharissi, 2015).

As we argue in Outside the Bubble, when scholars theoretically and empirically consider how social media are transforming the landscape of political participation, they must resist two opposite temptations: considering anything that happens on social media as lacking the basic properties of participation or, at the opposite end, counting as participation any form of political communication that involves citizens. In answering our reviewers' call to address broader dimensions of contemporary participation than the ones we consider, future research should avoid the trap of being too inclusive, which would risk making it ‘nearly impossible to distinguish between what we aim to explain (participation) and what we believe can at least in part explain it’ (Vaccari and Valeriani, 2021: 52). We hope that our conceptual approach will help avoid this risk. We distinguish between experiences that involve exposure to political content but do not imply any action; citizenship, which includes active political behaviours ostensibly lacking the intention to influence others' actions and/or attitudes; and participation, which comprises actions deliberately aimed at affecting politically relevant outcomes. We hope future research can build on this conceptual map to shed further light on both what participation is and what factors promote it.

The where (platforms)

Both Borge and Boulianne highlight the importance of investigating how political experiences may differ, in frequency and type, across different social networking platforms and what the implications of these differences might be for political participation. We do agree that one of the most intriguing challenges for present and future research is to move away from an approach, such as ours, in which we refer to (and study) ‘social media’ as a whole. Some scholars have already conducted relevant research in this mould, investigating political experiences (Barnidge, 2020), political campaigning (e.g. Bossetta, 2018), and implications for means of political participation (Theocharis et al., 2022) across different platforms. However, more work is needed in this direction, not least because the social media landscape is constantly evolving.

Admittedly, Outside the Bubble only partially addresses these issues. We do include frequency of use of four different platforms (Facebook, Twitter, Instagram, and Youtube) as control variables in all our models and, when relevant, we comment on their association with the political experiences we consider. However, we do not discuss the association between frequency of use of individual platforms and political participation. Moreover, as Boulianne suggests, we could go one step further and consider the levels of diffusion of these platforms in different countries as another ‘key contextual variable’ that may help explain cross-country variation. Investigating the relationship between platforms' affordances and national specificities may represent a promising path for future comparative research.

More broadly, research should continue interrogating the specificities characterizing the digital architectures of different platforms and their democratic implications. To this end, we hope that the theoretical approach to social media affordances that we developed in Outside the Bubble can be a useful starting point. Indeed, among the three fallacies that we invite researchers of social media and participation to resist, we enlist what we call ‘the affordances-as-destiny fallacy’, that is, the idea that the technical functioning of social media tends to automatically produce specific outcomes. To overcome this simplistic approach, we propose to treat the outcomes of platform affordances ‘as empirical questions rather than as fixed assumptions, and thus to measure those outcomes as variables rather than assume them as constants’ (Vaccari and Valeriani, 2021: 27). Such vision is not based on the idea that platforms' architecture does not matter, but on the belief that affordances are multiple and evolving (and this applies both between and within platforms), and that different users engage with affordances in different ways and in pursuit of different goals.

Let us think for example about a platform such as TikTok, which had not even been released when we started collecting our data but has become increasingly relevant in political communication across the world (see the episode of the US 2020 presidential campaign known as the ‘Tulsa prank’ that we discuss in the opening of Chapter 2). Patterns of TikTok usage, as well as the algorithmic design of the platform, are different from those of the social media we consider in Outside the Bubble, at least at the time we studied them (Bandy and Diakopoulos, 2020). The ‘endless scrolling’ mode facilitated by TikTok's affordances (Kaye et al., 2021) is combined with the aggressive algorithmic organization of the ‘For you’ section, where content is suggested based on multiple factors. In such a heavily curated setting, exposure to content becomes incidental by default, which might increase the likelihood of unsearched exposure to politics, but might also result in a different relationship with participation, depending on how individuals react to this incidental exposure (see the discussion in the ‘Who’ section above). Moreover, specific practices connected to platforms' uses must be considered. In the case of TikTok, Zulli and Zulli (2022: 1882) proposed the concept of ‘imitation publics’ to describe ‘a collection of people whose digital connectivity is constituted through the shared ritual of content imitation and replication’. When chains of imitative actions become the defining way of (actively) inhabiting a platform, researchers need to consider the implications for the relationship between social media experiences and (political) participation, as well as for what we should count as participation (see the discussion in the ‘What’ section above).

We offered this brief discussion of TikTok mainly to highlight how, for future research to effectively illuminate the role of specific platforms, scholars may want to heed our advice to avoid the ‘affordances-as-destiny fallacy’ that, in our view, has characterized much early theorizing on the political impact of social media, particularly in debates around ‘echo chambers’ and ‘filter bubbles’. For example, TikTok's success has prompted other platforms to emulate some of its features or redefine their own functionalities to better distinguish themselves, effectively changing the whole platform ecosystem (Markov, 2022). Adequately accounting for this complexity requires scholars to consider not only the particular ways of functioning and uses of specific platforms, but also to cope with the volatility of patterns (and thus the temporal validity of findings; see Munger, 2019) as well as the reciprocal influences and relationships between platforms.

The why (normative implications)

The most important challenge that we hope the next wave of research on social media and participation will address requires defining the values that should guide our research. In other words, if the final question we should answer is: ‘Do social media hurt or help democracy?’, then a broader discussion on what a healthy democracy is and what it needs is necessary and urgent.

Such goal, which we obviously do not aim to achieve here, is strictly intertwined with the empirical question of why people participate. Or, more precisely, based on what factors and in pursuit of which goals do citizens decide to participate? As Dennis contends in his review, the idea of ‘problematizing the notion that all participation is normatively good’ is increasingly central to contemporary debates. Shedding light on these issues is crucial to make sense of the problematic aspects of the relationship between social media and political engagement that have emerged more prominently in the second half of the 2010s and that earlier research was less attentive to. Let us think for example about the coordinated (and un-coordinated) circulation of problematic information (Giglietto et al., 2020), or the high levels of incivility (Kim et al., 2021) and intolerance (Rossini, 2022) characterizing political communication of some élite and non-élite actors on social media. And let us think about whether, and how, un-democratic content circulating on social media may eventually support un-democratic actions, such as the Capitol Hill riots in the United States in January 2021 (Gounari, 2021).

As we argue in the conclusions of Outside the Bubble, considering the effects on participation of normatively dysfunctional experiences such as exposure to disinformation, uncivil political talk, or hate speech must be central in the future agenda of political communication research. To this end, we hope that scholars develop more granular measures of the experiences we considered, for example by differentiating between prevalent exposure to supportive information with low or high levels of factual accuracy, or exposure to mobilization messages that feature or do not feature incivility.

At the same time, the main finding of Outside the Bubble is that social media can contribute to increasing democratic equality by exposing substantial numbers of citizens who are less politically engaged to meaningful political experiences, and that the association between these experiences and participation is stronger for these less involved citizens than for those who are highly involved. In this sense, we agree with Dennis when he invites researchers to consider social media's implications for participatory gaps across different dividing lines than those we focus on. Gender, ethnicity, class, and age are surely important dimensions to be investigated, not least in light of recent manifestations of connective action around race, sexual harassment, income inequality, and climate change, among others.

Understanding how political communication contributes to these movements, and to the interplay between the broader forces pushing for and against change in our societies, requires normative theories and research designs that go beyond the procedural approaches to democratic governance that informed the foundation of political communication as a field of research. The field of political communication is increasingly, albeit still slowly, reckoning with its increasingly shaky, outdated, and contested normative foundations. As we write in the concluding pages of Outside the Bubble, social science research must always strive to contemplate both what is and what should be, as neither makes sense without the other. This is perhaps the most important lesson we learned during the nearly ten years during which we researched social media and political participation.