Social media,porous boundaries,and the development of online political engagement among young citizens

Abstract

Concepts such as “porous boundaries” and “low thresholds” appear frequently in the literature on online political engagement. Social media, it is argued, are characterized by less distinct boundaries between non-political and political activities, thereby lowering thresholds into political engagement. This argument is analyzed and empirically tested. Relying on a five-wave panel study among Swedish adolescents, we provide unique insights into the levels and development of political engagement in online political information, interaction, production, and collective action. In sum, the findings show that, for a majority, social interaction in social media coincides with engagement in political information and interaction, while few are engaged in production and collective action. Second, the study provides limited support to the idea that low thresholds in social media promote patterns of tune-in, tune-out political engagement over time. Finally, social interaction in social media has clear effects on online political engagement beyond political socialization and motivation factors.

Keywords

Adolescents collective action longitudinal panel study online media political engagement porous boundaries social media young citizens

Introduction

Online media have changed the conditions for political engagement. Digital networks and social media add new dynamics to social movements and activism (Bennett and Segerberg, 2012; Dencik and Leistert, 2015), and the opportunities for citizens to express opinions in public and take part in collective actions have certainly expanded. But what distinguish these changing contexts and how do they affect patterns of citizen’s political engagement? In this study, we critically examine a key argument in the literature: that social media are characterized by porous and less distinct boundaries between private and public spaces, between modes of engagement, between non-political and political activities, thus lowering the thresholds assumed to shape conventional forms of political engagement (Bimber et al., 2012; Brundidge, 2010a; Carlisle and Patton, 2013; Xenos et al., 2014). For those frequently engaged in social media, a step into the political is assumed to be easy and incidental, rather than associated with critical decisions. As a consequence, social media have the potential to bridge gaps between the politically engaged and the disengaged—inequalities typically regarded as a democratic challenge (Norris, 2001; Schlozman et al., 2010). If this holds true, it seems to imply a radical transformation, given that online media no longer constitute alternative but central environments for citizens’ engagement in politics.

The idea of porous boundaries and lowered thresholds is frequently referred to in research on the impact of online media use on various aspects of political engagement (Brundidge, 2010b; Carlisle and Patton, 2013; Xenos et al., 2014). Extensive research has also explored the vigorous forms of online activism, assumed to attract not least younger citizens (Dencik and Leistert, 2015; Velasquez and LaRose, 2015). However, the implications of these transformations for patterns of political engagement among young people have rarely been empirically investigated. This is the main purpose of this study. More specifically, we use a longitudinal five-wave panel survey to analyze three possible implications of porous boundaries. While the first two implications relate to frequencies and patterns of online political engagement, the third focuses on the mechanisms assumed to affect individuals’ engagement.

The first, and most radical implication, suggests that among large groups of young people, social interaction in social media coincides with engagement in politics. To examine this in more detail, we analyze different modes of online political engagement. While some authors refer to porous boundaries in relation to online discussions, a more critical argument relates to involvement in collective political actions (Bimber et al., 2012; Earl and Kimport, 2011). Second, an alternative implication suggests that porous boundaries promote a significant level of occasional tune-in, tune-out engagement in politics. Here, we take advantage of a unique longitudinal research design to analyze how citizens move in and out of various modes of political engagement over time. Third, the idea of porous boundaries also implies that political engagement is partly independent of individuals’ political interest and resources. Just by being active in social media, individuals may be occasionally exposed to politics and drawn into easily accessible political activities. We critically examine this idea by studying the unique positive effects of social media use on different modes of online political engagement.

Boundaries, thresholds, and political engagement in online media

The expansion of online political activities has raised a number of questions regarding the distinctiveness of these activities. As Theocharis (2015) notes, online activities have been defined as both “equivalents of offline political acts” and “an entirely new way of participating in politics” (p. 2). Extensive research not only conceptualizes online and offline political engagement as distinct activities but also shows that these are related in young people’s orientations toward politics (Ekström et al., 2014; Kim et al., 2016; Kruikemeier et al., 2014). The idea explored in this study suggests that online media are distinctive with respect to the boundaries that shape citizens’ involvement in politics. Political activities are structured by boundaries (social, organizational, discursive) that enable and restrict participation. Taking a step into politics involves boundary-crossings that are more or less demanding, moving from private to public spaces, accepting membership of and commitments to organizations, joining new social groups or networks, or just shifting to a political topic in a conversation—a boundary-crossing considered as potentially risky (Ekström, 2016). The higher the thresholds, the stronger are the personal motivations required to become politically engaged.

The concept of porous boundaries has been used to explain how online environments increase exposure to political information and involvement in discussions on public matters—as well as involvement in various political initiatives (Brundidge, 2010b; Kim et al., 2013). The networks formed in social media are supposed to create a constant flow of information with opportunities to share news from diverse perspectives. The literature provides related arguments regarding the lower costs of accessing political information (Carlisle and Patton, 2013: 884; Vissers and Stolle, 2014: 940); the increased chances for individuals to come into contact with people engaged in political activities (Shah et al., 2001); and the easily available online tools for political discussion, public production, and joining of collective protests (Ekström and Östman, 2015; Boulianne, 2009; Brundidge, 2010b; Jenkins, 2006).

It is, however, in relation to collective actions that the idea of porous boundaries has the most far-reaching implications. Bimber et al. (2012) describe the weakening of boundaries between private and public spaces as fundamental in understanding the changing forms of collective actions. Communication technologies provide an infrastructure for “entrepreneurialism on the part of the individuals” and low-cost contributions without formal memberships. Public engagement emerges from private domains without deliberate decisions: “When people have a great capacity to act freely and without constraints by larger social structures or formalized rules of belonging, their negotiation of a boundary typically involves less intentionality and calculation” (Bimber et al., 2012: 62). Similar ideas are also discussed in research on protests and activism. Bennett and Segerberg (2012) suggest the concept of “connective actions” to characterize engagements based on “self-motivated (although not necessarily self-centered) sharing of already internalized or personalized ideas, plans, images and resources with networks of others” (p. 753). Social media enable individualized, flexible, and loose forms of connection to collective movements.

In addition, Earl and Kimport (2011) emphasize the decreasing costs of organizing and participating in online petitions, boycotts, and campaigns, although the investments in time and resources vary between participants and forms of activism. Frequent instances of ephemeral (“five minute”) activism, forwarded in expanding peer networks, are supposed to enable easy transitions from private opinions to public actions and thus increase the levels of citizen participation (Earl and Kimport, 2011: 72–76). Earl (2014) argues that low-cost activism is potentially important as such. It is well known, she notes, that “most people don’t act, and more people act when costs are lower than act when costs are higher” (Earl, 2014: 174). These analyses, thus, reevaluate activities that have been demonized and labeled as “clicktivism” and “slacktivism” (Halupka, 2014).

In sum, a review of the literature suggests that the porous boundaries are related to three dimensions of the transforming contexts of political engagement. First, the technological affordances of online media are assumed to provide an infrastructure for social networking, sharing, and publishing—widely accessible although certain skills are required. As Bimber et al. (2012: 64) note, the technologies provide boundary-crossing mechanisms equivalent to bridges in physical boundary-crossing. Social network sites include a diversity of tools designed for easy expansion of networks and shifts between activities. Second, online media are described as contexts of substantial convergence of spaces and activities. What characterizes these contexts, Papacharissi (2010: 76, 144) argues, is the convergence of self-centered, social, commercial, and political spaces. Activities separated in other contexts are integrated on social media platforms. The relations between private and public activities, non-political and political activities, are supposed to be fluid rather than distinct. Third, online media contribute to broader reorganizations of the institutional contexts for political engagement, most evident in the shifts toward more self-organized activities taking place in domestic settings and through mobile media. These trends reflect broader changes in political culture, toward increasingly personalized, cause-oriented, and ad hoc forms of engagement (Bennett, 2012; Dalton, 2009). Thus, online political engagement can be easily integrated into the rhythm of everyday life, at preferred times and places as they typically require limited commitments.

Implications and hypotheses

Extensive research thus provides strong arguments to conclude that online media provide new conditions for citizens’ engagement in politics—making it easier and less demanding to become involved. The ideas of low thresholds and porous boundaries have been related to a mix of arguments regarding low costs (Carlisle and Patton, 2013; Earl and Kimport, 2011; Vissers and Stolle, 2014), easy access and exposure (Boulianne, 2009; Brundidge, 2010b), loose forms of engagement (Bennett and Segerberg, 2012), and engagement requiring less motivation and commitments (Bimber et al., 2012). However, the assumed implications for patterns of online political engagement are rarely specified or empirically investigated. We suggest three distinct implications. Before doing so, however, a conceptualization of political engagement is required.

Online media have been characterized as contexts for incidental exposure to information, fluid transitions from private to political matters in discussions and sharing, easy accessible facilities for creative production and publishing, and self-organized and low-cost forms of collective actions in online protests. Inspired by previous research, we therefore distinguish between four modes of political engagement: Political information (reading and searching for political information); political interaction (posting comments and discussing public matters, sharing, and linking news and clips on politics); public production (writing about politics on a blog or creating video clips); collective actions (initiating and joining protests and activism). Two aspects of this conceptualization are worth emphasizing. First, modes of political engagement refers to the ways in which citizens can be involved and connected to politics and public concerns (Ekström et al., 2014; Couldry et al., 2007) and is thus a broader concept than political participation (Hooghe, 2014; Theocharis, 2015; Van Deth, 2014). Among the modes of engagement we are studying, collective action is the one that most clearly meets common criteria of political participation. Second, “modes” refers to categories of concrete activities that can be related in various ways. Interactive activities such as sharing and posting can, for instance, be a way of engaging in political activism. In our conceptualization, interaction, however, primarily refers to interaction among friends and in networks not explicitly embedded in political activism. Furthermore, the modes reflect general activities (to keep informed, to interact and comment, to produce, and to join and become member) not exclusive to, but developed within, online media.

The literature suggests that the idea of low thresholds applies generally to different modes of engagement. There are, however, reasons to assume that such thresholds vary. Engagement differs with respect to the efforts required from individuals and the extent to which an activity entails public disclosure of opinions or political affiliations. Reading and posting comments are less demanding than producing and publishing content. Reading also differs by being a private rather than a public activity. Among the public activities, commenting and discussing are likely associated with weaker commitments compared to engagement in collective protests. On the other hand, it is precisely those distinctions related to efforts, costs, and commitments that are assumed to be reduced in online media. Online collective actions are, for instance, widely recognized as promising forms of participation in contemporary democracy, partly because they have the potential to involve citizens in low-cost, easily accessible, and ephemeral protests. To what extent the various modes of engagement are marked by different thresholds into politics is thus an empirical question. Based on this conceptualization, we critically examine three possible empirical implications of the general idea of porous boundaries and low thresholds.

The first implication suggests that most young people who are active in social media are also engaged in politics. This is a far-reaching implication, given the fact that disengagement from politics has been described as typical among young people. However, if the boundaries are really porous, almost liquidized, we can expect that social engagement and political engagement to a large extent coincide in online media. By also taking the expected differences between modes of political engagement into account, two hypotheses are addressed—the first being more radical than the second:

H1. A majority of young people using social media for social interaction is engaged in online politics, and this applies to all four modes of political engagement.

H2. A majority of young people using social media for social interaction is engaged in online politics, but this applies primarily to information and interaction, that is, activities assumed to be less demanding than production and collective action.

The second—alternative—implication suggests that porous boundaries promote a tune-in, tune-out form of online political engagement. This implication is in contrast to the first. Rather than a typical pattern of permanent involvement in politics, citizens move in and out of political engagement over time:

H3. A tune-in, tune-out form of online political engagement over time constitutes the most common pattern of online engagement, compared to those being either permanently engaged or disengaged.

Finally, the idea of porous boundaries suggests that involvement in online political activities comes with the very attendance and engagement in social media, independently of political interest and socialization factors. This is assumed to apply to different modes of engagement (Bimber et al., 2012: 62; Brundidge, 2010b: 684). We test this idea by addressing the following hypothesis:

H4. Social interaction in social media has a unique positive effect on online political engagement independent of political interest and socialization factors.

Methods and data

This study uses data from five annual waves of a longitudinal panel survey. Two cohorts are included: junior high school students, aged 13–14 years at the first data collection and surveyed over 5 years/waves; senior high school students, aged 16–17 years at the start and surveyed in four waves. The study is carried out in a region in Sweden with a population of approximately 275,000 people, which is close to the national average with respect to relevant variables such as unemployment rate, income level, population density, and political leanings. The target sample includes all students in the selected age groups at 13 of 26 schools in the region, and therefore approximately 50% of the students in the population. The schools were carefully selected to include theoretical and vocational programs: schools in the city center, suburbs, and rural areas and schools in neighborhoods with differences in the distribution of social, economic, and ethnic background. Within the selected schools, all students in the age groups were invited to participate. Data collections were conducted in the schools at scheduled school hours and was administered by trained research assistants who distributed the questionnaires after having informed about the study, the confidentiality of individuals’ responses, and that participation is voluntary. The school-based survey was supplemented by a postal survey (also available as a web questionnaire) for participants who had left primary and secondary school, which applies mainly to the final wave of data collection for cohort two. Full information on sample size and participation rates covering both cohorts across panel waves is provided in Table 1.

Table 1.

Sample and participation rate for each panel wave.

Year	Cohort 1 (13- to 14-year-olds)		Cohort 2 (16- to 17-year-olds)
Year	Sample	Participation rate	Sample	Participation rate
2010	960	94% (N = 904)	1052	85% (N = 892)
2011	987	89% (N = 883)	996	81% (N = 807)
2012	954	88% (N = 843)	914	81% (N = 740)
2013	934	77% (N = 723)	–	–
2014	935	74% (N = 695)	939	50% (N = 467)

The main reason for the region-based design is thus to make a large-scale school-based panel study achievable. The strength of the study is the exceptionally high response rates across the five waves of data collection. The on-site administration of the survey in the class rooms also contributes significantly to the quality of the survey. The non-random sample is a limitation and could raise concerns about generalizability. However, considering the response rates and the fact that the target sample reflects the national average on key sociodemographic variables, we believe the results reported have at least as high generalizability as a random national sampling would provide.

Measures

To analyze adolescents’ online political engagement, we rely on an extensive battery of survey questions asked at each wave of the panel. The online activities reflect the five dimensions discussed above: social interaction, political information, political interaction, public production, and collective action.

Social interaction was measured using three items following the overall question “How often do you use the Internet/computer to do the following?” The specific items were as follows: (1) Talking to friends using MSN, Myspace, or similar; (2) Keeping in touch and staying informed with my friends through Facebook, or similar; and (3) Sharing information about myself on Facebook, an online picture diary, or similar. The response scale for these items ranged from 1 (Daily) to 5 (Never)—and the additive index showed high reliability (Cronbach’s α = .77, wave 1).

Our measures of online political activities are based on a large battery of items related to the following overall question: “Have you done any of the following during the past 2 months?” The first dimension of online political engagement, political information, was measured using four items: (1) Visited a political website; (2) Searched for information about politics or societal issues on the Internet; (3) Read about politics on a blog; and (4) Watched videos or clips about societal issues or politics (Cronbach’s α = .81, wave 1). Political interaction was measured using the following four items: (1) Linked news to friends; (2) Discussed societal or political issues with friends online; (3) Linked video clips with political content; and (4) Shared music with an important political or societal message (Cronbach’s α = .70, wave 1). Public production was measured by three items: (1) Written about politics or societal issues on my own blog or website; (2) Produced/uploaded video clips with an important message on YouTube, or similar; and (3) Contributed with texts, pictures, or information to an online newspaper (Cronbach’s α = .65, wave 1). Finally, collective action was measured based on the following four items: (1) Organized an online protest or boycott; (2) Participated in an online protest; (3) Joined a Facebook group, or similar, devoted to societal issues; and (4) Started a Facebook group, or similar, devoted to a societal issue I think is important (Cronbach’s α = .73, wave 1). The response scale for these items ranged from 1 (Yes, several times) to 3 (No).

In addition to these variables of primary interest, our analyses will make use of certain control variables in order to remove the influence of key socialization contexts (family and friends) and motivation factors. These include political interest, which was measured based on two items: (1) “How interested are you in politics?” as well as (2) “How interested are you in what is happening in society?”, with response categories ranging from 1 (Not at all interested) to 5 (Very interested). These items were recoded to range between 0 and 1 before being averaged into a single political interest scale (Pearson’s r = .54, at wave 1). Political interest among parents and friends is recognized as important socialization factors and these concepts were measured separately for parents and friends using a set of agree–disagree statements, such as the following: (1) My parents (friends) are interested in what happens in society, and (2) My parents (friends) follow the news, with response categories ranging from 1 (Fully agree) to 4 (Fully disagree). The scales were inverted and recoded to range between 0 and 1 before being added into a parent (wave 1 Cronbach’s α = .65, M = 0.69, standard deviation [SD] = 0.18) and peer (wave 1 Pearson’s r = .55, M = 0.48, SD = 0.22) political interest index. Finally, family socioeconomic status is based on five items: (1) If you want things that cost a lot of money (e.g. a computer, skateboard, cell phone), can your parents afford to buy them if you want them? (2) If you compare with others in your class, do you have more or less money to buy things? (3) Does your family have more or less money than other families where you live? (4) How are finances in your family? (5) How often do you and your family go for vacations? Responses to these four- and five-level items were recoded to range between 0 and 1 and summed to form a single family socioeconomic status index (Cronbach’s α = .76, wave 1).

Data analysis

To a large extent, our analyses will be based on descriptive data presentations that visualize the levels, the combinations, and temporal stability of adolescents’ online social and political activities. In these descriptive presentations, we have opted for a generous operational definition of online political “engagement”: adolescents who either report (1) being engaged in one of the activities reflecting each mode of engagement several times in the past 2 months or report (2) being engaged in two such activities at least once in the past 2 months. Thereby, our ambition was to strike a balance between having a very generous criterion regarding the frequency of engagement (at least once in the past 2 months) and not being too sensitive to single items.

Apart from presenting descriptive statistics, we also use advanced panel regression models to provide a strict test of the third hypothesized implication (H4) of the porous boundaries and low threshold argument. More specifically, we use two-way fixed-effects panel models based on all five waves of the panel survey. These models control for unobserved heterogeneity between units (unmeasured differences between adolescents) by relying on within-person variation over time only (Allison, 2009; Finkel, 2008). All stable (time-invariant) characteristics are thereby accounted for in the estimation procedure, focusing instead on how within-person changes in the independent variables are related to within-person changes in the dependent variable. In the present case, we are interested in whether increases in online social interactivity are related to increases in online political engagement—beyond what is accounted for by classic socialization (family and friends) and political motivation factors—as suggested by the porous boundaries argument. Since the fixed-effects models only account for omitted stable characteristics, we will control for time-variant factors that may change within persons over time: family socioeconomic status, as well as adolescents’, parents’, and friends’ political interest.

Results

In this section, we present our findings in three steps, speaking to each of the hypotheses outlined above. Before doing so, however, Figure 1 provides a first glimpse at adolescents’ online activities by presenting descriptive statistics based on responses from the first panel wave conducted in 2010. As can be seen, social interaction is by far the most frequent online activity. An overwhelming majority reports using social media to talk to friends (95%), keep in touch with friends (92%), and share information about themselves (80%) at least “once a month.” This compares to the much lower numbers who say they have engaged in any form of online political activities at least “once in the past 2 months.” The most common activities are linking news to friends (48%) and joining a political Facebook group (37%). Least common is a collective action activity: organized a protest or boycott (7%).

Figure 1.

Engagement in social and political online activities (%).

Figure 2 displays the number of adolescents who engage in social interaction and each mode of online political engagement (either one activity several times in the past 2 months or two activities at least once in the past 2 months). The graph distinguishes between the two cohorts since there is a clear age effect and since we have data from five panel waves for the younger group but not for the older. Several findings are worth highlighting. First, Figure 2 confirms the importance of social interaction as an everyday online activity. Almost everyone is engaged in social interaction, and while there is a slight increase over time, this pattern is evident from the very outset of the study. Second, political information and political interaction are the most frequent political activities. The number of adolescents who engage in these activities also grows significantly over time—a pattern that is evident in both cohorts. Among cohort 1 respondents, 27% report being engaged in political information activities in 2010. After 4 years, the number is 56%. The corresponding increase within cohort 2 is from 46% to 65%. A similar trend is found for political interaction activities. Third, the figure also reveals that engagement in public production and collective action is substantially less common. No more than 13% report being engaged in public production activities during the past 2 months, while the number engaged in collective action activities is slightly higher (18–23% at its highest in both cohorts). Furthermore, compared to the frequency of political information and political interaction activities, there is no trend of public production or collective action becoming more popular over time.

Figure 2.

Frequency of online activities (%).

Are adolescents who are active in social media engaged in online politics?

So far, we have seen that social interaction is without doubt the most common type of online activity among adolescents and that some forms of political activities are more popular than others. The idea of porous boundaries and low thresholds, however, suggests that social interaction easily turns into political engagement. Hypothesis 1, based on a radical implication of the porous boundaries argument, predicted that a majority of young people using online media for social interaction would also be engaged in online politics. To test H1, Figure 3 displays the extent to which adolescents who are socially interactive also engage in each mode of online political engagement, based on cross-sectional data from wave 1. Put simply, for H1 to receive support, we should expect a large overlap between social interaction and political engagement.

Figure 3.

The frequency of different combinations of social and political online activities (%).

The top panel of Figure 3 reveals, for instance, the overlap between social interaction and political information. These findings reveal that a majority of cohort 1 adolescents (66%) is socially interactive only, while a smaller share (24%) is both socially interactive and uses online media for political information purposes. Among cohort 2 adolescents, there is somewhat more support for the notion of significant overlap between social interaction and political information (44%), with approximately half of respondents being socially interactive only (49%). As expected, the size of the group of “only socially interactive” increases when the less frequent types of political activities are considered. Looking at public production, for instance, approximately 80% of adolescents are socially interactive only, while less than 11–12% are combining social interaction with public production—and a very similar pattern is found for collective actions. From these data, it appears relatively clear that while an overwhelming majority uses online media for social interaction, a minority of the respondents combine their social media use with online political activities, and this is particularly the case for public production and collective action. The differences between political information and interaction, on one hand, and public production and collective actions, on the other hand, become even more striking if the analyses above are replicated using data from the last panel wave, when the adolescents have become older (not shown in figure). While a majority (or close to) of respondents are combining social interaction with political information and interaction at this point, both public production and collective action have become even less popular than in the first panel wave. Taken together, H1 is therefore not supported by the data, while there is some support for H2 regarding differences between modes of online political engagement.

Does tune-in, tune-out constitute the most common pattern of online political engagement?

Using cross-sectional data, the evidence in support of a radical implication of the porous boundaries and low threshold argument is thus not overwhelming. Most adolescents tend to engage in social interaction without being political. At the same time, there is a larger degree of overlap when we consider political information and political interaction activities than public production and collective action. Our third hypothesis builds upon a more modest interpretation of the porous boundaries argument and suggests a certain longitudinal pattern to online political engagement. More specifically, H3 predicted that a tune-in, tune-out form of online political engagement over time constitutes the most common pattern of online engagement compared to those being either permanently engaged or disengaged.

We start addressing this issue in Figure 4, which utilizes the panel component in the data. The bars display the number of times over the years that adolescents report being politically engaged—ranging from those who never report engaging in a specific activity to those being active at every wave of the panel survey. It should be noted that these analyses are based on respondents who participated in all waves of the study. Looking at the most frequent form of activity among cohort 1, political information reveals that 20% never use the Internet for such purposes, while 8% report doing so at each wave of the panel study. The large majority, however, 71%, is active some but not all of the years. The pattern is very similar among cohort 2 adolescents, as shown in Figure 4. Across the board of modes of engagement, the pattern of findings in Figure 4 highlights a number of interesting conclusions. First, political information is the most common mode of political engagement and the one that most adolescents return to most regularly. Second, very few adolescents can be considered “constantly engaged”—in fact, only a small minority report being active more than 2 years. Third, some of the activities are very infrequent, particularly public production and collective actions. The large majority never or seldom (only once) uses social media for such purposes. For instance, 83% of cohort 1 adolescents and 79% of cohort 2 adolescents are not engaged in collective actions more than once. The numbers for public production are even lower.

Figure 4.

The stability of online political activities over time (%).

At the same time, the numbers in Figure 4 say little about how adolescents move in and out between private and political online activities. Figure 5 speaks directly to this question, revealing the share of adolescents who tune in and out from one wave to another, compared to those who remain either active or passive between years. Looking first at political information, Figure 5 reveals that approximately 30% of cohort 1 adolescents tune either in or out between panel waves (movements), while 70% remain active or passive (stability). The numbers are very similar with respect to political interaction, with 28%, on average, reflecting between-wave movements in and out and 72% reflecting stability. Approximately, 13% and 18% tune in and out from public production and collective actions, while 88% and 82% remain either passive or active from one year to another (although those staying passive constitutes the overwhelming majority). Although Figure 5 only presents data for cohort 1, the pattern across modes of engagement is almost identical among cohort 2 adolescents. Based on these findings, H3 receives very little support. Rather than a large majority tuning in and out over time, online political engagement is characterized much more by temporal stability.

Figure 5.

The share of tune-in, tune-out and stability between years (%, cohort 1).

The effects of online social interactivity on online political engagement

Finally, to test our fourth hypothesis related to the porous boundaries argument—that social interaction has a unique effect on online political engagement—Table 2 presents the results from a series of two-way fixed-effects panel models. As argued above, the theoretical argument implies that social interaction has unique effects on each mode of online political engagement, independently of socialization and motivation factors. It is worth reiterating that these fixed-effects models control for stable unobserved heterogeneity between respondents, focusing only on within-person variation over time. Thereby, these models implicitly control for all constant factors that may be related to online political engagement. The time dummy variables neutralize the over time growth in some of the political online activities, as noted above. In addition, by including time-variant political interest, parents’ and friends’ political interest, respectively, as well as socioeconomic status, we also control for individual-level changes in these variables over time. To utilize the full amount of variation in our measures, the original scales of both our main independent variable (social interaction) and each dependent variable (modes of online political engagement) are used. Due to the skewed distribution of the dependent variables, we use a Poisson specification. Since these models are based on within-person variation, all respondents who do not change over time are automatically excluded from the analyses.

Table 2.

Effects of online social interaction on four types of online political engagement (estimates from two-way fixed-effects Poisson models).

	Political information	Political interaction	Public production	Collective action
Social interaction	.36 (.20)	.82*** (.22)	.96*** (.36)	1.36*** (.33)
Time-variant controls
Political interest	1.61*** (.25)	.95*** (.25)	.84* (.38)	1.13** (.35)
Political interest (parents)	.20 (.29)	−.06 (.30)	−.74 (.47)	−.57 (.44)
Political interest (friends)	.45* (.21)	.49* (.22)	.67 (.35)	.39 (.31)
Socioeconomic status	−.30 (.37)	−.18 (.38)	−.18 (.54)	−.37 (.52)
Time controls
Wave 2	.05 (.08)	.06 (.09)	−.02 (.14)	−.06 (.12)
Wave 3	.32*** (.08)	.40*** (.09)	−.28 (.15)	−.21 (.13)
Wave 4	.63*** (.11)	.38** (.12)	−.35 (.21)	−.46* (.20)
Wave 5	.51*** (.11)	.36** (.12)	−.64** (.23)	−.69** (.21)
Log-likelihood	−1140.069	−996.37009	−388.36495	−491.32026
N (Individuals/Observations)	1508/5173	1556/5324	829/2889	1107/3785

Unstandardized b-coefficients with standard errors in parenthesis.

p < .05; **p < .01; and ***p < .001.

The findings lend overall support to the basic argument. While there is no effect of more frequent online social interaction on political information (b = 0.36, p > .05), there are distinguishable effects on political interaction (b = 0.82, p < .001), public production (b = 0.96, p < .001), and collective action (b = 1.36, p < .001), controlling both for all stable between-unit characteristics and within-person changes in political interest, parents’ and friends’ political interest, and socioeconomic status over time. Put differently, irrespective of important motivational factors, adolescents who increase their online social interactivity are significantly more likely to engage in several political activities, including collective action. Thereby, H4 receives support from the data.

Conclusion and discussion

This study aimed to contribute to the literature on young citizens’ online political engagement. While much research has explored the dynamics of online activism (Bennett and Segerberg, 2012; Kavada, 2015) and identified significant effects of online media use on political participation (Boulianne, 2009, 2015), questions related to the actual patterns of young citizens’ online engagement remain unanswered. To the best of our knowledge, this is the first large-scale longitudinal study investigating the implications of the assumed lower thresholds to political engagement in social media. The five-wave study comprising rich data on various online activities and the exceptionally high response rates has enabled a robust test of four hypotheses.

We formulated a radical hypothesis (H1), assuming that, for a majority, social interaction in online media coincides with occasional involvement in four different modes of political engagement, and a modest version (H2) suggesting that this pattern applies only to less demanding modes of engagement. While H1 was falsified, H2 received empirical support—and the differences across modes of engagement became even clearer over time. The study clearly indicates that thresholds into politics vary between modes of engagement. Put simply, almost every adolescent uses social media for social interaction. For a fairly large—and increasing—group, this coincides with an engagement in political information (reading, watching, searching) and political interaction (discussing, commenting, sharing), while very few combine social interaction with public production (writing, producing video clips, etc.) or collective actions (signing or organizing groups, protests, and boycotts). The fact that both public production and collective action lose popularity over time is striking given the fact that general political interest—reflected in political information and interaction activities—increases during the same period. Although the reasons behind these trends remain to be analyzed in future research, such developments may very well reflect growing public awareness concerning integrity on social media and reluctance to reveal personal political identities publicly (Ekström, 2016).

Our longitudinal data also enabled a unique test of a third hypothesis (H3), namely that porous boundaries and low thresholds in social media promote patterns of tune-in, tune-out political engagement over time. Although H3 received no general support, the differences between modes of engagement were reflected also in these analyses. Approximately 30% of all adolescents move in and out from political information and interaction activities, while less than 20% do so when it comes to public production and collective action.

Our fourth hypothesis (H4), however, received support. We observed statistically significant effects of online social interaction on different modes of online engagement, while controlling for individual-level changes in political interest and socialization factors—as well as all stable individual-level characteristics. This indicates that low thresholds into politics help engage daily users of social media independently of their interest in political matters. Taken together, these findings clearly suggest that while the mechanisms behind the porous boundaries argument appear to operate on a micro level, these processes do not turn large groups of adolescents into politically active citizens.

The fact that young people are more frequently engaged in reading, sharing, and commenting on politics, than producing and publishing texts or video clips, signing petitions, and joining collective protests in social media, is perhaps not surprising. It seems easier and less demanding and decisive in terms of public articulations of opinions and political affiliations. The general ideas of low-cost engagement and liquid boundaries between non-political and political activities in social media are, however, frequently discussed in relation to a diversity of activities, including connections to protests and activism. We have developed and applied a theoretically anchored typology which made it possible to identify important differences in the frequencies and patterns of online political engagement, thus indicating different boundaries into politics. It is a challenge for future research to study the boundaries and the related boundary-crossings in more detail.

Although the flexible low-cost connections to activism have been recognized as crucial to the ongoing transformations of political engagement in digital media (Bennett and Segerberg, 2012; Dencik and Leistert, 2015; Earl and Kimport, 2011), our study clearly indicates that such activities engage only a minority of adolescents. How can such discrepancies be understood? To be sure, there is no doubt that digital network media contribute to high-level participation in online campaigns and protests. Most likely, these seemingly contradictory pictures relate to differences in research objects and data. While many previous studies focus on the dynamics and aggregations of participants within protest movements, our study focuses on a general population of young citizens using survey data. This is crucial not only from a methodological point of view but also when assessing the democratic implications of online media in terms of the levels and distribution of political engagement in society.

A reasonable objection to our conclusions is that this study fails to capture the effects of low thresholds due to the fact that young people may have a restrictive conception of the “political”—reflected in an underreporting of political activities in survey responses. Thus, citizens may frequently cross boundaries by joining groups in social media, without necessarily considering this as a step into politics. That is, we cannot exclude the possibility that respondents to some extent underestimate their political engagement. On the other hand, several of our survey questions asked about politics and social issues, and the items used to measure collective actions did not at all include the term “politics” (but protest, boycott, societal issues, etc.)

In sum, our findings indicate that taking a step into politics is conditioned and that structuring boundaries (related to cultural, social, and psychological factors) continue to exist in social and online media. This is not to question the significance of incidental exposure, the fluid transitions from private to public, and the low-cost engagement in social media. However, what this means for citizens with different social background, habits, motivations, and identities very much remains to be investigated.

As regards, in particular, young people, it has been argued that the widespread concerns about decreasing political participation are exaggerated due to a limited focus on conventional modes of participation. An updated understanding of participation, including online forms of activism and protests, is assumed to offer a more optimistic view (Dalton, 2009; Loader, 2007; Shakuntala and Buckingham, 2013). This research may give the impression that large groups of young people are active in joining global protests or local Facebook initiatives. This study indicates that this is not the case.

To what extent can we assume the patterns observed in this study to be generalizable beyond the Swedish context? Cross-cultural data are required to provide a clear answer to this question. Young people’s online political engagement is likely to vary depending on political cultures, institutions, and concrete circumstances. Trust in political institutions is, for example, relatively high in Sweden. However, there is nothing to suggest that Sweden is so unique that the patterns of online political engagement would be radically different in other countries in Scandinavia or Western Europe—or that the mechanisms behind porous boundaries would operate differently.

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was made possible by access to data from thePolitical Socialization Program, a longitudinal research program at YeS (Youth & Society) at Orebro University, Sweden. Professors Erik Amna, Mats Ekstrom, Margaret Kerr and Hakan Stattin were responsible for the planning, implementation and financing of the collection of data. The data collection and the study were supported by grants from Riksbankens Jubileumsfond.

Author biographies

Mats Ekström is a professor in the Department of Journalism, Media and Communication, University of Gothenburg. His research focuses on media discourse, political communication, and young people’s political engagement.

Adam Shehata is an assistant professor in the Department of Journalism, Media and Communication, University of Gothenburg. His research focuses on media use and effects.

References

Allison

(2009) Fixed Effects Regression. Thousand Oaks, CA: SAGE.

Bennett

(2012) The personalization of politics: political identity, social media, and changing patterns of participation. The Annals of the American Academy of Political and Social Science 644: 30–38.

Bennett

Segerberg

(2012) The logic of connective action: digital media and the personalization of contentious politics. Information, Communication & Society 5: 739–768.

Bimber

Flanagin

Stohl

(2012) Collective Action in Organizations: Interaction and Engagement in an Era of Technological Change. New York: Cambridge University Press.

Boulianne

(2009) Does Internet use affect engagement? A meta-analysis of research. Political Communication 26: 193–211.

Boulianne

(2015) Social media use and participation: a meta-analysis of current research. Information, Communication & Society 18: 524–538.

Brundidge

(2010a) Encountering “difference” in the contemporary public sphere: the contribution on the Internet to the heterogeneity of political discussion networks. Journal of Communication 60: 680–700.

Brundidge

(2010b) Toward a theory of citizen interface with political discussions and news in the contemporary public sphere. International Journal of Communication 4: 1056–1078.

Carlisle

Patton

(2013) Is social media changing how we understand political engagement? An analysis of Facebook and the 2008 presidential election. Political Research Quarterly 66: 883–895.

10.

Couldry

Livingstone

Markham

(2007) Media consumption and public engagement. Basingstoke: Palgrave Macmillan.

11.

Dalton

(2009) The Good Citizen: How a Younger Generation Is Reshaping American Politics. Washington, DC: QC Press.

12.

Dencik

Leistert

(2015) Critical Perspectives on Social Media and Protest. London: Rowman & Littlefield International.

13.

Earl

(2014) Something old and something new: a comment on “new media, new civics.” Policy & Internet 6: 169–175.

14.

Earl

Kimport

(2011) Digitally Enabled Social Change: Activism in the Internet Age. Cambridge, MA: MIT Press.

15.

Ekström

(2016) Young people’s everyday political talk: A social achievement of democratic engagement. Journal of Youth Studies 19: 1–19.

16.

Ekström

Olsson

Shehata

(2014) Spaces for public orientation: Longitudinal effects of Internet use in adolescence information. Communication and Society 17: 168–183.

17.

Ekström

Östman

(2015) Information, interaction and creative production: The effects of three forms of internet use on youth democratic engagement. Communication Research 42: 796–818.

18.

Finkel

(2008) Linear panel analysis. In: Menard

(ed.) Handbook of Longitudinal Research: Design Measurement and Analysis. Burlington, MA: Academic Press, pp. 475–504.

19.

Halupka

(2014) Clicktivism: a systematic heuristic. Policy & Internet 6: 115–132.

20.

Hooghe

(2014) Defining political participation: how to pinpoint an elusive target? Acta Politica 49: 338–341.

21.

Jenkins

(2006) Convergence Culture: Where Old and New Media Collide. New York: New York University Press.

22.

Kavada

(2015) Creating the collective: social media, the occupy movement and its constitution as a collective actor. Information, Communication & Technology 18: 872–886.

23.

Kim

Hsu

Gil

Zuniga

(2013) Influence of social media use on discussion network heterogeneity and civic engagement: the moderating role of personality traits. Journal of Communication 63: 498–516.

24.

Kim

Russon

Amnå

(2016) The longitudinal relation between online and offline political participation among youth at two different developmental stages. New Media & Society. 19 (6):899–917.

25.

Kruikemeier

Van Noort

Vliegenthart

et al . (2014) Unraveling the effects of active and passive forms of political Internet use: does it affect citizens’ political involvement? New Media & Society 16: 903–920.

26.

Loader

(ed.) (2007) Young Citizens in the Digital Age: Political Engagement, Young People and New Media. London: Routledge.

27.

Norris

(2001) Digital Divide? Civic Engagement, Information Poverty and the Internet Worldwide. Cambridge: Polity Press.

28.

Papacharissi

(2010) A Private Sphere: Democracy in a Digital Age. Malden, MA: Polity Press.

29.

Schlozman

Verba

Brady

(2010) Weapon of the strong? Participatory inequality and the internet. Perspectives on Politics 8: 487–509.

30.

Shah

Kwak

Holbert

(2001) Connecting and disconnecting with civic life: patterns of Internet use and the production of social capital. Political Communication 18: 141–162.

31.

Shakuntala

Buckingham

(2013) The Civic Web. Cambridge, MA: MIT Press.

32.

Theocharis

(2015) The conceptualization of digitally networked participation. Social Media + Society 1 (2):1–14.

33.

Van Deth

(2014) A conceptual map of political participation. Acta Politica 49: 349–367.

34.

Velasquez

LaRose

(2015) Youth collective activism through social media: the role of collective efficacy. New Media & Society. 17 (6):899–918.

35.

Vissers

Stolle

(2014) The Internet and new modes of political participation. Information, Communication & Society 17: 937–955.

36.

Xenos

Vromen

Loader

(2014) The great equalizer? Patterns of social media use and youth political engagement in three advanced democracies. Information, Communication & Society 17: 151–167.