The psychology of online activism and social movements

We review online activism and its relations with ofﬂine collective action. Social media facilitate online activism, particularly by documenting and collating individual experiences, community building, norm formation, and development of shared realities. In theory, online activism could hinder ofﬂine protests, but empirical evidence for slacktivism is mixed. In some contexts, online and ofﬂine action could be unrelated because people act differently online versus ofﬂine, or because people restrict their actions to one domain. However, most empirical evidence suggests that online and ofﬂine activism are positively related and intertwined (no digital dualism), because social media posts can mobilise others for ofﬂine protest. Notwithstanding this positive relationship, the internet also enhances the visibility of activism and therefore facilitates repression in repressive contexts. Using interviews and participant observation, the author shows that peace movement actors in Ambon, Indonesia, employ social media to sustain their movement. This is a rare example of a study that follows a social movement’s online behaviours after its initial rise and peak.

Oftentimes, when thinking of activism people envision mass behaviours such as demonstrations and disruptive activities (blockades, riots). This contrasts sharply with online activism by individuals united in their worldviews yet dispersed in time and/or place. This review of online activism investigates the role of online activities in the larger repertoire of contention. We first briefly review collective action as it occurs online. Many studies on online activism also consider its relations to offline protest. We discuss evidence for, respectively, negative, inconsistent, and positive relations between online and offline actions and their outcomes. We conclude with a consideration of protest in repressive contexts, and a discussion including directions for future research.

Online activism
Online activism takes many forms, from symbolic signalling of one's stance on a politicised issue (e.g. changing one's social media profile picture) to more complex engagement (e.g. writing detailed posts about a social issue [1]). Social media facilitate online activism in three key ways. First, they allow individuals to express experiences and opinions, relating them to collective causes (see #metoo [2,3]). Second, they allow online community members to provide support, organise activities, and challenge negative responses to their activities [4]. One example is 'digilantism', where perceived norm transgressions (e.g. misogyny) are exposed and publicly sanctioned [5]. Within in-groups, this can raise awareness and nourish activism. Nevertheless, it has downsides similar to vigilantism and can invite an inter-group backlash [6]. Third, social media allow people to involve others outside their online community to collectively negotiate new shared realities and spread these [4,7]. This can empower communities, as exemplified by women's #freethenipple posts of topless photos to normalise unsexualised representations of breasts and reclaim the female body [4]. In sum, three types of communication via social media can boost activism: Relating individual perspectives to activist causes, organising activist communities, and negotiating shared realities with outsiders.

Relations between online and offline action
The formation of online activist communities is rarely isolated: The online and offline are typically closely integrated. Indeed, online activism facilitates offline protest by advertising and organising it [8]. Increasingly, this means that mass protests can occur without formal structures (e.g. trade unions). Some suggest we are witnessing the birth of an entirely new form of connective action [9,10]: Bottom-up mobilisation that occurs when calls to action cascade through interconnected personal networks. Of course, social media vastly increase communication faculties, but throughout history comparable bottom-up protests have regularly occurred. Pre-19th century, these were probably predominant, and studies of food riots and riots during the reformation [11,12] suggest that communications of the day (pamphlets, town square assemblies, rumours) played a role in the group dynamics of mobilisation similar to that played by social media today. Notwithstanding these similar communication functions, the literature provides a mixed view on the links between online and offline action -supporting, respectively, a negative, no unequivocal, or a positive relationship between online and offline activism. We will now discuss the empirical basis of these three perspectives.
Negative relation: the trade-off hypothesis Especially when it was unfamiliar, online activism was dismissed as 'slacktivism' that was supposedly effort-free, unproductive, and inhibiting more effortful, effective offline protest (the latter is essentially a trade-off hypothesis). Increasingly, this reasoning is seen as simplistic [13]. Several factors moderate whether online and offline activism relate negatively. For instance, online activism does not inhibit offline protest if activists perceive their actions as effective [14 ]. Other moderators are age (for older users online engagement is not sufficient) and network heterogeneity (homogeneity increases carry-over between online and offline activism through social support [15]). Furthermore, effortful online actions (producing videoclips, managing events) cross over to offline action [16]. Other mechanisms can also cause negative relations between the unfolding of collective action online and offline. For example, activists can online distance themselves from offline riots [17], illustrating how online and offline actions may react to each other by contrasting away from the other domain. Alternatively, online and offline activities can be complementary over phases of action: Planning and mobilisation, real-time reporting and framing, and aftertalk 'reviewing' actions and demobilisation. To recap, a few isolated studies suggest that online activism occasionally substitutes offline activism, but this appears to be rare. The relationship appears more complex than the trade-off hypothesis suggests.

Inconsistent relation: digital divide, echo chambers, and digital dualism
Other sources indicate that in some contexts, online and offline protests are neither negatively nor positively related. Three processes can explain this finding: (1) digital divides, (2) spiral of silence and echo chamber effects, and (3) digital dualism. First, people engaging in online action may differ from those acting offline -that is, digital divides. For instance, working-class people are less politically active online because they feel less technology savvy [18]. Furthermore, some evidence suggests that younger people engage more online and older men engage more offline [19]. This is contradicted by findings that digital divides do not play a role in online petition signing [20]: The effort required for actions may play a moderating role.
Second, relations between online and offline activism become unreliable if processes that encourage or dampen activism evolve differently online versus offline. One such a process is the spiral of silence [21]: People selfcensor opinions that they expect to be unpopular. But meta-analytic evidence suggests that the spiral of silence is equally strong online and offline [22]. Also, selfsegregation into like-minded networks allegedly would cause online activism to be different. Social media characteristics (e.g. ease of 'unfriending') facilitate echo chambers [23], in which the same shared realities are echoed and socially validated, encouraging the formation of monocultures. The resulting perceived sharedness can strengthen people's world views [24]. But the literature is not clear whether this is a greater problem online than offline and, moreover, evidence indicates that opinion heterogeneity (the opposite of echo chambers) can also fuel collective action [25].
Third, digital dualism suggests that people enact different personae online versus offline. Relatively anonymous online environments free people from concerns to be positively evaluated and consequent social restrictions to their behaviour [2, 26,27]. This may facilitate online activism without fear of social repercussions. Online disinhibition becomes particularly likely if people lack self-control [28], are low in avoidant or anxious attachment [29], or suffer psychological distress [30]. A persistent misconception regarding online (relative) anonymity is that when people feel less individually identifiable they become deindividuated and, hence, less responsive to all social norms. Anonymity to outsiders instead empowers people to behave more consistently with the norms of their own group of 'insiders' [31 ]. That is, pseudonymised online community members are only more likely to riot if that community consists of violent activists but disorderly behaviour is less likely if their community consists of pacifists. Thus, online activism potentially diverges from offline activism but the exact nature of this divergence is context-dependent.

Positive relation: intrapersonal consistency and interpersonal mobilisation
Ample evidence supports positive relations between online and offline activism [8,32,33,34]. Online activism participation can stimulate individuals to also protest offline -an intrapersonal effect. Small online actions can ease people into more costly offline action (although this foot-in-the-door technique may backfire especially for non-profit movements [35]). Besides this compliance technique, other psychological mechanisms may play a role. For instance, social media might encourage transition from online to offline activism by facilitating social identity formation -albeit recent meta-analytic evidence is mixed [36 ]. Online activism may thus cultivate the psychological preconditions to embolden individuals to embrace more burdensome offline protest. These preconditions include tightly knit, thick social identities characterised by online and offline interest alignment [37][38][39][40], morality, solidarity, or shared belief regarding the issue at hand [40][41][42][43][44], self-efficacy [44], and unfairness [45]. In addition to such gateway effects from online to offline action, the reverse may also occur; when one's offline action spills over into the online domain [46]. And finally, intrapersonal concurrence between online and offline activism may result from the intertwining of one's offline and online lives (e.g. incorporation of Tinder in people's intimate 'offline' life [47]). Thus, online and offline activism seem strongly related within persons -arguing against digital dualism [48].
Alternatively, interpersonal effects occur when individuals coordinate, recruit, develop social identities and shared realities, and share information online before, during, and after movements' initial rise [14 ,43,49,50,51 ,52,53]. Indeed, social media and online activism have been heralded as instrumental (albeit not without obstacles) in mobilising potential new participants for offline action [54]. Both intrapersonal and interpersonal consistency between online and offline activism paint a general picture of collective action as positively related across the two domains.

Internet as technology for democratisation or repression
Most research on relationships between online and offline activism concerns western democracies. The few studies analysing non-democratic, repressive contexts mostly focus on macro-level cross-country analyses [55][56][57] on how internet access or use influences protests. The internet has a two-faced function [58], as liberation technology supporting activism [57] or repressive control technology [56]. Online actions can be subject to horizontal surveillance (social control among citizens, digilantism [5,6]). Repressive regimes can also use the internet for vertical surveillance, controlling citizens and suppressing protests that threaten their power. Although the internet may support online activism and its spread to offline protest, such increases in (online and offline) protest can invigorate repression [55]. Thus, at the macro level online activism may initially stimulate offline activism under repressive regimes while the relation subsequently becomes complicated by the regimes' responses to these actions.
Micro-level analyses in these matters are rare, mainly due to the lack of individual-level data on activism in repressive contexts. As one exception, recent panel data indicate that Iranian Green Movement supporters who are more active online are also more active offline, and vice versa [59]. Additional micro-level support for positive relations between online and offline activism comes from a cross-national survey in Muslim-majority countries around the Arab Spring [60]. Notably, individuals' general internet use was unrelated to offline protest. In sum, what people do online is more important than mere internet access in the relationship between online and offline activism.

Discussion and conclusion
We have reviewed online activism and its relations with offline protest. To recap, several social media characteristics facilitate online activism: Particularly its role in documenting and collating individual experiences, in community building and norm formation, and in the development of shared social realities. There is mixed empirical evidence that online protest prevents offline protests, resulting in ineffective slacktivism. Other evidence suggests that in some cases, online and offline actions are relatively unrelated because people act differently online versus offline (intrapersonal effect) or because different people engage in online versus offline action (interpersonal effect). Overall, the literature currently suggests that in many cases online and offline activism correlate, either because people's online and offline behaviours are intertwined or because one person's online activism can mobilise others for offline protest. That is, the current evidence argues against digital dualism. In repressive contexts, macro-level analyses indicate that the internet can stimulate activism and revolutions, but also facilitate top-down repression. Micro-level evidence supports a positive relation between online activism and offline protest among citizens under repressive regimes.
Together, these findings suggest valuable avenues for future research. More research is needed on understudied phenomena such as restricted communication and repression. Future research could also focus on relations between technology and psychological outcomes, by exploring differences between online platforms (Facebook, Twitter), different online behaviours (commenting, sharing, liking), or new technologies (e.g. live streaming, asynchronous video-sharing [61]). Furthermore, research could cover more completely the life-cycle of online movements. Specifically, it could move beyond the predominant focus on the initial stages of (online) action development (cf. [62 ]) by investigating unsuccessful social movements or cycles between online and offline action.
In conclusion, the current state of the literature paints an intriguing picture about how social media are utilised for collective action. The internet is widely used for emancipatory actions to raise awareness, rally people, set activist agendas, to debate and evaluate actions, but also antagonistically (by groups and authorities) to polarise, misinform, and repress unwanted actions. Unmistakably, minority groups can more easily make contact and make themselves heard through social media. This gives social media a great vibrancy and pluralism, but it may also divide and polarise societies. Increasingly, online and offline activism are inseparable and complementary social-psychological instruments for politicisation, debate, mobilisation, and conflict.

Conflict of interest statement
Nothing declared.

Author note
This research was partly funded by a grant of the National Coordinator for Security and Counterterrorism at the Dutch Ministry of Justice and Security (grant number 40100013112) to Tom Postmes and Hedy Greijdanus. The funding source did not play a role in design, analysis, interpretation, writing, or any of the other stages of this research.
The authors thank Patrick Dü lsen, Wouter Kiekens, and Inka Papenfuss for their help with the literature search.

References and recommended reading
Papers of particular interest, published within the period of review, have been highlighted as: of special interest of outstanding interest