This is an open access article distributed under the terms of the Creative Commons
Attribution License (In recent years we have witnessed an increase in social protests around the world (Akaev et al., 2017; Routledge, 2017). From uprisings in the Middle East against oppressive regimes, to movements in different European countries against austerity policies, a global movement in favour of women's rights, the outbreak of popular uprisings in several countries of South America, or anti-racism protests in the US in the midst of the Covid-19 Pandemic. These events have generated a growing interest on the part of researchers, governments, legislators and citizens, to understand the emergence of social movements and the motives why people are involved in them.
Social Psychology has made significant contributions to understanding the factors that motivate collective action (van Stekelenburg & Klandermans, 2017; van Zomeren et al., 2008), but has paid less attention to the different forms it can take (Wright et al., 1990), although it can range from moderate and non-violent actions (e.g., participating in marches and signing petitions), to radical or violent forms (e.g., destruction of public or private property). As Becker and Tausch (2015) note, systematic work on the psychosocial processes that underpin participation in different types of action has been, until recently, rather scarce.
An important consideration in understanding the choice of action may be the sociopolitical context in which activism takes place (Moghaddam, 2018, 2019) and a key element of that context is how the police respond to the protests (Drury & Reicher, 2000; Drury et al., 2020; Saavedra & Drury, 2019). Recent events have showed that repression of social protest is not an exclusive problem of newer democracies or authoritarian contexts, and abuse of police authority can take place anywhere (Bonner et al., 2018; Regilme, 2019).
It is widely acknowledged that repression of social mobilization can influence the intensity of protests, but there is no consensus on the direction of this relationship, and comparatively little data is available (Earl & Braithwaite, 2022; Honari, 2018; Lichbach, 1987; Norris, 2022). Several examples show how coercive power is effective in suppressing collective action (Bellin, 2012; Ellefsen, 2016; Wood, 2007), but other cases put in evidence that use of force against unarmed citizens can have the opposite effect, leading to counter mobilization (Adam-Troian et al., 2020; Ayanian & Tausch, 2016; El-Ghobashy, 2011).
This inconsistency in the relation between repression and participation has been called the “punishment puzzle” (Davenport, 2007) or the “protest-repression nexus” (Lichbach, 1987), and it remains a mystery. Some plausible explanations for these contradictory results suggest that repression may decrease protests in general (Norris, 2022), but increase illegal actions (Ellefsen, 2016); that scholars have failed to distinguish different types and targets of repression (Earl & Braithwaite, 2022); that the use of consistent repressive tactics tends to reduce protest while inconsistent responses can increase it (Belgioioso et al., 2018); and that repression can incentivize protest if it is communicated effectively and considered excessive by relevant audiences (Hess & Martin, 2006). Honari (2018) argues that most studies on the effects of repression tend to understand it from a macro perspective, without paying attention to the fact that individuals, although integrated into similar structural networks and contexts, perceive and interpret repression differently. Therefore, the failure to explain the variation in the effects of repression could be due to a lack of attention to individual responses to repression.
Recent studies carried out from a psychosocial perspective, that is, taking the individual as the unit of analysis, show that the risks attributable to state repression instead of decreasing resistance incentivize it, by increasing indignation, politicized identity, efficacy and moral obligation (Ayanian et al., 2021); that soft repression could increase the fear of social sanction in some activists, but it can also have a backfiring effect by facilitating contacts among radicals and increasing distrust of institutional politics (Jämte & Ellefsen, 2020); that threat and mistrust interact to motivate non-normative responses (Gulliver et al., 2023); and that exposure to police violence would have direct positive effects on the intention to attend future demonstrations and to sacrifice for the movement, and that these effects are mediated by the perception of loss of significance and identification (Adam-Troian et al., 2020).
Chile is especially interesting to the study of collective action facing police repression since the country is often lauded for its successful democratic transition and political and economic stability (Delamaza, 2015; Letelier & Dávila, 2015). According to The Freedom House score (2019) and Polity score (Marshall et al., 2016), Chile is one of the countries with the highest levels of democracy in the world, at the same level as the USA, Canada and the United Kingdom, for example. It is described by The Freedom House (2019) as “a stable democracy that has experienced a significant expansion of political rights and civil liberties since the return of civilian rule in 1990”. However, in the last decade there has been a strong reactivation of repressive police violence against protesting citizens, which has involved criminalization of social protest and police violence (Donoso & Salinero, 2016; Donoso & von Bülow, 2017). Police handling of public protests has failed to meet international human rights standards and the use of anti-riot equipment such as water cannons and tear gas is often indiscriminate and disproportionate (Human Rights Watch, 2015, 2020).
Hence, the aim of this work is to contribute to understanding the motivations to participate in collective action when it faces repression within the framework of a democratic society.
Forms of Participation
In the literature on political participation, numerous proposals can be found to classify the different forms of collective action (Sabucedo & Arce, 1991), such as: conventional and unconventional action (Barnes et al., 1979); political action inside and outside the system (Sabucedo & Arce, 1991); constitutional and extra-constitutional action (Hayes & McAllister, 2005); and activism and radicalism (Moskalenko & McCauley, 2009). In this paper, the definition of Wright et al. (1990) is used, which distinguishes between normative and non-normative action, because although it is very similar to the above taxonomies mentioned (Tausch et al., 2011), it places greater emphasis on the social context in which the action is carried out, since this defines the limits between the categories. The normative actions would be those that conform to the norms of the dominant social system in a democratic regime (e.g., participation in marches or rallies), and the non-normative ones, those that transgress the norms defined by that social system (e.g., confrontations with the police).
Predictors of Participation
Among the predictors of collective action identified by social psychology, identification with the ingroup (e.g., Tajfel & Turner, 1979; Thomas et al., 2012) stands out, alongside injustice appraisals (e.g., Becker & Wright, 2011), emotions such as anger or indignation (e.g., Hayward et al., 2018; Leonard et al., 2011), and the perceived efficacy of collective action (e.g., Hornsey et al., 2006). Drawing these predictors together, the Social Identity Model of Collective Action (SIMCA, van Zomeren et al., 2008), one of the most prominent models of collective action inspired by social identity approaches, proposes that identification with the group provides the psychological basis for the experience of group-based anger and group efficacy beliefs that motivate collective action. Thus, SIMCA identifies two pathways to participation, which are based on social identification: an instrumental pathway through collective efficacy, as a stronger sense of identity empowers individuals and predicts their participation (Drury & Reicher, 2005; van Zomeren et al., 2008); and an emotional pathway through perceived injustice and anger, where group-based anger motivates collective action (van Zomeren et al., 2008).
Notwithstanding, numerous studies have shown that not only anger predict participation, but also other emotions, especially positive emotions, such as pride, hope and joy (e.g., Asún et al., 2021; Bar-Tal et al., 2007; Poma & Gravante, 2016; Sabucedo & Vilas, 2014; Tausch & Becker, 2013; Wlodarczyk et al., 2017). Moving beyond anger alone, as Sabucedo and Vilas (2014) note, challenges scholars to identify what other emotions are relevant for participation in collective actions, and what roles the different factors play in participation in different forms of collective action. Despite some promising progress, showing, for example, that different emotions are associated with normative versus non-normative action (e.g., Tausch et al., 2011), and that different emotions are also associated with activism versus charity (Thomas & McGarty, 2018), there is still no consensus or clarity about the role of some emotions in the participation in collective actions.
Measurement of Participation in Collective Action
Given the ethical concerns regarding asking people to participate in violent non-normative actions in the context of an experiment, in the laboratory what is usually done is to evaluate intentions to perform in such actions in imagined scenarios (Becker & Tausch, 2015). On the other hand, given the low frequency of participation in non-normative protests in natural contexts, most past research has examined either willingness to engage in collective action or support for different forms of collective action as dependent measures, rather than actual participation (Tausch et al., 2011). This approach limits the understanding of the great complexity of processes involved in real situations, since there are important differences between the intentions of participation provoked in an experimental situation, the intentions of participation generated in a real situation, and the actual participation (e.g., Beyerlein & Hipp, 2006). This is why the present research seeks to make a contribution to the literature by measuring self-report of actual participation in diverse forms of collective action.
The Case Study
The present investigation was carried out in the context of the social movement called “Aysén your problem is my problem”, that took place in 2012, in the Aysén region, in Chilean Patagonia. Aysén is located in the south of Chile and is a large region with rugged geography. The isolation and great distances between the region and the centre of the country have generated problems in access to basic services such as health or education (Ministerio de Desarrollo Social, 2013).
The social movement was a broad coalition fighting for longstanding demands for improving regional conditions, such as better health and education facilities, lower prices for fuel and staples, and citizen consultation on major hydroelectric dams. The movement united fishermen, small business owners, various social and civic organizations, and militants from the left and the right. The protests paralyzed air, land, and water transport, cutting traffic into and out of the region (Rein, 2014). To deal with the protests, the government employed a strategy of violent repression, and the movement implemented a communication strategy to publicize this violence, obtaining great support from Chilean and international public opinion (Durston et al., 2016). According to our systematic review of national and regional press, the movement lasted 56 days, during which time 99 protest events were held, of which 25% ended in violent and unprecedented clashes between the police and the demonstrators (Sandoval et al., 2020).
We have selected this case study for three reasons: the movement suffered severe repression within a democratic regime; the movement deployed a diverse set of collective actions, including peaceful normative actions, such as motorcades and rallies, and non-normative ones, such as roadblocks and confrontation with police forces; and furthermore, it was a regional movement, which tend to be more heterogeneous in terms of the socio-demographic characteristics of the participants than other kind of movements, and tends to bring together people of different ages, genders, occupations, ideologies and socio-economic status (Asún & Zúñiga, 2013a). Testing an established model in this novel context with real self-report normative and non-normative actions in a representative community sample is thus of great value (Ayanian & Tausch, 2016; van Zomeren, 2015). Comparing the performance of well-studied predictors (identity, efficacy, anger), as well as comparatively understudied positive emotions, for normative and non-normative actions, is also of interest. To this end, two studies were carried out, one quantitative and one qualitative, which are described below. Both studies were approved by the Ethics Committee of the Faculty of Social Sciences of the University of Chile.
In the present work, we do not consider repression as a variable but rather as a characteristic of the context in which collective action takes place, and we understand it as those actions developed by government authorities to prevent mobilization through police violence, harassment, surveillance, prohibitions, intimidation of activists, among others (Davenport, 2007).
Study 1
Study 1 had as its first objective, to determine if participation in collective actions of the social movement "Aysén your problem is my problem" can be understood as a unitary construct, or if different types of collective actions, carried out by different groups of people, can be clearly distinguished. Secondly, we aimed to test the hypothesis that identity, collective efficacy, anger and positive emotions predict people's participation in the different protest actions developed by this social movement in a framework of strong repression. The present research is part of a broader program of research which sought to answer the question of why people engaged in the violent actions of this social movement (see also, Medel et al., 2022).
Method
Participants and Procedure
The participants were 500 people above the age of 18, selected through a multistage cluster random sampling from the two main districts1 in the region, including urban and rural areas. The surveys were conducted face-to-face by trained interviewers in participants’ homes in early 2015. 45.4% of participants were male (n = 227) and 54.6% female (n = 273). The age range was 18 to 94 with a mean of 48 (SD = 17.4). Regarding socioeconomic status2, 25.8% were low, 56.6% middle, and 17.6% high.
Instruments
Collective Action
To measure participation in the protests, a set of 15 items were presented (ordinal α = .97), asking the respondents if they had participated in the different actions that took place during the movement, including both normative and non-normative forms of action. The response alternatives were never, once, or more than once. These items represent actions developed within the framework of this and other Chilean territorial social movements, and were identified from qualitative studies and press analysis carried out by our research team. They have been tested in the general population and university population in different regions of the country, with excellent indicators of reliability and construct validity (Medel et al., 2022; Muñoz et al., 2020). Items are presented in Table 1.
Table 1
Factor Loadings of Exploratory Factor Analysis
| Item | Factor Loading |
|---|---|
| 1. March | .830 |
| 2. Gathering | .842 |
| 3. Motorcade | .819 |
| 4. Raise money | .704 |
| 5. Support to demonstrators | .831 |
| 6. Cultural performance | .750 |
| 7. Cacerolazoa | .756 |
| 8. Strike | .896 |
| 9. Buildings occupation | .873 |
| 10. Barricades | .839 |
| 11. Funasb | .866 |
| 12. Meetings interruption | .922 |
| 13. Confrontation with police | .875 |
| 14. Destruction of public or private property | .820 |
| 15. Throw stones at the police | .835 |
aA group of people creating noise by banging pots, pans, and other utensils in order to call for attention. bA demonstration of public repudiation against a person.
Identification With the Regional Group
Group identity was measured by the Regional Identity Scale (RIS-2; Asún, Zúñiga, & Morales, 2018), in its version of 16 items (ordinal α = .85). Examples of items are: “I feel proud to be part of this region” and “When I talk about the people of this region, I often say ‘we’ instead of ‘they’”. Respondents indicated their agreement with these items using 5-point scales (strongly disagree to strongly agree).
Efficacy
We understand group efficacy as the perception that the members of a group are able to work together to solve collective problems or address collective needs (Bandura, 1997, 2000). Hence, group efficacy was measured by four items (ordinal α = .67): “The inhabitants of the region had the ability to organize ourselves to achieve the changes we wanted”, “There were many people willing to make sacrifices for the benefit of the entire region”, “There were enough leaders to conduct a social movement that sought to improve conditions in this region”, and “We could achieve social changes by mobilizing and protesting”. Respondents indicated their agreement with these items using 5-point scales (strongly disagree to strongly agree).
Anger
Anger was measured by three items (ordinal α = .73): “Thinking about the problems that motivated the movement, to what extent did they make you feel angry?”, “The reaction of the national authorities toward the social movement, to what extent did it make you feel angry?", and “Remembering what was experienced during the mobilization, to what extent did you feel anger?”. The response options were: None, Little, Some and A lot.
Positive Emotions
Positive emotions were measured with three items (ordinal α = .831) that began with the phrase "Remembering what was experienced during the mobilization, to what extent did you feel ...?", and then the emotions of hope, joy and pride were presented. The response options were: None, Little, Some and A lot.
Analyses
Exploratory factor analysis, latent class analysis and structural equation model were conducted, with the software M-plus 7.4. Considering that we had very few missing values (they constituted only 0.5% of the responses, and 97.2% of the participants responded to all items), we excluded them from the pairwise analysis.
Results
A large part of the sample, 45.7%, participated in at least one protest action. The three most frequent actions were: providing support to demonstrators, such as taking them food or giving first aid to those injured (37.9% of the respondents did it at least once); marches (33.6% took part in at least one); and gatherings (23.9%).
Structure of Participation in Collective Action
Two analyses were conducted: an exploratory factor analysis of the actions, to determine whether it is possible to clearly distinguish between two or more types of actions (for example, normative and non-normative); and a latent class analysis, to determine whether those actions were performed by two or more different groups of people.
The exploratory factor analysis was carried out with polychoric correlations and ULSMV estimation were used, given the ordinal nature of the data and to avoid spurious factors arising derived from the participation asymmetry in the different types of actions (Asún et al., 2016).
The results show that a single factor solution, which includes behaviours ranging from marches or gatherings, to confrontations against the police and destruction of public property within a single repertoire, fits the data well. The goodness-of-fit indices show an excellent fit, χ2(90, N = 499) = 177.122, p < .001; χ2/df = 1.97; RMSEA = .044; CFI = .976; TLI = .972. In Table 1, the standardized factor loadings of each item are presented and as seen, the lowest value is .704. That is, all the items are closely related to the measured dimension, and therefore for parsimony the single factor model was retained.
Second, a latent class analysis was carried out (Table 2) to determine if the individuals who participated in normative actions are a group different from those who carried out non-normative actions (see also Medel et al., 2022, a related paper elaborated in the discussion).
Table 2
Goodness-of-Fit Indices of Latent Class Analysis
| Class | Loglikelihood | Akaike | BIC | Vuong-Lo-Mendell-Rubin | Lo-Mendell-Rubin Adjusted LRT |
|---|---|---|---|---|---|
| 1 | -3383.8 | 6831.6 | 6966.5 | ----- | ----- |
| 2 | -2600.7 | 5331.3 | 5605.3 | 23.2, p < .001 | 1558.5, p < .001 |
| 3 | -2440.6 | 5077.3 | 5490.3 | 151.2, p = .084 | 318.5, p = .085 |
This analysis indicates that when the existence of two classes of subjects is assumed, the goodness-of-fit improves compared to assuming the existence of a single class. If three groups are considered, the goodness-of-fit improves but not significantly (p > .05), so the most parsimonious solution is to suppose the existence of two groups.
This is consistent with previous results (Medel et al., 2022) showing that when attempting to divide the sample into three groups, separating those who did not participate in the protests, those who only participated in peaceful actions, and those who took part in violent actions, the latter two groups form a continuum of participation, rather than being totally differentiated.
When the two groups’ profiles are compared, as seen in Figure 1, one of them (Class 1: 74.6% of the sample, solid line) includes people who had low levels of participation in the protests and those with no participation at all, while the other group (Class 2: 25.4% of the sample, dotted line), is formed by the people who had higher levels of participation in the protests, without any distinguishable subset of protest behaviours identified. As can be observed, individuals in this class have a very high probability of having participated at least once (sum of the probabilities of having participated once and more than once) in marches and support to the demonstrators, but they also have a 50% chance of having participated at least once in barricades and even more than 30% chance of having been involved in confrontations with the police.
Figure 1
Latent Class Analysis
Note. Solid line = Class 1, non-participants; Dotted line = Class 2, participants.
The data are consistent with the approach generally taken by many scholars of collective action (e.g., van Zomeren et al., 2008), to identify one repertoire of actions in a movement. Therefore, in the analyses below to explain the participation of people in collective action, participation was aggregated into a single dependent variable of reported participation.
Explaining Participation in Protest
We conducted two analyses to predict participation: one testing the SIMCA model and a second examining an extended model which included positive emotions.
For the first test, we conducted mediation analyses by means of structural equations, using the WLSMV estimation method for ordinal data (Li, 2016), with participation predicted as a function of regional identification via two mediators, anger and efficacy (the SIMCA model: van Zomeren et al., 2008). Bootstrapping of 2,000 iterations was performed in order to estimate the confidence intervals.
The model (presented in Figure 2) obtained excellent goodness-of-fit indices, χ2(662, N = 499) = 1118.787, p < .0001; χ2/df = 1.69; RMSEA = .037; CFI = .974; TLI = .972, and explained 12% of the variance of the dependent variable.
Figure 2
Social Identity Model of Collective Action (SIMCA)
Nevertheless, the results were slightly different from what was predicted. There was a significant effect of group efficacy and anger on collective action, and no significant direct effect of identification with the regional group on individual participation in collective action in favour of this group (direct effect = -.129; SE = .072; p = .075). The indirect effects of identification were significant via group efficacy (indirect effect = .132, SE = .041, p < .001) and anger (indirect effect = .082; SE = .030; p = .006). The total effect of identification was not significant (total effect = .085, SE = .061, p = .160), although the total indirect effect was (IE = .082; SE = .030; p = .006).
Next, to evaluate the role that positive emotions can play in the prediction of collective action, we added them to the SIMCA model, as presented in Figure 3. The results indicate that this new model has excellent goodness-of-fit indices, χ2(772, N = 499) = 1522.405, p < .0001; χ2/df = 1.97; RMSEA = .044; CFI =.958; TLI = .955, and explained 33.7% of the variance of the dependent variable, which implies 21.7% more variance explained compared to the previous model, a strong effect size according to Cohen (1988). In other words, positive emotions increase the explanatory capacity of the model much more than could be expected just due to its increased complexity, highlighting the relevance of positive emotions to explain people's participation in protests.
Figure 3
Social Identity Model of Collective Action Plus Positive Emotions
As in the previous model, identification did not have a significant total effect in explaining collective actions (total effect = .085, SE = .057, p = .137). However, identification did have a significant positive indirect effect (total indirect effect = .443, SE = .096, p < .001) via group efficacy (indirect effect = .140, SE = .043, p < .001) and anger (indirect effect = .098, SE = .041, p = .015) and also via positive emotions (indirect effect = .205, SE = .044, p < .001). Finally, a significant negative direct effect was found (direct effect = -.358, SE = .102, p < .001), i.e., a suppression effect when positive emotions are considered alongside efficacy and anger.
Discussion
Having a representative sample of the population of the two main cities where this social movement took place, allowed us to confirm that even in mobilizations widely supported by the community, only a minority of the population actively takes part in the protests, since the latent class analysis only classified 25% of the respondents as participants. This study also confirms that even in a context of high conflict, where non-normative participation has been legitimized as a strategy to defend the ingroup against unjust aggression, radicalization is not the norm, but the exception, as evidenced by the fact that normative protest actions were the most used by participants. Study 1 aimed to understand the repertoire of collective action in a community sample, in a context where there was an important social movement that suffered severed repression and included normative and non-normative forms of protest. We found that, in the case studied, it is not easy to distinguish between participants who carried out normative or non-violent actions and those who engaged in non-normative or violent actions. That is, people who participated in non-normative actions, such as barricades or confrontations against the police, also participated in normative actions such as marches or cultural performance. Or, in other words, among those who participated in normative actions, some also participated in non-normative actions, which would indicate that we are rather in the presence of a continuum of participation, instead of two different participation types. This finding is different from what has been reported in other studies that have shown a clear distinction between normative and non-normative participation and between those people who take part in one or another type of action (Becker & Tausch, 2015; Tausch et al., 2011). The implication is interesting both theoretically and socially, highlighting that radicalization or the development of repertoires that include violent non-normative actions can occur when people feel that protest actions, which are a form of legitimate political participation in democratic contexts, face disproportionate and unfair police repression (Drury & Reicher, 2009), which has so far been reported mainly in qualitative studies.
In the next analysis, participation in collective action was directly predicted by group efficacy and anger, and indirectly by regional identification, consistent with the SIMCA approach. However, we found that positive emotions were also important positive predictors of participation, and mediated indirect effects of identification along with group efficacy and anger. The strong role of positive emotions in understanding the protest is notable, especially considering that these protests took place amid high police violence. This is consistent with the qualitative results obtained by Drury and Reicher (2005), which show that people value their ingroup more positively and create a stronger identification with other demonstrators when they perceive that protests are unfairly repressed by the police.
In order to understand in greater depth the results of Study 1, we carried out a new study, this time a qualitative one, reported below as Study 2. The present findings were also explored further quantitatively in Medel et al. (2022; a paper written after the present manuscript, but in print first due to the vagaries of the publication process). For the benefit of interested readers, we note here that when the most radical (6%) minority is examined in relation to other activists, it is distinguished by younger age, strong activist networks, and higher self-efficacy. However, in Study 2, we pursue a different research question: understanding the emergence of a non-normative repertoire for the broader community sample.
Study 2
Study 2 aimed to describe people's motives to participate in different kinds of collective actions in a repressive context, and the role played by positive emotions and group identity, from the perspective of the actors themselves. We were particularly interested in understanding how participants felt that both normative and non-normative actions came to be a legitimate part of their repertoire of protest, and the role played by emotions and identification with the regional group in participation.
Method
Participants and Instruments
The participants were selected by a purposeful sampling method, according to their role in the protests. Specifically, eight leaders of the movement, representing different types of organizations, were interviewed individually; and 16 people who participated in at least one protest action of the movement (identified through snowball sampling from the leaders), were interviewed in two focus groups (9 and 7 participants respectively).
A semi-structured interview guideline was used, covering the following topics: participation in the movement, feelings about participating, motives to participate, identification with the region and movement, reaction of significant others to own participation, evaluation of the movement, costs and benefits, previous participation, and willingness to participate again.
The interviews and focus groups were conducted by the first author in November 2014. Interviews lasted between 50 and 90 minutes each. The focus group lasted between 90 and 120 minutes. All were audiotaped and later transcribed verbatim. To protect the identity of the participants, no sociodemographic variables were recorded. A copy of the full set of interview questions is given in the Appendix (see Supplementary Materials).
Analysis
A qualitative content analysis (Schilling, 2006) was carried out to define categories and subcategories. The first stage of analysis was textual, selecting paragraphs, fragments and significant quotes from the transcriptions. The second stage was conceptual, to identify categories and subcategories which could be interrelated. Investigator triangulation was applied, so two researchers analysed the data independently. The software Atlas.ti 7 was used.
Results
Regarding the study objectives, no differences were observed between leaders and ordinary citizens, therefore, the results are presented jointly. The results that coincide with the SIMCA model show that people were motivated to participate by a very strong social identification with the regional group, and by appraisals of injustice, anger, and group efficacy. Nonetheless, in line with what was proposed by Hornsey et al. (2006), in this case, group efficacy is not limited to whether collective action would influence decision makers and win the desired outcomes for the group. These results show that rather than focusing only on short-term expectations of influence, group efficacy considerations also included how successful the collective action is in building a social movement for the longer term, and expressing values that are important to the participants. To illustrate those findings, quotes from the interviewees themselves (lightly edited for diction by a native English speaker) are presented in Table 3.
Table 3
Categories and Examples of Quotations on Variables Coinciding With SIMCA Model
| Category | Description | Examples |
|---|---|---|
| Regional Identity | Identification with the regional group, including territory and culture | …this is a super important identity topic for us, the territory... the Patagonian if he doesn't have his territory is not Patagonian, he lives longing it, for that reason, because he lacks a super important part of himself. Here the territory, the landscape is very ... So big, and we are very small, those of us in the region, not only in population -we don't come to one inhabitant per square kilometre- but everything is big, the rivers are immense, the waterfalls, everything is big, the trees are big, then we are like very little, then we are part of the land, we are not its owners. (Leader 3) |
| It is my town, where I grew up, and I feel that there is nothing else, that is, this is the maximum possible love, that there is nothing better than that, you know? (Participant 1, FG 2) | ||
| Injustice appraisal | Perception that the group is unfairly treated by the outgroup | …we are tired of this situation (…) we are the most abandoned across the country. What you earn here is not even enough to support one person. The situation is really difficult (…) we are totally abandoned. (Participant 8, FG 1) |
| This region is very rich in natural resources and everyone uses it when they need something. If they need energy in the north, 'let's go to Aysén', if they need minerals, fishing, whatever. But they always have us as last in the queue, because we are few, because in the number of votes we don't amount to much (...) then we were always like the poor brother, holding on. If there is a budget cut, the first region that is cut out is this region. We were always being harshly punished, and not only by the government in power at that one time, but in the past historically too. (Leader 7) | ||
| Group based anger | Anger felt in relation to the unfair treatment received by the group | The people were very angry, furious... (Participant 6, FG 1) |
| … and you get angrier and angrier when one arrives at Chile, as it is said of Puerto Montta to the north, and you find multi-lane roads, cloverleaf interchanges, immense infrastructures, and here they are not able to make even a trail? Are we worth so little? (Leader 4) | ||
| People were outraged, they were outraged and when you are outraged, you always tend to react, you get outraged, you get annoyed and you do something, or you stay at home raging, or go out to the street, and in this case, people went out to the street. (Leader 5) | ||
| Group efficacy | Evaluation that the collective action was successful for the achievement of certain goals | The people of Aysén are proud (...) people responded and they affirmed their worth. Many of us had the conviction that important things were going to be obtained for our community. I don't know the specific things that were asked for, but it could be something fundamental for the country. (Participant 6, FG1) |
| …whatever the government is in power, from now on, they will not see us with the same eyes that they saw us before (…) but they see a region that is tenacious, that is strong (…) some people can say ‘nothing was gained’, but here there is an achievement and a much more intangible value, which is this respect that is achieved, I believe, towards the region and towards the people. (Leader 4) | ||
| A couple of concrete things were obtained, but I think there are also some contextual things that help too, like the social fabric that was healed, it got better and where people came back to shake hands, I think that's fundamental. (Participant 2, FG1) |
aCapital of the region just north of Aysén.
Additionally, results highlight three specific elements relevant to involvement of people in actions that differ strongly from their everyday behaviour as citizens and violate the rules of the social system. These are: perception of repression as disproportionate, indiscriminate and unjust; positive emotions that involve social cohesion, positive emotional climate3 and a strong sense of group solidarity, that arise as a result of the experience of participation in itself and of coping collectively with repression; and legitimization of the active opposition to the police, due to the police actions are considered illegitimate, a finding similar to that reported by Adam-Troian et al. (2020), Ayanian et al. (2021) and Drury and Reicher (2005). In the Table 4 some quotes of the interviewees (lightly edited for English language diction) are presented as examples of these results.
Table 4
Categories and Examples of Quotations on Variables not Considered by SIMCA Model
| Category | Description | Example |
|---|---|---|
| Repression experience | Experience that the group was attacked by police | ...the attack during the night was a bombing, it was literally a bombing, all night (...) I heard the bullets, the bullets, the tear gas, but it was all night, all night. It was a feeling of horrifying helplessness. It made you want to cry because you felt you were at war. (Participant 1, FG1). |
| …they were police forces of your own country, who’re there to protect you and are attacking you. Then, who can protect you? Only you and for that you have to organize together. (Leader 3) | ||
| The repression, I felt that it was too oppressive, as if they thought themselves, I don't know, unaccountable. It was too brutal. I saw things too terrible from the pacosa here, how they beat people or they fired pellets. People screaming, crying. I think it was too much. It was notorious that these guys were from the outside, they were bad, they did it with pleasure, it seems. (Participant 3, FG2) | ||
| Positive emotions | Emotions of pride, joy, hope and solidarity | …There were daily marches, where the families expressed themselves (…), and they were huge marches! We are talking about thousands of people in a very tiny town, who marched every day… (Participant 6, FG1) |
| It was like everyone was like friends or brothers (...) they were all unified by the movement, drinking mateb there in the barricades (...) the unification of the people was very special, as they were all concerned about each other. Then it was quite emotional... (Participant 3, FG2) | ||
| I'm so proud of being a Patagonian, of being an Ayseninac and having fought for what we wanted. (Participant 4, FG1) | ||
| I met people who were such great value, all of them deeply committed, it was a spectacular thing. It was something that impressed me a lot… When they were to charge us with this anti-terrorist law, it was said: «anyone who wants to quit, can quit» Nobody quit. (Participant 8, FG1) | ||
| Legitimization of confrontation with the police | Violent actions carried out by ordinary people | … the people were joining this defence, helping, because it challenged the powerlessness of how they were treating people (…) everyone joined, there was not one of us who at some point didn't go to help throw stones for the police to back off, and it was always achieved. (Participant 1, FG1) |
| I saw people that I never thought I was going to see in a barricade. I never thought they were going to be holding wood to set fire to it and block the road. (…) I have female friends who work… you know, in a bank, I saw them outraged (…) children, grandparents, you know? teachers, students, everyone was there. (Leader 3) | ||
| The kids were also going to fight and helping to gather and throw stones, the younger ones gathered stones, or arrived alone with their trays with stones, and their moms watched them from a distance. (Leader 8) |
aDerogatory word for policeman. bA traditional South American caffeine-rich infused drink. cInhabitant of Aysén Region.
Discussion
Study 2 affirms what has been found in Study 1, that violent non-normative actions together with normative ones, were considered as part of a pool of options that people used in this particular intergroup conflict context: citizen mobilizations facing strong police repression in a democratic society. Results also affirm that, consistent with SIMCA (van Zomeren et al., 2008), the participation in the social movement arose from anger and perceptions of injustice in addition to a very strong regional identity and group efficacy, in the form of building a social movement and expressing values. As Study 1, Study 2 also affirmed the role of the positive emotional climate, arising from a strong sense of solidarity in a community under attack. Finally, Study 2 sheds light on how the group came to develop a repertoire of violent non-normative actions, which from actors’ perspective emerged as a legitimate response to illegitimate police brutality and state repression (Drury & Reicher, 2009).
This qualitative study develops our understanding of two motives of participation that are not considered by the SIMCA model, which are the experience of repression and a positive emotional climate. For respondents, participation led to social cohesion and a strong sense of group solidarity. Faced with repression, the neighbours formed a homogeneous group united against the outgroup, and confrontation became normative for the ingroup (Drury & Reicher, 2000). While perceptions of repression were not measured in Study 1, Study 2 suggest that the repression could have motivated participation by reinforcing the perception of injustice and strengthening regional identity and social cohesion, leading ordinary people to get involved in violent and non-violent collective actions. There is also an interesting dynamic to be explored of how violent responses are proposed, legitimised, and spread: a group-level analysis of norm change which must include both top down and bottom up processes (see also, González et al., 2021; Louis et al., 2020; Smith et al., 2020; Tausch et al., 2011; Thomas et al., 2014).
That participation in collective action could be related to positive emotions is consistent with studies about the outcomes of participation (Boehnke & Wong, 2011; Drury & Reicher, 2000; Vestergren et al., 2018). But with few exceptions (e.g., Klar & Kasser, 2009), most of past research addresses participation in normative actions, so the novelty, in this case, is that the positive emotional climate has occurred in a context where violent non-normative actions were being carried out in response to repression. The experience of strong camaraderie invoked by outsiders’ attack has been explored in studies of combat soldiers’ identity fusion (Whitehouse et al., 2014) and is consistent with a social identity approach that sees illegitimate, unstable intergroup relations as an antecedent of identification (e.g., Wright et al., 1990). Despite some advances (e.g., Asún et al., 2021, 2022), exploring the unique role of positive emotions and their effects in combination with anger, remains an important direction for future collective action research, especially in quantitative research.
General Discussion
Two studies examined the antecedents of collective action within the framework of a social movement that faced severe repression in a democratic context. The results of both studies stand in contrast to past research that has argued that distinct processes (e.g., Tausch et al., 2011) or distinct types of protestors (e.g., Kruglanski & Fishman, 2006), are associated with engagement in different forms of action. We certainly acknowledge that normative and non-normative actions and actors may diverge in other contexts, and it will be important to understand why. However, in the present data, we argue that it is useful to examine the actions carried out within the framework of this movement, considering the different actions as forming part of the same repertoire. Rather than being two different types of actions there was a continuum of high and low participation motivated by anger, efficacy and positive emotions, as well as indirectly by identification.
Another interesting result of this work is that, although the most accepted models to explain individual participation in collective actions only include negative emotions, Study 1, with a representative community sample, shows evidence that positive emotions have a significant and strong effect (see also, Sabucedo & Vilas, 2014). Study 2, with a sample of activists, also speaks to the importance of positive emotions. In a movement faced with police violence, as is the case studied, it could be assumed that the most present emotions would be anger, sadness, fear, anxiety (Fredrickson et al., 2003) and contempt (Tausch et al., 2011), and indeed, those emotions were present in the discourse of the interviewees. Nonetheless, at the same time they tell a story of a very positive emotional climate, which was of great relevance in dealing with repression. This co-occurrence of strong positive and negative emotions in stressful situations is not unknown (Folkman & Moskowitz, 2000; Fredrickson et al., 2003; Páez et al., 2007). Blends of contrasting feelings may coexist in the same moment in the same social group, as exemplified by Rimé (2007), referring to the case of September 11th attacks in New York. But as far as the authors of this paper know, this is the first mixed study in which positive emotions play such a relevant role in the prediction of participation in protests, surpassing the predictive capacity of anger.
In the present case, positive feelings towards the ingroup, such as pride, joy and empathy, bonded protestors together with each other and the community (see also, Goodwin & Pfaff, 2001; Poma & Gravante, 2016). Participation led to social cohesion, positive emotional climate and a strong sense of group solidarity. And, as pointed out by Drury and Reicher (2000), being common group members could lead to an expectation of common support. In this case, the expected support was the help to face the attacks of the outgroup (the police and the government), and to mobilise in self-defence.
The quotations allow us to understand from the activists’ perspective how the repertoire of forms of action broadened to include more radical or violent choices. Some individuals may be more prone to radicalize, such as younger activists (Medel et al., 2022). In the present analysis, however, ordinary citizens, without a prior history of violence, became involved in this type of collective action as a reaction to perceived repression and illegitimate policing. This results are consistent with studies conducted in different contexts, which have shown that perceived risk of government repression positively predicted anger, which, in turn, predicted increased willingness to engage in collective action (Ayanian & Tausch, 2016), and that police violence against peaceful protesters can trigger backlash (Adam-Troian et al., 2020; Drury & Reicher, 2000, 1999; Jetten et al., 2020). In turn, resistance to police violence by protesters can empower them and increase their well-being (Selvanathan & Lickel, 2019).
On an applied level, these results show that in democratic contexts the police repression of the protest may lead to processes of civilian radicalization. In consequence, it is possible to conclude that political leaders, and especially, governmental security agents, must be trained to improve their ability to allow freedom of expression and peaceful assembly in public areas (Anisin, 2016; Stott & Reicher, 1998). While future research must test the generalizability of the findings, we hypothesize that if the intergroup context in which collective action takes place is perceived as violent and unjust (as peaceful normative demonstrations facing repression in democratic countries), it could promote an intense emotional climate, which integrates positive emotions towards the ingroup and negative emotions towards the outgroup, and motivates collective action. This dynamic of mutual radicalization (Drury & Reicher, 2000; Moghaddam, 2018) could lead to a reconfiguration of the boundaries between what is considered normative and non-normative.
Regarding the role of regional identity, numerous studies have shown that politicized identity is one of the strongest predictors of collective action (van Zomeren et al., 2008), but that identification alone is not a sufficient condition for participation to emerge (Sabucedo et al., 2010; Simon & Klandermans, 2001). Consistent with these results, Chilean regional identities have also shown inconsistent links to protests (Asún & Zúñiga, 2013a; Muñoz et al., 2020), so it is possible to raise two hypotheses that could explain these findings. One possible contributing factor is that the identification of the participants with their region is so high (mean of 4.27 on a scale of 1 to 5) that there is not enough variance to distinguish between people who participated in the collective actions of the movement and those who did not. Another alternative hypothesis is that the Chilean regional identities, which are based principally in the appreciation of nature and landscape, and strongly associated with values of universalism and benevolence (Asún & Zúñiga, 2013b; Zúñiga & Asún, 2010, 2013), are not strongly politicised, or they are more politicized in some individuals than in others. From the social identity perspective, when people identify with a group, they enact the norms of that group, whether for collective action or inaction. The present null relationship between regional identity and political action could therefore mean that there is no consensus on the social norms to promote or inhibit political actions associated with belonging to the Aysen region. However, given that the qualitative data (Study 2) speaks of powerful community norms for action at the time, measuring norms explicitly and testing the moderation quantitatively may be useful to explore the boundary conditions of the present finding.
Limitations and Future Directions
Regarding the limitations of this work, the research has been carried out within the framework of a particular type of social movement, such as regionalist movements, so it is a pending challenge to test the hypotheses raised here in other types of movements. However, it is a great strength of the study to examine a real understudied mass movement using a representative community sample, and identifying past normative and non-normative actions rather than hypothetical intentions.
Despite the relevance granted to the experience of repression in this work, it has not been considered a variable in quantitative or comparative analyses. It would be of great interest to carry out similar studies including the variable perception of repression, or comparing contexts with and without repression, and examining whether and how the factor structures of normative and non-normative collective action change.
Another limitation is that data collection was conducted over two years after the movement took place. While this may have influenced the accuracy of the participants’ memories, given the high emotional impact these events had, the experiences were still very salient for the individuals. On the other hand, the fact that sufficient time had passed since the events for the police repression to have ceased, may have empowered participants to feel confident enough to talk openly about their participation in both normative and non-normative actions. Where conditions allow, conducting contemporaneous interviews with collective actors facing violent police repression would be a valuable direction of future research.
In summary, this work highlights the importance of conducting future research looking at the participation in different form of actions in different movements that face and do not face repression, paying special attention to group-based positive emotions and not just anger or negative emotions.
Conclusion
Using mixed methods, this research examined collective action within the framework of a social movement that faced severe repression in a democratic context, with samples of general population (quantitative study) and activists (qualitative study), measuring reported participation. The key findings are: 1) violent and peaceful forms of collective action appear here on a continuum rather than as distinct sets of behaviours; 2) traditional predictors of collective action (anger, efficacy and social identity) are also predictors of collective action in this context; but 3) positive emotions are relevant to this process as well; and 4) experiences of police violence and repression appear central to motivate people to move from peaceful to violent protest tactics.
By exploring these processes among the protesters of a real social movement and within the framework of state repression, the research contributes to understanding radicalization and collective action more broadly.