Hope and Anger as Mediators Between Collective Action Frames and Participation in Collective Mobilization : The Case of 15M

The study set out to integrate collective action models and emphasize the role of emotions. Whereas the importance of anger is indisputable, relatively little attention has been paid to the role of positive emotions, such as hope, in collective action research. Hence, the aim of the study was to explore the role of hope and anger as drivers of participation and involvement in collective mobilizations. A cross-sectional field study (N = 638) conducted right after the emergence of the 15-M socio-political protest movement in Spain assessed the emotions and beliefs of both demonstrators and those who took no part in the active mobilization. We hypothesized that anger and hope would sequentially mediate the relationship between collective action frames and participation in collective action. Furthermore, to test this premise, we ran two alternative sequential mediation models based on the social identity model of collective action (SIMCA) and the encapsulated model of social identity in collective action (EMSICA), but with emotions as mediators between collective action frames and intensity of participation. Both models fit the data well, suggesting the importance of considering multiple causal pathways, and showing that anger and hope sequentially mediate the relation between these frames and involvement in collective action. The results support the crucial role of hope in mobilizing individuals to take part in collective action.

) is an exceptionally relevant context for studying the different psychological motivators and processes that foster collective action.
Since the emergence of the 15-M movement in Spain, both activists and scholars have stressed its plural, inclusive and transversal character and the role of shared emotions.Indignation and hope played a pivotal role in transforming crises into protest (Castells, 2012;Laraña & Diez, 2012;Perugorría & Tejerina, 2013).Trust was destroyed, provoking a wave of indignation, but hope for a better future was crucial in turning grievances into action and motivated large numbers of individuals from diverse backgrounds to gather within physical spaces.Slogans like "Organize your anger", "Without hope there is no future" and "Yesterday angry, today hopeful" pointed out that although anger was necessary to mobilize protesters, it was hope that offered possibilities for change and the creation of new spaces within civil society (Páez et al., 2013;Sabucedo & Vilas, 2014).
Recent literature on collective action aims to integrate the constructs offered by different theoretical approaches and to include multiple predictors of engagement in collective action (e.g.Drury & Reicher, 2009;Gamson, 1992;Stürmer & Simon, 2004;Turner, Hogg, Oakes, Reicher, & Wetherell, 1987;van Zomeren, Postmes, & Spears, 2008).After having neglected the importance of emotions in protest movements for many years, extant models of collective action (Stürmer & Simon, 2009;Thomas, McGarty, & Mavor, 2009a;van Stekelenburg, Klandermans, & van Dijk, 2011;van Zomeren, Leach, & Spears, 2012;van Zomeren, Spears, Fischer, & Leach, 2004) have begun to incorporate group-based emotions.Particularly anger have been identified as separate but complementary pathway to collective action.At the same time, surprisingly few studies have directly examined the role of positively valenced group-based emotions in inspiring individuals to undertake collective action.Although many recent authors have pointed out the importance of considering positive emotions when explaining group dynamics (e.g.Mackie, Devos, & Smith, 2000;Smith & Mackie, 2008;Thomas, McGarty, & Mavor, 2009b), the majority of studies on collective action fail to sufficiently account for their crucial role in encouraging engagement in political protest.
One positive emotion which has recently aroused most interest is hope.In our view, hope is a motivating emotion capable of fueling resistance and protest.We ground our argument for incorporating hope into explanatory models of collective action in a wide range of sociological studies on the capacity of hope for generating change (see Braithwaite, 2004a;Courville & Piper, 2004;McGeer, 2004).From this perspective, hope can be viewed as a bridge between collective action frames and actual action.Thus, a belief in hope for a better future provides the foundation upon which individuals find it worthwhile to engage in efforts to improve their personal and social situation (Foster-Fishman et al., 2007;Hanna, 2002).More precisely, as argued by Aminzade and McAdam (2001) " [A]nger in the case of perceived injustice and hope regarding the prospects for change [...] appears to serve as the necessary affective bedrock on which many movements are built" (p.31).This suggests that negative emotions, although important (van Zomeren et al., 2004(van Zomeren et al., , 2012)), are not the only emotional predictors of collective action.
According to Frijda (1988), "anger implies hope", further hope is "yearning for better" even "when the odds do not greatly favor it" (Lazarus, 1991, p. 283).Therefore, we argue that the potential effectiveness of anger in generating social change could be explained by feelings of hope it lights up, which in turn mobilize people to act.
In this paper, building on the theoretical set up offered by theories of collective action and social identity, we propose a sequential mediation model that places hope at the core of the mechanism promoting participation and involvement in collective mobilizations.In this sense, and in light of increasing empirical evidence (Greenaway, Cichocka, van Veelen, Likki, & Branscombe, 2016;Páez et al., 2013;Sabucedo & Vilas, 2014;Smith & Leiserowitz, 2014), we attempt to examine the role of hope in accounting for the motivational potential of anger to integrate this perspective into other prominent social psychological perspectives.Thus, in the present paper, we extend existing research by examining the mediating role of hope and anger in the relationship between factors such as the perception of injustice, collective efficacy and group identification and intensity of participation within the context of the 15-M sociopolitical protest movement.We report the results of a field study that tests the ecological validity of current explanatory models of collective action (SIMCA and EMSICA) in a specific real-life context and explores the power of emotions for transforming beliefs into actual action.
First of all, we briefly review recent psychological perspectives on collective action and the importance of emotions in mobilizing collective action, taking into account those amplifiers of existing motivations.Subsequently we describe the relationship between anger and hope and hope and efficacy, and its particular importance in collective action research.Finally, we propose that hope is the key element in explaining the relationship between collective action frames, anger and commitment to social movements such as 15-M.

Frames of Collective Action
According to the literature on the social and psychological dynamics of collective action there are at least three components that constitute collective action frames: perceived injustice, social identity and efficacy (Gamson, 1992;Klandermans, 1997;van Zomeren et al., 2008).Therefore, collective action is likely to emerge when the social situation is interpreted as unjust (Touraine, 1978).Moreover, in line with relative deprivation theory, this appraisal must be shared within the members of a group and elicit emotional arousal (Kawakami & Dion, 1993;Leach, Snider, & Iyer, 2002).Thus, perception of the injustices themselves and moral indignation in connection with one's adversaries can incite feelings of mutual identification that enable the construction of an "us", a collective identity.According to Touraine's (1978) conceptualization of social movements, social identity theory (Tajfel, 2010;Tajfel & Turner, 1979), Klandermans' (1988) concept of collective action and the neo-Durkheimian perspective (Páez, Rimé, Basabe, Wlodarczyk, & Zumeta, 2015), social or collective identity plays a central role in encouraging collective mobilization.The salience of social identity has the quality of increasing members' identification with the group's values and beliefs, the perception of similarity, positive feelings (attraction, empathy, etc.) and prosocial relations of cooperation and solidarity (Turner et al., 1987).This is important for the movement, as it generates group cohesion and promotes unity of beliefs and coordinated action (Javaloy, Rodríguez, & Espelt, 2001).Finally, instrumental motivation and individuals' expectations concerning the costs and benefits of engagement in collective mobilization are important.As proposed by resource mobilization theory (McCarthy & Zald, 1977), people have to share a belief that the situation is unstable and that as long as certain essential resources are mustered, it can be effectively changed.In this sense, the perception of group efficacy is defined as the belief that the group is capable of changing the current situation through organized action (van Zomeren et al., 2004).
Recently, van Zomeren and colleagues (2008) carried out a meta-analysis of studies on collective action.Their results show that perceived injustice, identity and efficacy predict collective action in a near-similar fashion (and with a moderate effect size); and at the same time, they are positively inter-correlated.Furthermore, the authors found a greater effect size for politicized identity.On this basis, van Zomeren and colleagues (2008) proposed a comprehensive model called the social identity model of collective action (SIMCA).
The SIMCA suggests that while the collective action frames (efficacy, injustice and identity) are direct predictors of collective action, social identity reinforces the perception of injustice, emotions and the perception of efficacy (thereby acting as an indirect predictor of collective action).According to van Zomeren and colleagues (2008), identity functions as a bridge between efficacy and the perception of injustice, facilitating and boosting the group experience of injustice and efficacy.Furthermore, the SIMCA allows for the direct effect of identification to be stronger when identity is politicized.
An alternative model called the encapsulated model of social identity in collective action (EMSICA) has been proposed by Thomas, McGarty, and Mavor (2009a).This model inverts the causal relations established in the SIMCA.The EMSICA suggests that social identity mediates the effect of perceived injustice and efficacy on collective action, with perceived injustice and efficacy providing the basis for the emergence of social identity.Since experiences of injustice and efficacy tell us about who "we" are, and social identity can capture (and mediate) these two emotional reactions, their direct effects are not so substantial in the model (Thomas et al., 2009a).Although both models assign social identity a central role in driving collective action, it is important to emphasize that they cater to different contexts.Whereas the SIMCA was derived from many primary studies on traditional forms of collective action (involving low-status or otherwise deprived groups, including social movements), EMSICA refers to more emergent, opinion-based social identities.
The causal relation is disputable: The experimental studies by Swaab, Postmes, van Beest, and Spears (2007) suggest that the incorporation of one of the frames can lead to the development of the others, and that shared social identity can be both a product and a precursor of shared cognition (Swaab et al., 2007) or collective action (Thomas et al., 2009a).While van Zomeren et al.'s (2008) meta-analysis did not include the opinion-based groups closest to the profile of participants in new social movements, the empirical study by Thomas, Mavor, and McGarty (2012) compared the SIMCA and EMSICA in three independent samples of shared-opinion groups.The authors concluded that neither of the models is clearly superior to the other, and that either model's superiority may depend on the socio-cultural context.Given the importance of confirming the generalizability of these previous findings, we attempted to investigate the SIMCA's and the EMSICA's applicability within the context of the emergence of the 15-M socio-political protest movement in Spain.

The Importance of Emotions in Social Mobilization
Since Gamson's (1992) proposal regarding the important motivating role of anger in political protest, emotions are no longer seen as irrational.Recent studies have stressed the importance of group emotions in people's willingness to form part of a collective action (Mackie et al., 2000;Musgrove & McGarty, 2008;van Zomeren et al., 2008).Also, as demonstrated by van Stekelenburg and Klandermans (2007), emotions may be viewed as amplifiers of existing motivations, rather than as a separate pathway to collective action (van Zomeren et al., 2004).Moreover, human behavior is likely to be influenced by automatic emotional judgments, as well as by conscious cognitive calculation (see Bargh, 1984;LeDoux 1996;Loewenstein, Weber, Hsee, & Welch, 2001;Slovic et al., 2002).Thus, intense emotions are closely linked to action (Frijda, 1988;Maitner, Mackie, & Smith, 2006) and likely to have a powerful direct impact on actual behavior, complementing deliberative decision-making or cost-benefit calculations (Loewenstein & Lerner, 2003;Pearlman, 2013).According to intergroup emotions theory (Smith, 1993), shared appraisal leads to group emotions which in turn foster action tendencies.Thus, the causal link between collective action frames and group emotions may be recursive.On the one hand, being part of a specific group may facilitate the experience of group-based emotions (see Mackie et al., 2000;Postmes, Haslam, & Swaab, 2005;van Zomeren et al., 2008), something which is particularly true in the case of social identities based on established long-term social categories.On the other hand, however, shared emotional experiences and social sharing of emotion may trigger psychological group formation (see Livingstone, Spears, Manstead, Bruder, & Shepherd, 2011;Swaab et al., 2007).Indeed, both experimental (Kirschner & Tomasello, Wlodarczyk, Basabe, Páez et al. 203 2010;Wiltermuth & Heat, 2009) and longitudinal studies (Páez et al., 2015;Rimé, Páez, Basabe, & Martínez, 2010) have confirmed that actual participation in coordinated collective gatherings reinforces group identity and cohesion.According to the available literature (see Kessler & Hollbach, 2005;Smith & Mackie, 2008;Thomas, McGarty, & Mavor, 2009a), the causal relationship between group identity and group emotions should be seen as dynamic and bidirectional, since it is likely to be influenced by social context and will probably vary over time.

Hope and Anger
Injustices are a constant feature of society, and even though people may be aware of them, they do not necessarily mobilize themselves (Goodwin & Jasper, 2006).One of the ways to counter apathy is through anger which grows out of a perception of the illegitimacy of the status quo and the discrepancy between the actual and the desired situation.According to Gamson (1992), anger fuels the injustice frame in order to counter a sense of legitimacy.Group emotions of anger are also activated when people perceive threats and identify with the disadvantaged group.In accordance with Smith and Mackie (2008) and Fischer and Manstead (2008), directing people's anger towards another group may be associated with a tendency for aggression towards that group.Furthermore, anger accounted for reinforcing mobilization in the case of farm workers in Holland and Spain as shown by Sabucedo, Durán, Fernandez, Romay, and Dorna (2007).According to Tangney, Stuewig, and Mashek (2007) anger about injustice and moral indignation can also motivate people who are not so strongly affected by a crisis to mobilize themselves out of solidarity with their more severely affected fellow citizens.Moreover, it has been widely acknowledged that anger can act as a bridge between sensitivity to social problems and active commitment to a social movement (Traïni, 2009;van Stekelenburg et al., 2011;van Zomeren et al., 2004van Zomeren et al., , 2012)).
Despite all the empirical evidence, it seems like anger may not always be enough to "fire up" the energy necessary to act.Certain positive emotions can also prompt people to fight for a better future.According to Jasper (1998), anger, indignation and fear are transitory responses that are shaped by underlying affects which can lead people into collective action or drive them away.Pearlman (2013) suggested that emboldening emotions that promote optimistic assessment, acceptance of risk and feelings of personal efficacy may encourage people to prioritize dignity and increase their willingness to engage in action.The motivational role of positive emotions was examined by Sabucedo and Vilas (2014) in a survey of university students carried out in a setting of cut-backs in education.
This study found that positive emotions, such as hope, pride and optimism, acted as important mediator variables and explained the relationship between anger and intention to participate in a protest action.Furthermore, social movement scholars emphasize the coexistence of positive emotions and anger in real-life collective actions.Castells (2012) claims that the movements which spread across the Arab countries, Europe and the USA in 2011 were triggered by emotions of outrage and hope.More specifically, Pearlman (2013) argues that pride, anger and solidarity played an important role in the 2011 uprisings in Tunisia and Egypt.Davou and Demertzis (2013) underline the importance of hope and perceived political efficacy in motivating Greek citizens to become involved in political action.These findings highlight the fact that negative and positive emotions are interrelated and may both directly or indirectly inspire engagement in collective action.
Of all the emotions that can help people to find the courage to overcome fear and apathy, perhaps the most important is hope.As stated by Ahmed (2004), "hope is what allows us to feel that what angers us is not inevitable, even if transformation can sometimes feel impossible" (p.184).Anger alone lacks prospects for future change; therefore, we argue that it may be the component of hope that prevents anger from leading to resignation.The study by Aminzade and McAdam (2001) suggests that both anger and hope are necessary for the mobilization of individual and collective forms of protest.These authors assert that: "[I]n and of itself, anger is not likely to produce organized collective action, but rather other (usually individual) forms of resistance [...].It is only when anger gets joined with hope that the forms of action we normally associate with social movements and revolutions are apt to take place" (p.31-32).Despite the theoretical associations between hope and anger, to our knowledge, no previous empirical study has directly assessed the relationship between these two constructs.In the present study, we seek to redress this situation by examining both the direct effect of anger and hope on the intensity of participation in socio-political movements, and the sequential indirect effect of anger through feelings of hope.
In our view, hope is crucial to the act of protest.Moreover, a substantial amount of recent research on social movements emphasizes the role of hope in influencing political behavior.From a philosophical perspective, the mere act of hoping is already an initiation of "a new beginning".In this sense, hope is not only a belief that the present circumstances can be different, but also that they should and will be different (O'Brien, 2008).In a similar approach, Billias (2008) defines hope "as a future orientation of ongoing agency" (p.17), implying that one cannot act without hope.Social scientists define hope as more than a simple desire for a certain outcome.Rather, it is a desire that is inseparable from agency (McGeer, 2004, p. 103).More importantly, hope is necessary for social change as it sustains action and prevents disengagement in the face of difficulties or lack of control.According to Larsen et al. (1993), a belief that the present or future can only have negative outcomes would make people act according to these expectations, thus proving themselves correct.When there is no hope, agency is relinquished, resulting in a belief that action will not make any difference.Furthermore, hope requires anticipation and expectation that the desired outcome will be achieved (Drahos, 2004), and therefore provides a plan of action.Thus, hope could be seen as "a renewable resource for social change" (Courville & Piper, 2004, p. 58), as it turns individual hopes into group hopes, while at the same time spreading from the group to the individuals in that group, increasing the probability of successfully altering the status quo.Collective hope has the power to move people towards a "social script […] in which we are expected to be active and responsible participants contributing to vibrant civil society" (Braithwaite, 2004b, p. 7).
Hope has been the object of considerable empirical research in psychology, yet little attention has been paid to its relationship with collective action frames and participation in social movements.Some scholars have begun to theorize on the role of collective hope in overcoming adversity, emphasizing particularly that it is linked to the aspiration and expectation that, with a reasonable degree of probability, certain negative circumstances can and will be changed (Bar-Tal, Halperin, & De Rivera, 2007;Fredrickson, 2009).Hope has been found to be positively associated with moral emotions and negatively associated with shame (Williamson, Sandage, & Lee, 2007) and distress (Scioli, Ricci, Nyugen, & Scioli, 2011).Hope has also been identified as an important element of the coping processes, as it arises in situations in which people expect the worst but still try to make things better anyway (Lazarus, 1991;Snyder, Rand, & Sigmon, 2005).As such, hope was found to be associated with cognitive flexibility and creativity, which in turn facilitate planning, goal setting and further action.Accordingly, Hobfoll et al.
(2007) found that hope was crucial for achieving favorable recovery outcomes after trauma exposure and to forge resilience to adversity in the long term.Moreover, as suggested by Braithwaite (2004a), hope fosters social inclusion and empowerment among marginalized groups according to Courville and Piper (2004).Similarly, researchers have begun to investigate the role of hope in intergroup contexts.More specifically, Halperin and Gross (2011) found that emotion regulation and cognitive positive reappraisal were positively associated with hope, which motivated humanitarian support for others.These findings are especially important as they provide evidence of the positive effect of hope in a context of intractable conflict.Recent studies have confirmed that hope and positive emotions, even when coexisting with negative ones, are an important predictor of the endorsement of positive  & Gross, 2013).Furthermore, research has shown that as an emotion, hope is associated with social engagement and cohesion (Gee, Khalaf, & McGarty, 2007), as well as with collective efficacy and social mobilization (Páez et al., 2013;Sabucedo et al., 2007).Therefore, as Sabucedo and Vilas (2014) point out, sharing the same objectives and believing that they can be accomplished fosters both hope and a sense of pride in being a part of a movement capable of making the desired change possible.

Hope and Efficacy
The importance of hope that the current situation can be improved cannot be underestimated.As mentioned above, the most important features of hope are positive expectations about the possibility of achieving certain objectives in the future and positive feelings of outcome anticipation, regardless of the negativity of the present situation.Therefore, hope implies goal setting, planning and actual mobilization (Cohen-Chen et al., in press).
Hope has the power to encourage people to successfully overcome adversity, or as Snyder (1994) puts it: "hope is more than distancing oneself from and delimiting the impact of failures; hope is the essential process of linking oneself to potential success" (p.18).Although, the concept of hope is closely related to the concept of collective efficacy, each focuses on different aspects.According to Bandura (1986), efficacy is related to the belief that certain goals "can" be achieved, while hope focuses on the belief that "yes" the goals "will" be achieved in the future.
For example, research by van Zomeren et al. (2004van Zomeren et al. ( , 2008) ) shows that group efficacy beliefs increase collective action tendencies and group identification, pointing out that "can" leads to "we".Likewise, empirical evidence (Magaletta & Oliver, 1999) suggests that the essential components of hope, "will" and "ways", are related to efficacy and optimism, but are not identical constructs.In our view, the belief that certain goals can be achieved has the power to instill the will to achieve them and therefore leads to increased engagement in protest.Given that hope concerns future goals and the plausibility of achieving them, we suspect it may be seen as a potential key element for understanding engagement in collective action.In this sense, we conceptualize collective efficacy as the perception of the collective capacity to gather the necessary recourses (McCarthy & Zald, 1977), for example the capability to organize and execute the specific actions required to enforce social change through joint effort.
However, being capable of changing the current situation does not necessarily imply being hopeful or optimistic about the future.Therefore, we argue that the group members can perceive that the group is capable of changing the current situation through organized action (van Zomeren et al., 2004); in consequence, this objective evaluation may foster the expectancy of success and positive feelings, which characterizes hope, which in turn fosters stronger commitment to norms for action.Evidence from previous research supports this pathway: A study on the 15-M movement found that the effect of collective efficacy on participation was mediated by hope (Páez et al., 2013).In short, given that hope often serves as the spark that triggers action, we hypothesize that it fuels the collective efficacy frame and transforms it into actual mobilization.

The Present Study
Based on the above considerations, in the present study we seek to make headway regarding the identification of direct and indirect predictors of participation in collective mobilizations, such as 15-M.We highlight the importance, coexistence and interrelatedness of positive and negative group emotions (Halperin et al., 2013;Páez et al., 2013, Sabucedo & Vilas, 2014) and their relationship with collective action frames (Thomas et al., 2009a;van Zomeren et al., 2008).On the basis of the theoretical and philosophical sources cited earlier, we hypothesize a sequential mediator effect of hope in the positive relationships between anger, collective action frames and participation in collective action.Thus, the general aim of the study is to provide an empirical mediational test that explores

Participants and Procedure
Participants in this study were 638 individuals (43.4% men) aged between 18 and 73 (M = 31.66,SD =11.39) years, of whom 48.7% were in paid employment, 33.4% were university students, 14.4% were unemployed or retired, and 1.1% were housewives.Around half (52.8%) had a university degree, and roughly 70% had some kind of higher education.Spanish nationals accounted for 90.4% of the sample.Participants were residents of Madrid (37.2%),Barcelona (20.3%),San Sebastián (19.2%),Bilbao (11.1%) and Vitoria (7.7%).A total of 61% of those who participated in the survey claimed to have been involved in activities related to the 15-M movement on one or more than one day (see Table 1 for descriptive data). i We interviewed participants directly in the 15-M movement camps or during demonstrations in different Spanish cities (San Sebastián, Vitoria, Bilbao, Madrid and Barcelona), as well as at universities and in the areas surrounding the 15-M camps during the month of June, 2011.The interviews were carried out by psychology PhD students.
The research team instructed the interviewers to spread out through the camps or protest areas once the crowd had deployed fully and select every n-th participant in order to guarantee that every person had an equal chance of being selected.This ratio was based on the estimated number of people gathered in the camps or protest areas, and determined how many rows of respondents were skipped before every n-th participant was selected (see Klandermans et al., 2011).Next, the interviewers administrated a short face-to-face survey to those who voluntarily agreed to participate in the study.All study participants signed written informed consent.The research team was responsible for alphanumerically coding the registered data, ensuring anonymity.All procedures were approved by the University of the Basque Country's Ethics Committee for Research Involving Human Beings.

Measures Socio-Demographic and Socio-Political Variables
We gathered data on participants' sex, age, nationality, employment situation (Employed, Unemployed, Student, Other), educational level (1 = Primary and Middle School Education only, 2 = Further Education, 3 = Diploma or Graduate level) and ideological position: "Please circle a number (from 1 to 7) indicating your political views", the range being from 1 (Extreme left) to 7 (Extreme right).We also inquired about how they had voted in the elections held on May 15 th , 2011 (see Table 1).

Perception of Injustice
As indicators of perceived injustice in relation to the current social situation we developed four items to assess endorsement of the specific goals and grievances of the 15-M movement.We asked participants to what extent they agreed with the need to protest against: "the current political, economic and social scenario"; "corruption among politicians"; "corruption among bankers and business people"; and "the vulnerability of ordinary citizens vis-à-vis the 'powers that be' in general and the holders of socioeconomic and political power in particular".Response options ranged from 1 (Not at all) to 7 (Very strongly).Cronbach's alpha coefficient of reliability was high (α = .88).

Perception of Efficacy of Collective Action
We asked participants to think about the experience of participation in demonstrations/concentrations related to 15-M or in others group activities.We inquired about the perception of group capability for changing the current situation through the organized action of the 15-M movement with the following statement, created ad hoc: "We realized we were perfectly capable of achieving our aims", with a response range from 1 (Not at all) to 7 (Very strongly).

Group Identity
We used three items (based on Páez et al., 2013) to assess the extent to which respondents shared beliefs of group unity, satisfaction and empowerment.Participants evaluated to what extent they felt that: "We acted as a single person.We experienced and shared a moment of unity", "We felt proud to belong to this group", "We felt transformed and much more convinced of our ideas".The response range was from 1 (Not at all) to 7 (Very strongly).Cronbach's α coefficient was .90.

Emotions
We asked respondents to evaluate to what extent they felt anger and hope when thinking about the current social situation.Respondents were asked: Indignation, righteous anger -"Do you feel indignation or anger about the current social situation?"and hope -"Do you have high hopes that the current social situation will improve?"(response range from 1 [Not at all] to 7 [A great deal]).

Personal Participation
To measure the intensity of participation we asked: "How many days did you spend in the demonstration?"We used this continuous variable, which ranged from 0 to 24 days (M = 3.75; SD = 5.74), for all subsequent analyses described in this paper. ii

Analytical Strategy
With the aim of exploring the relationships between the frames of collective action and the mediating role of emotions, we used structural equations modeling analyses with Mplus 6.11 (Muthén & Muthén, 2010).The estimation procedure applied was maximum likelihood.Considering the lack of multivariate normality, we applied the bootstrap method (N = 5000), which consists of the repeated extraction of samples from the dataset and the estimation of the desired statistic in each of the resampled datasets.It is a non-parametric approach to parameter estimation, and hence free from assumptions about the normality of the variables' distribution or the sampling distribution of the statistics (see Efron & Tibshirani, 1993).Standard errors and confidence intervals based on a bootstrap sampling distribution are calculated for each of the parameters or statistics.If the values of the estimated frames of efficacy and injustice and collective action, with hope and anger being included as mediators as in the previous model.We also allowed a separate path from anger to hope.Table 3 shows the fit indexes for the proposed models.The TLI and CFI and the SRMR were highly satisfactory for both Model 1 and Model 2. The AIC was very similar for both models, suggesting that neither of them can be considered clearly superior in terms of representing the relationships posited.We now report the results for both models.As can be seen in Figure 1, identity was positively related to perceived efficacy and injustice.Moreover, the direct effect of identity on intensity of participation did not quite reach conventional levels of statistical significance (p = .086);however, it was significant based on 95% bootstrapped confidence intervals (B = .10,SE = .06,p = .086,95% CI [0.004,0.193]).The direct effect of the frames of injustice and efficacy on intensity of participation did not emerge as significant.
On the other hand, the effects of efficacy and injustice on emotions were significant and large; more importantly we found a direct relationship between hope and intensity of participation.Considering the lack of any direct effects of efficacy and injustice on intensity of participation we examined indirect effects.We observed that efficacy had a significant indirect effect on intensity of participation through hope (B = .07,SE = .02,p < .001,99% CI [0.052, 0.398]) and injustice had an indirect effect on intensity of participation sequentially through anger and hope (B = .01,SE = .01,p = .023,95%CI [0.010, 0.177]).We also found a significant indirect effect of identity through efficacy and hope (B = .05,SE = .01,p < .001,99% CI [0.048, 0.371]) and through injustice, anger and hope (B = .01,SE = .01,p = .038,95%CI [0.001, 0.033]).Furthermore, the results revealed an indirect effect of anger on intensity of participation through hope (B = .03,SE = .01,p = .011,95%CI [0.029, 0.265]).
On the other hand, and contrary to our expectations, the indirect effect of injustice through identity did not reach significance based on bootstrapped confidence intervals (B = .01,SE = .01,p = .121,95% CI [-0.001, 0.027]).
Nevertheless, we did find significant sequential indirect effects for injustice on intensity of participation through anger and hope (B = .01,SE = .01,p = .023,95% CI [0.002, 0.023]), as well as through identity, anger and hope (B = .01,SE = .01,p = .096,99% CI [0.0001, 0.001]).Just like in Model 1, the direct effect of identity on intensity of participation was marginally significant (B = .10,SE = .06,p = .086,95% CI [0.004, 0.193]), and identity was indirectly associated with intensity of participation through anger and hope (B = .01,SE = .01,p = .076,95% CI [0.0001, 0.009]).As in the previous model, anger had an indirect effect on intensity of participation through hope (B = .03,SE = .01,p = .011,95% CI [0.000, 0.052]).The results revealed that identity partially encapsulated the effect of the perception of injustice and efficacy on intensity of participation.Moreover, the indirect effect of perceived efficacy through hope, and the indirect effect of injustice and identity sequentially through anger and hope, suggested that hope could be considered as one of the most important direct predictors of intensity of participation.
In the two models, both injustice and efficacy had substantial positive effects on the intensification of the emotions of anger and hope, respectively.More importantly, hope completely mediated the effect of anger on intensity of participation.Furthermore, in both models we found indirect effects of the predictor variables, suggesting that hope could be considered the key component that explains the relationship between the collective action frames, anger and the intensity of participation.

Discussion
The results of the present field study with a large sample of individuals contacted directly at either the 15-M movement camps or their surrounding areas generally support theoretical claims about the crucial role of hope in mobilizing individuals to take part in collective action.Hope triggers participation by framing the social issue as a solvable problem (Cohen-Chen et al., in press;Páez et al., 2013).These findings contribute to the limited yet growing body of empirical literature on hope in collective action and, to our knowledge, offer the first empirical test of associations between collective action frams, anger and hope, and actual participation in protest movements.
The results also provide scientific support for the theoretical works of Ahmed (2004) and Aminzade and McAdam (2001), and the empirical findings reported by Sabucedo and Vilas (2014), illustrating that hope accounts for the effect of anger in evoking protest participation.
Furthermore, our study replicates the findings of previous studies in other socio-cultural contexts (Thomas et al., 2009a;van Zomeren, Leach, & Spears, 2010;van Zomeren et al., 2008).The results support the dynamic, multifaceted and interactive conception of identity (Thomas et al., 2012), which is also in line with the neo-Durkheimian model (Páez et al., 2015).Collective action can be seen as both the product of some shared beliefs and identities, and as causing the emergence of a shared collective identity.Nevertheless, in the particular case of 15-M; and as pointed out by Thomas and colleagues (2009a); when people experience a strong reaction to injustice and at the same time believe that group efforts can help to achieve goals and combat the injustice, it is highly likely that an identity will emerge based on the shared emotional reaction (Mazzoni, van Zomeren, & Cicognani, 2015).Indeed, the comparison of recently proposed models of collective action (Thomas et al., 2009a;van Zomeren et al., 2008) permitted us an evaluation of the facilitating and encapsulating effect of collective identity in collective action.Although the two models in question fit the data well, we can conclude that group identity motivates participation not only directly, but also indirectly through feelings of anger and hope.When shared grievances are translated into shared identity and a vision that involves prospects for social change, hope is what inspires people to achieve their goals (Ahmed, 2004;Cohen-Chen et al., in press;Snyder, 1994;Snyder et al., 2005).
Our study extends previous findings by introducing emotions of hope and anger as amplifiers of collective action frames.Our results revealed that regardless of the model in question hope acts as a mediator reinforcing participation and (at least partly) explains the influence of a shared frame of injustice, collective efficacy and identity on intensity of participation.
These results were congruent with evidence from a field study conducted by van Stekelenburg et al. (2011), who found that emotional response and group-based anger mediated between instrumental motives and motivational strength, whereas identity and ideological motives had a direct effect on protest participation.Furthermore, our findings highlight the relative importance of affective variables over non-affective ones in predicting collective action (van Zomeren et al., 2008).More importantly, our findings revealed that hope needs to be taken into consideration when explaining relationships between motives and actual participation.On the one hand, we found that hope, together with identity and indirectly with efficacy, is strongly and positively related to intensity of collective partici-pation.On the other hand, anger and hope are intensified by both injustice and efficacy, respectively.Moreover the indirect effect of efficacy through hope and identity is greater than that of injustice through anger and hope.
These results confirm and extend recent claims by Sabucedo and Vilas (2014) that positively valenced emotions could be considered as mediators between anger and involvement in collective action.Therefore, we can conclude that increased engagement in collective action is not possible without anger, but anger is fueled by hope in order to be an effective agent of change.
In the case of the present study, we explored the role of hope as a potential direct predictor variable of participation, underlining at the same time that it fully mediates the effect of anger on participation.Therefore, we found empirical evidence for Castells' (2012) idea that the movements of 2011 were triggered by networks of outrage and hope.Moreover, such results may contribute significantly to recent literature on the effects of hope in conflictive contexts (Cohen-Chen et al., in press;Halperin et al., 2013;Páez et al., 2013;Sabucedo et al., 2007).
Furthermore, it is important to underscore that the study explores the dynamics of action in relation to social issues that are of defining significance within the recent contemporary social context.Moreover, it points to the vitality of collective action models developed in social psychological experiments, and extends the SIMCA and EMSICA by highlighting the role of emotional reactions as mediators of action and mobilization.
The present study also has some limitations that need to be taken into account when discussing the implications of our findings and designing new research directions.First, the measure of social identity that we implemented in the present study was based on situational identification with other demonstrators or the group, rather than on identification with a specific social category.We argue that the identity which emerges from interaction with the co-present participants elicits similar outwardly focused action tendencies aimed at overcoming an obstacle or injustice.That said, social interaction has been found to be crucial for identity formation (Postmes, Haslam, & Swaab, 2005;Postmes, Spears, Lee, & Novak, 2005) and to help identities become a powerful tool for promoting social change (McGarty et al., 2014).Accordingly, and in line with the predictions of self-categorization theory, intense interaction with the group of demonstrators would be associated with enhanced identification and positive emotions (Novelli, Drury, Reicher, & Stott, 2013;Páez et al., 2013).Thus, although we asked our participants about situational group identity or perceived unity with the group of participants, we believe that our measure accurately assesses the construct.Second, collective efficacy and hope were measured with a single item.In view of the fact that the collective efficacy measure showed no collinearity with the affective measure of hope (the correlation between the two variables was only moderate, see Table 2), we are confident that there is a difference between the affective and non-affective measures of predictors of collective action.
Of course an important limitation concerns the correlational nature of the study; this prevents us from inferring causality from our results.The cross-sectional and retrospective design does not allow for causal claims about the associations between collective action frames and participation, or the impact of emotions (anger, hope).
Nevertheless, we would like to emphasize that research on real-world events in general is very difficult to carry out.First, events like spontaneous protest demonstrations are sometimes unpredictable.In this regard, future research should test those associations in different samples and with longitudinal designs that incorporate measures before and after the participation in the protest demonstration, as well as follow-ups in order to evaluate positive and negative changes over time.
More precisely, in the case of our study, shared identity, perceived injustice and collective efficacy and emotions were postulated as predictors of collective action.However, participation in a demonstration may have consequences for the emotions and perceptions involved (Novelli et al., 2013;Valdesolo, Ouyang, & DeSteno, 2010) and may motivate engagement in future actions (Becker, Tausch, & Wagner, 2011).Nevertheless, our findings revealed that a reverse causal model (in which the effects of intensity of participation in collective action on group identity, perceived efficacy and injustice are mediated by hope and anger) did not produce a better fit and showed fewer significant mediation effects.We are aware of the fact that a longitudinal study of the role of positive and negative emotions in the relationship between collective action frames and participation would surely provide a more convincing assessment of causality.Therefore, given that the main aim of this study was to explain intensity of participation, we decided to follow our theoretical framework which postulated the role of hope and anger as amplifiers of collective action frames (Sabucedo & Vilas, 2014;van Stekelenburg & Klandermans, 2007;van Zomeren et al., 2004).
We acknowledge that we have necessarily limited our analysis to collective action frames and the emotions of anger and hope.In particular, we have limited our analysis to the positive emotion of hope.Despite the theoretical associations between hope and action tendencies directed towards achieving desirable goals, empirical evidence is still scare.A recent study by Greenaway et al. (2016) confirmed the motivating potential of hope, over and above the effect of other emotions, and illustrated the power of this emotion in increasing support for social change, although the authors did not focus their analyses on actual political behavior.We are aware that there may be other positive emotions that encourage participation.Pride, for example, may have the power to prompt individuals to act in accordance with their principles, irrespective of their perception of the efficacy of group actions (Tausch & Becker, 2013;Vilas & Sabucedo, 2012).In this sense, the pride generated by doing what one believes to be morally correct may transform apathy into action.In this case, we would be speaking of a moral obligation.Future research should determine the specific role of individual moral convictions, moral social identity and group-based pride for motivating political engagement.
Finally, it is also important to mention that we tested our hypothesis within the context of the emergence of a specific movement.We consider the 15-M protest movement an example of a new type of political participation that spread rapidly throughout the world in 2011.Nevertheless, future studies should determine the role of hope, or positive emotions in general, in the mobilization of political action in different mobilizing contexts.

Conclusion
This study sets out to help clarify the explanatory mechanisms of collective action and stresses the role of group emotions of different valence in shaping actual protest participation.We based our theoretical background on the assumptions and models developed within the frameworks of theories of collective action, social identity theory and the neo-Durkheimian perspective (Páez et al., 2015;Páez, Rimé, Basabe, & 2005).In brief, the results of this study, while supporting the widely accepted view that anger is activated when people perceive injustice, collective efficacy and identity, extend existing collective action research by providing initial evidence that positive emotions may be crucial for motivating political participation.Overall, we argue that despite the recognition of the importance of emotions in collective action, the current body of existing literature still does not sufficiently account for the role of positively valenced emotions.Whereas Thomas, McGarty, and Mavor (2009b) have argued for the importance of advantaged groups' prosocial emotions in transforming apathy into action for social change, we believe that protest participation, although initially motivated by emotions of a negative nature is, encouraged and sustained rather through its coexistence with positive emotions, such as hope.To the best of our knowledge, this is the first study to test the sequential mediating effect of anger and hope between collective action frames and actual participation.Importantly, our findings illustrate that self-transcendent emotional states have the power to transform beliefs and tendencies into actual action.Hence, organized action and social change can be successful if people's feelings of hope are strong enough to constructively channel the mobilizing force of negative emotions (arising from adverse social situations).

Notes
i) Although around 70% of Spaniards reported to share the grievances and sympathize with 15-M (Cols, 2011), this does not mean that they would declare their disposition to participate, nor that they would eventually play an active part in the movement.Therefore, we were interested in examining the attitudes, beliefs and emotions of those who were and who were not actively involved.Those who indicated that they did not take active part in any of the activities related to15-M were instructed to think of the most important group activity or mass gatherings in which they had participated in the last two weeks.
ii) In preliminary analyses we used a categorical measure of intensity of participation, which presented exactly the same patterns of results.Considering the reviewers' critiques and the fact that a continuous variable (number of days spent at the demonstrations) may represent more directly the mechanisms that drive people to participate we re-run all the analyses with the continuous measure "intensity of participation" as a dependent variable.Moreover, this outcome variable assessed actual involvement rather than a declaration of intention to participate, which should minimize a social desirability bias.Furthermore, in order to confirm our findings, we tested our model also with other measures of active participation which were included in the questionnaire but are not reported in this manuscript: questions on political coping, previous participation in demonstrations, and social sharing about social movements.All these outcome variables showed similar results, confirming the robustness of the findings presented in the current version of the manuscript.
whether hope and anger, triggered by perceptions of injustice, collective efficacy and identity, directly or indirectly explain participation and involvement in the collective actions of the 15-M movement.The present study extends collective action research by (a) analyzing relations between collective action frames, comparing the SIMCA and the EMSICA in a specific context (namely the emergence of the 15-M socio-political protest movement), (b) examining the sequential mediating role of emotions in the relationship between those frames and collective participation, and (c) directly exploring the association between anger and hope.

Figure 1 .
Figure 1.Model 1 -Sequential mediation models based on the Social Identity Model of Collective Action (SIMCA) with hope and anger as mediators between collective action frames and intensity of participation.Note.Path coefficients are standardized estimates.All the coefficients represented by continuous arrows in the graphs are statistically significant based on 95% bootstrapped confidence intervals, while the dashed lines indicate effects that are not statistically significant.

Figure 2 .
Figure 2. Model 2 -Alternative sequential mediation models based on the Encapsulated Model of Social Identity in Collective Action (EMSICA) with hope and anger as mediators between collective action frames and intensity of participation.Note.Path coefficients are standardized estimates.Significant covariances between latent factors are shown by double-arrowed solid lines.All the coefficients represented by continuous arrows in the graphs are statistically significant based on 95% bootstrapped confidence intervals, while the dashed lines indicate effects that are not statistically significant.

Table 1
Socio-Demographic Characteristics of 15-M Participants and Non-ParticipantsEach participant was asked to indicate whether he/she had voted / had intentionally abstained from voting / had intentionally changed his/her vote / voted in blank in the elections held on May 15th, 2011. a

Table 2
Correlations and Descriptive Statistics of the Variables Included in the Models ***p < .001.

Table 3
Fit Indexes for the Alternative Models