Acta Polit DOI 10.1057/s41269-017-0071-z

ORIGINAL ARTICLE

How government coalition affects demonstration composition. Comparing twin austerity demonstrations in

1 2 Ruud Wouters • Pauline Ketelaars • 2 3 Stefaan Walgrave • Nina Eggert

Ó Macmillan Publishers Ltd., part of Springer Nature 2017

Abstract Does the composition of a government affect the beliefs, motivations, and mobilization trajectories of protest participants addressing the government? We make use of a straightforward research design to test how the loss of a left-wing ally in power affected the individual-level characteristics of participants in two ‘twin’ demonstrations. Both demonstrations were staged by the same organizers (trade unions) who launched identical campaigns on the same issue (austerities) in the same country (Belgium) forwarding the same demands (fair taxation). The first demonstration was staged in 2011 against a newly formed center-left government. The second demonstration was staged in 2014 against a newly formed center-right government. Relying on protest survey evidence, campaign material and insights of political opportunity structure theory (POS), we mount evidence that the loss of a left-wing ally produced a threat that resulted in (1) bleaker perceptions of partici- pants (effectiveness, personal situation, trust), (2) the activation of informal mobi- lizing networks, and (3) different motivational dynamics (less instrumental). As such, this study contributes to a better understanding of macro–micro dynamics in contentious politics. Conclusion and discussion center on ways of studying the macro–micro link in protest participation research.

Keywords Protest Á Political opportunity structure Á Austerity Á Economic crisis Á Elite allies

& Ruud Wouters [email protected]

1 Amsterdam School of Communication Research (ASCoR), University of Amsterdam, Amsterdam, The Netherlands 2 University of Antwerp, Antwerp (Media, Movement & Politics, M2P), Antwerp, Belgium 3 University of Geneva, Geneva, Switzerland R. Wouters

Introduction

Does the political context in which a demonstration takes place affects its composition? Or, in more operational terms, does the presence of specific parties in a government coalition influence who demonstrates, what motivates these demonstrators, and how they were mobilized? In the study of social movements, contextual accounts of protest have reached canonical stature (Koopmans 1999; Meyer 2004). Political process theories and theories of political opportunity structures (POSs) have successfully increased the understanding of protest emergence, development, and impact (Andrews 2001; Kriesi et al. 1995; McAdam 1982; McAdam et al. 1996). Most of these studies link the macro-level of the political context with the meso-level of movement processes. Hardly any studies in this tradition, however, scrutinize whether and how the political context influences micro-level characteristics of individual participants in specific demonstrations (Walgrave and Verhulst 2009). In this article, we make use of a unique data opportunity to explore such a macro–micro relationship. Much as is the case with twin studies in psychology, we compare two demonstrations that share the same ‘genetic’ material. Our two demonstrations were organized in the same country (Belgium), by the same organizers (trade unions), who set in motion identical mobilization campaigns on exactly the same issue (austerity). The main difference between the twin demonstrations, we argue, and hence the reason why people with different perceptions, motivations, and mobilization trajectories showed up, is the political context in which they were staged. The first demonstration, in 2011, targeted a center-left government coalition with a left-wing, socialist prime minister. The second demonstration, in 2014, targeted a center-right government with a right-wing prime minister leading a cabinet without left-wing parties. We argue that this change in government coalition, and more specifically, the loss of a left-wing ally in power, is the main cause of the differences between the demonstrators at both events. To tackle this puzzle, we use campaign material of the organizers to give a detailed picture of the similarities between the cases at hand. And, we use protest survey data gathered at both events to draw a profile of the participants. We find significant differences between participants in these two largely similar demon- strations: perceptions of participants were bleaker, mobilization was broader, and motivations were less instrumental in the 2014 demonstration. Building on contextual theories in social movement studies, we find confirmation for the argument that the loss of a left-wing ally in power, compared to when protestors could still count on a left-wing ally in government in 2011, increased the threat for the organizing social movements and their constituents. Being confronted with a right-wing government, they were unsure how far the government would be willing to go in its austerity policy and there was no way to stop the government from within. We argue that this increased threat affected potential demonstrators’ propensity to participate and their underlying beliefs. As a consequence, also the How government coalition affects demonstration composition… demonstrations were composed differently in terms of demonstrators’ interpretation and mobilization processes. First, we describe both mobilization campaigns, stressing the many similarities and few differences between both episodes of contention. Next, we present the theoretical framework. We engage both with studies that tackle macro–micro links in protest studies and with work that deals with participation in times of austerity. After the presentation of our data and methods, we put a series of comparisons, mounting evidence that fits our theoretical expectations. The conclusion and discussion center on how future research might tackle the complex macro–micro puzzle.

Twin anti-austerity protests in Belgium

In 2008, the economic crisis hit Europe. The more than 7 years of economic downturn that followed became known as ‘‘The Great Recession.’’ European governments enacted austerity policies leading to severe budget cuts which propelled waves of anti-austerity protests across Europe (Giugni and Grasso 2015). Simultaneously, Belgium dealt with a severe political crisis related to issues of state reformation (Devos and Bouteca 2010). Against this background of political and economic crisis, the two demonstrations—the main events of two protest waves— took place. Both events were organized by the Belgian trade unions as a reaction to newly formed governments and their proposed austerity policies. In the next paragraphs, we give a detailed account of both demonstrations and of the context in which they took place. Table 1 summarizes the most important elements.

We have alternatives

The first demonstration, labeled ‘‘We have Alternatives,’’ was announced on November 15, 2011. The unions aimed to put pressure on the government formation process. Although an agreement about state reform was reached, the political crisis was still fresh and the negotiating coalition partners failed to find an agreement on the budget. In light of the economic crisis, budget cuts and austerity measures were urgently demanded by the European Commission. Aiming to prevent severe budget cuts, the unions announced a wave of protest, consisting of multiple events spread over a 2-month period. They demanded no blind austerity measures, no adaptation of the consumer price index mechanism and no dismantling of unemployment benefits. The unions proposed alternative measures: investment in new, sustainable jobs and pursuing fiscal justice by means of a capital gains tax and more actively tracing and sanctioning of fiscal fraud. On November 26, 2011, the Socialist, Liberal, and Christian Democratic parties reached an agreement on the budget and after 541 days the Di Rupo government was founded: a classic tripartite with strong socialist presence and a socialist prime minister. The unions claimed that the new government distributed austerity measures unequally across society. On December 2, 50,000 demonstrators walked the streets of Brussels, 80,000 if we follow the count of the organizers. The high R. Wouters

Table 1 Overview of the campaign and context of the two ‘twin’ demonstrations We have alternatives National manifestation

Date December 2, 2011 November 6, 2014 Organizers ACV-ABVV-ACLVB ACV-ABVV-ACLVB Campaign 15 November: announcement of event 15 October: announcement of event 2 December: National demo 6 November: National demo 22 December: strike public services 24 Nov–8 Dec: provincial strikes 30 January: National strike 15 December: National strike Demands ‘‘Hands off the index’’ ‘‘Hands off the index’’ ‘‘Fiscal justice’’ ‘‘Fiscal justice’’ ‘‘Fair taxation’’ ‘‘Fair taxation’’ Government Di Rupo Michel Composition PS-Cdh-MR ? MR ? Sp.a-CD & V-Open Vld N-VA-CD & V-Open Vld Inauguration December 6, 2011 October 11, 2014 Negotiations 541 days 139 days Political crisis High Low Economic crisis High Moderate Negotiations before No No Turnout 50,000 (police); 80,000 (organizers) 120,000 (consensus) Disruption No Yes Negotiations after Yes Yes turnout was considered a success by organizers and news media. The same day, trade union leaders met with Di Rupo and announced a national strike. The new government was given time to adapt the budget agreement in favor of trade union demands. If the government would not compromise, the unions would paralyze the country with a national strike on January 30, 2012. In February, the budget agreement was slightly adapted in favor of trade union demands, although the main austerity measures remained untouched.

National manifestation

The protest campaign in 2014 appears as a copy of the campaign in 2011. Again, a national manifestation was announced at a launching event, as part of a larger campaign with the very same temporal structure and action repertoire used in 2011 (see Table 1). The demands of the unions were the same: austerity measures are unequally distributed, the consumer index mechanism should remain intact, and the battle against fiscal fraud should be intensified. Yet, there were some important differences between both episodes of contention as well. The preceding elections of June 2014 were won in by the Flemish Nationalists, who campaigned exclusively on a socio-economic agenda and aimed to form a government without socialist parties. After 139 days of negotiations, for the first time in more than 20 years, a federal government without socialist parties was formed. The new How government coalition affects demonstration composition… center-right coalition became led by a francophone liberal prime minister (Charles Michel) and putted forward a strict austerity agenda. The government announced it would skip (for once) the adaptation of wages to the consumer price index, would further increase retirement age, freeze wages, and decrease unemployment benefits. The proposals pushed by the unions 3 years ago were being neglected. A national demonstration was announced, and in the following weeks mass media would put one issue on top of the news agenda: the ‘hete herfst’orhot autumn; the period of social unrest the unions were initiating. The national demonstration that followed is noteworthy in several respects. On November 6, more than 120,000 demonstrators marched and occupied Brussels, one of the largest demonstrations in Belgian history. Although most demonstrators marched peacefully, at the end of the demonstration a group of protesters started a riot. Cars were put on fire, property was damaged, and 30 demonstrators were arrested, a severity of violence rarely seen in Belgian demonstrations. The short- term result of the demonstration, however, was similar to the previous national demonstration: the union leaders were invited by the prime minister who called for social dialogue. In the coming months, a similar train of events, comparable to 2011, took place: the unions mobilized, a national strike paralyzed the country, employers and trade unions started the social dialogue, and an agreement—without the socialist union this time—was made. All of this, however, was of course not yet known by the demonstrators participating in the National Manifestation. In sum, both demonstrations were organized by the same organizations (the three main trade unions), dealt with the same topic (austerity) and were part of very comparable protest campaigns. Moreover, both demonstrations challenged newly formed governments and their plans (no concrete measures had been taken yet) and forwarded exactly the same claims. A closer comparison of both campaigns also highlights some differences. In 2011, the economic and political crises were still intensely present. This was less the case in 2014 when both crises had clearly withered. The main difference, we contend, is that in 2014 a new government without the socialist parties was formed, a government whose dominant party (N- VA; Flemish nationalists) explicitly campaigned on removing the socialist parties from the executive, with the aim of forming a right-wing government able to execute a program of strict austerity. In the next theoretical part, we build on POS theory to put forward the argument that this loss of an ally in government affected demonstration composition.

How context affects individuals

The political opportunity perspective is without doubt the most dominant paradigm in the field of social movement studies (Giugni 2009; Jasper and Goodwin 2011; Meyer 2004). Its basic premise is straightforward: the political context in which social movements are embedded influences their emergence, development, and impact (see, e.g., Eisinger 1973; Kitschelt 1986; McAdam 1982; Tilly 1978). As Tilly and Tarrow (2015, p. 59) aptly put it: ‘‘Features of a regime affect the opportunities and threats impinging on any potential maker of claims, and changes R. Wouters in those features produce changes in the character of contention.’’ Despite its obvious merits, the POS perspective has received its share of criticism (Gamson and Meyer 1996; Goodwin et al. 1999; Koopmans 1999). Two related criticisms are of interest here. First, POS researchers have been predominantly interested in explaining social movements’ tactical repertoires and the occurrence and strength of protest events; their favorite dependent variables are situated at the meso-level: decisions by movement leaders to stage a protest, leaders’ decisions about the protest repertoire to use, turnout at protest events, etc. How political circumstances affect individuals potentially participating in collective action has mostly been neglected (Opp 2009). Second, the POS approach rarely formulates clear propositions about how the macro political context exactly influences protest or protesters; the mechanisms that drive the macro–micro link are understudied (Christensen 2011; Koopmans 2005; McAdam et al. 2001). Here, we seek to contribute to both these lacunae. To be fair, in recent years, POS theory has been increasingly used to explain individual-level decisions of potential participants. Our study is notoriously different from these studies, however. First, almost all these studies use (cross- national) population survey data to examine the link between invariable aspects of the macro context and individual participation in protest activity in general (see, e.g., Braun and Hutter 2016; Corcoran et al. 2011; Dalton et al. 2010; de Moor 2016; Quaranta 2015;Vra´blı´kova´ 2014) or in other types of political participation (Morales 2009). These scholars argue that general features of a political system—its openness to challengers; its strength in implementing policies—influence potential participants’ estimation of how successful participation will be to reach a goal, and as such, their decision to participate or not. Second, in the wake of the 2008 economic crisis and its austerity measures, other accounts have focused on economic predictors of general protest participation in a country, revisiting the still undecided debate whether deprivation and grievances or rather resources and skills drive protest participation (e.g., Dalton et al. 2010). Among these scholars looking for aggregate-level, economic explanations, there has been a keen interest in the effect of economic crises on protest participation (e.g., Grasso and Giugni 2016; Kern et al. 2015;Ru¨dig and Karyotis 2014). These economic crisis studies have not fully looked at political context variables, though. They focused on the impact of individuals’ relative deprivation, sometimes in interaction with economic context variables, on protest participation. In sum, despite the obvious merits of the general population approach and the focus on stable regime characteristics, these studies cannot answer questions that dig into the more fine-grained effect that more variable aspects of POS have, especially on the motivations and features of protesters participating in specific protest actions. There is a handful of recent studies, however, that has scrutinized how temporary POS elements affect participants’ features in specific demonstrations. These studies rely on research designs that sample protest participants in specific demonstrations, allowing to link the macro context surrounding a specific protest event to micro- level features of protestors participating in that event. One of the first of these studies was carried out by Walgrave and Verhulst (2009). These authors compared protest crowds in eight different countries that took to the streets against the How government coalition affects demonstration composition… imminent war in Iraq in 2003. Their analyses show that the stance of the government vis-a`-vis the war affected the internal diversity of the protests: the more the government shared the claim of the protesters, the more diverse the crowd at the demonstration. Similarly, Ketelaars (2015) found government stance regarding austerity measures to affect protest participants’ reasons to take part. Both these studies compare different countries and the problem is that it is hard to argue that the only difference between countries would be government stance. Go´mez-Roma´n and Sabucedo (2014) compared participants in two protest demonstrations against reform of the Spanish labor law. Both demonstrations were organized by the same organizations, yet the first demonstration before and the second after the law was adopted. This change in political circumstances affected identity and efficacy beliefs of participants. A study in the US by Heaney and Rojas (2011) comes closest to what we aim to do here. They show how the election of Barack Obama, and thus a changing composition of government, had an impact on individuals in the antiwar movement. As the Democratic party achieved electoral success, Democratic activists left the movement, and they did so at a greater rate than third-party members and non-party members. In sum, some studies have suggested that macro political aspects can affect micro individual traits. The crucial question, however, is: How? POS scholars have distinguished the role of opportunities and threats in this regard. Tilly (1978) defines ‘opportunity’ as the likelihood that protesters will improve their situation if they act collectively. ‘Threat,’ in contrast, is the likelihood that existing benefits will be lost if protesters fail to act collectively. As such, groups may be either pulled into action by positive contextual cues, or pressed into action out of fear to get inflicted new harms (Almeida, 2003). In this body of research, the most common operationalization of opportunities and threats focuses on the presence or absence of elite allies—in this study, we define it based on the partisan composition of a coalition government. Interestingly, the literature expects both opportunities and threats to positively affect the propensity of individuals to partake in protest events. On the one hand, allies in government may spur turnout by increasing the tendency of individuals to take part in protest. Elite allies make it easier for challengers to access the political decision-making process. As elite allies provide institutional access points for disenfranchised groups, potential participants are expected to be more eager to protest as chances of influencing policy increase (Amenta et al. 2005; Andrews 1997; Kriesi 1995). According to Goldstone and Tilly (2001), on the other hand, people participate especially in response to threats; they are more likely to be driven by the wish to avoid losses compared to the odds of realizing gains. The rise of an unfriendly party increases the odds of hostile policies which intensifies the propensity to participate. Van Dyke (2003), for instance, finds a positive relationship between Republican presidents and campus protests in the United States. Republican presidents and the policies they enact created a threat for mostly liberal students, she argues, which boosted protest. Heaney and Rojas (2011) mount evidence for the opposite process, yet substantiating the very same claim. They show that an ally in government can dampen the willingness to take action. With Barack Obama in the White House, R. Wouters

Democrats became less motivated to take part in activities of the antiwar movement, because of conflicting party-movement identities. In all, the literature on the effect of elite allies on protest participation expects there to be a link between the absence or presence of allies, the perceptions and motivations of potential participants, the propensity of people to become protest activists, and the sort of people that eventually take to the streets. Note that the potential effect of elite ally presence/absence on participant features is twofold. Ally absence, compared to ally presence, has an impact on the perceptions of all potential participants: all potential participants are more anxious that they will suffer harms compared to a situation when they have a government ally. At the same time, as the ratio of pros and cons of participation shifts with the absence of elite allies (less opportunity, more threat), the participants who actually participate will differ from those who would participate if elite allies were present. Those who participate under adverse conditions are different from those who participate under favorable conditions. In this study, we draw on evidence of protest participants only. We have no information about movement supporters who eventually did not show up. Thus, our comparison of actual participants in two events only allows to test the second mechanism, the effect of elite allies on effective participants and not the effect on all potential protesters. The two Belgian anti-austerity demonstrations form an ideal case to examine perceptions and motivations of protesters in a context of opportunity (the presence of an elite ally) and threat (the absence or loss of an ally). Based on the literature review above, our argument is that the absence of elite allies in 2014, via its effect on the propensity to participate, affected the composition of the movement and its protest.

Hypotheses

Our two demonstrations present a case of diminished opportunities and increased threats. We already know that this threat resulted in one of the largest union demonstrations in recent Belgian history, which suggests that this is indeed a case in which a threat spurred rather than crippled mobilization (Almeida 2003). But how did this variation in government coalition affect protester traits? First, we expect the shift in government coalition and the increased threat to have resulted in other expectations and perceptions among the demonstrators. Like most political process scholars, we assume that citizens are capable of perceiving changes in the political environment (see, e.g., Kriesi et al. 1995, p. 245; McAdam 1982, p. 48) and that people adjust their expectations accordingly. Because of the loss of an ally in power who could defend the protesters position, we expect participants in the 2014 demonstration to feel more threatened and to have bleaker expectations compared to participants in 2011. In order to measure the threatening effect of the Michel government, we take a number of perceptions into account: the expectation that one’s personal situation will worsen, the belief in the political effectiveness of the demonstration, and trust in the current government. How government coalition affects demonstration composition…

H1 Compared to 2011 demonstrators, 2014 demonstrators have bleaker expec- tations and perceptions. Second, we expect the differential political context in 2014 to have produced a different mobilizing structure (Walgrave and Klandermans 2010). For networks to become mobilizing structures, they have to be activated (Boekkooi et al. 2011). Because of the ally in government in 2011, we expect the mobilization process to have been quite closed, limited to formal mobilization within the trade unions themselves. With the rise of a right-wing coalition in 2014, we expect wider, more open mobilization processes beyond typical trade union mobilization. The stronger threat of the right-wing government appealed to the whole ‘‘broad left,’’ activating not only the trade union mobilization machinery itself but more informal networks as well. Put differently, the right-wing threat in 2014 produced a feeling of ‘‘we are all in this together,’’ breaking identity boundaries and bringing together both trade union and non-trade union participants. In order to test these claims, we again take a number of variables into account: whether the protesters participated in the company of informal social ties like friends and family rather than with co-members and colleagues, whether they were (not) a member of a staging organization (the trade unions), and whether they were recruited via mass media channels, all elements that might point to a more open mobilization structure. We also measure the degree of protest experience and the extent to which protesters identified with parties left of center. This latter variable taps the party animosity that might have been triggered by the shift in government coalition. H2 Compared to the 2011 demonstration, the 2014 demonstration is characterized by a broader mobilizing structure. H3 Compared to the 2011 demonstration, the 2014 demonstration is characterized by stronger left-wing party identification. Finally, we expect participants’ motivations to differ across both demonstrations. Two studies of Van Stekelenburg and colleagues show how street demonstrations organized on the same day on the same topic—but staged by different organiza- tions—bring protesters with very different motivations to the streets (Van Stekelenburg and Klandermans 2014; Van Stekelenburg et al. 2009). We follow the same logic here, but in our case not the staging organizations vary, but the macro political context differs. Extant work on the motives of protest participants distinguishes between ideological, identity, and instrumental motives (e.g., Klan- dermans 2004). In order to distinguish the 2011 from the 2014 demonstrators, we believe especially instrumental and ideological motives are useful. People can protest in order to reach a goal (instrumental) or they protest to express themselves (ideological). Briefly put, the opportunity of the 2011 demonstration (left-wing party in government) should have led to more instrumental motivations, driven by the ambition to put pressure on the social-democratic ally in government to change policies. In 2014, however, the threat of the right-wing government coalition should have sparked more ideological motivations. In a threat situation, with no ally in government to put under pressure, there was far less hope to improve the situation as a right-wing government took office. R. Wouters

H4 Compared to 2011 demonstrators, 2014 demonstrators’ determination to participate is shaped less by instrumental and more by ideological motives.

Data and methods

The core of our design is a comparison between participants of two demonstrations, organized in the same country, by the same organizers, against similar (announced) austerity measures, and voicing the same central slogans. The key difference, we hold, is that the governments against which the two demonstrations protested were composed differently. In 2011, a friendly left-wing socialist party was leading the government, whereas in 2014 only right-wing or center parties were in government. We hold that, by affecting the propensity of potential protesters to take part, this difference in government party status is the key driver of the different type of protester that took to the streets in 2014 versus 2011. Our design, only relying on two demonstrations, does not allow to test alternative explanations. We see two potential alternate explanations of the difference between 2011 and 2014 demonstrators. First, it may be the case that the actual grievances potential demonstrators suffered had changed over time. In fact, in 2014, the crisis was withering, while in 2011 the economic downturn was still in full swing. If actual deprivation drove the results, we would expect the opposite of what we hypothesized above: more bleak expectations, a broader mobilization structure, more left-wing party identification, and a more expressive outcry in 2011 compared to 2014. Second, it may be the case that the planned government measures of 2014 were more threatening than those of 2011: the anticipated grievances of 2014 may have surpassed those of 2011. There are good reasons to assume that this was indeed the case as the Michel government installed in 2014 put forward a tougher austerity program than the Di Rupo government of 2011. If anticipated grievances drove the protest, we would expect the exact same patterns as those we hypothesized above. Hence, we cannot know for sure that it is the government’s composition or rather its planned policy measures that affected the demonstrators’ diverging profile, but the latter is without doubt a consequence of the former. In order to compare the participants of both demonstrations and to test our expectations, we sampled and questioned demonstrators at both events. We did so by using the standardized protest survey procedure, as spelled out and tested by Walgrave and colleagues (Walgrave and Verhulst 2011; Walgrave et al. 2016). In order to obtain a random sample of protest participants, protest surveying selects and interviews protest participants by means of pointers and interviewers. Pointers are senior researchers who, based on the estimated size of the demonstration, systematically count and skip rows of protesters and indicate which particular protestor in the middle, left, or right side of a row should be handed a postal questionnaire by the interviewers (booklet of 8 A5 pages). For each demonstration, one pointer with a group of interviewers starts at the head and one at the tail of the demonstration, as such covering the entire march. Sampling (pointers) and interviewing (interviewers) are disconnected as interviewers—if let free—tend to select approachable peers, causing selection bias. Protesters fill in the questionnaire at home and mail it back (postpaid envelope). How government coalition affects demonstration composition…

For each demonstration, we aimed to distribute 1000 questionnaires: 800 postal only questionnaires, 200 including a face-to-face interview on the spot. The face-to-face interviews allow for the calculation of potential response biases, as not all protestors might be equally prone to mail the survey back. We distributed about 800 questionnaires and conducted about 150 face-to-face interviews using a team of 15 interviewers in each demonstration. In case of the 2011 ‘‘We have alternatives’’protest, 169 surveys were completed and mailed back (22% response rate); 209 surveys were returned in the 2014 ‘‘National Manifestation’’ (24% response rate). These response rates are common in trade union events. A comparison between responsive and unresponsive respondents (Independent Mann–Whitney U tests) showed no response bias in case of the 2011 demonstration. In the 2014 demonstration, however, younger participants and participants with low political interest were significantly less responsive. We therefore calculated weights and weighed the 2014 respondents on both variables. As results of the weighed and unweighted dataset do not change the conclusions, we report unweighted data. In the appendix we provide an overview of the variables we use, their measurement and means, as well as some additional basic descriptives—not used in the analyses—to present an overall profile of the two protesting groups.

Results

In order to test our claim, we compare the individuals in both demonstrations. We do so by using Pearson’s Chi-square tests, which we report in text. Additionally, we present four logistic regressions in Table 2, using participation in the 2014 National Manifestation as the dependent variable. This way, we can gauge on which dimensions people who protested the Michel government differ from those who protested the Di Rupo coalition, while controlling for other variables. The first three models of Table 2 test each hypothesis separately. Model 4 presents the full model. Both bivariate and multivariate results provide strong evidence for the hypothesized differences between both demonstrations. With the loss of a left-wing ally in power, Hypothesis 1 predicted perceptions of participants in 2014 to be bleaker compared to 2011. For both demonstrations, we asked respondents to what extent they expected their personal situation to worsen as a consequence of the austerity measures. Results show a significant relation between demonstration and bad expectations (v2 (1; 358) = 37.341, p = 0.000). The exit of the socialist parties from the coalition seems to have increased the threat of the austerity measures for the demonstrators in 2014. In 2011, less than two demonstra- tors in three expected their situation to worsen (62.3%); in 2014, this held for nine out of ten demonstrators (89.4%), an increase of about 30%; note again that the crisis was already less pressing in 2014. Losing an ally in power also affected whether demonstrators expected to reach one of their most important goals: the installation of a capital gains tax (v2 (1; 362) = 5.335, p = 0.021). In 2011, about fifty percent of the demonstrators was convinced that the demonstration would be quite to very effective in putting pressure on the government to take steps in the direction of a capital gains tax (50.6%). Fewer demonstrators shared that belief in 2014 (38.5%). Also, the rise of Table 2 Logistic regressions explaining participation in National manifestation (2014) N = 291 Model 1 Model 2 Model 3 Model 4

Hypothesis 1 Personal situation (worse) 0.696 (0.149)*** 0.820 (0.163)*** Reach goal: taxing fiscal gains - 0.636 (0.134)*** - 0.528 (0.144)*** Trust in government - 0.472 (0.157)** - 0.483 (0.168)** Hypothesis 2 Member of staging organization - 0.280 (0.464) - 0.417 (0.570) Informed via mass media 0.107 (0.268) 0.218 (0.316) Informed via informal networks 0.647 (0.281)* 0.614 (0.333)° Informal company 0.495 (0.277)° 0.391 (0.323) Past protest participation - 0.306 (0.109)** - 0.410 (0.136)** Hypothesis 3 Party identification (ref. = right) Center 0.200 (0.524) 1.007 (0.648) Left 0.776 (0.435)° 1.327 (0.555)* None 1.305 (0.734)° 1.413 (0.941) Controls Education 0.412 (0.116)*** 0.272 (0.101)** 0.277 (0.096)** 0.369 (0.124)** Gender 0.495 (0.312) - 0.197 (0.282) 0.291 (0.274) 0.308 (0.335) Year born 0.018 (.013) - 0.003 (0.014) 0.016 (0.012) - 0.008 (0.016) Constant - 37.682 (26.320) 4.978 (27.186) - 33.898 (23.220) 12.888 (31.647) Prob [ chi2 0.000 0.000 0.000 0.000 Pseudo R2 0.2483 0.1096 0.0736 0.3055 .Wouters R. LR chi2 (df) 98.28 (6) 413.38 (8) 29.14 (6) 120.90 (14)

*** p [ 0.000; ** p [ 0.010; * p [ 0.050; ° p [ 0.100 How government coalition affects demonstration composition… a right-wing cabinet strongly eroded the trust in government among 2014 demonstrators (v2 (4; 371) = 20.182, p = 0.000). About a third of the 2014 participants indicated that they had no trust at all in the federal government (28.5%); in 2011, this was only about 15% (14.6%). All three variables testing Hypothesis 1 show significant correlations in the regression models of Table 2, both in Model 1 and in Model 4. Hypothesis 1, hence, gets strong confirmation by the data. Next, we compare the mobilization trajectories of the participants in the 2014 demonstration to those of the protesters in 2011. Hypothesis 2 predicted mobilization in 2011 to have remained confined to traditional trade union mobilization, while mobilization in 2014 was expected to have gone beyond trade union mobilization. We argue that the threat of the right-wing government activated a wider pool of participants, adherents who are not necessarily part of the core trade union constituency. Those people should be mobilized in a more informal way and are therefore also expected to be somewhat less protest savvy. Our results confirm these expectations. Although the vast majority of the demonstrators at both events were trade union members, twice as many demonstrators indicated to be not a trade union member in 2014: 14.7% compared to 7.7% in 2011(v2 (1; 371) = 4.375, p = 0.036). In addition, 2014 demonstrators indicated more often that mass media informed them about the upcoming demonstra- 2 tion (v (1; 378) = 5.167, p = 0.023; Media2014 = 73.6%, Media2011 = 62.7%). They also indicated that family, friends, and colleagues were more frequently sources of information about the demonstration (v2 (1; 378) = 5.167, p = 0.023; Infor- mal2014 = 73.6%, Informal2011 = 62.7%). As a consequence, the company in which the 2014 demonstrators participated was more informal as well, consisting to a larger extent of family, friends, and relatives (v2 (1; 378) = 18.164, p = 0.000; Com- pany2014 = 53.8%, Company2011 = 32.0%). And, finally, the 2014 demonstrators consisted to a larger extent of demonstrators that would normally not demonstrate, as evidenced by their more limited protest experience compared to the participants in 2011 (v2 (1; 370) = 21.809; p = 0.000). The results in Model 4 of Table 2 indicate that, when controlling for other variables, only two differences still stand out: the 2014 demonstrators were more informed via informal networks and they had less protest participation experience. Nevertheless, this evidence still largely corroborates Hypothesis 2, attesting of a quite different, more informal, and broader mobilization structure in 2014. Hypothesis 3 predicted the right-wing coalition to unify the left, attracting broad left-wing sympathy by means of a more inclusive mobilization. Indeed, in 2014 the share of participants who identify with a party left from the center (83.3%) increased compared to 2011 (70.2%) (v2 (1; 327) = 7.962; p = 0.005). Specifically, protesters in 2014 did identify more with the Green party and with more extreme left parties, rather than with the traditional socialist party. Identification with the center Christian Democratic party—which became the most leftist party in the right- wing coalition and received criticism from its left leaning constituency for partaking in this government—sharply declined. Interestingly, no differences between both demonstrations were found when using the classic left–right scale, suggesting that parties and the cues they presented in this shift to a more right-wing government really mattered for the demonstrators. The regression analysis of Model 4 (Table 2) confirms this result: there are significantly more protesters who identify with left- R. Wouters wing parties in the 2014 demonstration, even when controlling for all other variables. Whether participants shifted parties or whether certain party identifiers did not participate cannot be distinguished with the data at our disposal. Nevertheless, the results corroborate hypothesis 3, indicating more participants identifying with left-wing parties in 2014. Hypothesis 4, finally, predicted the motives of participants to differ between 2011 and 2014. Because of the lack of an elite ally, we expected the motivation of the 2014 participants to be less driven by instrumental reasons—i.e., to change government decisions and to defend one’s interests. Rather, their motivation to participate was expected to have been more ideological, shaped by feelings of worry because of the right-wing threat, the desire to vent anger, and motivations based on solidarity. In Table 3, we use respondents’ determination to participate as the dependent variable and we analyze which variables account for this in the 2011 and 2014. In the demonstration against the Di Rupo government, participants’ determination to participate correlates with the instrumental ambition to put pressure on politicians to change policies (B = 0.911; p = 0.007). Hence, in 2011, especially the people who wanted to influence government decisions were very determined to take part. In 2014, however, this correlation is turned upside down. People who were at the demonstration because they wanted to demand changes from the Michel government were significantly less motivated than people who

Table 3 Two ordinal logistic regressions explaining determination to participate N Di Rupo (2011) Michel (2014) 146 170

Instrumental motives Put pressure on politicians 0.911 (0.338)** - 1.384 (0.473)** Defend interests 0.161 (0.192) 0.388 (0.211)° Ideological motives Express opinion - 0.078 (0.332) - 0.175 (0.293) Make the public aware 0.168 (0.299) 0.306 (0.267) Identity motives Show solidarity 0.636 (0.399) 1.140 (0.323)*** Feeling morally obliged 0.162 (0.160) 0.256 (0.154)° Emotions Angry 0.625 (0.252)* 1.006 (0.256)*** Worried 0.395 (0.232) 0.632 (0.267)* Controls Education 0.231 (0.175) 0.066 (0.148) Gender - 0.102 (0.456) - 0.610 (0.395) Year born - 0.030 (0.021) - 0.026 (0.018) Prob [ chi2 0.000 0.000 Pseudo R2 0.2182 0.2610 LR chi2 (df) 52.79 (11) 71.86 (11)

*** p [ 0.000; ** p [ 0.010; * p [ 0.050; ° p [ 0.100 How government coalition affects demonstration composition… came to protest for other reasons (B =-1.384; p = 0.003). The loss of the socialist ally in government seems to have demotivated those who mainly protest to change policies. At this demonstration, the most motivated participants were fueled by feelings of solidarity (B = 1.140; p = 0.000). The threat of the right-wing coalition seems to have influenced demonstrators’ emotions as well: the people who were most determined in 2014 were very worried about austerity (B = 0.632; p = 0.018). In 2011, worry was not a predictor of someone’s motivation to participate. Finally, both in 2011 and in 2014 the determination to protest correlated with anger—but in 2014 this effect was stronger. The angrier someone felt about the upcoming cutbacks, the more this person was motivated to join the demonstration.

Conclusion and discussion

This study set out with a straightforward question: does the configuration of a government coalition affect a demonstration’s composition? Making use of a simple research design, we tried to single out the influence of a crucial element in POS theory: the presence or absence of elite allies. We studied two demonstrations, one in 2011 and one in 2014, staged under almost identical circumstances. Holding many aspects constant, the main moving part in our ‘most similar demonstrations’ design was the loss of an elite ally in power. We argued that the changed government coalition produced a macro-level threat, increasing the odds for the protest potential to lose existing benefits and to be confronted with new harms. Our findings show that those who participated indeed perceived such a threat, as evidenced by their bleaker personal and collective expectations. Additionally, we found that the threat produced a different mobilizing structure. The peril of the right-wing government unified the left and activated a pool of more informal, less experienced left-wing identifying networks, which was far less the case in 2011. Finally, we showed that the change in government coalition correlated with the motivations of the participants. When confronted with a right-wing government, determination to participate was fueled less by instrumental motives. Rather, ideological motives, like feelings of moral obligation and the intent to show solidarity, were driving protestors’ determination to participate. In all, we believe that these findings contribute to some of the puzzles raised in recent POS research. A small number of studies have made a beginning with disentangling how contextual characteristics influence individuals’ protest partic- ipation. Cross-national population studies have been used to study how stable char- acteristics of political systems affect aggregate-level protest participation. Studies relying on protest surveys, on the other hand, have started to scrutinize how more changeable contextual features such as government composition or government stance influence specific traits of those who take part in particular protests. Our study falls in this latter category. We brought new evidence to the table on the understudied macro–micro link, we did so for one of the first times with regard to a crucial POS variable—elite allies—and we elaborated theoretically on the distinct individual-level effects of opportunity and threat. Whereas extant literature shows that both opportunities and threats can spur mobilization, we find that consequences at the individual level are quite distinct. Both in times of opportunity and in times of R. Wouters threat people can be mobilized to take to the streets, but our study suggests that the mechanisms behind the mobilization—the protesters’ mobilization trajectories, their motivations, and their expectations—differ significantly. This confirms earlier research showing that different contexts give rise to different routes to movement support (see, e.g., Blackwood et al. 2003; Van Stekelenburg et al. 2009). Of course, more research is needed on the link between political context and individual-level protester features. For instance, we had no information on non- protesters and how political circumstances affected people who did not show up at the demonstrations. We argued above that it is likely that elites allies, or other POS elements for that matter, affect both the entire protest potential and those who effectively show up. We could not tell whether the change in government coalition caused certain people to stay home and whether this, for instance, was due to truncated expectations to reach the demonstration’s goals. Future research that combines the inclusion of non-participants and contextual variation will be able to scrutinize the full puzzle. Although the empirical scope of our findings is limited in terms of the number of issues, demonstrations, and countries deployed, our step forward is relevant. Studying the influence of the political context on protest participants in issue- specific demonstrations requires a rigorous research design, careful event selection, and a combination of contextual knowledge and precise quantitative demonstrator information. This is key to trace the influence of the macro context on individual protesters. Most studies to date, like ours, are rich comparisons of two events that share many things but are characterized by one important difference (a type of organization, the passage of a law, the loss of an ally). The problem with this research strategy is that it limits generalizability. One way forward lies in carefully selecting and adding paired comparisons, giving researchers the advantage of depth—deep background knowledge and an intimacy of analysis—and, across comparisons, scope (Tarrow 2010). Another approach lies in combining many more protest surveys—adding variation in nations, staging organi- zations, issues—and modeling the effect of the political context (in our case, the presence or absence of elite allies) at the higher level in a multilevel framework. In fact, such an approach would nicely bridge the currently more prevalent paired comparisons on specific protesters and the research using cross-national population surveys on protesters in general. The availability of data gathered under the banner of the ‘‘Caught in the Act’’ project (Klandermans et al. 2010; Van Stekelenburg et al. 2012)might present a useful starting point for researchers interested in further exploring both approaches: by cleverly selecting demonstrations and/or adding newly gathered protest survey data, the macro–micro knot of protest politics can be disentangled and a better understanding of the complex ‘dynamics of contention’ lies within reach.

Appendix

See Table 4. Table 4 Overview of the used variables composition demonstration affects coalition government How Variable name Survey question 2011 mean 2014 mean Both demonstrations

Mean SD Min Max

Hypothesis 1 Personal situation If the austerities are adopted, I expect my personal 3.54 4.31 3.98 1.16 1 5 situation to worsen (strongly disagree, disagree, neither, agree, strongly agree) Reach goal: taxing fiscal gains How effective do you think this demonstration will be 3.90 3.16 3.49 1.13 1 5 in reaching more fiscal justice through taxation of fiscal gains? (not at all, not very, somewhat, quite, very much) Trust in government Please indicate, in general, how much you would say 2.56 2.14 2.33 0.97 1 5 that you trust the national government Hypothesis 2 Member of staging organization Please list the main organizations staging this 0.92 0.85 0.88 0.32 0 1 demonstration. Are you a member of any of these organizations? (no, yes, don’t know/not sure) Information channels How did you find out about the demonstration? Was it … via (check as many as apply): radio or television; newspapers (print or online); alternative online media; online social networks (e.g., Facebook, Twitter); advertisement, flyers, and/or posters; (fellow) members of an organization; An organization (magazine, meeting, website, mailing list,…; partner and/or family; Friend and/or acquaintances; people at your school or work Informed via mass media If the answer to the question above was: radio or 0.56 0.63 0.68 0.46 0 1 television; newspapers (print or online); alternative online media; online social networks (e.g., Facebook, Twitter) Table 4 continued

Variable name Survey question 2011 mean 2014 mean Both demonstrations

Mean SD Min Max

Informal sources of information If the answer to the question above was: partner and/or 0.32 0.54 0.44 0.50 0 1 family; friends and/or acquaintances; people at your school or work. Informal company Were you at this demonstration (check as many as 0.30 0.41 0.41 0.49 0 1 apply): With your partner; with your children, with relatives, with friends? Other options were: alone; with acquaintances; with colleagues or fellow students; with members of an organization you are a member of Past participation How many times have you in the past taken part in a 3.51 2.92 3.18 1.33 1 5 demonstration? (never, 1–5, 6–10, 11–20, 21 ?) Hypothesis 3 Party identification With which party do you most closely identify right 3.45 3.57 3.52 0.89 1 4 now? Answers recoded to: (1) right, (2) center, (3) left, (4) none Hypothesis 4 Determination to protest How determined were you to participate in the 4.62 4.59 4.61 0.69 1 5 demonstration? (not very, rather, somewhat, quite, very much) Motivations to participate Please indicate to what extent you agree or disagree with the following statements? I participated in the demonstration in order to… (strongly disagree, disagree, neither, agree, strongly agree)

Pressure politicians Pressure politicians to make things change 4.70 4.76 4.74 0.60 1 5 Wouters R. Defend interests Defend my interests 4.13 4.18 4.16 1.06 1 5 Express opinion Express my views 4.46 4.48 4.47 0.76 1 5 Make the public aware Raise public awareness 4.44 4.41 4.42 0.79 1 5 Table 4 continued composition demonstration affects coalition government How

Variable name Survey question 2011 mean 2014 mean Both demonstrations

Mean SD Min Max

Show solidarity Express my solidarity 4.73 4.72 4.73 0.56 1 5 Feeling morally obliged Because I felt morally obliged to do so 3.58 3.89 3.75 1.34 1 5 Angry Thinking about budget cuts makes me feel angry (not at 4.40 4.30 4.40 0.79 1 5 all, not very much, somewhat, quite, very much) Worried Thinking about budget cuts makes me feel worried (not 4.32 4.48 4.34 0.86 1 5 at all, not very much, somewhat, quite, very much) Control variables Year born In which year were you born? 1964 1968 1966 11.66 1933 1996 Gender Are you…? Male (0)/female (1) 0.34 0.43 0.39 0.49 0 1 Education What is the highest level of education that you 4.85 5.45 5.18 1.47 1 8 completed? If you are a student, at what level are you studying? None, did not complete primary education Primary or first stage of basic … Lower secondary or second stage of basic Upper secondary Post secondary, non-tertiary First stage of tertiary (BA, University first degree) Second stage of tertiary (MA) Post tertiary (PhD) Additional measures Left–right scale In politics people sometimes talk of ‘‘left’’ and 2.97 2.88 2.92 2.30 0 10 ‘‘right.’’ Where would you place yourself on this scale, where 0 means the left and 10 means the right? Table 4 continued

Variable name Survey question 2011 mean 2014 mean Both demonstrations

Mean SD Min Max

Political values To what extent do you agree or disagree with the following statements? (Strongly disagree, disagree, neither, agree, strongly agree) Government should redistribute income from the better 4.22 4.11 4.16 0.95 1 5 off to those who are less well off Children should be taught to obey authority 3.51 3.30 3.39 1.04 1 5 Even the most important public services and industries 1.80 1.61 1.69 0.98 1 5 are best left to private enterprise People from other countries should be allowed to come 3.06 3.42 3.26 1.16 1 5 to my country and live here permanently if they want to Subjective class People sometimes describe themselves as belonging to 3.78 4.02 3.91 0.89 1 6 the working class, the middle class, or the upper or lower class. Would you describe yourself as belonging to the…? (Lower class, Working class, Lower middle class, Upper middle class, Upper class) Political interest How interested are you in politics? (Not at all, not very, 3.20 3.18 3.19 0.74 1 4 quite, very) Political efficacy To what extent do you agree or disagree with the 4.07 4.04 4.05 0.79 1 5 following statement? Organized groups of citizens can have a lot of impact on public policies in this country (Strongly disagree, disagree, neither, agree, strongly agree) .Wouters R. How government coalition affects demonstration composition…

References

Almeida, P.D. 2003. Opportunity Organizations and Threat-Induced Contention: Protest Waves in Authoritarian Settings. American Journal of Sociology 109 (2): 345–400. Amenta, E., N. Caren, and S.J. Olasky. 2005. Age for Leisure? Political Mediation and the Impact of the Pension Movement on US Old-Age Policy. American Sociological Review 70 (3): 516–538. Andrews, K.T. 1997. The Impacts of Social Movements on the Political Process. The Civil Rights Movement and Black Electoral Politics in Mississippi. American Sociological Review 62 (5): 800–819. Andrews, K.T. 2001. Social Movements and Policy Implementation: The Mississippi Civil Rights Movement and the War on Poverty, 1965 to 1971. American Sociological Review 66 (1): 71–95. Blackwood, L., G. Lafferty, J. Duck, and D. Terry. 2003. Putting the Group Back into Unions: A Social Psychological Contribution to Understanding Union Support. The Journal of Industrial Relations 45 (4): 485–504. Boekkooi, M., B. Klandermans, and J. Van Stekelenburg. 2011. Quarrelling and Protesting: How Organizers Shape a Demonstration. Mobilization: An International Journal 16 (2): 498–508. Braun, D., and S. Hutter. 2016. Political Trust, Extra-Representational Participation and the Openness of Political Systems. International Political Science Review 37 (2): 151–165. Christensen, H.S. 2011. Political Participation Beyond the Vote: How the Institutional Context Shapes Patterns of Political Participation in 18 Western European Democracies.A˚ bo: A˚ bo Akademi University Press. Corcoran, K.E., D. Pettinicchio, and J.T.N. Young. 2011. The Context of Control: A Cross-National Investigation of the Link Between Political Institutions, Efficacy, and Collective Action. The British Journal of Social Psychology 50 (4): 575–605. Dalton, R., A. Van Sickle, and S. Weldon. 2010. The Individual-Institutional Nexus of Protest Behaviour. British Journal of Political Science 40 (01): 51–73. de Moor, J. 2016. External Efficacy and Political Participation Revisited: The role of perceived output structures for state- and non-state oriented action forms. Parliamentary Affairs 69 (3): 642–662. Devos, C., and N. Bouteca. 2010. Belgische Politiek: chaotisch Voorspel Moet Uitmonden in Institutioneel Orgasme. Internationale Spectator 64 (9): 445–449. Eisinger, P.K. 1973. The Conditions of Protest Behavior in American Cities. The American Political Science Review 67 (1): 11–28. Gamson, W.A., and D.S. Meyer. 1996. Framing political opportunity. In Comparative Perspectives on Social Movements: Political Opportunities, Mobilizing Structures, and Cultural Framings, ed. D. McAdam, J.D. McCarthy, and M.N. Zald, 275–290. Cambridge: Cambridge University Press. Giugni, M. 2009. Political Opportunities: From Tilly to Tilly. Swiss Political Science Review 15 (2): 361–368. Giugni, M., and M.T. Grasso. 2015. Austerity and Protest. Popular Contention in Times of Economic Crisis. Farnham: Ashgate Publishing Limited. Goldstone, J.A., and C. Tilly. 2001. Threat (and Opportunity): Popular Action and State Response in the Dynamics of Contentious Action. In Silence and Voice in the Study of Contentious Politics, ed. R. Aminzade, 179–194. Cambridge: Cambridge University Press. Go´mez-Roma´n, C., and J.-M. Sabucedo. 2014. The Importance of Political Context: Motives to Participate in a Protest Before and After the Labor Reform in Spain. International Sociology 29 (6): 546–564. Goodwin, J., J.M. Jasper, and J. Khattra. 1999. Caught in a Winding, Snarling Vine: The Structural Bias of Political Process Theory. Sociological Forum 14 (1): 27–54. Grasso, M.T., and M. Giugni. 2016. Protest Participation and Economic Crisis: The Conditioning Role of Political Opportunities. European Journal of Political Research 55 (4): 663–680. Heaney, M., and F. Rojas. 2011. The Partisan Dynamics of Contention: Demobilization of the Antiwar Movement in the United States, 2007–2009. Mobilization: An International Quarterly 16 (1): 45–64. Jasper, J., and J. Goodwin (eds.). 2011. Contention in Context: Political Opportunities and the Emergence of Protest. Stanford, CA: Stanford University Press. Kern, A., S. Marien, and M. Hooghe. 2015. Economic Crisis and Levels of Political Participation in Europe (2002–2010): The Role of Resources and Grievances. West European Politics 38 (3): 465–490. R. Wouters

Ketelaars, P. 2015. Bridging the Protest Macro-Micro Gap: Investigating the Link between Motivations and Political Context. In Austerity and Protest: Popular Contention in Times of Economic Crisis, ed. M. Giugni, and M.T. Grasso, 111–125. Farnham: Ashgate Publishing Limited. Kitschelt, H.P. 1986. Political Opportunity Structures and Political Protest: Anti-Nuclear Movements in Four Democracies. British Journal of Political Science 16 (01): 57–85. Klandermans, B. 2004. The Demand and Supply of Participation: Social Psychological Correlates of Participation in Social Movements. In The Blackwell Companion to Social Movements, ed. D.A. Snow, S.A. Soule, and H. Kriesi, 360–379. Oxford: Blackwell Publishing. Klandermans, Bert, Jacquelien van Stekelenburg, Dunya Van Troost, Anouk Van Leeuwen, Stefaan Walgrave, Joris Verhulst, Jeroen van Laer, and Ruud Wouters. 2010. Manual for Data Collection on Protest Demonstrations. Caught in the Act of Protest: Contextualizing Contestation (CCC). Version 3.0. Amsterdam and Antwerp: VU University and University of Antwerp. Koopmans, R. 1999. Political. Opportunity. Structure. Some Splitting to Balance the Lumping. Sociological Forum 14 (1): 93–105. Koopmans, R. 2005. The Missing Link Between Structure and Agency: Outline of an Evolutionary Approach to Social Movements. Mobilization: An International Journal 10 (1): 19–33. Kriesi, H. 1995. The Political Opportunity Structure of New Social Movements: Its Impact on Their Mobilization. In The Politics of Social Protest, ed. C. Jenkins, and B. Klandermans, 167–198. Minneapolis: University of Minnesota Press. Kriesi, H., R. Koopmans, J.W. Duyvendak, and M. Giugni. 1995. New Social Movements in Western Europe: A Comparative Analysis. Minneapolis: University of Minnesota Press. McAdam, D. 1982. Political Process and the Development of Black Insurgency, 1930–1970. Chicago: University of Chicago Press. McAdam, D., J.D. McCarthy, and M.N. Zald. 1996. Comparative Perspectives on Social Movements: Political Opportunities, Mobilizing Structures, and Cultural Framings. Cambridge: Cambridge University Press. McAdam, D., S. Tarrow, and C. Tilly. 2001. Dynamics of Contention. Cambridge: Cambridge University Press. Meyer, D.S. 2004. Protest and Political Opportunities. Annual Review of Sociology 30 (1): 125–145. Morales, L. 2009. Joining Political Organisations: Institutions, Mobilisation and Participation in Western Democracies. Colchester: ECPR Press. Opp, K.-D. 2009. Theories of Political Protest and Social Movements: A Multidisciplinary Introduction, Critique, and Synthesis. New York: Routledge. Quaranta, M. 2015. Political Protest in Western Europe: Exploring the Role of Context in Political Action. Cham: Springer. Ru¨dig, W., and G. Karyotis. 2014. Who Protests in Greece? Mass Opposition to Austerity. British Journal of Political Science 44 (3): 487–513. Tarrow, S. 2010. The Strategy of Paired Comparison: Toward a Theory of Practice. Comparative Political Studies 43 (2): 230–259. Tilly, C. 1978. From Mobilization To Revolution. Reading: Addison-Wesley Pub Co. Tilly, C., and S. Tarrow. 2015. Contentious Politics, 2nd ed. New York: Oxford University Press. Van Dyke, N. 2003. Protest Cycles and Party Politics: The Effects of Elite Allies and Antagonists on Student Protest in the United States, 1930–1990. In States, Parties, and Social Movements, ed. J.A. Goldstone, 226–245. Cambridge: Cambridge University Press. Van Stekelenburg, J., and B. Klandermans. 2014. Fitting Demand and Supply: How Identification Brings Appeals and Motives Together. Social Movement Studies 13 (2): 179–203. Van Stekelenburg, J., B. Klandermans, and W.W. Van Dijk. 2009. Context Matters: Explaining How and Why Mobilizing Context Influences Motivational Dynamics. Journal of Social Issues 65 (4): 815–838. Van Stekelenburg, J., S. Walgrave, B. Klandermans, and J. Verhulst. 2012. Contextualizing Contestation: Framework, Design, and Data. Mobilization 17 (3): 249–262. Vra´blı´kova´, K. 2014. How Context Matters? MOBILIZATION, POLITICAL OPPORTUNITY STRUCTURES, and Nonelectoral Political Participation in Old and New Democracies. Compar- ative Political Studies 47 (5): 203–229. Walgrave, S., and B. Klandermans. 2010. Open and Closed Mobilization Patterns: The Role of Channels and Ties. In The World Says No to War. Demonstrations Against the War on Iraq, ed. S. Walgrave, and D. Rucht, 169–193. Minneapolis: University of Minnesota Press. How government coalition affects demonstration composition…

Walgrave, S., and J. Verhulst. 2009. Government Stance and Internal Diversity of Protest: A Comparative Study of Protest against the War in Iraq in Eight Countries. Social Forces 87 (3): 1355–1387. Walgrave, S., and J. Verhulst. 2011. Selection and Response Bias in Protest Surveys. Mobilization: An International Journal 16 (2): 203–222. Walgrave, S., R. Wouters, and P. Ketelaars. 2016. Response Problems in the Protest Survey Design: Evidence from Fifty-One Protest Events in Seven Countries. Mobilization: An International Journal 21 (1): 81–104.