A win or a flop? Measuring mass protest successfulness in authoritarian settings

Abstract

Previously rare events, mass protest movements have become popular vehicles for those seeking political, economic, and social change. How do we evaluate movement success? Most studies addressing movement outcomes are grounded in the goal attainment approach, where movement success is dependent upon fulfilling one’s stated demands. The models derived from this approach heavily rely on visibility and transparency in the policymaking process. These offer limited analytical utility for scholars studying movements in authoritarian states, where policymaking is shrouded and media is state-controlled. Evaluating movements solely on their fulfillment of mission goals is highly problematic, as movements produce more outcomes than their intended goals. Movements also produce unintended benefits: concessions unrelated to the movement’s mission. These include negative consequences, or societal costs. Since movements produce both positive and negative unintended outcomes, any evaluation of a movement should also incorporate the costs associated with new gains. I argue a cost–benefit approach improves scholarly conceptualization and measurement of protest success. I conceptualize protest success as multidimensional and comprised of protest gains and societal costs. I develop a 21-point scale of protest success using Mokken Scale Analysis. AISP diagnostics indicate gains and costs comprise separate subscales, which are collapsed to produce total sum scores. I score 34 nonviolent movements in authoritarian states between 2002 and 2013 on an additive scale. Protests in authoritarian settings attain considerable accomplishments; however, those gains come with significant cost. Most total success scores are negative, indicating considerable backlash is common during and immediately after the demobilization of movements in authoritarian states. Success scores improve upon the canonical binary measure by: (1) offering improved discrimination between movements, (2) identifying cases of regime ‘ignoring’, and (3) pinpointing misclassified cases. By incorporating negative consequences into our evaluations, we advance our understanding why movements deemed successful by scholars are disappointments to their home publics.

Keywords

anocracy authoritarian protest scaling social movements success

Introduction

How should we evaluate the ‘successfulness’ of mass movements? A decade after the Jasmine Revolution sparked a global surge in mass nonviolence movements, it is appropriate to re-examine how we evaluate movements. In this article, I posit a theoretical shift in the conceptualization of movement success. Traditional metrics of success focus on demonstrators’ ability to extract concessions from the government. However, concessions are not the sole external outcome of movements. Since movements produce both positive and negative external outcomes, any evaluation of a movement should also incorporate the costs associated with new gains. I propose a shift away from the goal attainment approach (which focuses solely on movement victories vis-á-vis their agenda) to a cost–benefit approach (where we analyze a movement’s scorecard of positive and negative consequences). I improve upon former measures of protest success by reconceptualizing success as multidimensional: comprised of gains and costs, and relative rather than absolute.

What constitutes success broadly interests conflict scholars and activists alike. While scholars of nonviolent conflict adapted models of political violence, few models examining the outcomes and intricacies of nonviolence travel back to the violence literature. This study attempts to fulfil that need by offering a reconceptualization of success grounded on dynamics and outcomes found in both violent and nonviolent movements (Abrahms, 2006; Pape, 2003). The incremental nature of success imbued in this model directly addresses the ‘partial success’ arguments of the terrorism and revolution literature. Scholars such as Thomas (2014) argue partial achievements, such as terrorist participation in negotiation settlements due to their terroristic activities, are indicative of success. By treating success as a multidimensional spectrum rather than discrete and unidimensional, I provide a general model of success with utility for both violent and nonviolent conflict scholars.

Despite the explosive growth of the nonviolent conflict literature, scholarly definitions of campaign ‘success’ are often at odds with societal realities. The same movements deemed wildly successful by scholars can appear to their home societies and even activists (Useem & Goldstone, 2022; Suh, 2014; Goodwin & Jasper, 2006; Giugni, McAdam & Tilly, 1999) as disappointments. The source of this disconnect is the exclusion of negative consequences from prevailing definitions of movement success. I bridge this disconnect by incorporating previously omitted positive and negative unintended consequences, producing an improved measure of protest success. Using Mokken Scale Analysis, I comparably measure the net successfulness of 34 antigovernmental movements in authoritarian regimes from 2002 to 2013. I find protests in authoritarian states effectively obtain positive outcomes, but often at significant cost. Once societal costs are accounted for, mass protests in authoritarian states have strongly negative scores for their societies. The prevalence of negative scores offers us insight into the disappointment literature, where dissatisfaction rates continue to rise even in success cases like Tunisia (Ianchovichina, 2018: 146). Simply analyzing the victories of the Jasmine Revolution, without accounting for the costs, obscures our ability to understand Tunisians’ persistent dissatisfaction.

Defining success

Mass protests are dangerous endeavors. When evaluating a movement, scholars are typically concerned with a movement’s internal or external consequences, or outcomes. The outcomes of anti-governmental demonstrations are vital to efficacy estimations of this type of political behavior. Internal outcomes, such as internal cohesiveness, mission creep, and the loss of nonviolent discipline are outside this study’s scope. Instead, I focus on movements’ political and economic external outcomes, representing shifts in a country’s polity, policies, or politics (Giugni, 1998; Giugni, Bosi & Uba, 2013). Protest outcomes may be intermediate or long-term, social, political, economic, cultural, ideological, and personal or biographical (Amenta, 2007; Giugni, 2008).

Studies which measure and compare movement outcomes remain surprisingly scarce. This is partially because outcomes are complex and dynamic (Snyder & Kelly, 1979), contain positive and negative unintended consequences (Rucht, 1998; Piven & Cloward, 1977), and are notoriously difficult to measure over time and space (Earl, 2000; Amenta & Young, 1999). Amenta et al. (2010: 295) argue outcome studies are scarce as they ‘require analysis over long time periods and across many different movements, issue areas, and countries’. As a result, the protest literature remains predominately preoccupied with protest initiations rather than their outcomes. This gap is problematic as campaign success is of vital interest to both academics and activists. Data collection of protest outcomes remains a significant barrier to quantitative studies of protests. Over the last 15 years, data collection efforts greatly reduced barriers to studying protests in repressive regimes. Publicly available datasets, such as NAVCO (Chenoweth & Shay, 2019), Mass Mobilization Protest Data (Salehyan et al., 2012), Social Conflict Analysis Dataset (SCAD) (Clark & Regan, 2016), and others track and report protest campaigns across the globe. With a wealth of new data collected on authoritarian states, it is appropriate to re-examine how we judge movements.

The canonical measure of a movement's success is the movement’s effectiveness in obtaining its stated demands (Wang & Piazza, 2016; Tarrow, 2011; Gamson, 2014). While parsimonious, the analysis of protest success is limited to movement outcomes matching the movement’s stated goals. This definition of campaign success remained remarkably static and firmly embedded in the resource and goal attainment approaches. The resource approach evaluates protests based on changes in the structural conditions and institutions of government. Movements are considered solely on their ability to extract resources from the regime (Giugni, 2008: 1584). Gamson (1975) firmly established the goal fulfilment definition of success in his foundational book, The Strategy of Social Protest. Gamson conditioned success upon a group’s ability to obtain their stated goals, establishing three magnitudes of success: (1) full success as the attainment of a group’s stated goals, (2) partial fulfilment of mission goals indicate partial success (new advantages), and (3) failure status applies to movements with non-appreciable progress in their stated goals (Goldstone, 1980).

Despite robust counter-arguments that success is not conditional upon acquiring new advantages (Meyer & Whittier, 1994), most analyses continue to utilize the goal attainment approach. The goal attainment approach spawned numerous policy studies evaluating movements’ influence on policymaking (Amenta et al., 2005, 2010; Rochon & Mazmanian, 1993; Kitschelt, 1986; Schumaker, 1975). Institutional studies expanded the policy models to include outcomes reflecting institutional changes hierarchically ordering institutional changes above policy outcomes (Kolb, 2007; Quandragno, 1992). Institutional scholars argued that institutional outcomes were ‘much harder to achieve […] they are also more difficult to reverse once […] realized’ (Kolb, 2007; 34). This hierarchical ordering between policy and institutional outcomes is helpful for scholars analyzing non-democracies, where differential attainment barriers to specific outcomes are particularly salient.

I deploy an alternative measure of success to include positive unintended consequences into the analysis (Tarrow, 2011). Under this technique, a count of a movement’s total number of concessions and new advantages becomes obtainable. This second measurement option offers an improved assessment of a movements’ success as it captures its entire range of positive external outcomes. This option challenges the goal attainment approach’s assessment of variable selection. The goal-attainment approach is ‘based on the assumption that organizations’ goals are identifiable and unambiguous’ (Forbes, 1998: 185) and has been contentious in the literature (Tilly, 1999; Gamson, 1990). Selecting only variables matching a movement’s stated goals ignores the dynamics of the conflict landscape. The fluid entrance and exit of internal and external actors produce multiple – and occasionally conflicting – goals over the duration of a movement. Therefore, we should evaluate the full range of outcomes produced by movements. Movements may have many effects unrelated to their stated demands and exceed or contradict mission goals.

The consequences literature is dominated by policy models, often embedded in the goal attainment approach. These models rely on visibility and transparency in policymaking and are largely based on Western democratic cases (Oliver, Cadena-Roa & Strawn, 2003). These model features are problematic as the wave of movements after the Great Recession were primarily located in authoritarian settings (Tilly & Tarrow, 2015; Tilly & Wood, 2013). Thus, we can expect limited utility for non-democracies with more shrouded policymaking, state-controlled media, and constrained opportunities for public opinion polls. Of concern is the common exclusion of low-level outcomes, such as recognizing the legitimacy of demonstrators’ anger. As movements in Western democracies easily attain governmental and public recognition for their grievance legitimacy and salience, the conflict scholarship moved ‘away from addressing whether movements of organizations are successful in gaining new benefits or acceptance (Gamson, 1990)’ (Amenta et al., 2010: 289). Low level outcomes, like recognition, were dropped from models as studies shifted to ‘examining the causal influence of movements on political outcomes and processes drawn from political sociology literature’ (Amenta et al., 2010: 289). However, these low-level outcomes are themselves success stories (Mirowsky & Ross, 1981), especially for demonstrators in settings where resistance movements have typically been repressed or ignored.

The re-incorporation of low-level outcomes is crucial for the applicability of models to non-democracies. Some concessions, such as state recognition of demonstrator legitimacy, are rarer in non-democracies as even minor improvements may signal demonstrators’ legitimacy and salience to the public, attracting new participants. As movements often obtain rounds of concessions, small concessions may act as gateways to larger ones, encouraging mobilization and endangering state stability. This domino effect of concessions is exemplified in Sudan, where demonstrators leveraged early concessions to bolster their public support and introduce significant regime changes. In January 2019, the Sudanese public engaged in sustained protests despite unrelenting and horrific state repression. The military acquiesced to public pressure and removed long-time dictator Omar al-Bashir in a coup. Demonstrators were emboldened and returned to the streets in October 2021, calling for civilian rule. Civilian rule was a radical proposal, as the Sudanese military has ruled almost continuously since independence. Despite harsh violence and delaying tactics, demonstrators persevered. An interim dual civilian-military government was established, with the military promising full civilian rule by 2024.

The goal attainment approach hinders our ability to fully understand a movement’s political and economic outcomes and impacts on society. The political outcomes literature enjoys a moderate number of studies, unlike the diminutive – but growing – economic outcomes literature. Most of these studies evaluated movements with a narrow lens. As Kolb (2007: 4) notes, external political outcomes are more than just campaign goals; movements produce ‘types of outcomes that are related to the state and changes in its policies, politics, and polity, as well as the consequences of these changes for the society at large’. While many studies focus on the first half of Kolb’s definition – state changes attributable to a movement – few address the second half of the definition – societal consequences. My study contributes to the literature by incorporating (unintentional) societal consequences into the definition and operationalization of movement success.

A movement’s intended societal consequences are typically plainly stated by its leadership. However, movements also produce a range of unintentional outcomes, both positive (Tilly, 1999) and detrimental (Amenta et al., 2005; McCarthy & McPhail, 1998; Rucht, 2007; Piven & Cloward, 1977). Positive unintentional consequences indicate when ‘challengers may fail to achieve their stated program – and thus be deemed a failure – but still win substantial advantages for their constituents’ (Amenta et al., 2005: 518). The 2014 revision of Tunisia’s constitution exemplifies a positive unintended consequence. At that time, a new constitution was neither an original nor attendant demand of demonstrators. Rather, the new Tunisian constitution was an offspring of the movement’s aftermath and not causally related to demonstrators’ stated demands, designating it an unintended positive outcome. These types of new benefits are hereafter treated as societal gains, comparable to governmental concessions.

This study’s major contribution is the insertion of negative consequences into the definition of success. Any new protest gains are likely to be accompanied by societal costs. Demonstrators may face ‘negative sanctions, such as restrictions on free speech, violations of life integrity rights, such as torture and political imprisonment, as well as widespread state terror in the form of genocide’ (Carey, 2006: 2). Under the law of coercive responsiveness (Davenport, 2007), regime retaliation against demonstrators is likely as mass mobilizations are particularly threatening. Lichbach’s 5% Rule (1995) posits that government collapse is inevitable once 5% of the population actively mobilizes. Chenoweth & Belgioioso (2019) found that as little as 3.5% mobilization of the populace is sufficient to bring about the demise of a regime. Regimes may then perceive large crowds as existential threats and respond with increasing ferocity to demobilize protests. Demonstrations may also inflict material costs, such as property damage or asset seizures. Economic consequences may especially impact public support of demonstrations. Disruptive campaigns may invoke backfire in the form of counter-protests, anti-demonstrator repressiveness by regime supporters, and regional and international interventions. The backfire literature traditionally focused on public reactions to violent campaigns (Aytaç et al., 2020; Simpson et al., 2018; McLeod, 2007; Stephan & Chenoweth, 2008; Hess & Martin, 2006; Carey, 2010). However, nonviolent movements can also spark public backfire, evidenced by increased polarization, civilian repression by police and military, and international intervention. Demonstrators can expect, even in democracies, to pay a heavy cost: whether jailing, surveillance, or police retaliation. The 2020 US Black Lives Matter movement exemplifies how marginalized groups in democracies can expect to face disproportionately harsh costs (Davenport et al., 2018; Simpson et al., 2018; Bandura et al., 1996).

Regime repression of mass movements is unsurprising. All regimes seek to maintain themselves. Davis & Ward (1990) found a reciprocal relationship between protest and repression. Various studies found an interdependency between regime concessions and state repression (Rasler, 1996; Muller & Opp, 1986) and the policy effectiveness literature treats movement success as interactive and conditional (Giugni, 2004, 2007, 2008); Amenta et al., 1992, 2005; Soule & Olzak, 2004; Cress & Snow, 2000). Despite the clear linkages (Amenta et al., 2010; Kolb, 2007; Giugni, 2007) between repression (costs) and concession (gains), these are typically treated separately, rather than in combination. By analyzing new gains and societal costs in combination, scholars can fulfil Forbes’s (1998: 186) admonition that we should measure effectiveness in ‘several different ways simultaneously’. The unit of analysis is the nonviolent campaign,¹ derived from NAVCO 2.1. I analyze the lifespan of movements for gains and costs, including the year after demobilization. I use Mokken Scale Analysis to systematically capture and analyze the dimensions of success, improving upon previous studies. In the next section, I address success’s dual dimensions of gains and costs.

Measuring success

Besides a more inclusive definition of success, new measures of protest success are needed. Protest success is traditionally measured as a unidimensional binary outcome. However, protest outcomes may be lagged, signed positively or negatively, and only partially observable. Cultural/ideological outcomes are often latent, as are personal/biographical outcomes, and require longer temporal periods to capture. Binary measurement masks important variation in the types of outcomes protests produce. The limitations of binary measurement emerge when comparing two measures of protest success within NAVCO 2.1. Chenoweth & Shay (2019) offer two measures of protest success: (1) a binary ‘success’ variable tracking whether campaigns attained most/all of their stated goals and (2) an ordered six-category² ‘progress’ variable capturing the progress a protest makes towards its campaign goals. The two measures present contradictory interpretations of movements success when evaluating 34 nonviolent movements that occurred in authoritarian states from 2000 to 2013.

The nominal success measure depicted in Figure 1 indicates that protests in authoritarian states overwhelmingly fail. However, the progress variable depicted in Figure 2 portrays the impressive array and level of concessions obtained by movements. The progress variable enhances our understanding of the types and distribution of outcomes obtained by movements (Figure 3). A floor effect, where campaigns consistently obtain only the lowest outcome levels, might be expected in a sample populated with only non-democracies. Rather, concessions present a bimodal distribution: protests are as likely to maintain the status quo as obtain success. Low level benefits, such as the limited outcome, were less frequent than more high-stake concessions such as verbal recognition. This seemingly positive glimpse at outcome acquisition by protests is tempered by the striking absence of any significant concessions. Regimes appear to offer everything (success) or nothing (status quo). Counter to common assumptions of governmental incrementalism, when governments do concede, they do so at full-throated strength. Increasing the level of measurement would afford us a richer understanding of the acquisition effectiveness of movements and the concession types offered by regimes.

NAVCO’s progress variable also implies important differences between anocracies and autocracies. Autocracies experience protest success more often than any other outcome, perhaps due to higher levels of scrutiny in their responses to demonstrators. In comparison, hybrid regimes, or anocracies, were equally likely to maintain the status quo as to succeed. This may be attributable to anocracies’ enhanced protection via third-party/foreign sponsorship (Carey, 2006), dampening the regime’s willingness to concede. Framing success as progressive, rather than absolute, offers enhanced insight into the types of concessions produced by mass movements. We can also disaggregate any accompanying societal costs in the same manner. Below, I explore the multiple dimensions of success in depth.

Measurement model

Protest success is treated in this study as a multidimensional latent concept scorable over benefits (gains) and detriments (costs) produced by protest campaigns. Gains and costs manifest as political, economic, intentional or unintentional (see Figure 4). Although regimes offer economic concessions more frequently than political concessions (Brancati, 2016; Chenoweth & Stephen, 2011), non-political outcomes are often absent in quantitative analyses. This is possibly due to the prevalence of regimes offering relatively low-level economic concessions, such as verbal promises for additional state hiring or reinstatement of food and fuel subsidies, in lieu of systematic economic changes. This was the case in Libya. In response to the unrest in Tunisia, Qaddafi offered generous tax relief to pre-empt potential protesting. Structural economic concessions addressing the developmental and economic grievances underlying protests rarely manifest. These types of concessions pose a severe threat to elite interests, more so than political concessions like executive turnover. The literature on economic concessions is small, but burgeoning. Economic concessions should be incorporated into our evaluations of success as losses such as property damage, asset seizure, and redistributive policies are key legacies of a movement

Figure 1.

Success differences by regime type

(Zemni, 2013; Beinin, 2012; Bishara, 2012). It is then appropriate to treat successfulness as a multidimensional concept capturing a protest campaign’s total scorecard of political and economic gains and costs. Future studies may then use these discrete gains and costs scorecards to compare movements across time and space.

Gains dimension

The gains dimension expands upon current conceptions and measures of protest effectiveness. NAVCO’s six-category progress variable is the template for the gains dimensions. The progress variable solely reflects the extent of a movement’s goal fulfilment. The gains dimension differs from its progress predecessor as intentional as well as unintentional new advantages are included. Unintentional benefits (causally unrelated to a movement’s demands) include gains such as positive media attention, lifting of emergency powers, increased police accountability, and new social services. The gains dimensions touts expanded category definitions, indicators, and an additional category refining term full success.

At the lowest end of the spectrum, status quo represents movements capable of emerging and resisting regime demobilization efforts, but unable to evoke discernable changes in state behavior or obtain observable benefits. This category may appear superfluous at first glance. However, the emergence of protests within heavily repressive regimes previously immune to democratization efforts is itself a feat – even if the protest obtains no discernable impact. This category unavoidably captures instances where regimes simply ignore protestors. Limited gains indicate conciliatory actions which avoid fulfilling any protestor demands. Regimes offer temporary or one-time handouts, such as food or fuel subsidies, intended to defuse public anger. These gains fall short of (1) substantively conferring legitimacy upon the demonstrators, (2) addressing any of the stated demands of demonstrators, (3) shifting the stated positions of legislators, or (4) producing new legislation. Limited gains also capture unintended benefits such as extensive state media coverage that provides outlet and opportunity for activists to promote their grievances and agendas.

Figure 2.

Campaign progress in stated demands across all regime types

Verbal recognition, derived from Gamson’s recognition variable, is governmental acknowledgement of protestor grievances. Verbal recognition may take the form of promises rectifying grievances and/or formal recognition of the movement’s leaders. This concession is an important legitimatization for protest movements, reflecting an important shift in the regime from denial to recognition. Recognition is dangerous for regimes ruling via fiat as it may embolden and legitimize future protests. Subsequently, regimes often offer a slew of concessions without ever recognizing the legitimacy of a movement’s claims. Significant gains represent the attainment of one or more of a movement’s stated demands via a formalized agreement. This category captures regime capitulation to one or more but not the entire slate of demands, nor the maximalist goal. While stated demands may vary in their intensity, scope, and magnitude, the capitulation of the regime to even one formally stated demand construes a sense of victory to the movement. In accordance with the value-expectations model, movement leaders strengthen their platform via these significant gains, using them as leverage for additional concessions and/or framing these concessions as indicative of the movement’s legitimacy and imminent success.

Partial success indicates the fulfilment of the campaign’s demands in relation to their maximalist goal. Movements are partially successful when obtaining executive removal even without attaining any other stated demand. This category is the counterpart of NAVCO’s full success category in the progress variable. NAVCO 2.1, like most studies and databases, considers regime overturn or executive removal as sufficient for full success. I counter that protests obtaining their ultimate demand are only partially successful without acquiring appreciable progress on their grievance source. In Tunisia, protests were a response to systemic employment disparities. Demonstrators’ focus shifted away from employment and towards Ben Ali’s removal once he activated the repressive state apparatus. While Tunisians

Figure 3.

Progress towards stated demands by regime type

Figure 4.

Measurement model

were successful in removing Ben Ali, successive governmental policies failed to substantively tackle the rampant underemployment and unemployment which sparked the protests. My model includes an additional category, full success, which captures structural, institutional, or cultural outcomes directly addressing the initializing grievance. While full success is included in the theoretical model, it should be noted that due to absence of observed events, it is not included in the reduced model.

Costs dimension

The costs dimension is intended to contain economic and political negative consequences commonly found in an active protest scenario. As repression against protestors is often a persistent feature of all regimes, especially in anocracies and autocracies (Carey, 2006), data on repressive state behavior is widely collected. Unfortunately, the same cannot be said of economic or financial fallouts of movements. This study leverages the ready availability of data on state repression, while expanding the definition of costs to include indicators of non-repressive backlash. To do so, the costs³ dimension is modeled after NAVCO 2.1’s repression variable but framed to capture more than repressive behaviors like surveillance, policing, and arrests. The costs dimension goes beyond simple repression to include adverse backlash such as post-movement democratic backsliding, and financial and economic instability. Thus, the costs dimension contains repression but is more than a repression variable. The inclusion of negative consequences like foreign direct investment flight, stock market crashes, and erosions in democratic values is novel for a quantitative success model. Most protest success models exclude such costs, as negative consequences are outside of a movement’s organizational control. However, costs, like concessions, indicate regime responsiveness to the presence and demands of demonstrators. It is difficult to justify omitting societal costs when analyzing campaign outcomes, especially for those in authoritarian settings; 88% of this sample reported extreme repression against demonstrators. Repression exerts heavy costs upon protestors and takes on material, physical, and psychological forms. Despite its persistence, repression is not static across time or space. The same actors may report differing state behavior across events. Under a cost–benefit approach, it is reasonable to include repression as a dimension of protest impact when considering the costs paid to obtain new gains.

Together, four categories capture the intensifying costs paid by society. The first three cost categories are modeled after the repression variable found in NAVCO 2.1 (Chenoweth & Shay, 2019), with an additional genocide category. Each category has been renamed to allow for the inclusion of non-repression indicators like democratic backsliding. Low backlash includes the ‘non-material or non-political backlash’ of verbal threats, elevated policing, and economic retaliation included in NAVCO’s repression variable. During the 2020 Black Lives Matter protests, US state and federal officials unleashed elevated policing to demobilize demonstrators. Indicators included are typically non-lethal deployment of tear gas, rubber bullets, batons, and tasing. Additionally, non-repressive retaliation such as workplace firings for participation or blackballing are included in this category. This differs from the formalized repressiveness and retaliation captured in the moderate backlash category. This category includes indicators of material and political backlash without the intent to kill. For repression, this includes regime and regional responses intended to cause significant harm beyond demobilizing street protests.⁴ This category addresses material and physical harm, such as destruction of property, restrictions on freedom of movement, and government sanctioned kidnappings. Foreign interventions like the deployment of regional troops to Bahrain are included, as are psychological and sexual violence against civilians. Examples include maiming of civilians by governmental officials (such as throwing acid in someone’s face). This category represents a wide range of dangerous and potentially, but not intentionally, lethal methods.

The reduced model’s high backlash category captures ‘material and/or political backlash intended to kill’. This category represents intentionally lethal policy/behavior by the government or third parties. Among the indicators are the deployment of death squads, militant groups known for enacting mass violence, deliberative and coordinated police violence, and deployment of military or security forces against civilians. Non-security backlash also includes deliberate and incidental deprivation of access to sanitation, food, health care and health resources and materials, shelter, and communications, as well as biological, chemical, or pathogen exposure, and deliberate ecological destruction that would cause starvation or famine, for example. The theoretical model contains a genocide category, where the regime deliberately and formally targets protestors as a group for eradication or expulsion. The reduced model excludes this category due to a lack of sample observations. As Ulfelder & Valentino (2008: 9) note, genocide typically occurs in the context of violent conflicts, such as civil wars or revolutions. In the next section, I discuss an application of the multidimensional success model.

Application

This study aims to produce a scorecard proof of concept usable for future analyses comparatively evaluating the outcomes of nonviolent movements. Using Mokken Scale Analysis, I evaluated 34 protests in strongly authoritarian regimes⁵ (Polity scores from 0 to –10) from NAVCO 2.1 (Chenoweth & Shay, 2019). The case set showcases highly mobilized campaigns as the NAVCO dataset only features mature campaigns. I cross-referenced participant counts with the NAVCO 1.3 dataset. Protests drop out of the analysis in the year they attain a violent flank or transition to civil war. As such, the dataset eliminates many separationist protests and anti-foreign occupation protests. While this reduces the potential number of cases, it offers a unique contribution to the literature. Most conflict studies compare violent to nonviolent campaigns. Few comparatively analyze solely nonviolent movements. This is important as violence changes the cost calculations of regimes when responding to demonstrators. A regime’s willingness to concede is likely to sharply differ when facing nonviolent as opposed to violent demonstrators. Additionally, the types and quality of concessions are likely to sharply diverge. To analyze the responsiveness of regimes to nonviolent campaigns, I explicitly limited the case selection to nonviolent movements.

To produce campaign scores, I employed a psychometric approach. Scaling analysis and procedures,⁶ such as Guttman scales, produce summed scores representing an individual’s position along some latent trait or ability. As cultural and personal impacts often remain hidden, delayed, or deeply embedded into other outcomes, it is appropriate to treat success as latent. I use observed success features to produce the success scale. To scale protests, both multidimensional and unidimensional scaling analyses are available. Item response theoretic (IRT) models allow for the joint estimation of protest campaign locations along the latent trait while simultaneously mapping between a protest’s latent positioning and observed outcomes. This mapping is otherwise known as item response functions (IRFs). IRT models, whether parametric or nonparametric, offer dimensionality analysis and ability mapping of cases. Non-parametric (NPIRT) models differ from their parametric versions as they cannot jointly estimate latent positions and IRFs. Despite this limitation, non-parametric models feature more relaxed assumptions that are helpful when dealing with binary data.

Mokken Scale Analysis, or MSA, is a non-parametric, probabilistic case of the Guttman scale and allows for the ‘ordering of persons on a scale by means of a person’s total scores on a set of items’ (Sijtsma & van der Ark, 2017: 137). While less familiar than other IRT models, cutting-edge projects such as the CIRI Human Rights Dataset utilize MSA (Cingranelli et al., 2014). MSA evaluates each subject’s pattern of responses to a set of items that act as indicators for a latent variable. MSA then creates a scale from the items and produces separate item and subject parameters.⁷ IRT models feature both item difficulty ordering and discrimination, the distinction between cases or persons. Higher scores then indicate a higher successfulness attribute for protests. MSA has several key features: data reduction allows scholars to treat multidimensional concepts as unidimensional; ‘includes an item parameter that shows how items differ in their distribution’; and ‘it can be applied in situations in which latent variables must be operationalized with only a small number of indicators’ (van Schuur, 2003: 139). The relaxed assumptions under MSA can handle mixed models consisting of continuous latent and non-continuous variables. MSA, unlike parametric procedures, allows for ‘exploratory dimensionality analysis not readily available for PIRT models’ or non-parametric models like Rasch models (Palmgren et al., 2018: 3). Unlike principal component analysis and factor analysis, Guttman scales and their special cases like MSA do not assume parallel measures. Factor analysis assumes continuous data and so struggles to interpret dichotomous data. MSA easily handles dichotomous data interpretation while avoiding the determinism and stricter assumptions of the Guttman scales.

Scaling costs and gains

I theorize success as a left–right continuum of gains and costs, ranging from abject failure to transformative success. A 21-point scale was constructed from –10 to 10 to reflect a protest’s position along the latent mapping from genocide to full success. Movements were scored on the highest gains and costs categories they fell on the theoretical model. None of the movements garnered concessions substantively addressing the initializing grievance, no movements were awarded the full 10 points allotted to fully successful campaigns. Similarly, none of the sample cases obtained the –10 scores allotted to instances of genocide. Success scores consist of a campaign’s gains minus the societal costs accrued. Costs are allowed to exceed gains to produce negative scores. I then estimated protests and item parameters (costs and gains variables) to position protests along the latent dimension via invariant item ordering. MSA analysis orders and positions items and subjects due to two submodels, the monotone homogeneity (MH) model and the double monotonicity (DM) model – a special case of the MH model. The MH model assigns an ordinal placement for a protest on the success dimension (θ) based on the observable test scores. The main assumptions of MSA (unidimensionality, monotonically non-decreasing IRFs, and local stochastic independence) come from the monotone homogeneity model. To establish a unidimensional scale of protest success, we must first identify any embedded scales, establish item scalability,⁸ and finally order items and subjects along the resulting success scale. See Figure 5.

Gains and costs were each found to comprise separate subscales, and so separately evaluated for their respective item, item-pair, and test scalability coefficients. I collapsed the now established subscales into a unidimensional scale of success. Summed scores of protest gains and reverse coded costs produced general campaign scores. Before scoring, I identified a true zero benchmark case to center the scale. Only a single case met the definition of a true zero, with an absence of gains and absence of costs. The NAVCO 2.1 dataset classified the 2010 Pro-Aristide protest as a failure. It was a one-day protest where demonstrators peacefully marched and dispersed without any response from the government. It is an excellent example of regime ‘ignoring’, as it failed to evoke any reaction at

Figure 5.

Dimensions of success

Figure 6.

Density plot of success scores

all from the regime, either positive or negative. Media coverage was marginal (a mere ten articles in the top Google hits), and demonstrators failed to gather any subsequent gains in momentum or mobilization. Unlike ‘toleration’, ignoring is not a form of attrition and is not intended to undermine demonstrator claims or group strength (Yuen & Cheng, 2017: 617).

Results

So how well do campaigns score when we consider their accomplishments and costs? Their total sum scores offer

Figure 7.

How did campaigns score?

insight into the magnitude of success produced by the protests during their respective campaign lifespans.

As indicated in Figure 6, the density of protests obtained negative total scores on success, suggesting that costs overwhelm the gains demonstrators extract from their home governments. Negative scores were obtained by 64.7% of the sample, with just under half of the protests obtaining highly negative scores of –5.9 or more. While one may argue we should expect repressive behavior towards demonstrators in authoritarian states, it is

Table I.

Comparative measure scores

Campaign name	Regime type	Binary	Gains score (out of 10)	Success scores
Anti-Al-Bashir Government	Hybrid	Fail	6	–2.7
Anti-Ali Abdullah Saleh	Hybrid	Fail	2	–5.9
Anti-Aristide	Hybrid	Success	8	0.5
Anti-Aziz Protests	Authoritarian	Fail	2	–5.9
Anti-Ben Ali Campaign (Jasmine Revolution)	Hybrid	Success	8	0.5
Anti-Gnassingbe/Coup Crisis	Hybrid	Fail	2	–7.5
Anti-Islamist Government Protests	Hybrid	Fail	4	–4.3
Anti-King Hamad Campaign	Authoritarian	Fail	6	–2.7
Anti-Military Government Campaign	Hybrid	Success	8	3
Anti-Morsi Protests	Hybrid	Success	8	0.5
Anti-Mubarak Movement	Hybrid	Success	8	0.5
Anti-Thaksin	Authoritarian	Success	8	8
Awami League Protests	Authoritarian	Success	8	0.5
Belarus Pro-Democracy Campaign	Authoritarian	Fail	6	–2.7
Civil Movement for Democracy	Hybrid	Fail	2	–5.9
Euromaiden	Hybrid	Fail	2	–5.9
Green Revolution and Day of Rage	Authoritarian	Fail	2	–5.9
Guinean Pro-Democracy Movement, 2009	Hybrid	Fail	6	0.5
Guinean Pro-Democracy Movement, 2007	Hybrid	Success	8	–2.7
Kifaya	Authoritarian	Fail	4	–4.3
Kyrgyz Anti-Interim Government	Authoritarian	Fail	2	–7.5
Kyrgyz Second Revolution	Authoritarian	Success	8	0.5
Let’s Save Togo (Anti-Gnassingbe)	Hybrid	Fail	2	–5.9
Pro-Aristide Campaign, 2010	Hybrid	Fail	2	0
Pro-Aristide Campaign, 2005	Hybrid	Fail	2	–5.9
Pro-Morsi Protests	Hybrid	Fail	2	–5.9
Pro-Ouattara Campaign	Hybrid	Fail	6	–2.7
Protest for Constitutional Reform	Hybrid	Fail	6	6.4
Saffron Rebellion	Authoritarian	Fail	2	–7.5
Southern Yemen Secessionist Movement, 2007	Hybrid	Fail	4	–2.7
Southern Yemen Secessionist Movement, 2013	Hybrid	Fail	6	–4.3
Swaziland Anti-Monarchy Protests	Authoritarian	Fail	2	–7.5
Syrian Civil Conflict	Authoritarian	Fail	2	–5.9
Tulip Revolution	Authoritarian	Success	8	0.5

notable that most protests obtained the same level of repression no matter which gains they attained. As such, protestors in these regimes may expect to experience the highest levels of societal sacrifices even if gains are relatively minimal. Protest events scored far worse than I expected. Rather than falling only slightly under zero, most protests scored between –7.7 and –2.7. These are not slight failures; significant societal harm occurred during protests.

As seen in Figure 7,⁹ campaigns with positive scores are only 35% of the sample and most barely succeed (scores of 0.5). This result offers us an enhanced understanding of campaign outcomes during the movement. Under the goal attainment approach, the binary measure paints a picture of protests as either wildly successful or abject failures based solely on their ability to garner new concessions. When we evaluate campaigns occurring in authoritarian settings on their full range of positive outcomes, they perform admirably in the gains dimension.¹⁰ Despite these accomplishments, the combined scores of gains and costs illuminates why afterwards publics may express disappointment or disillusionment with a popular movement.

The cost–benefit approach offers (1) enhanced benefits over the goal attainment approach in evaluating movements on their accomplishments and (2) a rationale for post-movement public disappointment. The goal attainment approach hinders our understanding and acknowledgement of the broad range of outcomes produced by movements and encourages suboptimal binary measurement. Binary measures fail to elucidate important distinctions between protests falling within the same failure/success category. While some clustering remains, success scores allow scholars to better discriminate within categories.

How do my success scores compare to the original binary measure?¹¹ There are several immediate differences. Table I compares the binary measure, the gains dimensions, and the combined success scores. Unlike the binary measure, the success scores identify cases of regime ‘ignoring’, such as Haiti’s 2010 Pro-Aristide protests. Regimes that ignore demonstrators do not actively repress demonstrators nor actively acquiesce, resulting in zero score cases. Binary measures are unable to identify these types of regime responses. Additionally, the success scores identify misclassified cases, where campaigns are categorized as failures despite obtaining high-level concessions. The NAVCO dataset codes Jordan’s Protest for Constitutional Reform as a failure despite receiving a ‘progress’ score of 3 and experiencing only moderate repression. The Jordanian protest is my model’s second highest-scoring protest (6.4/10). The issue of misclassification is essential, as 30% of the 24 movements coded as failures under the binary measure obtained verbal concessions (formal recognition and negotiations) or higher. This initial study is a small sample, and so misclassification instances under binary measurement could compound over a large sample.

Conclusion

In this article, I argue a cost–benefit approach offers an enhanced understanding of movement outcomes over the traditional goal attainment approach. The goal attainment approach limits our conceptualization of a movement’s full range of positive and negative outcomes. The binary success/failure typically deployed under a goal attainment approach is deficient in (1) capturing the full variation of outcomes produced by movements, (2) identifying cases of regime ‘ignoring’, and (3) identifying misclassified cases. The success scores combining negative and positive outcomes offer a snapshot of why the public is often disappointed after mass movements demobilize.

There are some caveats to the above conclusions. Designating an outcome as negative or positive is a normative endeavor. Evaluating a campaign by the stated demands is much simpler as it reduces the possibility of researcher bias. This study does not engage in this very relevant conversation over the conditionality of assigning a positive or negative sign to an outcome. Instead, this study’s purpose is to include societal outcomes outside of the goal attainment approach when quantifying protest success. Future research and evaluation of protests’ intermediate and long-term effects may ultimately switch an outcome from positive to negative. However, the identification and specification of oft-excluded outcomes provides a more robust evaluation of the multiple dimensions of consequences produced by mass protests.

A fundamental limitation of this study is the scope and intent of the underlying data. The NAVCO dataset, like many datasets, is intended to compare violent to nonviolent movements. My study is unique as it compares nonviolent to nonviolent movements. Gathering additional data and conducting within-variation studies of nonviolent resistance movements offers scholars and activists fruitful new information on the relationship between movement tactics and outcomes. This application is only one of many that could, and should, be conducted. Scholars should gather outcome data whenever they gather information on movements.

Measurement of protest success often excludes unintended consequences, particularly negative consequences. The resource and goal attainment approach produced measures which focused solely on the extractive ability of a protest campaign in terms of its stated goals. However, protests also produce important additional positive and negative consequences. I offer a new approach to conceptualizing and measuring success in the conflict literature. The success scores I produced using Mokken Scale Analysis indicate most of the sample’s protests have low levels of success despite garnering new advantages, and protests may incur tremendous costs to society. This finding has important implications for the conflict literature. As high levels of repression and backlash are common across all cases, demonstrators can expect to always pay extremely heavy costs. While this may simply be a feature of the regime type analyzed in this study, repression and backlash are also common features in democracies. The commonality of repression and backlash offers an additional metric of comparison between nonviolent and violent movements.

This piece furthers our understanding of movement outcomes. While we know movements may also produce social bads, negative consequences are rarely integrated into our success measures. By treating success as multidimensional, we can account for all outcomes movements produce, not just positive ones. From economic repercussions, societal instability, to democratic degradation: data on the negative impacts of movements must also be gathered for accurate and relevant measurement of movement success. The cost–benefit approach espoused in this study provides a more nuanced understanding of movement success and provides a foundation for future scholarship which maps the full post-movement aftermath.

Footnotes

Replication data

The dataset, codebook, and do-files for the empirical analysis in this article, along with the Online appendix, are available at . All analyses were conducted using R and Stata.

Acknowledgements

I would like to thank Erica Chenoweth, Carew Boulding, Christopher Shay, Jay Ulfelder, Erin Cikanek, Augusta Dell’Omo, Christine Bird, Adam Enders, and Tyler Girard for their support and insightful feedback. I also thank the participants of the International Security Program and the Political Violence Workshop at Harvard’s Belfer Center for their valuable comments.

ORCID iD

Kimberly Turner

Notes

References

Abrahms

Max

(2006) Why terrorism does not work. International Security 31(2): 42–78.

Amenta

Edwin

(2007) When Movements Matter: The Townsend Plan and the Rise of Social Security. Princeton, NJ: Princeton University Press.

Amenta

Edwin

Young

Michael P

(1999) Making an impact: Conceptual and methodological implications of the collective goods criterion. In: Giugni

Marco

McAdam

Doug

Tilly

Charles

(eds) How Social Movements Matter, New edition. Vol. 10. Minneapolis, MN: University of Minnesota Press, 22–41.

Amenta

Edwin

Caren

Neal

Chiarello

Elizabeth

Yang

(2010) Political consequences of social movements. Annual Review of Sociology 36: 287–307.

Amenta

Edwin

Caren

Neal

Olasky

Sheera Joy

(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.

Amenta

Edwin

Carruthers

Bruce

Zylan

Yvonne

(1992) A hero for the aged? The Townsend movement, the political mediation model, and US old-age policy, 1934–1950. American Journal of Sociology 98(2): 308–339.

Aytaç, S

Erdem

Rau

Eli Gavin

Stokes

Susan

(2020) Beyond opportunity costs: Campaign messages, anger and turnout among the unemployed. British Journal of Political Science 50(4): 1325–1339.

Bandura

Albert

Barbaranelli

Claudio

Caprara

Gian Vittorio

Pastorelli

Concetta

(1996) Mechanisms of moral disengagement in the exercise of moral agency. Journal of Personality and Social Psychology 71(2): 364–374.

Beinin

Joel

(2012) The rise of Egypt’s workers. Carnegie Endowment for International Peace (https://carnegieendowment.org/files/egypt_labor.pdf).

10.

Bishara

Marwan

(2012) The Invisible Arab: The Promise and Peril of the Arab Revolutions. New York: Nation.

11.

Brancati

Dawn

(2016) Democracy Protests: Origins, Features, and Significance. Cambridge: Cambridge University Press.

12.

Carey

Sabine

(2006) The dynamic relationship between protest and repression. Political Research Quarterly 59(1): 1–11.

13.

Carey

Sabine

(2010) The use of repression as a response to domestic dissent. Political Studies 58(1):167–186.

14.

Chenoweth

Erica

Belgioioso

Margherita

(2019) The physics of dissent and the effects of movement momentum. Nature Human Behaviour 3(10): 1088–1095.

15.

Chenoweth

Erica

Shay

Christopher Wiley

(2019) NAVCO 2.1 Dataset and Codebook (https://doi.org/10.7910/DVN/MHOXDV/WXVKGK, Harvard Dataverse, V2).

16.

Chenoweth

Erica

Stephan

Maria J

(2011) Why Civil Resistance Works: The Strategic Logic of Nonviolent Conflict. New York: Columbia University Press.

17.

Cingranelli

David L

Richards

David L

Chad Clay

(2014) The CIRI Human Rights dataset, Version 2014.04.14 (http://www.humanrightsdata.com).

18.

Clark

David

Regan

Patrick

(2016) Mass Mobilization Protest data (https://doi.org/10.7910/DVN/HTTWYL, Harvard Dataverse, V5, UNF:6: F/k8KUqKpCa5UssBbL/gzg== [fileUNF]).

19.

Cress

Daniel

Snow

David

(2000) The outcomes of homeless mobilization: The influence of organization, disruption, political mediation, and framing. American Journal of Sociology 105(4): 1063–1104.

20.

Davenport

Christian

(2007) State repression and political order. Annual Review of Political Science 10: 1–23.

21.

Davenport

Christian

McDermott

Rose

Armstrong

David A

(2018) Protest and police abuse: Racial limits on perceived accountability. In: Bonner

Michelle D

Seri

Guillermina

Mary

Rose Kubal

Kempa

Michael

(eds) Police Abuse in Contemporary Democracies. New York: Palgrave Macmillan, 165–192.

22.

Davis

David R

Ward

Michael D

(1990) They dance alone: Deaths and the disappeared in contemporary Chile. Journal of Conflict Resolution 34(3): 449–475.

23.

Earl

Jennifer

(2000) Methods, movements, and outcomes. In: Coy

Patrick G

(ed.) Research in Social Movements, Conflicts and Change, Volume 22. Bingley: Emerald, 3–25.

24.

Forbes

Daniel

(1998) Measuring the unmeasurable: Empirical studies of nonprofit organization effectiveness from 1977–1997. Nonprofit and Voluntary Sector Quarterly 27(2): 183–202.

25.

Gamson

William

(1990) The Strategy of Social Protest. 2nd edition. Belmont, CA: Wadsworth.

26.

Gamson

William

(2014) Measuring movement outcomes. Mobilizing Ideas 1 January (https://mobilizingideas.wordpress.com/2014/01/01/measuring-movement-outcomes/).

27.

Giugni

Marco G

(1998) Was it worth the effort? The outcomes and consequences of social movements. Annual Review of Sociology 24: 371–393.

28.

Giugni

Marco G

(2004) Personal and biographical consequences. In: Snow

David A

Soule

Sarah A

Kriesi

Hanspeter

(eds) The Blackwell Companion to Social Movements. Malden, MA: Blackwell, 489–507.

29.

Giugni

Marco

(2007) Useless protest? A time-series analysis of the policy outcomes of ecology, antinuclear, and peace movements in the United States, 1977–1995. Mobilization 12(1): 53–77.

30.

Giugni

Marco

(2008) Political, biographical, and cultural consequences of social movement. Sociology Compass 2(5): 1582–1600.

31.

Giugni

Marco

Bosi

Lorenzo

Uba

Katrin

(2013) Outcomes of social movements and protest activities. In: Oxford Bibliographies. Oxford: Oxford University Press. (https://www.oxfordbibliographies.com/view/document/obo-9780199756223/obo-9780199756223-0037.xml).

32.

Giugni

Marco

McAdam

Doug

Tilly

Charles

, eds (1999) How Movements Matter: Theoretical and Comparative Studies on the Consequences of Social Movements. Minneapolis, MN: University Minnesota Press.

33.

Goldstone

Jack A

(1980) The weakness of organization: A new look at Gamson’s The Strategy of Social Protest. American Journal of Sociology 85(5): 1017–1042.

34.

Goodwin

Jeff

Jasper

James M

(2006) Emotions and social movements. In: Stets

Jan E

Turner

Jonathan H

(eds) Handbook of the Sociology of Emotions. Boston, MA: Springer, 611–635.

35.

Hess

David

Martin

Brian

(2006) Repression, backfire, and the theory of transformative events. Mobilization 11(1): 249–267.

36.

Ianchovichina

Elena

Bader

Suleiman Abu

(2018) Unintended consequences? Foreign intervention, polarization, and conflict in MENA. Brookings 17 October (https://www.brookings.edu/blog/future-development/2018/10/17/unintended-consequences-foreign-intervention-polarization-and-conflict-in-mena/).

37.

Kitschelt

Herbert P

(1986) Political opportunity structures and political protest: Anti-nuclear movements in four democracies. British Journal of Political Science 16(1): 57–85.

38.

Kolb

Felix

(2007) Protest and Opportunities: The Political Outcomes of Social Movements. Frankfurt/New York: Campus.

39.

Lichbach

Mark I

(1995) The 5% rule. Rationality and Society 7(1): 126–128.

40.

McCarthy

John D

McPhail

Clark

(1998) The institutionalization of protest in the United States. In: Meyer

David S

Tarrow

Sidney G

(eds) The Social Movement Society: Contentious Politics for a New Century. Lanham, MA: Rowan & Littlefield, 83–110.

41.

McIver

John

Carmines

Edward G

(1981) Unidimensional Scaling. Newbury Park, CA: Sage.

42.

McLeod

Douglas M

(2007) News coverage and social protest: How the media’s protect paradigm exacerbates social conflict. Journal of Dispute Resolution 12(1): 185–194.

43.

Meyer

David S

Whittier

Nancy

(1994) Social movement spillover. Social Problems 4(2): 277–298.

44.

Mirowsky

John

Ross

Catherine E

(1981) Protest group success: The impact of group characteristics, social control, and context. Sociological Focus 14(3): 177–192.

45.

Muller

Edward N

Opp

Karl-Dieter

(1986) Rational choice and rebellious collective action. American Political Science Review 80(2): 472–487.

46.

Oliver

Pamela

Cadena-Roa

Jorge

Strawn

Kelley D

(2003) Emerging trends on the study of protest and social movements. Research in Political Sociology 12: 213–244.

47.

Palmgren

Per J

Brodin

Ulf

Nillson

Gunnar

Watson

Roger

Stenfors

Terese

(2018) Investigating psychometric properties and dimensional structure of an educational environment measure (DREEM) using Mokken Scale Analysis: A pragmatic approach. BMC Medical Education 18: 235.

48.

Pape

Robert A

(2003) The strategic logic of suicide terrorism. American Political Science Review 97(3): 343–361.

49.

Piven

Frances

Cloward

Richard A

(1977) Poor People’s Movements: Why They Succeed, How They Fail. New York: Pantheon.

50.

Quadagno

Jill

(1992) Social movements and state transformation: Labor unions and racial conflict in the war on poverty. American Sociological Review 57(5): 616–634.

51.

Rasler

Karen A

(1996) Concessions, repression, and political protest in the Iranian revolution. American Sociological Review 61(1): 132–152.

52.

Rochon

Thomas

Mazmanian

Daniel

(1993) Social movements and the policy process. Annals of the American Academy of Political and Social Science 528: 75–87.

53.

Rucht

Dieter

(1998) The structure and culture of collective protest in Germany since 1950. In: Meyer

David S

Tarrow

Sidney G

(eds) The Social Movement Society: Contentious Politics for a New Century. Lanham, MA: Rowman & Littlefield, 29–57.

54.

Rucht

Dieter

(2007) On the sociology of protest marches. In: Reiss

Matthew

(ed.) The Street as Stage: Protest Marches and Public Rallies since the Nineteenth Century. Oxford: Oxford University Press, 49–57.

55.

Salehyan

Idean

Hendrix

Cullen S

Hamner

Jesse

Case

Christina

Linebarger

Christopher

Stull

Emily

Williams

Jennifer

(2012) Social conflict in Africa: A new database. International Interactions 38(4): 503–511.

56.

Schumaker

Paul D

(1975) Policy responsiveness to protest-group demands. Journal of Politics 37(2): 488–521.

57.

Sijtsma

Klaas

Andries van der Ark

(2017) A tutorial on how to do a Mokken Analysis on your tests and questionnaire data. British Journal of Mathematical and Statistical Psychology 70(1): 137–158.

58.

Simpson

Brent

Willer

Robb

Feinberg

Matthew

(2018) Does violent protest backfire? Testing a theory of public reactions to activist violence. Socius: Sociological Research for a Dynamic World 4 . https://doi.org/10.1177/2378023118803189.

59.

Snyder

David

Kelly

William R

(1979) Strategies for investigating violence and social change: Illustrations from analyses of racial disorders and implications for mobilization research. In: Zald

Mayer N

McCarthy

John David

(eds) Dynamics of Social Movements. New York: Winthrop, 212–237.

60.

Soule

Sarah A

Olzak

Susan

(2004) When do movements matter? The politics of contingency and the equal rights amendment. American Sociological Review 69(4): 473–497.

61.

Stephan

Maria J

Chenoweth

Erica

(2008) Why civil resistance works: The strategic logic of nonviolent conflict. International Security 33(1): 7–44.

62.

Suh

Doowon

(2014) What happens to social movements after policy success? Framing the unintended consequences and changing dynamics of the Korean women’s movement. Social Science Information 53(1): 3–34.

63.

Tarrow

Sidney G

(2011) Power in Movement: Social Movements and Contentious Politics. 3rd edition. Cambridge: Cambridge University Press.

64.

Tilly

Charles

(1999) Conclusion: From interactions to outcomes in social movements. In: Giugni

Marco

Adam

Doug Mc

Tilly

Charles

(eds) How Movements Matter: Theoretical and Comparative Studies on the Consequences of Social Movements. Minneapolis, MN: University of Minnesota Press, 253–270.

65.

Tilly

Charles

Tarrow

Sidney

(2015) Contentious Politics. Oxford: Oxford University Press.

66.

Tilly

Charles

Wood

Leslie

(2013) Social Movements 1768–2012. Boulder, CO: Paradigm.

67.

Ulfelder

Jay

Valentino

Benjamin

(2008) Assessing risks of state-sponsored mass killing. SSRN Electronic Journal (http://www.10.2139/ssrn.1703426).

68.

Useem

Bert

Goldstone

Jack

(2022) The paradox of victory: Social movement fields, adverse outcomes, and social movement success. Theory and Society 51: 31–60.

69.

Van Schuur

Wijbrandt

(2003) Mokken scale analysis: Between the Guttman scale and parametric item response theory. Political Analysis 11(2): 139–163.

70.

Wang

Dan J

Piazza

Alessandro

(2016) The use of disruptive tactics in protest as trade-off: The role of social movement claims. Social Forces 94(4): 1675–1710.

71.

Yuen

Samson

Cheng

Edmund W

(2017) Neither repression nor concession? A regime’s attrition against mass protests. Political Studies 65(3): 611–630.

72.

Zemni

Sami

(2013) From socio-economic protest to national revolt: The labor origins of the Tunisian revolution. In: Gana

Nouri

(ed.) The Making of the Tunisian Revolution: Contexts, Architects, Prospects. Edinburg: Edinburgh University Press, 127–146.