gened essay 3 (final exam)

http://jpr.sagepub.com/ Journal of Peace Research

http://jpr.sagepub.com/content/50/3/305 The online version of this article can be found at:

DOI: 10.1177/0022343313476528

2013 50: 305Journal of Peace Research Victor Asal, Richard Legault, Ora Szekely and Jonathan Wilkenfeld

Gender ideologies and forms of contentious mobilization in the Middle East

Published by:

http://www.sagepublications.com

On behalf of:

Peace Research Institute Oslo

Journal of Peace Research Replication Data

can be found at:Journal of Peace ResearchAdditional services and information for

http://jpr.sagepub.com/cgi/alertsEmail Alerts:

http://jpr.sagepub.com/subscriptionsSubscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

What is This?

- May 16, 2013Version of Record >>

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

Gender ideologies and forms of contentious mobilization in the Middle East

Victor Asal

Department of Political Science, University at Albany, State University of New York

Richard Legault

Department of Homeland Security, Science & Technology Directorate, Washington, DC

Ora Szekely

Department of Political Science, Clark University

Jonathan Wilkenfeld

Department of Government & Politics, University of Maryland

Abstract This article explores those factors that shape a political organization’s choice of tactics in political mobilization with a particular focus on the influence of gender ideology on the choice of different type of contentious action. To under- stand why political organizations engaging in contentious politics choose to employ violent tactics, nonviolent tactics, or a mixture of both, current scholarship has tended to focus on factors such as relationship with the govern- ment, external support, and religious or leftist ideology. Far less attention has been given to the role of an organiza- tion’s ideology relating to gender when predicting its behavior. In addition, much of the analysis of contentious activity has analyzed the use of violence or protest separately and rarely examines the choice of a mixed strategy. We employ a time-series multinomial logistic regression analysis to examine the Middle East Minorities at Risk Organizational Behavior dataset (MAROB), including data over 24 years on 104 ethno-political organizations that have used a range of tactics including protest, violence, and/or a mix of the two, to investigate organizational and state-level variables that lead organizations to choose different strategies. We find that a number of variables can influ- ence a movement’s choice to engage in one strategy over another. Gender-inclusive ideology makes an organization more likely to engage in protest and less likely to choose a violent or mixed strategy.

Keywords

contentious politics, gender, protest, violence

Introduction

The choice to engage in contentious politics in an oppressive state carries with it certain risks; this is perhaps truer in the Middle East, where such protest has often proved fatal, than in many other regions given the extraordinarily high levels of repression that characterize many states in the region. Nonetheless, in many regions, including the Middle East, political organizations do

choose contention. Moreover, they engage in conten- tious politics in different ways including peaceful pro- tests, violent attacks on the state, and sometimes a mixture of the two. The question of what drives an orga- nization to contention and in what manifestation

Corresponding author: [email protected]

Journal of Peace Research 50(3) 305–318 ª The Author(s) 2013 Reprints and permission: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0022343313476528 jpr.sagepub.com

peaceR E S E A R C H j o u r n a l o f

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

remains an important puzzle. We argue that a key (and often overlooked) component at the organizational level is the gender ideology of the organization in question, and that it has a powerful effect on the likelihood that organizations engaged in contentious politics will use peaceful rather than violent tactics.

As the protests that have erupted during the ‘Arab Spring’ indicate, both individuals and organizations in the Middle East can and do choose to engage in multiple forms of contentious mobilization. Data on ethno- political organizations in the Middle East from the Minorities at Risk Organizational Behavior dataset (MAROB) indicate that a wide range of ethno-political organizations in the region have chosen to engage in peaceful protest, violence, or some combination of the two at various times in their existence.1 In fact, over 80% of organizations use contention at some point, and do so in 43% of organization-years.

This may be surprising given the particularly high costs of contention in the region, as demonstrated most recently by the responses of the Syrian, Libyan, and Bah- raini governments to popular mobilization. It also raises the question of why particular organizations choose to engage in particular types of contention. Many scholars have, of course, discussed the intersection between peaceful and violent mobilization (Gamson, 1990; Goodwin, 2002; Gurr & Moore, 1997; Lichbach, 1998; Saxton, 2005; Saxton & Benson, 2006; Scarritt & McMillan, 1995; Tarrow, 1998; Tilly, 1978; Pearl- man, 2011) including those in the quantitative literature (Moore, 1998; Regan & Norton, 2005; Scarritt, McMil- lan & Mozaffar, 2001). Much of this literature treats the use of violence or peaceful protest as an inherent, non- negotiable characteristic of the organization in question – that is, referring to them as ‘peaceful social move- ments’, ‘violent terrorist groups’ or ‘guerrilla move- ments’. We argue that this terminology misses a fundamental truth. The use of violence or protest is nothing more or less than a choice of tactics and many organizations can and do change their tactics frequently based on their goals, their ideologies, and the environ- ment in which they operate. Treating the use of violence, peaceful protest, or a combination of the two as a tactical decision rather than an innate characteristic allows us to

examine the variables that influence a specific movement to choose between violence and nonviolence, to opt for a combination of the two, or even to eschew contentious tactics altogether.

Moreover, while the difference between the use of non- contentious political channels, peaceful protest, violence, or a mixture of protest and violence is sometimes concep- tualized as a continuum or scale, we believe that the transition from one tactic to another is not simply a matter of escalating from nonviolence to violence or de-escalating in the other direction. Instead, we assert that organizations can choose from different tactics including violence, non- violence, or mixed tactics, choices that are motivated by a range of variables. If this behavior were a continuum, mixed strategies would be clear moves into violence, thus motivated primarily by the same influences that motivate violence. We do not expect this to be the case. Rather we anticipate that a mixed strategy is a distinct choice in its own right, motivated by a combination of internal features and external influences shared by both nonviolent and violent contention.2

We should note that our focus here is on the choice to use a particular form (or forms) of contention. While our base category is made up of organizations that employ only non-contentious activities, this does not mean that organizations using contention exclude traditional poli- tics. Rather, it means that they are, in an observed year, using some form of contentious political strategies. The Popular Front for the Liberation of Palestine (PFLP) in the West Bank and Lebanon is in many ways typical of this dynamic, shifting strategies many times between 1984 and 2004 (see Table I) (MAROB, 2009).

To understand why organizations shift between tac- tics we examine a range of variables, the majority of which are drawn from existing relationships in the liter- ature. However, we also expect that a less well-explored ideological characteristic will be important in predicting organizations’ choices: gender ideology. Although analyses of organizations’ choice of tactics include an important literature concerning the impact of gender equality (or inequality) on state behavior and/or internal political violence (see, for example, Caprioli, 2005), we are not aware of any work that attempts to determine the effect of gender-inclusive ideologies on organizational choices in tactics of contention. We assert that an organization’s perspective on gender will have an important effect on its contentious behavior. Specifically, organizations that

1 The MAROB dataset has data on political organizations (groups with a political goal and some organizational structure) that claim to represent a Minorities at Risk group and which have survived for at least three consecutive years, are active at the national or regional level, and are not umbrella organizations. For more detail see MAROB (2009).

2 Dudouet (2013) explores the shift from violence to nonviolence in her contribution to this issue.

306 journal of PEACE RESEARCH 50(3)

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

espouse a gender-inclusive ideology will be less likely to adopt violence and more likely to adopt a protest-only strategy.

What types of contention do organizations choose?

Why do some movements choose to remain within tra- ditional political channels while others choose to engage in contention? Why do those who choose contention elect to use violence, or peaceful protest, or a mix of the two? There is a range of existing answers. Violence is sometimes treated as a direct outgrowth of protest (della Porta, 1995) or as an unfortunate but logical choice for those excluded from traditional political channels (Piven & Cloward, 1977). There is also a small quantitative literature that examines protest and violence in an inte- grated fashion. Scarritt & McMillan (1995) find a strong but regionally varied relationship between democracy, protest, and rebellion. Building on work by Lichbach (1987), Moore (1998) examines the impact of state repression on non-state actors’ choices of either protest or violence. Focusing on opportunity structure and grie- vance, Saxton & Benson (2006) find that grievance has an impact on protest but that opportunity structure has an even stronger impact on both protest and violence. Furthermore, in contradiction to Moore (1998), they report that repression has a strong positive impact on rebellion. Regan & Norton (2005) find that discrimina- tion has a positive impact on political violence and no impact on protest, while repression encourages the former but discourages the latter. Despite the contradic- tions, taken together these findings suggest that various external influences affect organizations various regarding

contention, and that contention is not necessarily escala- tory in nature.

Broadly speaking, there are three dominant trends in the literature on contention. Explanations tend to focus on the political opportunity structure, resource mobiliza- tion, or grievance. The first two tend to focus more on ecological variables while the last is more rooted in the movement’s characteristics. Below, we relate the existing theoretical work to our efforts and the relationships we expect to see between organizational characteristics and choices of contention. We suspect that a movement’s gender ideology will prove noteworthy in shaping its behavior beyond other strong correlates that have been the main foci in previous investigations.

Grievance and ideology There is a large body of work on grievance as a motivator for mobilization (Collier & Hoeffler, 2004). Gurr (1970, 2000) argues that grievance moves groups towards rebellion while Wood (2003) argues that mobi- lization (violent or otherwise) is often motivated by genuine ideological commitment. Clearly, though, the content of mobilization-inducing grievance varies signif- icantly between and within organizations over time. It is therefore possible that different ideologies will produce different choices in terms of the form that mobilization takes at different times (Gurr, 2000). In the Middle East, the two most commonly cited ideologies are religious and leftist. The former dominated the region during the 1960s and 1970s, while the latter has become far more prominent since (Ashour, 2009). In addition to these, we examine the ideology of gender.3 While the impact of leftist and religious ideologies on contention has been widely documented, the impact of an organization’s beliefs about the correct place of men and women in polit- ical and public life has been largely unstudied (Ferree & Martin, 1995; Springer, 2005). In the Middle East, where the policing of gender norms and gendered behavior is framed by some scholars as a means of pushing back against colonial and Western influence, this is likely of particular interest (Ahmed, 2011; Abu Lughod, 2001). Yet, in work on the Middle East and elsewhere, gender and ideologies concerning gender are often treated as an

Table I. Changes in PFLP tactics (1980–2004)

PFLP West Bank PFLP Lebanon

1984–88 Violence 1980–81 Violence 1989–90 Mixed 1982–83 No contention 1991 Violence 1984 Violence 1992–93 Mixed 1985 No contention 1994 Violence 1986 Violence 1995 Mixed 1987 No contention 1996 Violence 1988–91 Violence 1997–2000 Nonviolent 1992 Mixed 2001–2003 Mixed 1993–1996 Violence 2004 Violence 1997–1999 No contention

2000–01 Nonviolent 2002 Violence 2003–04 No contention

3 The MAROB dataset focuses on ethnic organizations ruling out tests for the influence of ethnic or nationalist ideologies as this would be akin to analyzing a constant. Other ideologies in these data are much less prevalent than the three we are examining, and we are more interested theoretically in the impact of gender. It is also worth noting that there is no overlap between religious and leftist organizations in MAROB.

Asal et al. 307

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

entirely separate area of discussion rather than as a general feature of all organizations that can be measured alongside more frequently discussed ideological orientations.

There is a relatively extensive literature on the influ- ence of women in shaping the organizations in which they participate. There is evidence that the inclusion of women in political and organizational life has a strong and negative impact on interstate, intrastate, and state violence (Melander, 2005), although, as Kampwirth (2002) and Goldstein (2001) note, there is no historical shortage of female combatants. There is also a strong lit- erature at the state level suggesting that societies that are more gender-inclusive are much less likely to be violent (Caprioli, 2005; Enloe, 2007; Regan & Paskeviciute, 2003). Hudson et al. (2012) suggest that female partic- ipation in government can produce lower levels of cor- ruption and better economic planning and that social mechanisms that promote gender equality, such as later age of marriage, rejection of gender-based violence, and abolition of polygamy, are important in laying the groundwork for democracy. Of course, the participation of women in politics is not the same as an organization’s ideology about gender equality. There is a world of difference between a movement that includes women, while still advocating policies that may prove damaging to women’s interests in the long run, and one that actively advocates for their inclusion.

Much of the feminist literature also suggests that gender-inclusive ideologies should make organizations more peaceful (e.g. Warren & Cady, 1994; Zimmerman, McDermott & Gould, 2009). Likewise, Gleditsch et al. (2011) and Clark (2004b) find that misogyny is generally associated with higher levels of violence and acceptance of violence. Other theorists argue that patriarchal values pro- mote militarism and encourage violence within society (e.g. see Enloe, 2007). Enloe also argues that the choice of violence is related to patriarchal attitudes, while the choice of nonviolent resistance is explicitly connected to a feminist perspective (Enloe, 2000: 150). This is echoed by survey research in four Middle Eastern countries indicating that it is not necessarily an individual’s own gender but rather inclusive attitudes about gender in general that has an effect on his or her views concerning the use of violence (Tessler & Warriner, 1997).

Taken together, this suggests that views on gender should be treated as a specific ideology that is separate from the presence of women in the movement, and from other ideologies. The MAROB dataset allows us to test this by coding ‘gender-inclusive’ ideology as those organizations that have a statement in their declaration of principles or charter that specifically advocates for

inclusion of women. A wide range of militant groups in the Middle East do so for a number of reasons. These may include a genuine desire to increase the influence of women in political life, a pragmatic realization that women may be less likely to be arrested or assassinated than men, or a desire to demonstrate political sophistication to both domestic and international audiences. The research cited above, particularly Enloe (2007), suggests that gender-inclusive ideologies should make organizations much less likely to use contentious strategies that involve violence.

However, it does not suggest that such groups are likely to eschew contention altogether; militant groups that advocate gender-inclusive ideologies are often already swimming against the political tide, so to speak, and may be as or even more likely to engage in contentious politics as those who do not express such an ideology, both in the Middle East and elsewhere. We hypothesize, therefore, that:

H1: Organizations that espouse a gender-inclusive ideology will be more likely to use nonviolent strategies of contention.

Other ideologies, of course, may also affect an organiza- tion’s likelihood of using particular tactics of contention. Many argue that religious ideology increases organiza- tional violence (Berman, 2009; Juergensmeyer, 2003; Rapoport, 2001). Juergensmeyer (2003) suggests that reli- gious organizations are more likely to be violent because they believe that God is their primary audience, leading them to embrace a ‘cosmic war’ perspective, in which their adversary is objectively and fundamentally evil.

Empirical evidence also indicates that many secular organizations also use political violence. For instance, many leftist Palestinian organizations fit this description. The PFLP used violence in more than half of the years we exam- ine, and the DFLP (Democratic Front for the Liberation of Palestine) in 38% of those years. While recent research has focused on the link between religious ideology and violence, earlier work focused on leftist ideology as a motivation for terrorism because of the ideology’s push to overthrow the existing political structure (c.f. Wilkinson, 1977; Rapoport, 2001). Moreover, Feldmann & Perala (2004) point out that many of these organizations have continued to embrace violence since the end of the Cold War.

Therefore, we hypothesize that both organizations that are religious and those that are leftist will be more likely to mobilize, and more likely to mobilize violently. Many of the organizations advocating these ideologies exhibit similar desires to overthrow the existing structure of society and replace it with a utopian reality. And while

308 journal of PEACE RESEARCH 50(3)

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

there is no overlap between these two ideologies in our data, both ideologies often argue that social and political change necessitates violence. This leads us to expect that organizations that embrace these ideologies should be more likely to be both contentious and violent (Juergen- smeyer, 2003; Rapoport, 2001).

H2: Religious organizations will be more likely to engage in contentious politics in general and in violent politics in particular, and less likely to use traditional political channels or peaceful protest.

H3: Leftist organizations will be more likely to engage in contentious politics in general and in violent politics in particular, and less likely to use tradi- tional political channels or peaceful protest.

Political opportunity structure A second set of explanations for mobilization focuses on the role of the political structures in which movements exist (Tarrow, 1998). In this case, the constraints under which the organization operates are at least as important as the characteristics of the organization. These ‘constraints’ can be measured either by the character of the state itself or by the state’s interaction with the organization. Tilly argues that both regime type and state capacity influence mobili- zation, because they shape the space within which political movements can operate. Specifically, he views democrati- zation as conducive to mobilization and armed movements as more likely to arise in low capacity, authoritarian environments (Tilly & Tarrow, 2007). Similarly, human rights scholars have found that high levels of democracy constrain governments’ use of violence (Cingranelli & Richards, 1999; Davenport & Armstrong, 2004). This, in turn, enables organizations to engage in peaceful protest (or even violence) with less fear that their members will be shot in the streets. Therefore, high levels of democracy may reduce the risk of nonviolent contention, thus increasing the likelihood that an organization will embrace a non- violent approach. It should also have a smaller positive impact on the choice of a mixed strategy, and a negative impact on the adoption of a purely violent contentious strategy.4 Thus we hypothesize that:

H4: The more democratic a country, the more likely that an organization in that country will choose to adopt a protest or mixed strategy and the less likely that it will choose a strategy of only violence.

Regime type and state repression, however, are not the same. While the former is a characteristic of the state, the latter is a choice any state can make, regardless of its char- acteristics. Even high-capacity democracies can and do repress some movements (e.g. neo-Nazis in Germany), underscoring the importance of measuring repression separately from level of democracy (Gurr & Moore 1997: 1083; Davenport, 2007). Conversely, while democracy may open avenues of nonviolent contention, repression may prevent such mobilization entirely (Tilly, 1978; Gurr, 2000; Regan & Norton, 2005; Saxton & Benson, 2006). This leads us to hypothesize that targeted repression by the state increases the probability that an organization will choose violence as a tactic, while decreasing the likelihood that it will use other contentious strategies. Thus, dynamic changes in how a state acts toward an organization will influence the choices of that organization, whether violent or nonviolent. Given that a mixed strategy includes violence, we expect that repres- sion should have a small positive impact on such an approach. The use of violence increases the chances that participants will be killed, making it a risky act. Organiza- tions committed to mobilization in the face of these risks are also more likely to encourage their members to take the risks that nonviolent mobilization carries in a repressive environment. For example during the first intifada, Hamas encouraged its members to ‘cooperate in both violent and nonviolent actions. . . . clashing with Israeli forces, and attacking collaborators . . . [and] . . . to engage in civil disobedience’ (Mishal & Sela, 2006: 60).

Thus we hypothesize that:

H5: Organizations that are being actively repressed by the state will be much less likely to adopt tradi- tional politics or protest as a strategy, slightly more likely to adopt a mixed strategy, and much more likely to adopt violence as a strategy.

Resource mobilization There is a strong argument that resources and the ability to mobilize them are important in determining what organizations can and will do (Collier & Sambanis, 2002; McCarthy & Zald, 1977). Of course, not all organizations face equal constraints in this regard. An available source of support external to the arena of con- flict can empower organizations to engage in contentious

4 There are two additional causes not included in our analysis because of a lack of data. The first is the effect of non-state actors’ interactions with one another. While the MAROB dataset does not presently allow us to test for this effect, it remains an avenue for future research. The second is the impact of fragmentation that Pearlman (2011) argues is the most important factor in determining whether a movement will use violence or not. Because this is measurable in less than 4% of cases in our sample, we were unable to include it in this analysis.

Asal et al. 309

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

politics in a way inaccessible to those without similar sources. External sources can continue to support the organization while more vulnerable ‘local’ members, who may be targeted by the state, may be prevented from doing so. The civil war literature suggests that diaspora support in particular can be a powerful asset for conten- tious organizations (Fair, 2005; Arquilla & Ronfeldt, 2001). As Collier et al. (2003: 74) argue, ‘diasporas do not suffer the consequences of violence, nor are they in day-to-day contact and accommodation with ‘‘the enemy’’’. The advantages provided by diaspora support stand in sharp contrast to the limits faced by organiza- tions based entirely inside a single state, whose opportu- nities can be severely constrained (Keck & Sikkink, 1998; Tarrow, 2005; Salehyan, 2008).

At the local level, social service provision can be an important means of mobilizing resources. This can include institutions as diverse as orphanages, schools, road maintenance, provision of drinking water, or even major hospitals. Various scholars have argued that such services can improve relations with local constituents by rewarding members for their support (Clark, 2004a), improving the organization’s reputation (Flani- gan, 2006), and minimizing the risks of defection by ren- dering constituents dependent (Berman, 2009). Others have suggested that service provision represents a response by rebel groups to the demands of civilians in the areas under their control (Mampilly, 2011).

Whatever its intended purpose, the provision of social services by non-state entities clearly challenges the state’s authority by taking over a role previously assumed by the state and usurping the authority and legitimacy that role brings with it. It may also be an indication that the move- ment has established sufficient control over a particular area that, for all intents and purposes, it has assumed the role of the state with regard to everyday services (Weinstein, 2006; Mampilly, 2011). Social service provision thus both gener- ates tangible and intangible resources for the organization and challenges the state. We expect this will increase the likelihood that the organization will engage in contention through violent confrontation with the state.

In sum, the bases of support available domestically and abroad should have a strong effect on choices of con- tention. Because of the relative freedom from the costs of contention that diaspora supporters enjoy, we predict that such support will make contention in general more likely and have a particularly strong impact on the use of violence. Conversely, organizations with only domestic bases will be less likely to mobilize contentiously and even less likely to use violence. Social service provision should increase contention because of the resources it

generates and should have an effect similar to diaspora support. Thus we hypothesize that:

H6: Organizations with diaspora support will be less likely to adopt traditional politics, more likely to adopt contentious tactics, and most prone to adopt a violence-only strategy.

H7: Organizations that are located wholly within the state will be more likely to adopt traditional poli- tics, and less likely to adopt contentious strategies of any kind.

H8: Organizations providing social service support to their communities will be less likely to adopt tradi- tional politics, more likely to adopt contentious tactics, and most prone to adopt a strategy that includes violence.

Data and methods5

Our analyses are based on data from the Minorities at Risk Organizational Behavior dataset, representing 104 regional and/or national non-state organizations and their actions from 1980 through 2004 (MAROB, 2009). The dataset is an unbalanced panel at the organization-year level of analysis (n ¼ 1,789). These characteristics make it uniquely suited to test our hypotheses.

Because we propose that the choice of contention is not rank-ordered but rather a set of discrete choices, we employ a multinomial analytical approach. By apply- ing a multinomial logistic regression (MNLR) model, we test organizational characteristics that strongly shape its propensity to select discrete tactical options. Further, because we have the added advantage of using data that are arranged as a panel over a relatively long time period, we specify time order such that common concerns over endogeneity are addressed in the model itself.

At this point it is important to note several key limita- tions to our analysis. Because the level of analysis in the MAROB data is organization-years, daily or monthly interactive effects between government and organiza- tional behavior are not available or tested. Nonetheless, this is a significant improvement over previous analyses that focus only on the country level. Secondly, the data focus only on ethno-political organizations. (Indeed, MAROB represents the most complete collection of data on ethno-political organizations available today.) Although many ideologies are espoused by the

5 Detailed descriptions of the data coding, construction, and treatment; the estimator used, justifications, and model diagnostics; and additional analyses can be found in the online appendix.

310 journal of PEACE RESEARCH 50(3)

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

organizations herein, all are ethnically focused by defini- tion. Other ideological affiliations are in addition to ethno-communal affiliations. Finally, our data are geographically limited to the Middle East. Like every other region of the world, unique structural and cultural characteristics may impact organizational decisionmak- ing. The fact that the organizations are ethno-national in character does give them some common ground with similar groups in other parts of the world, but our ability to generalize to other areas from these data may be lim- ited. Nevertheless, the unique ability to directly compare violent and nonviolent organizations far outweighs the dataset’s potential limitations.

The variables Table II presents descriptive statistics for all variables used in this study. The primary outcome measured is the polit- ical behavior in which an organization participates to meet

its goals. We operationalize this concept using a single four-category measure to capture the nature of the discrete choices in behavior that are made by a given organization: Non-contentious political behavior, Protest/public demonstra- tion, Protest/public demonstration & violence, and Violence- only. We coded the variable 0–3, respectively. Of the organization-years in the sample, 57% exhibited only non-contentious political behavior, 9% only protests, 6% both protest and violence, and 28% only violence.

Several of our variables bear further discussion here. The country-level Democracy variable was coded using a variable that draws on both the Polity2 and Freedom House datasets (Freedom House, 2008; Marshall, Jag- gers & Gurr, 2011), combining the two-into-one 20- point measure (Hadenius & Teorell, 2005; Teorell, Holmberg & Rothstein, 2008). The variable was con- structed using an imputation scheme to estimate missing data, computing a single, valid measure that is more

Table II. Coding of variables, means, and standard deviations for all variables

Variable Coding of variable Mean SD Min Max N

Dependent Political behavior 0¼Non-contentious political behavior

1¼Protest/public demonstration 2¼Protest/public demonstration & violence 3¼Violence only

1.053 1.326 0 3 1,675

Non-contentious behavior

0¼ All other behaviors 1¼Non-contentious political behavior

0.571 0.495 0 1 1,675

Protest/public demonstration

0¼ All other behaviors 1¼Protest/public demonstration

0.088 0.283 0 1 1,675

Protest/public demonstration & violence

0¼ All other behaviors 1¼Protest/public demonstration & violence

0.059 0.236 0 1 1,675

Violence only 0¼ All other behaviors 1¼Violence only

0.282 0.450 0 1 1,675

Independent Democracyy Measure of nations’ level of democracy. 0¼least

democratic; 10¼most democratic 4.102 3.332 0 10 1,743

Domestic bases The organization has facilities and support only in its home country. 0¼no; 1¼yes

0.452 0.498 0 1 1,789

Diaspora support Organization receives significant support from members outside of the country of operation. 0¼no; 1¼yes

0.061 0.239 0 1 1,755

Gender inclusion Organization has formal policy supporting equal rights for women. 0¼no; 1¼yes

0.211 0.408 0 1 1,789

Religious organization

Organization is religiously based. 0¼no; 1¼yes 0.251 0.434 0 1 1,789

Leftist organization Organization supports leftist ideological goals. 0¼no; 1¼yes

0.315 0.464 0 1 1,764

Severe repression by state Organization is illegal and repressed by the state. 0¼no; 1¼yes

0.194 0.396 0 1 1,751

Social services Organization provides government-style social services to a population or subpopulation. 0¼no; 1¼yes

0.099 0.299 0 1 1,789

SD ¼ Standard deviation. y For full information on the construction, validity, and reliability of this variable, see Hadenius & Teorell (2005).

Asal et al. 311

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

complete than either of the component datasets while maintaining the integrity of both and ensuring that they are indeed comparable (Teorell, Holmberg & Rothstein, 2008).

Using the MAROB data, Domestic bases only was coded as 1 if the organization did not have a permanent presence overseas.6 This variable does not imply military bases. It simply means that the organization is located only in the country with which it is primarily engaged. The Diaspora variable from MAROB captures whether the organization had support from a diaspora group in a given year – this is coded as 1 if the organization received support from ‘close’ kindred groups located out- side the country (MAROB).

The ideology variables7 were coded using the corre- sponding variables from the MAROB data. Variables are coded as 1 if the organization advocates polices as follows:

Gender inclusion: Does the organization advocate the inclusion of women in public life?

Religious organization: Does the organization advocate policies that incorporate religion into public life?

Leftist organization: Does the organization advocate policies to redistribute wealth?

Severe repression by state was coded as 1 if an organization was both illegal and suffered either periodic or ongoing repression in the year in question. Social services was coded as 1 if the ‘organization provides social services in the area(s) of education, healthcare, poverty alleviation at a para-statal level (e.g., runs the equivalent of a school district, maintains networks of heath care facilities, etc.) in order to serve a large number of constituents on a sustained basis in a given year’ (MAROB, 2009).

Methods In an effort to describe variation in political behavior at the organizational level, a multinomial logistic regression (MNLR) is estimated on the panel of organizations.

Using this method allows a comprehensive depiction of the outcomes while the dependent variable is treated as a set of nominal, discrete behavioral choices in politi- cal action. That is, we are intentionally modeling these data with the understanding that none of the activities described here can be ranked or ordered in any meaning- ful way. MNLR also allows us to interpret the resulting coefficients as odds statements and does not assume a linear functional form (Aldrich & Nelson, 1984). There- fore, by using this method we are able to determine the odds of discrete choices to participate in each type of political activity over time, based on each of the indepen- dent variables and in reference to a baseline behavior (Long, 1997: 155).8 Other advantages of MNLR include the ability to model the panel as a time series and the ease with which one may interpret the outcomes of the equations.9 Finally, we will include measures of the modeled discrete change as probabilities by using a linear approximation of the variance function (Long, 1997; Xu & Long, 2005). Because of the nature of these data and the multinomial logistic model, the probabilities will represent an average absolute discrete change and can be summarized by:

�D ¼ 1 J

XJ

j¼1

D Pr y ¼ jjxð Þ Dxk

����

���� ð1Þ

where Pr (y ¼ j | x) is the probability that y ¼ j given x and with J values of change, one for each possible out- come (Long, 1997: 166).10

Estimating a complex model over time also requires a number of diagnostic tests in an attempt to ensure that MNLR is the correct methodological approach and that the model itself does not suffer from problems that are common to time-series estimators. Particular attention is paid to the Independence from Irrelevant Alternatives

6 More commonly examined in the literature is the impact of foreign support (Byman, 2005; Byman et al., 2001; Salehyan, 2008). It is precisely because there is such an extensive literature establishing the link between foreign support and a range of militant group behaviors that we decided to include the somewhat less well explored variable of having domestic bases only. Analysis of the role of foreign support is available in the online appendix. 7 Of the ideology variables in these data, religious and leftist organizations are mutually exclusive, but neither of these is exclusive to gender inclusiveness. This is not because of the coding scheme of the data, but because of the empirical observations of organizations in this region (MAROB, 2009).

8 Estimating effects over time is one of the most important reasons to select this estimator. Other potentially appropriate choices, such as the bivariate probit estimator, are not robust to time-series or panel considerations and will yield meaningless results. In any case, as we also diagnose for violations of the IIA assumption, we further contend that the main four conceptual categories here are substantively and empirically different from these categories in a measurable way and we offer tests of the model to confirm this assertion. 9 When the coefficients of a logistic regression are transformed through a simple formula, we may interpret the result as a percentage change in odds of the dependent variable for a one-unit change on the independent variable while holding all other variables constant (Legault, 2008: 85). 10 Details on the Delta Method for the computation of confidence intervals can be found at: http://www.indiana.edu/~jslsoc/stata/ ci_computations/spost_deltaci.pdf.

312 journal of PEACE RESEARCH 50(3)

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

(IIA) assumption to ensure that the categories of the dependent variable should not be collapsed. In addition, we also test to ensure that serial autocorrelation errors are not negatively affecting our analysis or interpretations of results (Greene, 2003). In conducting the test, we noted that a panel unit root was present at two and three lags. Therefore, two- and three-year lags of the dependent variable are included as independent variables and results represent short-term discrete change.

Results and findings

Multivariate – political behavior Table III represents our MNLR results as well as the sim- ple odds ratios for all cases of organizations that partici- pate in only protest and public demonstration in comparison to non-contentious political behavior.11

To ease interpretation of these effects, we have calculated discrete probability changes for the full equation, presented in Table IV (Long, 1997).

Hypothesis 1, on the impact of gender inclusion on the likelihood that an organization will engage in contention in general and nonviolent contention in particular, finds strong support. Organizations with gender-inclusive ideology were much more likely to adopt a protest-only approach (19.5%), much less likely to adopt violence-only (–19.3%), and mildly less likely to adopt a mixed strategy (–1.3%).

Our findings provide very different support for the hypotheses predicting the effects of other ideologies. While we argued in Hypothesis 2 that religious ideology should have a positive impact on all kinds of contentious strategies and a particularly strong impact on a violence- only approach, our analysis indicates that this is not the case with only a small positive impact (2.7%) on a mixed approach. Similarly, Hypothesis 3, arguing that leftist organizations should be more inclined to engage in all forms of contention versus regular politics, was strongly supported in relation to a violence-only approach (11.1%) but garnered no support for either a mixed strategy or a protest-only strategy. Hypothesis 4 is only partially supported in that organizations in more demo- cratic countries are more likely to be involved in protest (11.7%) or combined protest and violence (10.5%), but there is no significant impact on the choice of violence alone. We have similar partial support for Hypothesis 7 arguing that organizations with only domestic bases would be less likely to use any contentious strategy. This is true for the mixed strategy (–3.3%) and the protest- only strategy (–7.2%), but again this variable has no effect on the use of a violence-only strategy. Similarly our argument that social service provision should have a generally positive relationship with contention and a stronger impact on violence-only (Hypothesis 8) was supported. Social services provision increased participa- tion in violence-only (15.4%) and a mixed strategy (3.6%) but did not have a significant impact on a protest-only choice.

Hypothesis 5 argued that organizations facing repres- sion from the state will be less likely to adopt a non- contentious or protest-only approach but will be more likely to adopt a mixed or violence-only strategy. This

Table III. Multinomial logistic regression on political behavior

Variables Protest/Public demonstration Protest/Public demonstration & violence Participation in violence only

b (S.E.) Exp b b (S.E.) Exp b b (S.E.) Exp b

Democracy 0.120* (.038) 1.127 0.333* (.058) 1.395 –0.033 (.034) 0.967 Domestic bases –0.888* (.265) 0.411 –2.044* (.586) 0.130 –0.325 (.213) 0.722 Diaspora support 1.493* (.474) 4.450 2.803 (.471) 16.495 1.440* (.375) 4.220 Gender inclusion 1.263* (.308) 3.535 –0.920* (.457) 0.399 –1.378* (.288) 0.252 Religious organization 0.398 (.302) 1.489 1.151* (.499) 3.162 0.078 (.257) 1.081 Leftist organization –0.187 (.295) 0.830 0.648 (.444) 1.912 0.658* (.242) 1.932 Severe repression by state �0.888y (.534) 0.412 1.489* (.352) 4.432 1.420* (.236) 4.139 Social services –0.241 (.319) 0.786 1.455* (.413) 4.285 0.861* (.262) 2.365 Contention lag (2) 0.086 (.106) 1.090 0.712* (.144) 2.039 0.643* (.082) 1.902 Contention lag (3) 0.170 (.104) 1.185 0.498* (.152) 1.645 0.581* (.081) 1.787 Constant –2.477* (.287) 0.084 –6.371* (.663) 0.002 –2.723* (.267) 0.066

N ¼ 1,276; R12 ¼ .346; *a ¼.05; ya ¼.10; G2 ¼ –.888; P ¼ <.001; Wald w2 ¼ 605.08; P ¼ <.001.

11 For a discussion of the odds ratio please see our online appendix – in the article we focus on the probabilities using non-contentious political behavior as a baseline with all other independent variables held at their mean.

Asal et al. 313

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

is supported as the impact on violence-only (28%) and protest-only (–10%) is much larger than that on a mixed approach (3%). Hypothesis 6 suggested that diaspora sup- port will have a constraining impact on non-contentious politics but a strong positive impact on contention and particularly violence-only and is strongly supported as violence-only strategies are more likely by 17%.

Discussion

Our findings point to a number of interesting conclu- sions. The first is that organizations that pursue conten- tious politics have a choice in terms of the tactics that they use to achieve strategic goals. Rather than seeing the use of protest, protest and violence, or violence- only as a continuum of radicalization along which movements travel unidirectionally, our findings suggest that these should rather be conceptualized as items on a menu from which non-state movements can choose if they select the ‘contention’ menu at all – they may select the ‘non-contentious politics’ menu instead. Organizations can and do switch back and forth between tactics. This choice is situational, and no one concept (e.g. opportunity structure, movement ideol- ogy, or overall capability) is singularly predictive of the choices of contention that organizations make. Instead, combinations of variables predict whether an

organization will choose contention at all, and if so, which types of contention it will pursue.

When it comes to the question of whether we should treat a mixed strategy as being somewhere along a con- tinuum from nonviolence to violence or as a discrete choice of its own, our analysis strongly supports the latter. In Table V we compare the mixed approach to the ‘only’ strategies. This clearly indicates that a mixed strategy is not the same as either of the others. The effect of some variables on the choice of a mixed strat- egy is similar to the effect on the choice of violence; oth- ers have an effect similar to that on the choice of peaceful protest, and some are similar to those that pre- dict non-contentious political activities. The impact of three variables is similar between mixed strategies and violence; three are similar to nonviolence, one is similar to both, and one is different from both. These results support the idea that strategic choices are unique and discrete, not a continuum between nonviolence and violence.

With regard to the effects of specific variables, we find strong evidence that having some sort of external support network is a powerful enabling factor for contention in general. Having the support of a diaspora community that does not have to bear the costs of contention increases the likelihood that movements will turn toward contention. Conversely, being based only inside the

Table IV. Discrete change in probability for multinomial logistic model of political activity

Variable Change Average

D Non-contentious politi-

cal behavior Protest/public demonstration

Protest/public demonstration & violence

Violence only

0 1 2 3

Democracy Min!Max 0.111 –0.127 0.117 0.105 NS D1 0.009 –0.008 0.012 0.006 NS Ds 0.029 –0.029 0.038 0.020 NS Marginal

effect 0.009 –0.008 0.012 0.006 NS

Domestic bases 0!1 0.065 0.130 –0.072 –0.033 NS Diaspora support 0!1 0.189 –0.378 0.106 0.100 0.173 Gender inclusion 0!1 0.103 0.011 0.195 –0.013 –0.193 Religious

organization 0!1 0.031 –0.060 NS 0.027 NS

Leftist organization

0!1 0.061 –0.088 NS NS 0.111

Severe repression by state

0!1 0.156 –0.218 –0.095 0.029 0.284

Social services 0!1 0.095 –0.148 NS 0.036 0.154

0!1 is a discrete change from 0 to 1; D1 is the centered change of one unit around the mean; Ds is the centered change of one standard deviation around the mean. All other variables are held at their means. Average D is the average absolute change. All probabilities listed are statistically significant a ¼ .05.

314 journal of PEACE RESEARCH 50(3)

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

country, thus raising the cost of contention borne by the local movement, dampens any contentious choice.

Certain organizational characteristics also make particular strategies of contention more or less likely. Operating in a democratic system increases the likeli- hood of nonviolence, while state repression makes vio- lence more likely, as do diaspora support and leftist ideology, but religious ideology does not, at least among these ethno-political organizations.

Most interesting, though, is that having an ideology of gender inclusion as part of its political agenda has a strong impact on whether a contentious organization will choose violent or peaceful forms of contention: organiza- tions that promote a gender-inclusive ideology are more likely to employ peaceful tactics. This indicates that ideol- ogies of gender inclusion are at least as important as those centered on class, race or ethnicity in terms of their effects on movements’ organizational behavior. This sug- gests that theorists who have argued that gender policies and ideologies are powerful predictors of the level of vio- lence versus stability in society may well be correct, and this dynamic is echoed at the organizational level. Rather than treating gender ideology as a side issue or an artifact of other ideological preferences, this topic demands much greater focus and attention among those who study political violence.

Conclusion

As the editors of this special issue on nonviolent resis- tance have pointed out in their introduction, ‘violence is often viewed as a more pressing and troubling global problem, distracting researchers from the equally com- mon, civilian-led, unarmed struggles and revolutions that have always been present around the world’ (Cheno- weth & Cunningham, 2013). Organizations have a range of choices available to them. And while it is tempt- ing to view the use of protests, mixed use of protests and

violence, and the use of violence alone as a continuum along which organizations make strategic and tactical choices, the data we examine in the present analysis lead to the conclu- sion that the same movement can utilize all three options at various points, moving from violence to nonviolence and then back again depending on the particular array of inter- nal and external conditions that it finds itself in.

What, then, are some of the determinants of these strategic choices? Organizations differ in terms of their dominant political ideologies, the nature of the regime in which they operate and the level of repression directed against them, the presence or absence of outside (dia- spora) support, and perhaps most interestingly, an ideol- ogy of gender inclusion. The complex interactions among these factors help determine the strategic choices made by these organizations and movements as they attempt to promote their organizational goals and address grievances. Our analysis points to four inescap- able conclusions that beg for attention from both the academic and policy communities as we confront the spread of revolutionary fervor, not only in the Middle East and North Africa, but potentially in other regions as well. First, repression by governments is likely to fan the flames of resistance. Second, organizations with strong networks of ties to ethnic kin outside the region of conflict and contention are likely to adopt a more vio- lent set of strategies. Third, a democratic ideology makes the choice of violence less likely. And fourth, ideologies directly addressing gender inclusion as part of their polit- ical agenda are more likely to employ peaceful tactics.

This study leaves us with some interesting questions that may provide fruitful avenues for further research. To begin with, though our use of the MAROB dataset does allow us to analyze in considerable depth and directly compare the characteristics of both ‘mainstream’ organiza- tions and those that use violence (few other datasets allow this), the trade-off is that it does limit our analysis to ethno-nationalist movements. Future research on

Table V. Comparing mixed and singular strategies

Variable Protest/public demonstration Protest/public demonstration & violence Violence only

Democracy 11.7% 10.5%* NS Domestic bases –7.2 –3.3* NS Diaspora support 10.6 10y 17.3% Gender inclusion 19.5 –1.3** –19.3 Religious organization NS 2.7z NS Leftist organization NS NS* 11.1 Severe repression by state –9.5 2.9** 28.4 Social services NS 3.6** 15.4

*Similar to protest (three cases); **similar to violence (three cases); ysimilar to both (one case); zdifferent from both (one case).

Asal et al. 315

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

organizations that fall outside this subset, or operate in areas of the world outside the Middle East, may add important nuance to our understanding of these issues. More broadly, the project raises a number of important questions. What influence do organizations have on the political and social structures within which they mobilize? To what degree are they able to shape their own environ- ment (particularly with regard to the question of external support)? And, conversely, what influence do those struc- tures have on the organization’s own characteristics? While we have treated these factors as relatively static for the sake of measurement, there is certainly some interac- tion between them. The most interesting enhancement to our understanding of organizational choice and conten- tion, however, is the identification of gender ideology as an important, causal influence. This clearly demands fur- ther research, and should focus on both gender inclusion (examined here) and gender exclusion (for which our data were insufficient). Lastly, although the study of political movements and organizations has focused for some time on a particular litany of predefined ideologies that affect organizational behavior, gender policies should be included in this category in future work.

Replication data The dataset and do-files for the empirical analysis in this article as well as the online appendix can be found at http://www.prio.no/jpr/datasets.

Author note The analysis and conclusions and views presented and expressed in this study are those of the authors in indi- vidual capacities only and do not necessarily reflect the views of the Department of Homeland Security or the United States Government.

Acknowledgements We would like to thank the anonymous reviewers and the guest editors Erica Chenoweth and Kathleen Galla- gher Cunningham for their invaluable feedback.

Funding Support for this research was provided by the National Science Foundation through Award #0826886 and by the Science and Technology Directorate of the US Department of Homeland Security (grant number 2008ST061ST0004) through the National Consortium for the Study of Terrorism and Responses to Terrorism (START).

References Abu Lughod, Lila (2001) The marriage of feminism and Isla-

mism in Egypt: Selective repudiation as a dynamic of postcolonial cultural politics. In: Lila Abu Lughod (ed.) Remaking Women: Feminism and Modernity in the Middle East. Princeton, NJ: Princeton University Press, 243–269.

Ahmed, Leila (2011) A Quiet Revolution: The Veil’s Resurgence, from the Middle East to America. New Haven, CT: Yale University Press.

Aldrich, John H & Forrest D Nelson (1984) Linear Probabil- ity, Logit, and Probit Models. Newbury Park, CA: Sage.

Arquilla, John & David Ronfeldt (2001) Osama bin Laden and the advent of Netwar. New Perspectives Quarterly 18(4): 23–33.

Ashour, Omar (2009) The De-Radicalization of Jihadists: Transforming Armed Islamic Movements. New York: Routledge.

Berman, Eli (2009) Radical, Religious, and Violent: The New Economics of Terrorism. Cambridge, MA: MIT Press.

Byman, Daniel (2005) Deadly Connections: States that Sponsor Terrorism. New York: Cambridge University Press.

Byman, Daniel; Peter Chalk, Bruce Hoffman, William Rosenau & David Brannan (2001) Trends in Outside Sup- port for Insurgent Movements. Washington, DC: Office of Transnational Issues, National Security Research Division, RAND Corporation.

Caprioli, Mary (2005) Primed for violence: The role of gender inequality in predicting internal conflict. International Studies Quarterly 49(2): 161–178.

Chenoweth, Erica & Kathleen Gallagher Cunningham (2013) Understanding nonviolent resistance: An introduction. Journal of Peace Research 50(3): 271–276.

Cingranelli, David L & David L Richards (1999) Respect for human rights after the end of the Cold War. Journal of Peace Research 36(5): 511–534.

Clark, Janine (2004a) Islam, Charity and Activism: Middle Class Networks and Social Welfare in Egypt, Jordan and Yemen. Bloomington, IN: Indiana University Press.

Clark, Mary E (2004b) Rhetoric, patriarchy & war: Explain- ing the dangers of ‘leadership’ in mass culture. Women and Language 27(2).

Collier, Paul & Anke Hoeffler (2004) Greed and grievance in civil war. Oxford Economic Papers 56(4): 563–595.

Collier, Paul & Nicholas Sambanis (2002) Understanding civil war: A new agenda. Journal of Conflict Resolution 46(1): 3–12.

Collier, Paul; Lani Elliott, Håvard Hegre, Anke Hoeffler, Marta Reynal-Querrol & Nicholas Sambanis (2003) Break- ing the Conflict Trap: Civil War and Development Policy. Oxford: World Bank and Oxford University Press.

Davenport, Christian (2007) State repression and the tyranni- cal peace. Journal of Peace Research 44(4): 485–505.

Davenport, Christian & David A Armstrong (2004) Democ- racy and the violation of human rights: A statistical analysis

316 journal of PEACE RESEARCH 50(3)

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

from 1976 to 1996. American Journal of Political Science 48(3): 538–554.

della Porta, Donatella (1995) Social Movements, Political Violence, and the State: A Comparative Analysis of Italy and Germany. Cambridge: Cambridge University Press.

Dudouet, Véronique (2013) Dynamics and factors of transi- tion from violent to nonviolent resistance. Journal of Peace Research 50(3): 415–423.

Enloe, Cynthia H (2000) Maneuvers: The International Politics of Militarizing Women’s Lives. Berkeley, CA: University of California Press.

Enloe, Cynthia H (2007) Globalization and Militarism: Feminists Make the Link. Lanham, MD: Rowman & Littlefield.

Fair, Christine C (2005) Diaspora involvement in insur- gencies: Insights from the Khalistan and Tamil Eelam movements. Nationalism and Ethnic Politics 11(1): 125–156.

Feldmann, Andreas E & Maiju Perala (2004) Reassessing the causes of nongovernmental terrorism in Latin America. Latin American Politics & Society 46(2): 101–132.

Ferree, Myra Marx & Patricia Y Martin (1995) Feminist Organizations: Harvest of the New Women’s Movement. Phi- ladephia, PA: Temple University Press.

Flanigan, Shawn Theresa (2006) Charity as resistance: Connections between charity, contentious politics, and terror. Studies in Conflict & Terrorism 29(7): 641–655.

Freedom House (2008) Freedom in the World. Freedom House (www.freedomhouse.org).

Gamson, William A (1990) The Strategy of Social Protest. Belmont, CA: Wadsworth.

Gleditsch, Kristian Skrede; Julian Wucherpfennig, Simon Hug & Karina Garnes Reigstad (2011) Polygyny or miso- gyny? Reexamining the ‘First Law of Intergroup Conflict’. Journal of Politics 73(1): 265–270.

Goldstein, Joshua S (2001) Women and War. Cambridge: Cambridge University Press.

Goodwin, Jeff (2002) Are protesters opportunists? Political opportunities and the emergence of political contention. Working paper, Department of Sociology, New York University.

Greene, William H (2003) Econometric Analysis, 3rd edn. Sad- dle River, NJ: Prentice Hall.

Gurr, Ted R (1970) Why Men Rebel. Princeton, NJ: Princeton University Press.

Gurr, Ted R (2000) People vs. States. Washington, DC: United States Institute of Peace.

Gurr, Ted R & Will Moore (1997) Ethno-political rebellion: A cross-sectional analysis of the 1980s with risk assessments for the 1990s. American Journal of Political Science 41(4): 1079–1103.

Hadenius, Axel & Jan Teorell (2005) Assessing alternative indices of democracy. Committee on Concepts and Meth- ods Working Paper Series (August) (http://www.concepts- methods.org/Files/WorkingPaper/PC%206%20Hadenius% 20Teorell.pdf).

Hudson, Valerie M; Bonnie Ballif-Spanvill, Mary Caprioli & Chad F Emmett (2012) Sex & World Peace. New York: Columbia University Press.

Juergensmeyer, Mark (2003) Terror in the Mind of God: The Global Rise of Religious Violence. Berkeley, CA: University of California Press.

Kampwirth, Karen (2002) Women & Guerrilla Movements: Nicaragua, El Salvador, Chiapas, Cuba. University Park, PA: Pennsylvania State University Press.

Keck, Margaret E & Kathryn Sikkink (1998) Activists Beyond Borders: Advocacy Networks in International Politics. Ithaca, NY: Cornell University Press.

Legault, Richard L (2008) Trends in American Gun Ownership. New York: LFB.

Lichbach, Mark (1987) Deterrence or escalation? The puzzle of aggregate studies of repression and dissent. Journal of Conflict Resolution 31(2): 266–297.

Lichbach, Mark (1998) Contending theories of contentious politics and the structure-action problem of social order. Annual Review of Political Science 1(1): 401–424.

Long, J Scott (1997) Regression Models for Categorical and Lim- ited Dependent Variables. Thousand Oaks, CA: Sage.

Mampilly, Zachariah Cherian (2011) Rebel Rulers: Insurgent Government and Civilian Life during Civil War. Ithaca, NY: Cornell University Press.

Marshall, Monty G; Keith Jaggers & Ted R Gurr (2011) POLITY IV Project Dataset Users’ Manual. Centre for Sys- temic Peace (www.systemicpeace.org/polity/polity4.htm).

McCarthy, John D & Mayer Zald (1977) Resource mobiliza- tion and social movements: A partial theory. American Journal of Sociology 82(6): 1212–1241.

Melander, Erik (2005) Political gender equality and state human rights abuse. Journal of Peace Research 42(2): 149–166.

Minorities At Risk Organizational Behavior Project (2009) Minorities At Risk Organizational Behavior (MAROB) Code Book. College Park, MD: Center for International Development and Conflict Management.

Mishal, Shaul & Avraham Sela (2006) The Palestinian Hamas: Vision, Violence, and Coexistence. New York: Columbia University Press.

Moore, Will H (1998) Repression and dissent: Substitution, context, and timing. American Journal of Political Science 42(3): 851–873.

Pearlman, Wendy (2011) Violence, Nonviolence, and the Pales- tinian National Movement. Cambridge: Cambridge Univer- sity Press.

Piven, Frances Fox & Richard A Cloward (1977) Poor People’s Movements, Why They Succeed and How They Fail. New York: Vintage.

Rapoport, David C (2001) The fourth wave: September 11 in the history of terrorism. Current History 100(650): 419–424.

Regan, Patrick M & Daniel Norton (2005) Greed, grievance, and mobilization in civil wars. Journal of Conflict Resolution 49(3): 319–336.

Asal et al. 317

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from

Regan, Patrick M & Aida Paskeviciute (2003) Women’s access to politics and peaceful states. Journal of Peace Research 40(3): 287–302.

Salehyan, Idean (2008) No shelter here: Rebel sanctuaries and international conflict. Journal of Politics 70(1): 54–66.

Saxton, Gregory D (2005) Repression, grievances, mobilization, and rebellion: A new test of Gurr’s model of ethno-political rebellion. International Interactions 31(1): 87–116.

Saxton, Gregory D & Michelle A Benson (2006) Structure, politics, and action: An integrated model of nationalist pro- test and rebellion. Nationalism and Ethnic Politics 12(2): 137–175.

Scarritt, James R & Susan McMillan (1995) Protest and rebel- lion in Africa: Explaining conflicts between ethnic minori- ties and the state in the 1980s. Comparative Political Studies 28(3): 323–349.

Scarritt, James R; Susan McMillan & Shaheen Mozaffar (2001) The interaction between democracy and ethno-political protest and rebellion in Africa. Comparative Political Studies 34(7): 800–827.

Springer, Kimberly (2005) Living for the Revolution: Black Feminist Organizations, 1968–1980. Durham, NC: Duke University Press.

Tarrow, Sidney (1998) Power in Movements, 2nd edn. Cambridge: Cambridge University Press.

Tarrow, Sidney (2005) The New Transnational Activism. Cam- bridge: Cambridge University Press.

Teorell, Jan; Sören Holmberg & Bo Rothstein (2008) Quality of Government Dataset, version 7. Quality of Government Institute, Göteborg University. http://www.qog.pol.gu.se/ data/

Tessler, Mark & Ina Warriner (1997) Gender, feminism, and attitudes toward international conflict: Exploring relation- ships with survey data from the Middle East. World Politics 49(2): 250–281.

Tilly, Charles (1978) From Mobilization to Revolution. Read- ing, MA: Addison-Wesley.

Tilly, Charles & Sidney Tarrow (2007) Contentious Politics. Boulder, CO: Paradigm.

Warren, Karen J & Duane L Cady (1994) Feminism and peace: Seeing connections. Hypatia 9(2): 4–20.

Weinstein, Jeremy M (2006) Inside Rebellion: The Politics of Insurgent Violence. Cambridge: Cambridge University Press.

Wilkinson, Paul (1977) Terrorism and the Liberal State. New York: Halstead.

Wood, Elisabeth Jean (2003) Insurgent Collective Action and Civil War in El Salvador. Cambridge: Cambridge Univer- sity Press.

Xu, Jun & J Scott Long (2005) Confidence intervals for pre- dicted outcomes in regression models for categorical out- comes. Stata Journal 5(4): 537–559.

Zimmerman, Amber L; M Joan McDermott & Christina M Gould (2009) The local is global: Third wave feminism, peace, and social justice. Contemporary Justice Review 12(1): 77–90.

VICTOR ASAL, b. 1965, PhD in Government and Politics (University of Maryland, College Park, 2003); Associate Professor of Political Science at the University at Albany (2003– ); current main interest: political violence and oppression and the process of political inclusion.

RICHARD LEGAULT, b. 1971, PhD in Criminology and Criminal Justice (University at Albany, 2006); Post- Doctoral Fellow and Associate Research Faculty, University of Maryland (2006–09); Social Scientist, US Department of Homeland Security, Science & Technology Directorate (2009– ); current main interest: quantitative methodology; most recent book: Trends in American Gun Ownership (LFB, 2008).

ORA SZEKELY, b. 1977, PhD in Political Science (McGill University, 2011); Assistant Professor of Political Science at Clark University (2011– ); current book project is on the domestic and foreign policies of militant political movements in the Middle East.

JONATHAN WILKENFELD, b. 1942, PhD in Political Science (Indiana University, 1969); Professor of Government and Politics, University of Maryland (1969– ); current main interest: foreign policy and international crisis.

318 journal of PEACE RESEARCH 50(3)

at New School Digital Library on August 17, 2014jpr.sagepub.comDownloaded from