Reconceptualizing resistance organizations and outcomes: Introducing the Revolutionary and Militant Organizations dataset (REVMOD)

Abstract

In recent years, scholars of various forms of conflict involving revolutionary and militant organizations (such as terrorism, civil war, and nonviolent contestation) recognized that arbitrary organizational categories and typologies often leave large-N studies incomplete and biased. In moving away from nominal categorical boundaries that produce such selection biases and looking to a more generalized conception of resistance organizations, I constructed an original dataset that aims to bridge the gap between conflict literatures. Transcending traditional classifications, the Revolutionary and Militant Organizations dataset (REVMOD) consists of over 500 resistance organizations operative sometime between the years 1940 and 2014 and includes a diverse array of types of resistance organizations – many of which utilize a multitude of tactics, operate in various conflict contexts, and/or confront numerous target types. The dataset documents organizational attributes, allies, and adversaries at annual intervals (organization-years), making reliable time-series analyses possible. Tracking variables like organizational outcome-goal type and degree of achievement, political capacity, leader/s, constituent identity group, violence and demonstration levels, size, organization aliases, and several others, REVMOD breaks new ground in the collection of information on resistance organizations and can spur countless studies. A preliminary data analysis demonstrates that differences in organizational political capacity explain variation in resistance outcomes generally and in particular contexts such as civil war, terrorism, and nonviolent revolutions. REVMOD provides a unique opportunity to develop a new research paradigm for resistance studies that employs large-N empirical analyses to uncover generalities between different forms of political contention in the contemporary era, as well as to better understand why and how distinct resistance processes may produce specific outcomes.

Keywords

conflict outcomes dataset resistance organizations resistance studies

Research on civil war, terrorism, nonviolent contestation, and revolution commonly focus on the endeavors of non-state actors to resist and change the political status quo or resist and counter attempts to undo an existing system. In recent years, conflict scholars recognized that arbitrary organizational categories often leave large-N studies incomplete and biased. Looking to a more generalized conception of resistance organizations, I constructed an original dataset that aims to bridge the gap between conflict literatures. Transcending traditional classifications, the Revolutionary and Militant Organizations dataset (REVMOD) consists of over 500 resistance organizations operative sometime between the years 1940 and 2014 and includes a diverse array of types of resistance organizations – many of which use a multitude of tactics, operate in various conflict contexts, and/or confront numerous target types.¹ The dataset documents organizational attributes, allies, and adversaries at annual intervals (organization-years), making reliable time-series analyses possible. Tracking variables like organizational outcome-goal type and degree of achievement, political capacity, leader/s, constituent identity group, violence and demonstration levels, size, organization aliases, and several others, REVMOD breaks new ground in the collection of information on resistance organizations.

This overview showcases three sets of novel organizational measures: outcome-goal achievement, strategic approach, and political capacity. Outcomes necessarily remain at the center of contemporary resistance studies. And, many revolutionary thinkers point to political capacity as the key to winning conflicts. Specifically, a strong political command structure equips organizations with the ability to exert control over their cadres and supporters, enable strategic flexibility that exploits violent and nonviolent means, and institute political responsiveness that avoids costly actions that may be tactically successful yet strategically damaging (von Clausewitz, [1832]1984; Lenin, [1902]1969; Michels, [1911]1966; Mao, [1938]1965).

I proceed with four sections. First, I present the unit of analysis, making the case for bridging various conceptual frameworks in order to better understand resistance organizations and outcomes. Second, I review the dataset’s advantages over existing datasets and detail the data collection methodology. Third, I conduct a preliminary data analysis, showing that differences in organizational political capacity explain variation in resistance outcomes generally and in particular conflict contexts such as civil war, terrorism, and nonviolent revolutions. I conclude by discussing how REVMOD can uniquely help fuse disparate but inherently related literatures and advance resistance studies.

The unit of analysis: Resistance organizations

As resistance organizations may employ a range of tactics against multiple adversary types, it makes treating organizations as the unit of analysis optimal for cross-comparison of different forms of contentious political action. Scholars have increasingly worked to link analyses of organizations traditionally classified under different categories. Though well embedded in the security and conflict studies lexicon, traditional typologies such as terrorist, guerrilla, rebel, and revolutionary or domestic and international do not promote generalizable knowledge on contentious non-state actors (Abrahms, 2007; Cunningham, 2011). McAdam, Tarrow & Tilly (2001: 4) emphasize that ‘different forms of contention […] result from similar mechanisms and processes. It wagers that we can learn more about all of them by comparing their dynamics than by looking at each on its own.’ Accordingly, to overcome selection biases that derive from arbitrary categories and incomplete datasets (Hug, 2003), I reconceptualize resistance organizations along the lines suggested by the scholars noted above. This allows for collecting a broader universe of cases and sample and shifting relational typologies (e.g. insurgent or nonviolent revolutionary) to secondary groupings, which researchers can then test empirically against one another. REVMOD thus represents an effort to continue recent scholarship that seeks to advance the empirical analyses of theories that generalize across resistance typologies, as well as to better understand why and how distinct resistance processes may produce specific outcomes.

I operationalize the unit of analysis broadly as non-state organizations that employ non-institutionalized (i.e. illegal or extralegal) means to pursue political outcome goals.² The operationalization deconstructs into four constitutive parts. Non-state refers to an entity not officially representative of a recognized state. Organization indicates a group of persons who ‘cooperate to a common end’ (Barnard, [1938]1968: 104).³ Illegal/extralegal connotes that the organization uses means not sanctioned or approved by law within its area of operation.⁴ Notably, the illegal/extralegal or ‘non-institutionalized’ criterion is essential otherwise political opposition parties would fit the definition. Political outcome goals refer to organizational aims to alter a political system’s status quo or preserve or enhance existing political advantages.

Constructing REVMOD

REVMOD consists of 536 resistance organizations operative sometime between the years 1940 and 2014. The dataset’s uniqueness and comparative advantage stem not only from new measures but also its number of variables and the inclusion of more traditional organization types than other conflict datasets.⁵ Moreover, REVMOD tracks annual data for each organization over its lifespan. These dynamic data account for developments like changes in organizational outcome goals, degrees of achievement, size, leaders, command structure, strategic approach, allies, and adversaries.⁶

REVMOD’s comparative advantage

In designing REVMOD, I evaluated existing datasets – adopting their paramount characteristics and attempting to improve on their collective limitations. The Uppsala Conflict Data Program (UCDP)/Peace Research Institute in Oslo (PRIO)-based Armed Conflict⁷ and Non-State Actors in Armed Conflict (NSA)⁸ datasets remain premier examples of violent conflict datasets. UCDP/PRIO datasets offer several attractive qualities – chiefly the documentation of extensive, high-quality, dynamic data on particular conflicts. A focus on violent conflicts marks the main constraint of UCDP/PRIO datasets, preventing comparative analyses across the spectrum of resistance typologies. The Minorities at Risk Organizational Behavior (MAROB) dataset likewise does an excellent job at documenting variables at yearly intervals and was the first dataset to include both violent and nonviolent resistance organizations.⁹ It is important to test violent and nonviolent organizations together as they regularly pursue similar outcome goals and compete for support among a shared identity group. MAROB’s greatest limitations consist of confinement to the Middle East region and a brief time frame – resulting in a small and ungeneralizable sample. The Nonviolent and Violent Campaigns and Outcomes (NAVCO) dataset similarly contains violent and nonviolent actors and improves on the concept with a global scope and longer timeframe than MAROB.¹⁰ Yet, a mixture of units of analysis, neglect of randomization, wildly inaccurate conflict time frames due to unclear operationalizations, and questionable coding on various other fronts limit the dataset’s applicability and validity.¹¹ Avoiding NAVCO’s selection and coding issues, REVMOD adopts the important practice of including nonviolent resistance organizations. Additionally, unlike the UCDP/PRIO datasets and NAVCO that track conflicts/campaigns, REVMOD’s organizational unit of analysis facilitates tests of non-state actors that operate in multiple conflicts simultaneously in pursuit of a core outcome goal. For example, one can evaluate every conflict in which Palestinian Fatah participated – including those against Jordan, and Kata’eb and Amal in Lebanon – and not just its most prominent conflict with Israel. REVMOD aims to encompass the key conceptual qualities of the dynamic datasets noted above and exceptional static/aggregated datasets listed in Table A in the Online appendix, while offering novel variables and strengthening generalizability, operationalization, validity, and reliability.

Sourcing, coding protocols, and addressing potential biases

To construct REVMOD, I first established a list of the near universe/universe of known resistance organizations operative between 1940 and 2014. I built the list by mining numerous scholarly and historical sources, as well as by referring to existing lists in the Big, Allied and Dangerous database,¹² Global Terrorism Database (GTD),¹³ Global Nonviolent Action Database (GNAD),¹⁴ UCDP/PRIO datasets, Mapping Militant Organizations database,¹⁵ Suicide-Attack Network Database (SAND),¹⁶ Invisible Armies Database,¹⁷ and Schmid & Jongman (2008). Because resistance organizations frequently use various names or claim attacks or organize demonstrations under aliases, I took special care to avoid including duplicate organizations by recording organizational names and aliases.¹⁸ Subsequently, I selected organizations for inclusion in the dataset randomly from the extensive list. See the Online appendix for the list from which I drew the sample (Table B), the selection and randomization procedure, and the list of organizations included in the dataset (Table C). I then conducted exhaustive research on each organization using a multitude of materials and cross-referenced data entries with diverse sources. Next, I employed leading techniques (Gwet, 2014) to oversee an extensive intercoder reliability (ICR) exercise, which revealed a high degree of systematic reliability in the dataset.¹⁹ The codebook in the Online appendix details the coding of each variable and documents the sources employed systematically to code each variable. I cite sources used on an individual-data entry basis in the actual dataset files.

Attempting to prevent potential biases in source material (Salehyan, 2015), I cross-checked data entries from an array of sources and denote degrees of certainty in the coding. Data entries colored in black indicate high-quality sourcing, typically involving a specialist’s publication in a scholarly outlet. For example, Ron’s (2001) article on Sendero Luminoso serves as a high-quality source for data related to the organization’s attributes. Examples of variable-specific high-quality annual sources are established databases such as GTD and SAND for organization attack levels, Polity for regime typology of adversarial states,²⁰ and the World Bank for the gross domestic product (GDP) per capita of adversarial states.²¹ Data entries in blue depict moderate degrees of certainty, where sourcing entails more journalistic or institutional accounts. These include sources like O’Ballance’s (1998) book on Lebanon’s civil war or reports in wire services or major papers. Government documents and reports from think tanks and consulting firms detailing information on certain conflicts or organizations also fall within this category. Data entries in orange signify data with lower assurance that derive from a source not involving peer review or journalistic oversight, such as chronologies compiled in security blogs like the LongWar Journal.²² This color-coding schema affords researchers greater flexibility in utilizing the data. Black-blue-orange, black-blue, and black versions of both the dynamic and static versions of REVMOD are accessible at www.revolutionarymilitant.org.

Central limitation

A pioneer of resistance studies, James C. Scott underscores the historical importance of ‘everyday forms’ of resistance. Scott holds ‘that most subordinate classes throughout most of history have rarely been afforded the luxury of open, organized, political activity’ (1985: xv). Thus, the focus on organizations and exclusion of incidental – unorganized – and often individual forms of resistance represent the dataset’s central limitation.

REVMOD’s variables

REVMOD incorporates variables that various schools of thought expect to impact resistance processes and outcomes. Table I lists REVMOD’s dynamic variables.²³ I now turn to a coding overview of three sets of novel organizational measures that frame the preliminary empirical analysis: outcome-goal achievement, strategic approach, and political capacity.

Resistance outcomes

Following previous research, I treat outcome-goal achievement as ‘success’.²⁴ ‘Outcome goals’ signify the purpose of an organization’s genesis and its raison d’être. I code organizations at annual intervals as pursuing one of six outcome goals: Separate from an adversary (e.g. Janatantrik Terai Mukti Morcha’s objective to separate from Nepal and establish a Terai state), Overthrow an adversary (e.g. the Filipino People Power Revolution’s effort to topple the Marcos regime), Expel/Eliminate an adversary (e.g. Corelli Street 6’s aim to expel Nazi forces from the Netherlands), erect an Empire (e.g. ad-Dawlah al-Islamiyyah’s objective to re-establish the Caliphate), Reform an adversarial system (e.g. the Andijan Rights Movement’s aim to liberalize Uzbekistan’s legal system), and Counter-Revolution or an aim to block one of the previous five goals (e.g. the efforts of Protestant-Loyalist organizations to stymie Catholic-Republican separatists in Northern Ireland).²⁵ Therefore, an organization achieves success when its opponent makes concessions that fulfill the outcome goal or when an organization

Table I.

Summary statistics (dynamic–annual intervals)

Variable	N = 9,953	Measure	Mean
Organization	9,953
Identity group/constituency	9,805
Area of operation	9,950
Outcome goal	8,879
Separate	8,879	Binary
Overthrow/takeover	8,879	Binary
Expel/eliminate	8,879	Binary
Empire	8,879	Binary
Reform	8,879	Binary
Counter-revolutionary	8,879	Binary
Anti-system	8,997	Binary
Achievement	8,649	21-point	5.35
Complete success	8,914	Binary
Partial success	8,914	Binary
Political command	8,873	21-point	5.99
Political command	8,894	Binary
Strictly violent	6,693	Binary
Strictly nonviolent	6,693	Binary
Mixed approach	6,693	Binary
Attacks	5,214	Continuous	15.23
Kills	5,204	Continuous	50.61
Lethality	5,204	Continuous	3.32
Suicide attacks	9,945	Continuous	0.46
Demonstrations/protests (liberal estimate)	8,124	Continuous	0.13
Demonstrations/protests (conservative estimate)	8,124	Continuous	0.09
Number of demonstrations turned violent	8,114	Continuous	0.02
Civil war	9,018	Binary
Terrorism	8,246	Binary
Size	8,400	Logged	6.92
Excluded constituency	8,921	Binary
State sponsors	6,328	Continuous	0.24
Network ties	7,516	Continuous	2.14
Adversaries	8,260
Primary adversary polity	8,086	21-point	10.46
State adversary	8,227	Binary
Democratic adversary	8,174	Binary
Adversary GDP per capita	7,213	Continuous	5,348.66
Leader/s	8,028
Leadership decapitation (general)	8,028	Binary
Leader killed	7,742	Binary
Leader arrested	7,742	Binary
Leader arrested and executed	7,742	Binary
Organization age	9,953	Continuous	17.44
Conflict duration	9,953	Continuous	14.45

eliminates, topples, conquers, or outlasts its opponent resulting in the outcome goal’s accomplishment. Other potential objective measures of effectiveness, such as quantity of attacks, attack lethality, or size of protests, might gauge tactical efficacy yet they reveal little about the ability of organizations effectively utilizing capacity to achieve the goals that justify and make their existence possible.

REVMOD includes two measures for resistance outcomes annually. The first measure scores achievement using three dichotomous variables, with organizations scoring Complete success, Partial success, or No success. Complete success refers to an organization achieving the entirety or near entirety of its stated outcome goal. Partial success occurs when an organization reaches its outcome goal in a limited way. Examples of the difference between complete and partial success include: governmental power-sharing with other organizations/parties rather than enjoying total control, gaining autonomy instead of full self-determination, seizing a portion but not the entirety of a coveted territory, or succeeding in changing a regime’s leader but not the regime. The line between partial success and failure is the absence of any gained autonomy, governmental power-sharing, territorial control, or shift in regime makeup.²⁶

REVMOD’s main outcome measure captures organizational success at a high grade of granularity.²⁷ For this graduated measure, I code organizational Achievement on a scale of 0 to 20 at yearly intervals.²⁸ Table II reports the coding questions for each marker. A score of 20 refers to ‘complete victory’, wherein an organization achieves the entirety of its stated outcome goal and defeats the primary adversary. An 18 indicates an organization achieves its outcome goal and the adversary does not currently seek to reverse the organization’s gains. A 16 means an organization achieves its outcome goal yet the adversary still poses an active threat to reverse gains. The nuances of the victory scores of 20–16 signify degrees of irreversibility and are essential for assessing the likelihood of conflict recurrence. Scores between 14 and 10 signify

Table II.

Achievement scores

Score	Coding questions
0	Organization is defunct?
2	Organization does not initiate violent operations or illegal nonviolent actions?
4	Organization initiates operations/actions solely in a defensive posture?
6	Organization initiates operations/actions offensively?
8	Organization makes/maintains marginal advancements?
10	Organization partially achieves outcome goal but the adversary controls most remaining components?
12	Organization partially achieves outcome goal but the adversary controls around half the remaining components?
14	Organization partially achieves outcome goal but the adversary controls some remaining components?
16	Outcome goal achieved but the adversary still poses an active threat to reverse gains?
18	Outcome goal achieved and the adversary does not currently pose an active threat to reverse gains?
20	Outcome goal achieved completely and the adversary is outright defeated?

degrees of partial success, depicting levels of autonomy, regime reform, territorial control, or power-sharing – depending on the outcome goal. A 14 denotes the organization nears achieving its outcome goal but the adversary controls some remaining components. A 12 means the organization achieves around half of its outcome goal and the adversary controls the remaining components. A 10 conveys that an organization achieves some of its outcome goal and the adversary controls most remaining components. An 8 marks an organization’s marginal advancement toward its outcome goal, often characterized by a challenge in a public space (e.g. a longstanding mass demonstration, occupying a capital square, controlling a small portion of territory, or the adversary begins to make overtures). A 6 means that an organization actively resists by initiating attacks or nonviolent confrontational operations. A 4 reflects that an organization conducts operations/actions solely in a defensive posture (e.g. it trains/organizes but only actively resists or fights when attacked or repressed). A 2 signifies that an organization claims to exist yet shows no signs of activity. And 0 indicates the organization is dead as it does not function on any level. For detailed examples of the achievement scoring broken down by organizational outcome-goal type, see Table J in the Online codebook.

Strategic approach

The binary variable Only violent refers to organizations that strictly employ violence in pursuing their outcome goals. Only nonviolent denotes organizations that strictly adhere to nonviolent techniques. To code ‘nonviolent-resistance’ techniques, I use Chenoweth & Stephan’s (2011: 12) definition as a starting point: the employment of ‘boycotts, strikes, protests, sit-ins, stay-aways, and other acts of civil disobedience’. I add other facets of sociopolitical life and governance that impact an entity’s capability of winning: community and political organizing operations, distributing social services, and constructing/operating public works. Many such actions aid organizations in building a ‘shadow government’ and undermining the adversary’s credibility. Recent scholarship shows the importance of analyzing organizations’ diversification of resistance strategies and tactics (Cunningham, Dahl & Frugé, 2017). Learning from the mistakes of previous datasets that rely on a problematic dichotomy that categorizes organizations as either ‘violent’ or ‘nonviolent’ (Chenoweth & Stephan, 2011; Chenoweth & Lewis, 2013), I account for organizations that apply both violent and nonviolent means by devising the category of Mixed-approach organizations.

Political command

The graduated measure of Political command aims to identify the degree of political command and control over the entirety of organizational operations. Table III displays the coding questions for each marker.²⁹ A score of 20 signifies complete political command over organizational operations both military and nonviolent, with political leaders comprising all of an organization’s top echelon. An 18 conveys that political leadership commands the organization in a unified political-military/nonviolent action hierarchy. A 16 corresponds to a strong political command overseeing compartmentalized military/nonviolent action wing/s. A 14 signifies an organization’s political wing maintains equal sway within the overall organization along with the military/nonviolent action wing/s and the organization’s primary leader comes from the political wing. A 12 depicts an equitable amount of command between military/nonviolent action and political elements. A 10 indicates an organization incorporates its political entity into the command structure. An 8 means that an organization has a political wing or party, but it remains subordinate to the military/nonviolent

Table III.

Political command scores

Score	Coding questions
0	Organization has no political component?
2	Organization engages in low-levels of political mobilization?
4	Organization fields political operatives informally?
6	Organization has formalized a role for political operatives?
8	Organization has distinct political wing or party that remains subordinate to the military/nonviolent action command?
10	Organization incorporates its political entity into the command structure?
12	Organization’s political entity has equal sway within the overall organization along with the military/nonviolent action wing/s?
14	Organization’s political entity has equal sway within the overall organization along with the military/nonviolent action wing/s and organization’s leader comes from the political entity?
16	Organization’s political leadership commands the organization though the military/nonviolent action wing/s remains compartmentalized?
18	Organization’s political leadership commands the organization in a unified hierarchy?
20	Organization’s political command represents the entire top echelon of leadership?

action hierarchy. A 6 denotes an organization’s limited formalized role for political operatives. A 4 refers to an organization’s informal political operatives, who lack an official role within the organizational structure. A 2 marks an organization’s involvement in low-level forms of political mobilization, such as distributing political propaganda or writing political graffiti. And 0 corresponds to an organization’s absence of political infrastructure, political operatives, and political messaging and outreach.³⁰

Survival and success

Of all 536 organizations in REVMOD, the mean age is 17.4 years.³¹ The mean conflict duration is 14.5 years.³² The mean for years during a conflict that an organization actively conducts operations or demonstrations as opposed to merely maneuvering defensively is 11.1 years.

Figure 1 exhibits the success of resistance organizations. The number of organizations that partially or completely succeeded in achieving an outcome goal is 128 (23.9%), and 70 (13.1%) completely achieved an outcome goal. Notably, organizations with embedded political command succeeded in 103 (81.8%) of 126 efforts, whereas organizations without embedded political command succeeded in just 25 (6.1%) of 410 endeavors (see Figure 2).³³

Testing the effect of political command on organizational success

To demonstrate the dataset’s utility, I conduct preliminary empirical analyses of the relationship between political command and conflict outcomes. Table IV displays the results of Prais-Winsten estimations that analyze REVMOD’s annual time-series data. Accounting for potential autocorrelation, Model 1 provides strong support – at a general level – for the Leninist/Maoist hypothesis that political command over organizational resistance operations propels organizations to success. Models 2 through 5 support the hypothesis vis-à-vis traditional organizational categories, restricting analyses to violent organizations, nonviolent organizations, terrorist organizations, and organizations engaged in a civil war. Model 6 again tests the full dataset, including a one-year lag of achievement as an independent variable and showing that success begets success but not to the degree that political capacity precipitates (and maintains) success.³⁴ In nearly every context, as expected from previous research (DeNardo, 1985), organizational size boosts prospects for success. Except in civil war, conflict duration has an inverse relationship with success.³⁵

Concerning strategic approach, strictly nonviolent organizations operate from the greatest disadvantage – likely

Figure 1.

The success of resistance organizations (1940–2014)

Figure 2.

Organizational command and resistance success

stemming from an initial difficulty to establish a credible threat. As a result of REVMOD’s more comprehensive sample, the finding that nonviolence negatively affects achievement contrasts studies that ‘omit […] armed campaigns’ (Chenoweth & Schock, 2015: 432) or test a significantly smaller and non-randomized sample (Chenoweth & Stephan, 2011). Moreover, the misidentification of nonviolence as a predictor of success in previous studies likely derives from the testing of a false dichotomy that categorizes organizations as either ‘violent’ or ‘nonviolent’ (Chenoweth & Stephan, 2011).³⁶ After correcting for this with the inclusion of mixed-approach organizations and the addition of a political command measure,³⁷ it becomes clear that an organization’s command structure shapes resistance performance rather than its use of violence or lack thereof. As such, due to embedded political command, some nonviolent organizations succeed despite their nonviolence – not because of it. Taking Models 3 and 6 together, the nonviolent disadvantage could reflect Michels’s ([1911]1966) notion of the ‘incompetence of the masses’ absent their subordination to a political vanguard (Lenin, [1902]1969).³⁸ The preliminary analyses illustrate that revolutionaries like Lenin and Mao correctly understood that political command infrastructure

Table IV.

Prais-Winsten regression results

y=Achievement	Model 1	Model 2 Violent organizations	Model 3 Nonviolent organizations	Model 4 Terrorist organizations	Model 5 Organizations in a civil war	Model 6
Political command	0.424***	0.346***	0.544***	0.273***	0.322***	0.415***
	(0.027)	(0.026)	(0.062)	(0.036)	(0.032)	(0.030)
Achievement t–1						0.046*
						(0.022)
Mixed approach	–0.085					–0.034
	(0.144)					(0.145)
Only nonviolent	–1.803***					–1.732***
	(0.180)					(0.173)
Conflict duration	–0.019***	–0.014**	–0.022*	–0.010*	0.023	–0.016***
	(0.004)	(0.004)	(0.011)	(0.004)	(0.012)	(0.004)
Size (logged)	0.285***	0.286***	0.241**	0.153***	0.327***	0.240***
	(0.039)	(0.040)	(0.091)	(0.039)	(0.068)	(0.037)
Anti-system	0.049	–0.327	1.638***	–0.507*	–0.436	–0.037
	(0.251)	(0.256)	(0.417)	(0.240)	(0.395)	(0.267)
Adversary polity	–0.004	–0.005	–0.021	–0.006	–0.023	–0.008
	(0.006)	(0.006)	(0.021)	(0.006)	(0.013)	(0.006)
Constant	1.996***	2.540***	–1.375	4.246***	1.836**	2.069***
	(0.288)	(0.279)	(0.948)	(0.300)	(0.571)	(0.325)
Observations	5612	4729	883	2389	1923	5188
R²	0.35	0.36	0.39	0.44	0.24	0.37
F-statistic	93.70***	106.00***	22.87***	45.99***	48.16***	74.94***
Root MSE	1.63	1.53	1.84	1.22	1.54	1.48
ρ	0.83	0.81	0.88	0.86	0.88	0.85

Coefficients with semi-robust standard errors in parentheses; ***p < 0.001, **p < 0.01, *p < 0.05.

regularly equips organizations with the necessary political responsiveness and strategic flexibility to optimally pursue their outcome goals.

Advancing resistance studies

REVMOD uniquely facilitates testing the central questions of resistance studies, examining extant competing theories, and breaking and forging empirical ground for new theoretical frameworks. Resistance studies – as broadly conceptualized in this article – tend to emphasize five core dependent variables: (1) conflict onset, (2) conflict intensity, (3) conflict duration, (4) conflict outcome, and (5) post-conflict transition and development. For potential onset and duration studies, REVMOD’s careful attention to organizational names, aliases, and operational continuation offers superior assurance regarding organizational inception and the onset and conclusion of conflicts. REVMOD’s integration of yearly political violence data from the GTD, SAND, historical sources on insurgencies, Acosta & Ramos’s (2017) fix for GTD’s missing 1993 attack events, protest/demonstration data from GNAD, and many others make possible novel and reliable dynamic analyses on conflict intensity. The analysis of the relationship between embedded political command and organizational achievement highlights that REVMOD’s detailed documentation of degrees of organizational success enables comparative analyses of resistance outcomes from multiple dimensions. REVMOD’s conflict outcome measures and intensity variables can not only aid in assessments of who wins conflicts and how but also how specific processes and outcomes may predict post-conflict environments like democratization or conflict recurrence.

Moving forward, resistance studies should focus on harnessing the vast knowledge already accumulated from fields as diverse as civil wars studies to social movement and terrorism studies. REVMOD’s construction represents an effort to further unify such disparate but related literatures. Scholars have long studied the onset, intensity, duration, outcomes, and transitions of various types of revolutionary and militant organizations independently of one another. REVMOD provides a unique opportunity to develop a new research paradigm that employs large-N empirical analyses to uncover generalities across different forms of contemporary contentious politics. Researchers can use the dataset to continue answering general questions like why conflicts emerge, why conflicts involve particular types of resistance and not others, how conflicts end, who wins, and whether the nature of conflicts explains variation in post-conflict development. REVMOD can also facilitate answering questions of cross-typology variation. For example, are the forces that spawn nonviolent revolutions unrelated to those that spark terrorism campaigns? Do special factors drive rebels? After reaching some level of intensity, does political violence cease assisting organizations in achieving their outcome goals? REVMOD can help scholars advance new research agendas and test resistance organizations operating in different contexts against one another or together in search of generalizability.

Footnotes

Replication data

The dataset, codebook, and do-file for the empirical analysis in this article, along with the Online appendix, are available at www.prio.org/jpr/datasets and .

Acknowledgments

For impactful comments, I thank Chris Gelpi, Tricia Sullivan, Dan Silverman, Alex Wendt, Bear Braumoeller, Jacek Kugler, Lissa Rogers, Mark Abdollahian, Jeremy Wallace, Jan Box-Steffensmeier, Noa Naftali, Michal Miller, JPR’s editors and anonymous reviewers.

ORCID iD

Benjamin Acosta

Notes

References

Abrahms

Max

(2007) Why democracies make superior counterterrorists. Security Studies 16(2): 223–253.

Abrahms

Max

(2012) The political effectiveness of terrorism revisited. Comparative Political Studies 45(3): 366–393.

Acosta

Benjamin

(2014) From bombs to ballots: When militant organizations transition to political parties. Journal of Politics 76(3): 666–683.

Acosta

Benjamin

Ramos

Kristen

(2017) Introducing the 1993 Terrorism and Political Violence Dataset. Studies in Conflict & Terrorism 40(3): 232–247.

Aksoy

Deniz

Carter

David

(2014) Electoral institutions and the emergence of terrorist groups. British Journal of Political Science 44(1): 181–204.

Barnard

Chester

([1938]1968) The Functions of the Executive. Cambridge, MA: Harvard University Press.

Bayer

Markus

Bethke

Felix S

Lambach

Daniel

(2016) The democratic dividend of nonviolent resistance. Journal of Peace Research 53(6): 758–771.

Chenoweth

Erica

Lewis

Orion

(2013) Unpacking nonviolent campaigns: Introducing the NAVCO 2.0 dataset. Journal of Peace Research 50(3): 415–423.

Chenoweth

Erica

Schock

Kurt

(2015) Do contemporaneous armed challenges affect the outcomes of mass nonviolent campaigns? Mobilization 20(4): 427–451.

10.

Chenoweth

Erica

Stephan

Maria

(2011) Why Civil Resistance Works. New York: Columbia University Press.

11.

Cunningham

David E

Gleditsch

Kristian Skrede

Salehyan

Idean

(2013) Non-state actors in civil wars. Conflict Management and Peace Science 30(5): 516–531.

12.

Cunningham

Kathleen Gallagher

(2011) Divide and conquer and divide and concede. American Political Science Review 105(2): 275–297.

13.

Cunningham

Kathleen Gallagher

Dahl

Marianne

Frugé

Anne

(2017) Strategies of resistance: Diversification and diffusion. American Journal of Political Science 61(3): 591–605.

14.

DeNardo

James

(1985) Power in Numbers. Princeton, NJ: Princeton University Press.

15.

Gelpi

Christopher F

Griesdorf

Michael

(2001) Winners or losers? American Political Science Review 95(3): 633–647.

16.

Gleditsch

Nils Petter

Wallensteen

Peter

Eriksson

Mikael

Sollenberg

Margareta

Strand

Håvard

(2002) Armed Conflict 1946–2001. Journal of Peace Research 39(5): 615–637.

17.

Goldstone

Jack

(2001) Toward a fourth generation of revolutionary theory. Annual Review of Political Science 4: 139–187.

18.

Gwet

Kilem L

(2014) Handbook of Inter-Rater Reliability, 4th edition. Gaithersburg, MD: Advanced Analytics.

19.

Hug

Simon

(2003) Selection bias in comparative research: The case of incomplete data sets. Political Analysis 11(3): 255–274.

20.

Lenin

Vladimir I

([1902]1969) What Is To Be Done? New York: International.

21.

Mao

Tse-Tung

([1938]1965) Problems of war and strategy. In: Selected Works of Mao Tse-Tung, Volume II. Peking: Foreign Languages.

22.

McAdam

Doug

(1999) Political Process and the Development of Black Insurgency, 1930–1970. Chicago, IL: University of Chicago Press.

23.

McAdam

Doug

Tarrow

Sidney

Tilly

Charles

(2001) Dynamics of Contention. New York: Cambridge University Press.

24.

Melander

Erik

Petersson

Therése

Themnér

Lotta

(2016) Organized Violence, 1989–2015. Journal of Peace Research 53(5): 727–742.

25.

Michels

Robert

([1911]1966) Preface. In: Paul

Eden

Paul

Cedar

(eds) Political Parties. New York: Free.

26.

O’Ballance

Edgar

(1998) Civil War in Lebanon, 1975–1992. New York: St Martin’s.

27.

Price

Bryan

(2012) Targeting top terrorists. International Security 36(4): 9–46.

28.

Ron

James

(2001) Ideology in context: Explaining Sendero Luminoso’s tactical escalation. Journal of Peace Research 38(5): 569–592.

29.

Salehyan

Idean

(2015) Best practices in the collection of conflict data. Journal of Peace Research 52(1): 105–109.

30.

Scott

James C

(1985) Weapons of the Weak: Everyday Forms of Peasant Resistance. New Haven, CT: Yale University Press.

31.

Schmid

Alex

Jongman

Albert J

(2008) Political Terrorism: A New Guide to Actors, Authors, Concepts, Data Bases, Theories, and Literature. London: Transaction.

32.

Steedly

Homer

Foley

John

(1979) The success of protest groups. Social Science Research 8(1): 1–15.

33.

Sullivan

Patricia L

(2012) Who Wins? New York: Oxford University Press.

34.

Themnér

Lotta

Wallensteen

Peter

(2014) Armed Conflicts, 1946–2013. Journal of Peace Research 51(4): 541–554.

35.

von Clausewitz

Carl

([1832]1984) On War, trans. Howards

Michael

Paret

Peter

. Princeton, NJ: Princeton University Press.

36.

Wilkenfeld

Jonathan

Asal

Victor

Pate

Amy

(2008) Minorities at Risk Organizational Behavior (MAROB) Middle East, 1980–2004 (https://dataverse.harvard.edu/dataset.xhtml?persistentId=hdl:1902.1/15973&studyListingIndex=2_fcee301dba3be9dc90023375a288).

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB