Principal Why Civil Resistance Works: The Strategic Logic of Nonviolent Conflict

Why Civil Resistance Works: The Strategic Logic of Nonviolent Conflict

For more than a century, from 1900 to 2006, campaigns of nonviolent resistance were more than twice as effective as their violent counterparts in achieving their stated goals. By attracting impressive support from citizens, whose activism takes the form of protests, boycotts, civil disobedience, and other forms of nonviolent noncooperation, these efforts help separate regimes from their main sources of power and produce remarkable results, even in Iran, Burma, the Philippines, and the Palestinian Territories. Combining statistical analysis with case studies of specific countries and territories, Erica Chenoweth and Maria J. Stephan detail the factors enabling such campaigns to succeed and, sometimes, causing them to fail. They find that nonviolent resistance presents fewer obstacles to moral and physical involvement and commitment, and that higher levels of participation contribute to enhanced resilience, greater opportunities for tactical innovation and civic disruption (and therefore less incentive for a regime to maintain its status quo), and shifts in loyalty among opponents' erstwhile supporters, including members of the military establishment. Chenoweth and Stephan conclude that successful nonviolent resistance ushers in more durable and internally peaceful democracies, which are less likely to regress into civil war. Presenting a rich, evidentiary argument, they originally and systematically compare violent and nonviolent outcomes in different historical periods and geographical contexts, debunking the myth that violence occurs because of structural and environmental factors and that it is necessary to achieve certain political goals. Instead, the authors discover, violent insurgency is rarely justifiable on strategic grounds.
Columbia University Press
ISBN 10:
ISBN 13:
Columbia Studies in Terrorism and Irregular Warfare
PDF, 5.52 MB
Descarga (pdf, 5.52 MB)

You may be interested in


Most frequently terms

You can write a book review and share your experiences. Other readers will always be interested in your opinion of the books you've read. Whether you've loved the book or not, if you give your honest and detailed thoughts then people will find new books that are right for them.


Bruce Hoffman, Series Editor

This series seeks to fill a conspicuous gap in the burgeoning literature on terrorism, guerrilla warfare, and
insurgency. The series adheres to the highest standards of scholarship and discourse and publishes books
that elucidate the strategy, operations, means, motivations, and effects posed by terrorist, guerrilla, and
insurgent organizations and movements. It thereby provides a solid and increasingly expanding foundation
of knowledge on these subjects for students, established scholars, and informed reading audiences alike.



The views expressed in this book
do not represent those of the United States Government.

Publishers Since 1893
All rights reserved
E-ISBN 978-0-231-52748-4
Library of Congress Cataloging-in-Publication Data
Chenoweth, Erica, 1980–
Why civil resistance works : the strategic logic of nonviolent conflict / Erica Chenoweth and Maria J.
p. cm. — (Columbia studies in terrorism and irregular warfare)
Includes bibliographical references and index.
ISBN 978-0-231-15682-0 (cloth : alk. paper) — ISBN 978-0-231-52748-4 (electronic)
1. Civil disobedience. 2. Nonviolence. I. Stephan, Maria J. II. Title. III. Series.
JC328.3.C474 2011
A Columbia University Press E-book.
CUP would be pleased to hear about your reading experience with this e-book at
References to Internet Web sites (URLs) were accurate at the time of writing. Neit; her the author nor
Columbia University Press is responsible for URLs that may have expired or changed since the manuscript
was prepared.

—E. C.
—M. J. S.


UPRISING, 1988-1990



SUCCESSES, 1940–2006


















COMPARED, 1987–1992













to recollect all the people from whom I
received inspiration, assistance, and unwavering support while re-searching and
writing this book. But I wish to recognize a few, with additional thanks to those
not mentioned here.
First are my colleagues at the International Center on Nonviolent Conflict—
Peter Ackerman, Jack DuVall, Hardy Merriman, Althea Middleton-Detzner,
Maciej Bartkowski, Daryn Cambridge, and Vanessa Ortiz—all of whom have
believed in and supported this project from the very start. They introduced me to
the topic and to Maria, and I gratefully acknowledge the financial support that
made the study possible. I also thank Stephen Zunes, Doug Bond, Cynthia Boaz,
and Kurt Schock for their comments on the research.
The cohort of scholars I met during two years at the Belfer Center at
Harvard’s Kennedy School of Government helped the project take off. To Ivan
Arreguín-Toft, Boaz Atzili, Kristin Bakke, Emma Belcher, Nik Biziouras, Tom
Bielefeld, Jonathan Caverley, Fotini Christia, David Cunningham, Kathleen
Cunningham, Erik Dahl, Alexander Downes, Ehud Eiran, Emily Greble, Kelly
Greenhill, Mike Horowitz, Matthew Kocher, Sarah Kreps, Matthew Kroenig,
Adria Lawrence, Jason Lyall, Steve Miller, Assaf Moghadam, Jonathan Monten,
Harris Mylonas, Wendy Pearlman, Phil Potter, Scott Radnitz, Elizabeth
Saunders, John Schuessler, Tammy Smith, Monica Toft, and Stephen Walt: your
brilliance continues to awe and humble me.
Matthew Fuhrmann gave up four days of his vacation during July 2009 to fly
across the country and help me resolve seemingly intractable problems from data
structure to simultaneous equations. I can only hope to emulate his selflessness
and clarity of mind as my career progresses.
I also appreciatively acknowledge the continued support of colleagues at the
University of Colorado. Colin Dueck, Steve Chan, David Leblang, and Jennifer
Fitzgerald are excellent mentors. Special thanks go to Susan Clarke. Everyone
lucky enough to know Susan is familiar with her dedication to mentoring young
scholars and the enthusiasm with which she challenges us intellectually while
simultaneously advocating for us professionally. My classmates at the University

of Colorado have also proved to be some of my most valued colleagues. I owe
Jessica Teets, Orion Lewis, Michael Touchton, Helga Sverrisdóttír, and Marilyn
Averill a great debt for helping me mature intellectually, and I look forward to
our continued collaborations. Thanks also to my earlier mentors at the University
of Dayton, including Margaret Karns, David Ahern, Jaro Bilocerkowycz, Gerald
Kerns, and Mark Ensalaco.
The Institute of International Studies at the University of California, Berkeley,
provided me with a scholarly home away from home from 2007 to 2009. I am
especially indebted to Ned Walker, Regine Spector, and Brent Durbin, who
provided useful feedback at various stages of the project.
I also thank my colleagues at Wesleyan University. Don Moon has been a
relentless advocate of the project, and a fellowship at Wesleyan’s Center for the
Humanities, under the headship of Jill Morawski, provided me with useful
feedback and time to complete the manuscript. I also thank my colleagues in the
Government Department for their friendship and support, especially Mary Alice
Haddad, Peter Rutland, Mike Nelson, Erika Fowler, and Doug Foyle for
commenting on various versions of the manuscript. I am indebted to several
terrific students, especially Jeremy Berkowitz for helping with data collection
and Nicholas Quah for his assistance in proofreading the manuscript. Elizabeth
Wells, at American University, provided research assistance during the early
stages of data collection.
We benefited from outstanding feedback from seminar and panel participants
at Georgetown, Rutgers, Yale, Harvard, Wesleyan, the University of Dayton, the
United States Institute of Peace, and King’s College, as well as at meetings of
the International Studies Association, the American Political Science
Association, and the World International Studies Committee.
Our editor at Columbia University Press, Anne Routon, has been extremely
helpful throughout the preparation of the manuscript, as has her assistant Alison
Alexanian. We thank them both for their responsiveness and guidance and for
securing top-notch reviews that helped us to improve the manuscript. We also
thank Mike Ashby for his stellar copyediting.
My family’s generosity is what has made everything possible. All the
Chenoweths and Abels have inspired and encouraged me throughout my life and
career. My parents, Richard and Marianne, have been my most persistent
advocates, and even read and commented on draft chapters. My sister Andrea
and her fiancé Phil are terrific friends and brilliant communicators; I thank them
for their support and inspiration. I also thank my brother, Christopher, and his
wife, Miranda. In the past year, Christopher and Miranda have blessed us all
with William, my only nephew, whose few months on this earth have made me

even more dedicated to helping to end violent conflict wherever it is
unnecessary. I also thank the Petty family—Kathy, Linda, Mattie Jean, and
Warren—as well as Tyler, Elizabeth, Stephanie, and Adam for supporting me
through various stages of this project. I owe a debt I can never fully repay to
Kathe, Angi, Joyanna, Melody, Kathy, George, Tommy, Scott, Rachel, Vic,
Marc, Nadia, and Gelong Tashi for all that they have given to me. And finally,
there is Allison. The daily joys of sharing our lives together have kept me afloat
through this and many other endeavors. I thank her for her wisdom, patience,
kindness, humor, and enduring eagerness for adventure.

I did not expect to write a book on people power with a domestic terrorism
expert who takes delight in running regression analyses! But after our chance
meeting at Colorado College four summers ago, Erica and I realized that we
needed to bring together our respective expertise to produce this book. And it
has been a great ride together. I would like to thank first my mentors from the
Fletcher School, including Richard Shultz, Eileen Babbitt, and Hurst Hannum
for supporting my initial foray into the study of civil resistance. Professor Shultz
and Steve Miller, from Harvard’s Belfer Center for Science and International
Affairs, recognized the weighty international security implications of popular
struggles involving different weapons and enthusiastically encouraged me to
pursue this line of research.
Dr. Peter Ackerman, one of the world’s leading experts on strategic nonviolent
action, became my Fletcher dissertation adviser, mentor, and friend. Peter
understood, when writing his own doctoral dissertation four decades ago with
Gene Sharp, a pioneer in the field of nonviolent action to whom we all owe a
great debt of gratitude, that eventually the academy would catch on to the
remarkable albeit underappreciated track record of popular nonviolent struggles
around the world. As the founding chair of the International Center on
Nonviolent Conflict (ICNC), Peter, along with his partner, Jack DuVall, have
shepherded the global expansion of knowledge and practical know-how about
the waging of nonviolent struggle. During my tenure at the ICNC I had the
opportunity to interact with some remarkable and courageous nonviolent
activists from all over the world. Their determination, bravery, and will to win
using nonviolent methods have been a source of profound inspiration for me.
I thank Peter and Jack, along with the incredibly dedicated people at the ICNC

and at Rockport Capital, including Hardy Merriman, Shaazka Beyerle, Vanessa
Ortiz, Berel Rodal, Althea Middleton-Detzner, Nicola Barrach, Maciej
Bartkowski, Jake Fitzpatrick, Daryn Cambridge, Suravi Bhandari, Deena
Patriarca, Ciel Lagumen, and Kristen Kopko for their hard work, support, and
friendship. Hardy Merriman, in particular, has been an editing rock star. The
ICNC’s diverse team of academic advisers, including Stephen Zunes, Kurt
Schock, John Gould, Mary Elizabeth King, Larry Diamond, Doug McAdam, Les
Kurtz, Cyndi Boaz, Janet Cherry, Howard Barrell, Roddy Brett, Kevin Clements,
Barry Gan, Scott O’Bryan, Lee Smithey, Victoria Tin-bor Hui, Brian Martin,
Senthil Ram, April Carter, and Howard Clark have provided Erica and me with
good advice and prompt and thoughtful feedback on earlier iterations of this
work. Mubarak Awad and Michael Beer, from Nonviolence International, have
also been great supporters over the years. Through their own interdisciplinary
work, the aforementioned scholars and scholar-practitioners have made
significant strides to advance the study and practice of civil resistance.
Some of my most enjoyable and amusing moments at the ICNC were spent in
the company of “the Serbs”—the young guns who formed Otpor and helped
mobilize the Serbian population to nonviolently oust “the butcher of the
Balkans” in 2000. Srdja Popovic, Ivan Marovic, Slobo Djinovic, and Andrej
Milojevic went on to found the Center for Applied Nonviolent Action and
Strategies, a Belgrade-based NGO that trains nonviolent activists throughout the
world. May they continue to grow a global cadre of nonviolent conflict veterans
and help transfer skills and hope to a new generation of civic leaders.
Ambassador Mark Palmer, who has been a great mentor of mine, showed me a
different side of the U.S. State Department and encouraged me to be a friend of
nonviolent-change agents from within the U.S. government. Through his work
with the Council for a Community of Democracies, Mark is helping
institutionalize global solidarity with those who are fighting against huge odds to
defend basic rights and freedoms. I greatly admire Mark and hope to follow in
his footsteps.
I would also like to extend thanks to my Pol-Mil colleagues at the U.S.
embassy in Kabul, particularly the Civ-Mil Plans and Assessments team. Phil
Kosnett, JoAnne Wagner, Melanie Anderton, Jen Munro, Emilie Lemke, and
Tammy Rutledge have listened to me expound on the virtues of civic
mobilization while supporting my efforts to engage with Afghan civil society
and speak publicly about civil resistance in Afghanistan. I hope that organized
civic action led by Afghans will help transform this war-torn society and lead it
to a more peaceful future.
Finally, I would like to thank my parents, Marianne and Phil, and my brother

Peter, whose love, encouragement, and insistence that I maintain a sense of
perspective (and humor) while working on this book helped see me through. A
girl could not ask for a more supportive and caring family. I am also grateful for
the friendships of those in Vermont who continue to serve as my “prayer
warriors.” They know who they are.

Why Civil Resistance Works

Nonviolence is fine as long as it works.


Indonesian president Suharto ordered a
full-scale invasion of East Timor, claiming that the left-leaning nationalist group
that had declared independence for East Timor a month earlier, the
Revolutionary Front for an Independent East Timor (Fretilin), was a communist
threat to the region. Fretilin’s armed wing, the Forças Armadas de Libertação
Nacional de Timor-Leste (Falintil), led the early resistance to Indonesian
occupation forces in the form of conventional and guerrilla warfare. Using
weapons left behind by Portuguese troops,1 Falintil forces waged armed struggle
from East Timor’s mountainous jungle region. But Falintil would not win the
day. Despite some early successes, by 1980 Indonesia’s brutal counterinsurgency
campaign had decimated the armed resistance along with nearly one third of the
East Timorese population.2
Yet nearly two decades later, a nonviolent resistance movement helped to
successfully remove Indonesian troops from East Timor and win independence
for the annexed territory. The Clandestine Front, an organization originally
envisaged as a support network for the armed movement, eventually reversed
roles and became the driving force behind the nonviolent, pro-independence
resistance. Beginning in 1988, the Clandestine Front, which grew out of an East
Timorese youth movement, developed a large decentralized network of activists,
who planned and executed various nonviolent campaigns inside East Timor, in
Indonesia, and internationally. These included protests timed to the visits of
diplomats and dignitaries, sit-ins inside foreign embassies, and international
solidarity efforts that reinforced Timorese-led nonviolent activism.
The Indonesian regime repressed this movement, following its standard
approach to violent and nonviolent challengers from within. But this repression
backfired. Following the deaths of more than two hundred East Timorese

nonviolent protestors at the hands of Indonesian troops in Dili in November
1991, the pro-independence campaign experienced a major turning point. The
massacre, which was captured on film by a British cameraman, was quickly
broadcast around the world, causing international outrage and prompting the
East Timorese to rethink their strategy (Kohen 1999; Martin, Varney, and
Vickers 2001). Intensifying nonviolent protests and moving the resistance into
Indonesia proper became major components of the new strategy.
Suharto was ousted in 1998 after an economic crisis and mass popular
uprising, and Indonesia’s new leader, B. J. Habibie, quickly pushed through a
series of political and economic reforms designed to restore stability and
international credibility to the country. There was tremendous international
pressure on Habibie to resolve the East Timor issue, which had become a
diplomatic embarrassment, not to mention a huge drain on Indonesia’s budget.
During a 1999 referendum, almost 80 percent of East Timorese voters opted for
independence. Following the referendum, Indonesianbacked militias launched a
scorched-earth campaign that led to mass destruction and displacement. On
September 14, 2000, the UN Security Council voted unanimously to authorize
an Australian-led international force for East Timor.3
The United Nations Transitional Administration in East Timor oversaw a twoyear transition period before East Timor became the world’s newest independent
state in May 2002 (Martin 2000). Although a small number of Falintil guerrillas
(whose targets had been strictly military) kept their weapons until the very end,
it was not their violent resistance that liberated the territory from Indonesian
occupation. As one Clandestine Front member explained, “The Falintil was an
important symbol of resistance and their presence in the mountains helped boost
morale, but nonviolent struggle ultimately allowed us to achieve victory. The
whole population fought for independence, even Indonesians, and this was
Similarly, in the Philippines in the late 1970s, several revolutionary guerrilla
groups were steadily gaining strength. The Communist Party of the Philippines
and its New People’s Army (NPA) were inspired by Marxist-Leninist-Maoist
ideologies and pursued armed revolution to gain power. State-sponsored military
attacks on the NPA dispersed the guerrilla resistance until the NPA encompassed
all regions of the country. The Philippine government launched a concerted
counterinsurgency effort, and the NPA was never able to achieve power.
In the early 1980s, however, members of the opposition began to pursue a
different strategy. In 1985 the reformist opposition united under the banner of
UNIDO (United Nationalist Democratic Organization) with Cory Aquino as its
presidential candidate. In the period leading up to the elections, Aquino urged

nonviolent discipline, making clear that violent attacks against opponents would
not be tolerated. Church leaders, similarly, insisted on discipline, while the
National Citizens’ Movement for Free Elections trained half a million volunteers
to monitor elections.
When Marcos declared himself the winner of the 1986 elections despite the
counterclaims of election monitors, Cory Aquino led a rally of 2 million
Filipinos, proclaiming victory for herself and “the people.” The day after
Marcos’s inauguration, Filipinos participated in a general strike, a boycott of the
state media, a massive run on state-controlled banks, a boycott of crony
businesses, and other nonviolent activities.
A dissident faction of the military signaled that it favored the opposition in
this matter, encouraging the opposition to form a parallel government on
February 25 with Aquino at its head. Masses of unarmed Filipino civilians,
including nuns and priests, surrounded the barracks where the rebel soldiers
were holed up, forming a buffer between those soldiers and those who remained
loyal to Marcos. President Ronald Reagan’s administration had grown weary of
Marcos and signaled support for the opposition movement. That evening, U.S.
military helicopters transported Marcos and his family to Hawaii, where they
remained in exile. Although the Philippines has experienced a difficult transition
to democracy, the nonviolent campaign successfully removed the Marcos
dictatorship. Where violent insurgency had failed only a few years earlier, the
People Power movement succeeded.

The preceding narratives reflect both specific and general empirical puzzles.
Specifically, we ask why nonviolent resistance has succeeded in some cases
where violent resistance had failed in the same states, like the violent and
nonviolent pro-independence campaigns in East Timor and regime-change
campaigns in the Philippines. We can further ask why nonviolent resistance in
some states fails during one period (such as the 1950s Defiance Campaign by
antiapartheid activists in South Africa) and then succeeds decades later (such as
the antiapartheid struggle in the early 1990s).
These two specific questions underline a more general inquiry, which is the
focus of this book. We seek to explain two related phenomena: why nonviolent
resistance often succeeds relative to violent resistance, and under what
conditions, nonviolent resistance succeeds or fails.5
Indeed, debates about the strategic logic of different methods of traditional
and nontraditional warfare have recently become popular among security studies
scholars (Abrahms 2006; Arreguín-Toft 2005; Byman and Waxman 1999, 2000;

Dashti-Gibson, Davis, and Radcliff 1997; Drury 1998; Horowitz and Reiter
2001; Lyall and Wilson 2009; Merom 2003; Pape 1996, 1997, 2005; Stoker
2007). Implicit in many of these assessments, however, is an assumption that the
most forceful, effective means of waging political struggle entails the threat or
use of violence. For instance, a prevailing view among political scientists is that
opposition movements select terrorism and violent insurgency strategies because
such means are more effective than nonviolent strategies at achieving policy
goals (Abrahms 2006, 77; Pape 2005). Often violence is viewed as a last resort,
or a necessary evil in light of desperate circumstances. Other scholarship focuses
on the effectiveness of military power, without comparing it with alternative
forms of power (Brooks 2003; Brooks and Stanley 2007; Desch 2008; Johnson
and Tierney 2006).
Despite these assumptions, in recent years organized civilian populations have
successfully used nonviolent resistance methods, including boycotts, strikes,
protests, and organized noncooperation to exact political concessions and
challenge entrenched power. To name a few, sustained and systematic nonviolent
sanctions have removed autocratic regimes from power in Serbia (2000),
Madagascar (2002), Georgia (2003), and Ukraine (2004–2005), after rigged
elections; ended a foreign occupation in Lebanon (2005); and forced Nepal’s
monarch to make major constitutional concessions (2006). In the first two
months of 2011, popular nonviolent uprisings in Tunisia and Egypt removed
decades-old regimes from power. As this book goes to press, the prospect of
people power transforming the Middle East remains strong.
In our Nonviolent and Violent Campaigns and Outcomes (NAVCO) data set,
we analyze 323 violent and nonviolent resistance campaigns between 1900 and
2006.6 Among them are over one hundred major nonviolent campaigns since
1900, whose frequency has increased over time. In addition to their growing
frequency, the success rates of nonviolent campaigns have increased. How does
this compare with violent insurgencies? One might assume that the success rates
may have increased among both nonviolent and violent insurgencies. But in our
data, we find the opposite: although they persist, the success rates of violent
insurgencies have declined.
The most striking finding is that between 1900 and 2006, nonviolent
resistance campaigns were nearly twice as likely to achieve full or partial
success as their violent counterparts. As we discuss in chapter 3, the effects of
resistance type on the probability of campaign success are robust even when we
take into account potential confounding factors, such as target regime type,
repression, and target regime capabilities.7
The results begin to differ only when we consider the objectives of the

resistance campaigns themselves. Among the 323 campaigns, in the case of
antiregime resistance campaigns, the use of a nonviolent strategy has greatly
enhanced the likelihood of success. Among campaigns with territorial
objectives, like antioccupation or self-determination, nonviolent campaigns also
have a slight advantage. Among the few cases of major resistance that do not fall
into either category (antiapartheid campaigns, for instance), nonviolent
resistance has had the monopoly on success.
The only exception is that nonviolent resistance leads to successful secession
less often than violent insurgency. Although no nonviolent secession campaigns
have succeeded, only four of the forty-one violent secession campaigns have
done so (less than 10 percent), also an unimpressive figure. The implication is
that campaigns seeking secession are highly unlikely to succeed regardless of
whether they employ nonviolent or violent tactics. We explore various factors
that could influence these results in chapter 3. It is evident, however, that
especially among campaigns seeking regime change or liberation from foreign
occupation, nonviolent resistance has been strategically superior. The success of
these nonviolent campaigns—especially in light of the enduring violent
insurgencies occurring in many of the same countries—begs systematic





This book investigates the reasons why—in spite of conventional wisdom to
the contrary—civil resistance campaigns have been so effective compared with
their violent counterparts. We also consider the reasons why some nonviolent
campaigns have failed to achieve their stated aims, and the reasons why violent
insurgencies sometimes succeed.

Our central contention is that nonviolent campaigns have a participation
advantage over violent insurgencies, which is an important factor in determining
campaign outcomes. The moral, physical, informational, and commitment
barriers to participation are much lower for nonviolent resistance than for violent
insurgency. Higher levels of participation contribute to a number of mechanisms
necessary for success, including enhanced resilience, higher probabilities of
tactical innovation, expanded civic disruption (thereby raising the costs to the
regime of maintaining the status quo), and loyalty shifts involving the
opponent’s erstwhile supporters, including members of the security forces.
Mobilization among local supporters is a more reliable source of power than the
support of external allies, which many violent campaigns must obtain to
compensate for their lack of participants.
Moreover, we find that the transitions that occur in the wake of successful
nonviolent resistance movements create much more durable and internally
peaceful democracies than transitions provoked by violent insurgencies. On the

whole, nonviolent resistance campaigns are more effective in getting results and,
once they have succeeded, more likely to establish democratic regimes with a
lower probability of a relapse into civil war.
Nestling our argument between literatures on asymmetrical warfare,
contentious politics, and strategic nonviolent action, we explain the relative
effectiveness of nonviolent resistance in the following way: nonviolent
campaigns facilitate the active participation of many more people than violent
campaigns, thereby broadening the base of resistance and raising the costs to
opponents of maintaining the status quo. The mass civilian participation in a
nonviolent campaign is more likely to backfire in the face of repression,
encourage loyalty shifts among regime supporters, and provide resistance leaders
with a more diverse menu of tactical and strategic choices. To regime elites,
those engaged in civil resistance are more likely to appear as credible negotiating
partners than are violent insurgents, thereby increasing the chance of winning
However, we also know that resistance campaigns are not guaranteed to
succeed simply because they are nonviolent. One in four nonviolent campaigns
since 1900 was a total failure. In short, we argue that nonviolent campaigns fail
to achieve their objectives when they are unable to overcome the challenge of
participation, when they fail to recruit a robust, diverse, and broad-based
membership that can erode the power base of the adversary and maintain
resilience in the face of repression.
Moreover, more than one in four violent campaigns has succeeded. We briefly
investigate the question of why violent campaigns sometimes succeed. Whereas
the success of nonviolent campaigns tends to rely more heavily on local factors,
violent insurgencies tend to succeed when they achieve external support or when
they feature a central characteristic of successful nonviolent campaigns, which is
mass popular support. The presence of an external sponsor combined with a
weak or predatory regime adversary may enhance the credibility of violent
insurgencies, which may threaten the opponent regime. The credibility gained
through external support may also increase the appeal to potential recruits,
thereby allowing insurgencies to mobilize more participants against the
opponent. International support is, however, a double-edged sword. Foreign-state
sponsors can be fickle and unreliable allies, and state sponsorship can produce a
lack of discipline among insurgents and exacerbate free rider problems (Bob
2005; Byman 2005).

We bring to bear several different types of evidence to support our argument,

including statistical evidence from the NAVCO data set and qualitative evidence
from four case studies: Iran, the Palestinian Territories, Burma, and the
It is appropriate here to briefly define the terms to which we will consistently
refer in this book. First, we should distinguish violent and nonviolent tactics. As
noted earlier, there are some difficulties with labeling one campaign as violent
and another as nonviolent. In many cases, both nonviolent and violent
campaigns exist simultaneously among competing groups. Often those who
employ violence in mass movements are members of fringe groups who are
acting independently, or in defiance of, the central leadership; or they are agents
provocateurs used by the adversary to provoke the unarmed resistance to adopt
violence (Zunes 1994). Alternatively, often some groups use both nonviolent and
violent methods of resistance over the course of their existence, as with the ANC
in South Africa. Characterizing a campaign as violent or nonviolent simplifies a
complex constellation of resistance methods.
It is nevertheless possible to characterize a campaign as principally nonviolent
based on the primacy of nonviolent resistance methods and the nature of the
participation in that form of resistance. Sharp defines nonviolent resistance as “a
technique of socio-political action for applying power in a conflict without the
use of violence” (1999, 567). The term resistance implies that the campaigns of
interest are noninstitutional and generally confrontational in nature. In other
words, these groups are using tactics that are outside the conventional political
process (voting, interest-group organizing, or lobbying). Although institutional
methods of political action often accompany nonviolent struggles, writes
sociologist Kurt Schock, nonviolent action occurs outside the bounds of
institutional political channels (2003, 705).8
Our study focuses instead on a type of political activity that deliberately or
necessarily circumvents normal political channels and employs noninstitutional
(and often illegal) forms of action against an opponent. Civil resistance employs
social, psychological, economic, and political methods, including boycotts
(social, economic, and political), strikes, protests, sit-ins, stay-aways, and other
acts of civil disobedience and noncooperation to mobilize publics to oppose or
support different policies, to delegitimize adversaries, and to remove or restrict
adversaries’ sources of power (Sharp 1973).9 Nonviolent resistance consists of
acts of omission, acts of commission, and a combination of both (Sharp 2005).10
We characterize violent resistance as a form of political contention and a
method of exerting power that, like nonviolent resistance, operates outside
normal political channels. While conventional militaries use violence to advance
political goals, in this book we are concerned with the use of unconventional

violent strategies used by nonstate actors.11 These strategies are exhibited in
three main categories of unconventional warfare: revolutions, plots (or coups
d’état), and insurgencies, which differ according to the level of premeditated
planning, protractedness, and means of overthrowing the existing order.12 The
weapons system available to an armed insurgent is very different from that of its
nonviolent analogue. Violent tactics include bombings, shootings, kidnappings,
physical sabotage such as the destruction of infrastructure, and other types of
physical harm of people and property. However, the cases we examine do not
include military coups, since we are primarily interested in substate actors that
are not part of the state. Both violent and nonviolent campaigns seek to take
power by force, though the method of applying force differs across the different
resistance types.
The list of nonviolent campaigns was initially gathered from an extensive
review of the literature on nonviolent conflict and social movements. Then these
data were corroborated with multiple sources, including encyclopedias, case
studies, and a comprehensive bibliography on nonviolent civil resistance by
April Carter, Howard Clark, and Michael Randle (2006). Finally, we consulted
with experts in the field, who suggested any remaining conflicts of note. The
resulting list includes major campaigns that are primarily or entirely nonviolent.
Campaigns where a significant amount of violence occurred are not considered
Violent campaign data are derived primarily from Kristian Gleditsch’s (2004)
updates to the Correlates of War (COW) database on intrastate wars, Jason Lyall
and Isaiah Wilson’s (2009) database of insurgencies, and Kalev Sepp’s (2005)
list of major counterinsurgency operations. The COW data set requires all
combatant groups to be armed and to have sustained a thousand battle deaths
during the course of the conflict, suggesting that the conflict is necessarily
This study makes a further qualification. Nonviolent and violent campaigns
are used to promote a number of different policy objectives, ranging from
increasing personal liberties to obtaining greater rights or privileges for an ethnic
group to demanding national independence. However, this project is concerned
primarily with three specific, intense, and extreme forms of resistance:
antiregime, antioccupation, and secession campaigns. These campaign types are
chosen for several reasons. First, they provide a hard case for civil resistance.
Antiregime, antioccupation, and self-determination campaigns are typically
associated in the literature with violence, whereas civil rights and other strictly
human rights movements are more commonly associated with nonviolent
methods. However, in this study we argue that nonviolent resistance can be used

to achieve political objectives most commonly identified with violent
Success and failure are also complex outcomes, about which much has been
written (Baldwin 2000). For our study, to be considered a “success” a campaign
had to meet two conditions: the full achievement of its stated goals (regime
change, antioccupation, or secession) within a year of the peak of activities and a
discernible effect on the outcome, such that the outcome was a direct result of
the campaign’s activities (Pape 1997).13 The second qualification is important
because in some cases the desired outcome occurred mainly because of other
conditions. The Greek resistance against the Nazi occupation, for example, is not
coded as a full success even though the Nazis ultimately withdrew from Greece.
Although effective in many respects, the Greek resistance alone cannot be
credited with the ultimate outcome of the end of Nazi influence over Greece
since the Nazi withdrawal was the result of the Allied victory rather than solely
Greek resistance.
The term campaign is also somewhat contentious as a unit of analysis.
Following Ackerman and Kruegler (1994, 10–11), we define a campaign as a
series of observable, continual tactics in pursuit of a political objective. A
campaign can last anywhere from days to years. Campaigns have discernible
leadership and often have names, distinguishing them from random riots or
spontaneous mass acts.14 Usually campaigns have distinguishable beginning and
end points, as well as discernible events throughout the campaign. In the case of
resistance campaigns, beginning and end points are difficult to determine, as are
the events throughout the campaign. In some cases, information on such events
is readily available (e.g., Northern Ireland from 1969 to 1999); however, in most
cases, it is not. Therefore, our characterization of the beginning and end dates of
campaigns is based on consensus data and multiple sources.15
Some readers may be tempted to dismiss our findings as the results of
selection effects, arguing that the nonviolent campaigns that appear in our
inventory are biased toward success, since it is the large, often mature campaigns
that are most commonly reported. Other would-be nonviolent campaigns that are
crushed in their infancy (and therefore fail) are not included in this study. This is
a potential concern that is difficult to avoid.
We adopted a threefold data-collection strategy to address this concern. First,
our selection of campaigns and their beginning and end dates is based on
consensus data produced by multiple sources. Second, we have established
rigorous standards of inclusion for each campaign. The nonviolent campaigns
were initially gathered from an extensive review of the literature on nonviolent
conflict and social movements. Then these data were corroborated with multiple

sources, including encyclopedias, case studies, and the bibliography by Carter,
Clark, and Randle (2006).
Finally, we circulated the data set among experts in nonviolent conflict. These
experts were asked to assess whether the cases were appropriately characterized
as major nonviolent conflicts, whether any notable conflicts had been omitted,
and whether we had properly accounted for failed movements. Where the
experts suggested additional cases, the same corroboration method was used.
Our confidence in the data set that emerged was reinforced by numerous
discussions among scholars of both nonviolent and violent conflicts.
Nonetheless, what remains absent from the data set is a way to measure the
nonstarters, the nonviolent or violent campaigns that never emerged because of
any number of reasons. Despite this concern, we feel confident proceeding with
our inquiry for two main reasons. First, this bias applies as much to violent
campaigns as to nonviolent ones—many violent campaigns that were defeated
early on are also unreported in the data. Second, this study is not concerned
primarily with why these campaigns emerge but with how well they perform
relative to their competitors that use different methods of resistance. We focus on
the efficacy of campaigns as opposed to their origins, and we argue that we can
say something about the effectiveness of nonviolent campaigns relative to
violent campaigns. We do concede, however, that improved data collection and
analysis and finding ways to overcome the selection bias inherent in much
scholarship on conflict are vital next steps for the field.

Generally, scholars have eschewed the systematic comparison of the outcomes of
violent and nonviolent movements. One notable exception is William Gamson,
whose seminal work (1990) on American challenge groups discovered that
groups employing force and violence were more successful than groups
refraining from violent tactics (McAdam, McCarthy, and Zald 1996, 14). Not
only does he seem to conflate force with violence, but also his conclusions,
while perhaps pertinent to certain types of groups within the American political
system, do not necessarily apply to all countries during all times.16
Hence scholarship on this question rightly investigates whether such
generalizations are applicable to other places and periods. In attempting to
understand the relationship between nonviolent and violent tactics and the
outcomes of resistance campaigns, however, scholars have tended to focus on
single case studies or small-n comparisons in what has become a rich
accumulation of research and knowledge on the subject (Ackerman and DuVall
2000; Ackerman and Kruegler 1994; Boudreau 2004; Schock 2005; Sharp 1973,

2005; Wehr, Burgess, and Burgess 1994; Zunes 1994; Zunes, Kurtz, and Asher
1999). What has been missing, though, are catalogs of known campaigns and
systematic comparisons of the outcomes of both nonviolent and violent
resistance campaigns, although this trend has begun to shift (Shaykhutdinov
2010; Stephan and Chenoweth 2008).
As one might expect, there are several good reasons why social scientists have
avoided comparing the dynamics and outcomes of nonviolent and violent
campaigns, including their relative effectiveness. First, the separation of
campaigns into violent and nonviolent for analytical purposes is problematic.
Few campaigns, historically, have been purely violent or nonviolent, and many
resistance movements, particularly protracted ones, have had violent and
nonviolent periods. Armed and unarmed elements often operate simultaneously
in the same struggle. Still, it is possible to distinguish between different
resistance types based on the actors involved (civilians or armed militants) and
the methods used (nonviolent or violent).17 Scholars have identified the unique
characteristics of these different forms of struggle, and we feel comfortable
characterizing some resistance campaigns as primarily violent and others as
primarily nonviolent. We are furthermore careful to avoid characterizing a
campaign as violent merely because the regime uses violence in an attempt to
suppress the protest activity.
Second, security studies scholars seem to have eschewed the study of
nonviolent action because nonviolent action is not typically viewed as a form of
insurgency or asymmetrical warfare (Schock 2003). Groups deliberately
adopting nonviolent tactics are commonly understood as doing so for moral or
principled reasons (Howes 2009). Since some key authors promoting strategic
nonviolent action have also been pacifists, this characterization has not been
wholly unfounded. Nonetheless, among some security studies scholars, the idea
that resistance leaders might choose nonviolent tactics as a strategic choice may
be considered naive or implausible. Although the topic of civilian-based defense,
a type of unconventional defense involving civilian populations defending their
nations from military invasions and occupations using organized noncooperation
and civil disobedience, received the attention of security and strategic studies
(including the RAND Corporation) during the Cold War, interest in the subject
from the security studies community has waned since the fall of the iron curtain
(Sharp 1990).18 Hence the serious study of strategic nonviolent action has
remained something of a pariah within security studies despite decades of
scholarship on the subject.
Finally, the questions of interest in this book—whether nonviolent resistance
methods are more effective than violent resistance methods and under which

conditions civil resistance succeeds or fails—are by nature extremely difficult to
study. It is not by accident that few authors have been able to compile large-n
data sets on the subject despite important efforts to do so.19 The measurement of
effectiveness itself is difficult to gather and defend, and the independent effects
of resistance methods on the outcomes are not always easy to discern given the
complexity of these contentious episodes.
Despite the challenges associated with studying this subject, we argue that the
theoretical and policy implications of the research questions at hand are too
important to avoid. Sidney Tarrow has argued that investigating the reasons why
movements succeed and fail is one of the main foci of the entire contentious
politics research program (1998). Our book demonstrates that scholars can take a
reasoned look at the relative effectiveness of nonviolent and violent resistance,
even if the measures of such terms are imperfect. We undertake such an
exploration by examining 323 cases from 1900 to 2006 of major nonviolent and
violent campaigns seeking regime change, the expulsion of foreign occupiers, or
secession. This research is the first to catalog, compare, and analyze all known
cases of major armed and unarmed insurrections during this period. From this
data, we find support for the perspective that nonviolent resistance has been
strategically superior to violent resistance during the twentieth and twenty-first
centuries. Because the data are highly aggregated, we provide only a first look at
these trends. But our findings point to a powerful relationship that scholars and
policy makers should take seriously.

This research is situated among several distinct albeit related subfields of
political science and sociology. We are explicit in conceptualizing civil
resistance as a form of unconventional warfare, albeit one that employs different
weapons and applies force differently. The literature on contentious politics has
long explored the relationship between methods and outcomes. Recent
scholarship in security studies has explored similar questions.20 Others in the
discipline deal with the concept of strategic effectiveness in an indirect, if
somewhat peripheral, way. For instance, in his seminal work on the political
economy of rebellion, Jeremy Weinstein (2007) argues that activist rebellions are
more likely than opportunistic rebellions to achieve their strategic objectives.
Activist rebellions, which are dependent on social support, are more likely to
target opponents selectively. Opportunistic rebellions target indiscriminately,
thereby undermining their public support.
Wood (2000, 2003) argues that transitions to democracy are likely when
insurgents are able to successfully raise the costs to economic elites of

maintaining the status quo, a process that emerges when labor unions and worker
parties strike over an extended period. DeNardo’s work (1985) on mass
movements also demonstrates that methods and outcomes of revolutions are
related, with disruption and mass mobilization being key determinants of
revolutionary success. However, Weinstein (2007), Wood (2000, 2003), and
DeNardo (1985) all remain agnostic as to how the methods of resistance—
nonviolent or violent—could affect the outcomes of resistance campaigns.
Following those who have analyzed nonviolent campaigns through the lens of
strategic theory, we are similarly interested in the relationship between strategy
and outcome (Ackerman and Kruegler 1994; Ganz 2010; Helvey 2004; Popovic
et al. 2007; Sharp 1973). Our perspective does not assume that nonviolent
resistance methods can melt the hearts of repressive regimes or dictators.
Instead, we argue that as with some successful violent movements, nonviolent
campaigns can impose costly sanctions on their opponents, resulting in strategic
gains. We join a long line of scholars concerned with the strategic effectiveness
of different tactical and operational choices (Ackerman and Kruegler 1994;
Sharp 1973; Zunes 1994).
What is perhaps obvious is our voluntaristic approach to the study of
resistance. In this book, we make the case that voluntaristic features of
campaigns, notably those related to the skills of the resistors, are often better
predictors of success than structural determinants. On the surface, this argument
immediately puts us at odds with structural explanations of outcomes such as
political opportunity approaches. Such approaches argue that movements will
succeed and fail based on the opening and closing of opportunities created by the
structure of the political order. As Tarrow has argued, “political opportunity
structures are ‘consistent dimensions of the political environment which either
encourage or discourage people from using collective action’” (Tarrow 1998,
18). Let us briefly discuss how our perspective differs from this approach.
In our study, a political opportunity approach might suggest that nonviolent
campaigns succeed so often because the regime is undergoing a transition,
signaling to the opposition that the time is right to go on the offensive. McAdam
argues that “most contemporary theories of revolution start from much the same
premise, arguing that revolutions owe less to the efforts of insurgents than to the
work of systemic crises which render the existing regime weak and vulnerable to
challenge from virtually any quarter” (1996a, 24).21
What we have found, however, is that the political opportunity approach fails
to explain why some movements succeed in the direst of political circumstances
where chances of success seem grim, whereas other campaigns fail in political
circumstances that might seem more favorable. Such explanatory deficiencies

leave us wondering how the actions of the groups themselves shape the
outcomes of their campaigns.
For instance, a common misperception about nonviolent resistance is that it
can succeed only against liberal, democratic regimes espousing universalistic
values like respect for human rights. Besides the implicit and false assumption
that democracies do not commit mass human rights abuses, the empirical record
does not support this argument. As Kurt Schock writes, the historical record
actually points to the opposite conclusion:
In fact, nonviolent action has been effective in brutally repressive contexts,
and it has been ineffective in open democratic polities. Repression, of
course, constrains the ability of challengers to organize, communicate,
mobilize, and engage in collective action, and magnifies the risk of
participation in collective action. Nevertheless, repression is only one of
many factors that influence the trajectories of campaigns of nonviolent
action, not the sole determinant of their trajectories. (Schock 2003, 706)
The claim that nonviolent resistance could never work against genocidal foes
like Adolph Hitler and Joseph Stalin is the classic straw man put forward to
demonstrate the inherent limitations of this form of struggle. While it is possible
that nonviolent resistance could not be used effectively once genocide has
broken out in full force (or that it is inherently inferior to armed struggle in such
circumstances), this claim is not backed by any strong empirical evidence
(Summy 1994). Collective nonviolent struggle was not used with any strategic
forethought during World War II, nor was it ever contemplated as an overall
strategy for resisting the Nazis. Violent resistance, which some groups attempted
for ending Nazi occupation, was also an abject failure.
However, scholars have found that certain forms of collective nonviolent
resistance were, in fact, occasionally successful in resisting Hitler’s occupation
policies. The case of the Danish population’s resistance to German occupation is
an example of partially effective civil resistance in an extremely difficult
environment (Ackerman and DuVall 2000).22 The famous case of the
Rosenstraße protests, when German women of Aryan descent stood for a week
outside a detention center on the Rosenstraße in Berlin demanding the release of
their Jewish husbands, who were on the verge of being deported to concentration
camps, is a further example of limited gains against a genocidal regime brought
about by civil resistance. The German women, whose numbers increased as the
protests continued and they attracted more attention, were sufficiently disruptive

with their sustained nonviolent protests that the Nazi officials eventually
released their Jewish husbands (Mazower 2008; Semelin 1993; Stoltzfus 1996).
Of course, the civil resistance to Nazi occupation occurred in the context of an
Allied military campaign against the Axis powers, which was ultimately decisive
in defeating Hitler.
Regardless, the notion that nonviolent action can be successful only if the
adversary does not use violent repression is neither theoretically nor historically
substantiated. In fact, we show how, under certain circumstances, regime
violence can backfire and lead to the strengthening of the nonviolent challenge
A competing approach, resource mobilization theory, suggests that campaigns
succeed when resources converge around given preferences, allowing for
mobilization to occur regardless of political opportunities. A resource
mobilization approach would suggest that “the dynamics of a movement depend
in important ways on its resources and organization,” with a focus on
entrepreneurs “whose success is determined by the availability of resources”
(Weinstein 2007, 47). However, this perspective does not account for the ways in
which the actions of the opponent may account for the success or failure of
campaigns when they deploy their own resources to either counter or
outmaneuver the challenge group.
Instead of attempting to fit our explanation within one of the two prevailing
approaches, we instead view our approach as an interactive one that draws on a
contentious politics approach. Such a perspective can be justified by the fact that
the structure of the political environment will necessarily shape and constrain the
perceptions of resistance leaders, whereas the actions of resistance movements
will often have distinguishable and independent effects on the structure of the
system. This approach follows from a number of recent works in social
movement studies and security studies (Arreguín-Toft 2005; Schock 2005;
Weinstein 2007; Wood 2000, 2003).
Civil Resistance Research in Context

Readers familiar with the literature on civil resistance may wonder how our
work differs from the canonical literature in this field. The seminal works on
nonviolent resistance by Gene Sharp, Robert Helvey, Peter Ackerman and
Christopher Kruegler, Ackerman and Jack DuVall, Stephen Zunes, Adam
Roberts and Timothy Garton Ash, Kurt Schock, Mary E. King, and others have
all advanced our understanding of strategic nonviolent action in important ways.
Sharp’s three-volume opus, The Politics of Nonviolent Action, established the
theoretical foundation for nonviolent action. It reads as a handbook of nonviolent

resistance, explaining the theory of power and the different methods of
nonviolent action and the ways that nonviolent action can affect the adversary
(conversion, persuasion, accommodation, and coercion). Sharp’s work is
seminal; it provides a unified theory on the strategic mechanisms through which
civil resistance can work.
Robert Helvey builds on much of Sharp’s original foundation in his work on
how to act strategically during the prosecution of a nonviolent conflict (2004).
He identifies similarities between civil resistance and military strategy, providing
a handbook of sorts for how to identify campaign goals, develop strategic plans,
and operational problems movements face during a campaign.
Our book is distinct in several ways. First, although Sharp’s and Helvey’s
volumes provide a theoretical gold mine, they do not attempt to test their
assertions empirically. Our book is the first attempt to comprehensively test
many of the ideas Sharp and Helvey have developed. Second, Sharp’s and
Helvey’s comparisons with violent resistance are implicit; they simply present
nonviolent resistance as an effective strategy in asymmetrical conflict. In our
study, we explicitly compare nonviolent and violent resistance to test the
hypothesis that nonviolent resistance is indeed a more effective strategy.
In Ackerman and Kruegler’s Strategic Nonviolent Conflict, the authors
develop a framework informed by strategic theory for analyzing the outcomes of
nonviolent resistance campaigns. The book features multiple case studies of
successful and failed nonviolent action, from which the authors generalize
twelve principles of successful nonviolent action. Although the book is highly
analytical, the case studies are inductive in nature: their purpose is to find
patterns about why nonviolent campaigns succeed rather than to test hypotheses.
Ackerman and DuVall’s book A Force More Powerful has been perhaps the
most widely read book on nonviolent action. The book is empirical, featuring
descriptive accounts of nonviolent campaigns ranging from Russia to South
Africa. One of the most accessible books on nonviolent conflict, it was adapted
into an Emmy-nominated documentary series. Recently the authors have
sponsored the development of a video game named after the book, the purpose
of which is to train scholars and activists in the tactics and strategy of nonviolent
resistance. The book is not intended to be an analytical exploration of why
nonviolent resistance succeeds compared with violent resistance, nor does it
attempt to control for other factors that might predict the success or failure of
movements. Our study expands the universe of cases, explicitly compares
nonviolent and violent resistance, tests theoretical hypotheses concerning the
mechanisms that lead to success, and controls for other factors that might
account for different outcomes. We do, however, focus far less on the dynamics

of violent unconventional warfare, such as guerrilla warfare and violent
Stephen Zunes, Adam Roberts and Timothy Garton Ash, and Kurt Schock
have all contributed to the academic understanding of the conditions under
which nonviolent resistance succeeds and fails. Their works share a comparative
case study approach to explaining individual cases or illuminating patterns in
nonviolent resistance activity (Roberts and Garton Ash 2009; Schock 2005;
Zunes, Kurtz, and Asher 1999). Much of our argument is compatible with
findings in Zunes’s various works, although our aim is to explain broad patterns
rather than individual cases. Roberts and Garton Ash similarly attempt to explain
the dynamics of nonviolent resistance in a diverse range of cases. Other authors
have examined single case studies and associated phenomena in great depth
(Bleiker 1993; Clark 2000; Dajani 1994; Eglitis 1993; Huxley 1990; Martin
2007; McCarthy and Sharp 1997; Miniotaite 2002; Parkman 1988, 1990;
Roberts and Garton Ash 2009; Sharp 2005; Stephan 2010; Stoltzfus 1996). The
goal of these contributors, however, is not always to explain campaign success or
failure but rather to explore a number of social movement problems and
questions related to their cases. Thus their works demonstrate some important
lessons but not necessarily about why and when civil resistance works.
In Unarmed Insurrections, sociologist Kurt Schock compares successful and
failed nonviolent, prodemocracy campaigns against nondemocratic regimes.
This work comes much closer to the analytical purposes of our book. Schock
compares six nonviolent campaigns in nondemocracies to identify patterns
among the trajectories of these campaigns. He challenges the political
opportunity approach, and argues that strategic factors can help explain the
outcomes of the campaigns. Most important, Schock’s work bridges the
structure-agency divide and analyzes the iterative, interactive nature of political
opportunities and strategic choice. Specifically, Schock argues that tactical
innovation, resilience, and the shifting between methods of concentration and
methods of dispersion can help to explain the divergent outcomes of different
Vincent Boudreau also analyzes the outcomes of prodemocracy movements in
Southeast Asia, using a compelling contentious politics model (2004). However,
he does not focus on the relative effectiveness of nonviolent and violent action,
instead exploring the interaction between different modes of repression
employed by dictators in Burma, the Philippines, and Indonesia and the impact
of these forms of repression on the protestors. He is explicitly critical of the
possibility of accurately representing these conflicts using quantitative analysis,
instead arguing in favor of viewing each conflict as a complex system of its own

(2004, 3).
Our findings are highly compatible with Schock’s and share much in common
with Boudreau’s as well, notwithstanding methodological differences. But our
argument about the primacy of participation in nonviolent resistance appears
unique in this literature. Moreover, as with the Ackerman and Kruegler book,
our study expands the universe of cases to include antioccupation and secession
campaigns in addition to regime-change campaigns. Our study is not limited to
Southeast Asia, nor are our cases restricted to nondemocratic targets. Instead, we
attempt to comprehensively examine major nonviolent and violent campaigns all
across the globe, against all types of targets, from 1900 to 2006.
Readers familiar with Ivan Arreguín-Toft’s argument in How the Weak Win
Wars may see some similarities to our argument. In his book, Arreguín-Toft
argues that weak powers sometimes win wars when they employ indirect
strategies against stronger powers. That is, if the stronger power is employing
conventional war strategies, a weaker power that uses unconventional or
guerrilla war will be likely to succeed. For instance, the British conventional
army succumbed to the guerrilla war waged by American colonists during the
Revolutionary War (though, as mentioned earlier, the armed insurgency followed
years of nonviolent civil resistance). On the other hand, a weaker power that
uses conventional strategies against a stronger power relying on conventional
strategies will fail. The 1991 Gulf War demonstrates that point: the militarily
inferior Iraqi army was unable to successfully take on Coalition forces.
Conversely, if a stronger power employs unconventional strategies against a
weaker power’s conventional strategies, the weaker power will win. For
instance, Hitler’s air bombing of British civilian targets did not force the British
into compliance. Instead, the attacks emboldened the British against the
Germans (Arreguín-Toft 2001, 108). But when a stronger power employs
unconventional strategies against a weak power also using unconventional
strategies, the stronger power will win. The Russian government has used
“barbaric” strategies against Chechen rebels, effectively crushing the Chechen
While we do not dispute Arreguín-Toft’s findings, we illuminate a new
dimension in his typology, which is the use of strategic nonviolent action as an
indirect strategy against a militarily superior opponent. When Arreguín-Toft
describes indirect strategies for weaker powers, he refers to two types of
strategies: direct defense, which he defines as “the use of armed forces to thwart
an adversary’s attempt to capture or destroy values such as territory, population,
and strategic resources,” and guerrilla warfare, defined as “the organization of a
portion of society for the purpose of imposing costs on an adversary using armed

forces trained to avoid direct confrontation” (2001, 103). We argue that unarmed,
civil resistance can be even more effective than direct defense or guerrilla
warfare, both of which are armed strategies against militarily superior
Our results are also consistent with Max Abrahms’s findings, which suggest
that terrorist activities that target civilians are less effective than guerrilla
warfare strategies that target policy and military personnel (2006). But our
findings extend his thesis further, in that we argue that in most cases all types of
violent campaigns are likely to be less effective than well-managed nonviolent
What all these works, including ours, have in common is a call for scholars to
rethink power and its sources in any given society or polity. Although it is often
operationalized as a state’s military and economic capacity, our findings
demonstrate that power actually depends on the consent of the civilian
population, consent that can be withdrawn and reassigned to more legitimate or
more compelling parties.
Squaring the Circle: The Effectiveness of Violence?

Some scholars, such as Robert Pape, have developed recently theses on the
efficacy of violent conflict. In particular, some argue that terrorism—especially
suicide terrorism—is an effective coercive strategy, especially against
democracies (2003, 2005). Jason Lyall and Isaiah Wilson have also discovered
that violent insurgency is growing in effectiveness—against democracies in
particular (2009). Given these authors’ findings, there are some surface
discrepancies with our findings. We address each of these arguments in turn.
First, Pape argues that suicide terrorism is an effective punishment strategy
against democracies (2003, 2005). Suicide bombers convey both capability and
resolve to soft targets in democracies, demonstrating to these countries that
continued occupation will result in protracted, escalating, indiscriminate war
against the country’s civilian population. Such acts lead to a decline in morale in
the democracy, which ultimately judges that withdrawal from the occupied
territory is less costly than the occupation. In his study, five out of the eleven
suicide bombing campaigns since 1980 have achieved at least partial success.
Pape’s argument and empirics have been widely criticized (see, for instance,
Ashworth et al. 2008). Yet if we take his argument at face value, we can offer yet
another criticism, which could be applied to almost all scholars whose research
tests the efficacy of different violent methods. Such scholars often assume or
argue that violence is effective, but compared with what? In particular, Pape
makes no attempt to compare the relative efficacy of suicide terrorism against

alternative strategies. Even in some of his most prominent cases—Lebanon and
the Palestinian Territories—we have seen mass, nonviolent resistance perform
effectively where violent insurgencies have failed. In the Lebanese case, the
2005 Cedar Revolution involved more than a million Lebanese demonstrators
forcing Syria to withdraw its forces from Lebanese soil. And, as shown in
chapter 5, the First Intifada moved the Palestinian self-determination movement
further than the Palestine Liberation Organization’s violent campaign that
preceded it, or the Al-Aqsa Intifada that succeeded it.
In another example, Lyall and Wilson argue that violent insurgencies are
becoming more effective against highly mechanized militaries, which prove to
be unwieldy in urban settings against well-camouflaged insurgents (2009). They
show that since 1975 states have succeeded in crushing insurgencies only 24
percent of the time. In their study, they determine success from the state’s
perspective, such that complete defeat of the insurgents is considered a success,
whereas a draw or a loss to insurgents is considered a failure. When one looks
more closely, however, one can see that their primary finding—that violent
insurgencies have succeeded in over 75 percent of cases since 1976—is based on
data in which nearly 48 percent of the cases were stalemates. Thus only 29.5
percent of their insurgencies since 1976 actually succeeded in defeating their
state adversaries, a statistic that is much closer to our own. Lyall and Wilson also
exclude ongoing campaigns from their findings, whereas we code such cases as
failures through 2006.23
The difference in measurement is one way that our findings diverge from
Lyall and Wilson’s. But perhaps the most important difference is that they do not
compare the relative effectiveness of violent insurgency with nonviolent
campaigns. If we analyze the success rates of nonviolent campaigns since 1976,
we see a much higher rate of nonviolent campaign success (57 percent).
Thus our study represents a departure from techniques used by those arguing
that violent insurgency is effective. As Baldwin argues, “Only comparative
analysis of the prospective success of alternative instruments provides policyrelevant knowledge” (2000, 176). Our approach involves the relative comparison
of nonviolent and violent campaigns, which sheds more light on how
unsuccessful violent campaigns really are.24

Beyond scholarly contributions, this research possesses a number of important
implications for public policy. Research regarding the successes and failures of
nonviolent campaigns can provide insight into the most effective ways for
external actors—governmental and nongovernmental—to aid such movements.

From the perspective of an outside state, providing support to nonviolent
campaigns can sometimes aid the movements but also introduces a new set of
dilemmas, including the free-rider problem and the potential loss of local
legitimacy. This study strongly supports the view that sanctions and state support
for nonviolent campaigns work best when they are coordinated with the support
of local opposition groups; but they are never substitutes.
For instance, although there is no evidence that external actors can
successfully initiate or sustain mass nonviolent mobilization, targeted forms of
external support have been useful in some cases, like the international boycotts
targeting the apartheid regime in South Africa. The existence of organized
solidarity groups that maintained steady pressure on governments allied with the
target regimes proved to be very helpful, suggesting that “extending the
battlefield” is sometimes necessary for opposition groups to enhance their
leverage over the target. Lending diplomatic support to human rights activists,
independent civil society groups, and democratic opposition leaders while
penalizing regimes (or threatening penalties) that target unarmed activists with
violent repression may be another way that governments can improve the
probability of nonviolent campaign success. Coordinated multinational efforts
that used a combination of positive and negative sanctions to isolate egregious
rights violators supported successful civil resistance movements in South Africa
and Eastern Europe.

The remainder of the study examines the specific mechanisms by which
nonviolent campaigns succeed and fail. It does so by interchanging quantitative
and qualitative analyses of nonviolent and violent campaigns in the Middle East
(Iran and the Palestinian Territories) and Southeast Asia (the Philippines and
Burma). Each of the four cases features periods of both violent and civil
resistance against repressive regimes, but with varying degrees of success. This
allows us to more closely examine the conditions under which nonviolent and
violent campaigns succeed and fail, both within and across the cases.
The book proceeds as follows. First, in chapter 2, we introduce the general
argument of the study and explore how this argument converges and diverges
with the findings of other scholars. We argue that civil resistance campaigns are
more successful than violent campaigns at overcoming barriers to participation,
an important prerequisite of success.
In chapter 3, we explore the major alternative arguments—that regime
features may independently affect the outcomes of the nonviolent or violent
conflicts, or that the origins and outcomes of resistance campaigns are

endogenous. First, we test whether opponent regime type (i.e., democracy or
nondemocracy), capabilities, or use of violent repression against the challenge
group reduces the likelihood of success for nonviolent resistance. We also test
the effects of time, region, and campaign goal on the probability of success. We
find that even when taking into account structural features, nonviolent resistance
is still a more effective strategy than violent resistance.
Chapter 3 also addresses the issue of endogeneity head-on, that is, whether
violent campaigns fail because they emerge in conditions in which failure is
extremely likely, thus explaining their poor success rates relative to nonviolent
campaigns. We find that nonviolent and violent insurgencies are likely to emerge
in very similar circumstances, such that their outcomes cannot be explained
exclusively on the basis of endogeneity.
In part 2, we compare nonviolent and violent resistance campaigns and their
outcomes in Iran, the Palestinian Territories, the Philippines, and Burma. We
explain the case selection in detail before the substantive chapters begin. Chapter
4 discusses the Iranian Revolution (1977–1979). In this case, violent campaigns
failed to dislodge the Shah, whereas the nonviolent campaign succeeded.
Chapter 5 explains why violent Palestinian campaigns orchestrated by an exiled
leadership achieved little or no success before the First Intifada (1987–1992),
whereas the mass popular uprising that originated inside the occupied territories
achieved partial success through some important Israeli concessions.
Chapter 6 deals with the successful case of the People Power movement in the
Philippines (1983–1986), which ousted Ferdinand Marcos from power. This
mass uprising achieved what the Maoist and Muslim-led insurgencies in that
country had been unable to achieve. Chapter 7 identifies a case of failed
nonviolent resistance: the Burmese prodemocracy uprising of 1988. Both
nonviolent and violent campaigns failed in this case, which provides a useful
deviating outcome for comparison.
Part 3 explores the implications of this research across multiple dimensions.
First, in chapter 8, we discuss the consequences of violent insurgency,
particularly violent insurgent success. Our statistical evidence suggests that
countries in which violent insurgencies exist are more likely to backslide into
authoritarianism or civil war than countries where nonviolent campaigns exist,
which often become more stable, democratic regimes.
Finally, the concluding chapter summarizes the key findings, highlighting how
these findings make a contribution to the literature. This chapter also argues for
the incorporation of nonviolent conflict into security studies inquiry and suggests
ways to improve and expand upon our study. The last section identifies the
policy implications derived from this research.

Although not the final word in any sense, we hope that this book challenges
the conventional wisdom concerning the effectiveness of nonviolent struggle and
encourages scholars and policy makers to take seriously the role that civilians
play in actively prosecuting conflict without resorting to violence.

What is a rebel? A man who says no.

of nonviolent resistance
campaigns relative to violent campaigns? We argue that a critical source of the
success of nonviolent resistance is mass participation, which can erode or
remove a regime’s main sources of power when the participants represent
diverse sectors of society. All resistance campaigns—violent and nonviolent—
seek to build the personnel bases of their campaigns. Personnel are recruited for
their special skills, knowledge, material resources, and their willingness to fight
and support the resistance. The quantity and quality of campaign participation is
a critical factor in determining the outcome of resistance struggles (DeNardo
1985; Lichbach 1994; Weinstein 2007; Wickham-Crowley 1992).
This chapter has two aims. First, we establish that nonviolent campaigns are
more likely to attract higher levels of participation than violent campaigns
because the barriers to participation are lower. Second, we argue that high levels
of participation in resistance campaigns can activate numerous mechanisms that
improve the odds of success. Such mobilization is not always manifested in the
form of mass rallies and street demonstrations but rather can manifest in
numerous forms of social, political, and economic noncooperation. The tactical
and strategic advantages of high levels of diverse participation explain—in large
part—the historical success of nonviolent campaigns.


We define participation in a resistance campaign as the active and observable
engagement of individuals in collective action. As such, when measuring
campaign participation, we use estimated counts of observed individuals.1
Instead of constructing cumulative counts, which would be nearly impossible,
we count the maximum number of estimated participants that participated in
peak events in the campaign. For example, if a resistance campaign holds mass
protests in, say, September with 12,000 people, November with 24,000 people,

and December with 20,000 people, we use the November figure for our estimate.
That is, we code that particular campaign as having 24,000 participants. We use
estimates of armed participants to generate figures about the level of
participation in violent insurgencies.2 Of the 323 resistance campaigns analyzed
in this book, we were able to collect reliable membership data for 259 campaigns
—80 nonviolent and 179 violent—by referencing multiple sources that estimated
the maximum number of participants in each campaign.3
This is a rather strict conceptualization of participation, and we recognize that
many forms of participation are impossible to observe, such as providing
sanctuary, food, and supplies to guerrillas, raising funds, communicating
messages, acting as informants, or refusing to cooperate with government
attempts to apprehend insurgents. For instance, for some individuals, simply
refusing to report the presence of guerrillas in one’s village to state police may
be a form of participation in a resistance campaign, albeit one that is more
passive and impossible for us to quantify. Recent studies have identified multiple
and complex levels of such participation. As Roger Peterson writes, “there are
collaborators, neutrals, locally based rebels, mobile fighters, and gradations in
between” (2001, 8).
We do not dispute that our definition likely misses many unobserved
participants, but we find the definition both necessary and justified for two
reasons. First, in our definition of nonviolent resistance participation, civilians
are the active and primary prosecutors of the conflict, executing nonviolent
methods against the adversary with varying degrees of risk. This is quite
different from the typical conception of civilians as serving a supportive role to
Second, we assume that some types of unobservable participation occur in
approximately equal measure in both nonviolent and violent resistance
campaigns. Out of necessity, we focus exclusively on the participants that make
themselves visible to observers and opponents as a rough measure of campaign
mobilization. The risks of visibility should be similar for both nonviolent and
violent resistance campaigns, which in our study often involve illegal and at
times high-risk actions against powerful and repressive adversaries.
We do wish to avoid the misconception, however, that civil resistance always
assumes the form of mass protests in the streets. Nonviolent resistance is just as
likely to take the form of stay-aways, sit-ins, occupations, economic boycotts,
and so forth, in which the numbers of participants are extremely difficult to
estimate. When such estimations are possible because of reliable recording of
such events, we include them in our figures.


Mass mobilization occurs for many different reasons, which multiple scholars
have analyzed in great depth (see, for instance, Kalyvas 2006; Peterson 2001). In
this chapter, we do not seek to explain why mobilization occurs. Rather, we
argue that once mobilization begins, a nonviolent resistance campaign has wider
appeal than a violent one, thereby enlarging the personnel base of the former and
bringing more assets and resources to the fight against a state opponent.
Skeptics may disagree. It is often argued, for instance, that violent
insurgencies provide immediate results—such as loot, prestige, score settling, or
territorial gains—that give them more appeal than nonviolent resistance. Beyond
the prospect of achieving political objectives, the potential to obtain material
payoffs from resistance leaders, to seize territory and weapons, to gain control
over lucrative extractive industries, trade, and trafficking routes, to inflict
casualties, or to exact revenge are factors that may attract some recruits to
violent resistance.
The psychosocial dimensions of participation in armed conflict have similarly
attracted a great deal of attention. Frantz Fanon famously advocated armed
resistance on the grounds that it bestows feelings of communal solidarity
through actively fighting against injustice while being willing to die for a cause
greater than self (Boserup and Mack 1974; Fanon 1961).4 Violence may have its
own attraction, especially for young people, for whom the allure may be further
perpetuated by cultural references and religious defenses of martyrdom
(Breckenridge 1998).5
Despite its supposed appeal, however, the resort to violence is rare at both
individual and group levels and therefore may not have the allure that some
theorists ascribe to it (Collins 2008, 20). On the whole, physical, informational,
commitment, and moral considerations tend to give nonviolent campaigns an
advantage when it comes to mobilizing participants, which reinforces the
strategic benefits to participation.
We have found strong evidence suggesting that nonviolent campaigns have
been, on average, more likely to have a larger number of participants than
violent campaigns. The average nonviolent campaign has over 200,000 members
—about 150,000 more active participants than the average violent campaign. A
look at the twenty-five largest campaigns yields several immediate impressions.
First, twenty of the largest campaigns have been nonviolent, whereas five have
been violent. Second, of the nonviolent campaigns, fourteen have been outright
successes (70 percent), whereas among the five violent campaigns, only two
have been successful (40 percent). In other words, among these massive
campaigns, nonviolent campaigns have been much more likely to succeed than

violent campaigns.6

The Iranian Revolution of 1977–1979 is illustrative. Although violent
insurgencies such as those of the fedayeen and mujahideen had resisted the Shah
since the 1960s, they were able to attract only several thousand followers.
Pahlavi’s regime crushed the armed groups before they produced meaningful
change in the regime. The nonviolent revolution that emerged between 1977 and
1978, however, attracted several million participants and included nationwide
protests and boycotts involving all sectors of society that paralyzed the economy
and eroded the Shah’s most important pillars of support.
These trends are further borne out in the data set. Nonviolent campaigns are
persistently associated with higher levels of membership, even when controlling
for the population size of the entire country. Consider table 2.2, which shows the
effects of a nonviolent resistance type on the number of participants, controlling
for population size.7 Thus nonviolent resistance campaigns have been associated
with higher levels of participation. In this section, we argue that the physical,
informational, and moral barriers to participation are lower in nonviolent
campaigns than in violent campaigns.
Physical Barriers

Active participation in a resistance campaign requires variable levels of physical
ability. The physical risks and costs of participation in a violent resistance
campaign may be prohibitively high for many potential members. Actively
joining a violent campaign may require physical skills such as agility and
endurance, willingness to train, ability to handle and use weapons, and often
isolation from society at large. While certain of these qualities, including
endurance, willingness to sacrifice, and training are also applicable to
participation in nonviolent resistance, the typical guerrilla regimen may appeal
only to a small portion of any given population.

Resistance is Primarily Nonviolent
Population, logged
Prob > F

Number of Participants, logged
2.26*** (.29)
.23* (.13)
6.70*** (1.17)




Physical barriers to participation may be lower for nonviolent resistance since
the menu of tactics and activities available to nonviolent activists is broad and
includes a wide spectrum, ranging from high-risk confrontational tactics to lowrisk discreet tactics.8 Generally, participation in labor strikes, consumer boycotts,
lockdowns, and sit-ins does not require strength, agility, or youth. Participation
in a nonviolent campaign is open to female and elderly populations, whereas
participation in a violent resistance campaign is often, though not always,
physically prohibitive. Although female operatives—such as female suicide
bombers and guerrillas—have sometimes been active in violent campaigns in Sri
Lanka, Iraq, Pakistan, Palestine, El Salvador, and East Timor, they are
nevertheless exceptions in most cases.
Informational Difficulties

Scholars have found that individuals are more likely to engage in protest activity
when they expect large numbers of people to participate (Goldstone 1994;
Granovetter 1978; Kuran 1989; Kurzman 1996, 2004; Lichbach 1994; Marwell
and Oliver 1993; Oberschall 1994; Olson 1965; Rasler 1996; Schelling 1978;
Tullock 1971). To successfully recruit members, campaigns must publicize their
activities to demonstrate their goals, abilities, and existing numbers to potential
recruits. Because of the high risks associated with violent activity, however,
movement activists may be limited in how much information they can provide.
They may need to remain underground, thereby exacerbating informational
problems. Although violent acts, including assassinations, ambushes, bombings,
and kidnappings, are public and often attract significant media attention
providing signals of the campaign’s abilities, the majority of the campaign’s
operational realities—including information about the numbers of active
members—often remain unseen and unknown.9 The absence of visible signs of
opposition strength is, therefore, problematic from the perspective of
recruitment. Thus violent resistance may be at a disadvantage in this regard,
since the actual number of activists may not be explicit. The counterargument, of
course, is that dramatic acts of violence achieve a bigger bang for the buck.
Whereas nonviolent organization requires communication and coordination
involving larger numbers of people, a single suicide bomber can wreak great

damage while attracting significant media attention at relatively little cost.
Violent campaigns often rely on propaganda materials that try to exaggerate their
size and strength to attract recruits. In the propaganda realm, violent campaigns
may have a tactical advantage over many nonviolent campaigns.
On the other hand, nonviolent, public tactics have important demonstration
effects, which help address the informational problem. Nonviolent campaigns
sometimes include clandestine activities (e.g., the use of samizdat underground
publications during the Polish Solidarity struggle, or the actual planning of
nonviolent campaigns by the leadership), particularly during the early stages
when the resistance is most vulnerable to regime repression and decapitation.
Typically, however, nonviolent campaigns rely less on underground activities
than do armed struggles.10 When communities observe open, mass support and
collective acts of defiance, their perceptions of risk may decline, reducing
constraints on participation. This contention is supported by critical-mass
theories of collective action, which contend that protestors base their perceptions
of protest opportunities on existing patterns of opposition activity (Kurzman
1996, 154). Courage breeds courage, particularly when those engaged in protest
activities are ordinary people who would be conformist, law-abiding citizens
under typical circumstances. Media coverage amplifies the demonstration effects
of their acts of defiance.
Another factor that enhances participation in nonviolent campaigns is the
festival-like atmosphere that often accompanies nonviolent rallies and
demonstrations—as exemplified by the recent nonviolent campaigns in Serbia,
Ukraine, Lebanon, and Egypt—where concerts, singing, and street theater
attracted large numbers of people (particularly young people) interested in
having fun while fighting for a political cause. Humor and satire, which have
featured prominently in nonviolent campaigns (less so in armed campaigns),
have helped break down barriers of fear and promote solidarity among victims of
state-sponsored oppression (Kishtainy 2010).
Moral Barriers

Moral barriers may constrain potential recruits to resistance campaigns, but such
constraints may inhibit participation in nonviolent resistance far less than
participation in violent activities. Although an individual’s decision to resist the
status quo may follow a certain amount of moral introspection, taking up
weapons and killing adds a new moral dimension. Unwillingness to commit
violent acts or to support armed groups necessarily disqualifies segments of the
population that sympathize with the resistance but are reluctant to translate that
sympathy into violence.11 For violent resistance campaigns, the leadership may

need to rely on the proportion of the population that is willing to use violence
against the adversary and its supporters, while settling for sympathy and passive
support from the rest of the population.
Nonviolent resistance campaigns, however, can potentially mobilize the entire
aggrieved population without the need to face moral barriers. Although the
moral quandaries associated with nonviolent resistance might involve putting at
risk one’s freedom, family well-being, life and livelihood, joining such a
campaign “requires less soul-searching than joining a violent one. Violent
methods raise troublesome questions about whether the ends justify the means,
and generally force the people who use them to take substantial risks” (DeNardo
1985, 58).
Commitment Problems

Beyond physical, informational, and moral barriers, nonviolent resistance
campaigns may offer an opportunity to participate to people with varying levels
of commitment and risk tolerance. Campaigns that rely primarily on violent
resistance must depend on participants who have high levels of both
commitment and risk tolerance for four principal reasons.
First, the new recruit to a violent campaign may require more training than a
recruit to a nonviolent campaign, creating a lag between volunteering and
participation. This lag—and the strenuous requirements for participation in a
violent campaign—may reduce the number of people who join a violent
campaign on a whim.12
Second, violent campaigns typically enforce higher levels of commitment at
the outset. Screening potential participants is much more intense in violent
movements. Often new recruits to violent movements must undertake a violent
act to demonstrate their commitment. This is a further inhibition to participation
in armed struggles, because potential recruits may wish to eschew drastic
screening processes or movement leaders may find it hard to trust new recruits.
Third, during the prosecution of a conflict, participants in nonviolent
campaigns can often return to their jobs, daily lives, and families with lower risk
than a participant in a violent campaign.13 Compared with those in armed
struggle, participants in civil resistance can more easily retain anonymity, which
means that they can often commit acts of resistance without making major life
sacrifices. This is particularly true when a campaign uses nonviolent methods of
dispersion (a concept we elaborate on later), such as stay-at-home strikes or a
consumer boycott, in which cooperation is withdrawn without providing the
state with a tangible target for repression (Burrowes 1996, 224–25; Schock 2005,
52). The commitment required by people who join violent campaigns often

prevents them from resuming their lives during or after the conflict, and they are
more likely to go underground to evade state security.
Fourth, nonviolent resistance offers a greater repertoire of lower-risk actions.
Although nonviolent struggle is rarely casualty-free, as the nonviolent struggle
in Egypt recently demonstrated, the price of participating (and being caught) in
armed struggle is often death. The possibility of accidental death during training
exercises or through friendly fire is omnipresent as well. Thus the likelihood of
being killed while carrying out one’s duties as an armed insurgent is high,
whereas many lower-risk tactics are available to participants in a nonviolent
resistance campaign. The wearing of opposition insignia, the coordinated
banging of pots and pans and honking of horns, the creation of underground
schools, participation in candlelight vigils, and the refusal to obey regime orders
are a few examples of less-risky nonviolent tactics that have been used by
groups around the world (Sharp 1973).
Mobilization during the Iranian Revolution demonstrates the latter point.
Notwithstanding the Shah’s deep unpopularity among large numbers of Iranians,
many Iranian citizens were unwilling to participate in protest activity until the
revolution had attracted mass support, which occurred only after nonviolent
popular struggle replaced guerrilla violence as the primary mode of resistance
(Kurzman 1996). A similar dynamic could be seen in the 1988 popular ouster of
General Augusto Pinochet in Chile, and the 1986 People Power revolution
against Ferdinand Marcos in the Philippines, where armed challenges to the
dictatorships invited harsh regime reprisals without attracting mass support or
threatening the regime’s grip on power, whereas nonviolent actions opened up
space for broad-based, multisectoral participation (Ackerman and DuVall 2000;
Boudreau 2004; Schock 2005).
The dynamics of participation discussed thus far point in one direction. They
suggest that nonviolent campaigns will be more successful at generating large
bases of participants. When large numbers of people in key sectors of society
stop obeying and engage in prolonged acts of social, political, and economic
disruption, they may fundamentally alter the relationship between ruler and
ruled. If mass participation is associated with campaign success, then nonviolent
campaigns have an advantage over violent ones.

We have established how and why nonviolent resistance campaigns are able to
attract a larger number of active participants than violent struggles. But is mass
participation truly important? After all, many regimes specialize in controlling
large populations. Some might suspect that a smaller number of well-armed

comrades competing against an unsuspecting military and government could
have better odds than a million unarmed protestors engaging a repressive
opponent (see, e.g., DeNardo 1985). This expectation is certainly corroborated
by several empirical examples: the Cuban Revolution shows the success of
small, armed bands, whereas the massacre at Tiananmen Square demonstrates
the failure of a large-scale nonviolent campaign.
The data, however, reveal a different pattern. Over space and time, large
campaigns are much more likely to succeed than small campaigns. A single unit
increase of active participants makes a campaign over 10 percent more likely to
achieve its ultimate outcome.14 Consider figure 2.1, which shows the effects of
number of participants per capita on the predicted probability of campaign
success. The trend is clear that as membership increases, the probability of
success also increases.15
We recognize, however, that numbers alone do not guarantee victory in
resistance campaigns. As some cases demonstrate, a high number of participants
does not automatically translate into success. Some enormous campaigns—like
the anticommunist campaigns in East Germany in the 1950s (boasting about four
hundred thousand participants) and the anti-Japanese insurgency in China during
the 1930s and 1940s (with over 4 million participants)—failed utterly.
Thus, numbers may matter, but they are insufficient to guarantee success. This
is because the quality of participation—including the diversity of the resistance
participants, strategic and tactical choices made by the opposition, and its ability
to adapt and innovate—may be as important as the quantity of participants. As
proposed in the preceding, lower barriers to participation enjoyed by nonviolent
campaigns will increase not only the size of the campaign but also the diversity
of the campaign. The more diverse the participation in the resistance—in terms
of gender, age, religion, ethnicity, ideology, profession, and socioeconomic
status—the more difficult it is for the adversary to isolate the participants and
adopt a repressive strategy short of maximal and indiscriminate repression. Of
course, this does not mean that nonviolent campaigns are immune from regime
repression—typically they are not—but it does make the opponent’s use of
violence more likely to backfire, a point we return to later.
Moreover, thick social networks among members of the resistance and regime
actors, including members of the security forces, may produce bonds that can
become very important over the course of the resistance. Diverse participation
also increases the likelihood of tactical diversity, since different groups and
associations are familiar with different forms of resistance and bring unique
skills and capacities to the fight, which makes outmaneuvering the opponent and
increasing pressure points more plausible.

As with any campaign, strategic factors like achieving unity around shared
goals and methods, establishing realistic goals, assessing opponent
vulnerabilities and sources of leverage, sequencing tactics, and navigating
structural constraints (including regime repression) are also likely to be crucial
determinants of campaign outcomes. These strategic factors are independent of
the mechanisms we develop in the following but can affect whether the
mechanisms actually translate into effectiveness. We emphasize these features
more prominently in our case studies. In the meantime, however, we suggest that
the execution of any resistance strategy—violent or nonviolent—and the ability
to stay in the contest with the adversary depend on the availability of willing

As such, large-scale and diverse participation may afford a resistance
campaign a strategic advantage, which, in turn, increases the pressure points and

enhances the leverage that the resistance achieves vis-à-vis its state adversary.
The ability of nonviolent campaigns to more easily exploit these advantages of
broad-based mobilization, and the high costs of prolonged disobedience and
noncooperation by large numbers of dissenters, explain in part why civil
resistance has been