Abigail A. Fuller

What Works? Research on the Outcomes of Nonviolent Action

Abigail A. Fuller

Associate Professor of Sociology

Manchester College

June 5, 2005

Abstract: While there is much written on the philosophy and practice of nonviolence, there has been little systematic analysis of the effectiveness of nonviolent action for social movements. This is problematic, as it inhibits nonviolent activists from learning from their collective experiences. However, an increasing number of studies by social scientists have investigated the effectiveness of different social movement strategies and tactics, most of which are nonviolent, though they are not explicitly identified as such in the literature. This article summarizes the available evidence for what works (the conditions under which nonviolent social movements are or are not effective) from the scholarly literature in the social sciences. It concludes with suggestions for future research that can aid activists in nonviolent social movement in maximizing their chances for success.


Despite the diffusion of nonviolence as a conscious strategy through movements around the world in recent decades, little is understood about how or why nonviolence works as a technique for securing social change. (Zunes 2000)

As a student of peace studies in the 1980s, I discovered nonviolence, and it completely changed how I viewed the world: here was a method of waging conflict that could be effective and was consistent with the moral values of justice and compassion. Studying sociology in graduate school, I learned something else: the power of using social scientific methods to understand how the world really works, so that we can be more effective at improving it. And yet, I found that there is little social scientific research on nonviolent action. Most of the literature on nonviolence is either philosophical or persuasive or, if it is empirical, it is descriptive rather than explanatory--case studies of nonviolent action, or A how to@ books that represent the accumulated wisdom of nonviolent activists but do not systematically evaluate various strategies and tactics. Without systematic, empirical research, we do not know for sure under what conditions various nonviolent strategies and tactics are effective. In the absence of such research, activists risk repeating the use of less effective techniques or shunning more effective ones, based on tradition or ideology. For example, since the 1970s mass civil disobedience in the United States has become very scripted: protest organizers inform local police exactly what law they plan to break, how, and when; at the appointed time and place, they do so (sit down/form a blockade/cross a line); and as long as they remain polite, generally the police are polite in return as they conduct arrests. But what effect does this have, if any? Does it impress decision makers? Does it strengthen the commitment of participants? In another example, we know very little about the effects of fasting as a nonviolent tactic. When, how, and why does it work to strengthen nonviolent movements? To soften the hearts of opponents?

Let me note that what we might call peace and justice activist culture in the U.S. contains a wealth of useful information from seasoned activists, which I do not wish to denigrate. This information has been useful to countless groups in planning and carrying out nonviolent actions. Nevertheless, social scientists know that personal experience and A common sense@ are not always sources of accurate information about the social world. Sociology is full of counter-intuitive research findings (that random drug testing in high schools does not reduce the rate of drug use by students; that women on welfare do not bear more children in order to increase their benefits; that changes in behavior tend to produce changes in attitudes, not the other way around).

What We Know: Current Research on Nonviolent Action

Most research on the effects of nonviolent action has been conducted by sociologists and political scientists who study social movements. Virtually all social movements use nonviolent tactics, while some use violent tactics as well. In addition, the vast majority of nonviolent actions are collective undertakings, by people involved in social movements. Hence the study of social movements provides the richest source of empirical research findings on nonviolent action.

Social movement scholars have looked at two sets of factors that might influence the effectiveness of social movements: internal or movement-controlled factors and external factors. There is conflicting evidence on the relative importance of each set of factors, but both have been found influential.

Internal Factors

1. Factionalism decreases chances for success.

Gamson= s The Strategy of Social Protest (1975) is most cited work in social science on the outcomes of social movements. Gamson studied fifty-three social movement organizations, randomly selected from hundreds that existed in the United States from 1800 to 1945. He found that 49 percent of the groups were successful, defined as achieving most or all of the new advantages that the group sought. One clear finding was that when movements experience internal schisms, they are less likely to succeed. (This was confirmed by later analyses of his data (see Frey et al. 1992).) Most likely this is because factions must compete for resources and end up spending more time fighting each other than opponents (Frey et al.1992).

2. Diverse and accountable leadership teams optimize the chances for success.

A movement headed by a A leadership team@ (Ganz 2000), rather than a single leader, is likely to produce better strategic decisions. Formulating strategy is a creative process, and so it is enhanced by multiple inputs. A team of leaders brings together more information and ideas, and hence deliberations can be more thorough. Both Martin Luther King, Jr. and Cesar Chavez surrounded themselves with leadership teams, even though in the popular imagination their successes are attributed to individual genius and charisma (Morris and Staggenborg 2002). A King hadY an unexcelled ability to pull men and women of diverse viewpoints together and to keep their eyes focused on the goalY King demonstrated...a rare talent for attracting and using the skills and ideas of brilliant aides and administrators@ (Bennett 1970:32-33, cited in Morris and Staggenborg 2002). For example, in Birmingham, Alabama in 1963, King and other leaders of the SCLC (Southern Christian Leadership Conference) were planning a mass disruption of the city. Their plan was to hold demonstrations and an economic boycott that would lead to mass arrests and fill the city jails. However, they were encountering difficulty mobilizing people to take part, and debated whether to allow children to participate in such a potentially dangerous action. King left town for a speaking engagement, and while he was gone other leaders began including children in the demonstrations. Confronted with this reality, King accepted the decision upon his returnB and the campaign succeeded in disrupting the city and inspiring movement activists in other cities (Morris and Staggenborg 2002). As brilliant as he was, if King alone had been responsible for the planning, the campaign may well have failed.

When two or more organizations or groups are working together within a movement, connections among leaders facilitate the sharing of information, forms of organization, and tactics. There are numerous examples from the civil rights movement. Hattie Kendrick was a longtime local movement leader who was crucial in recruiting young leaders and putting leaders in contact with one another (Herda-Rapp 1989, cited in Morris and Staggenborg 2004).

In addition, the composition of the leadership team matters. Leadership of individuals with a variety of backgrounds, viewpoints, and skills is more likely to formulate effective strategies (Ganz 2000). In particular, the most effective leadership teams consist of both A insiders,@ or indigenous leaders, and A outsiders,@ who do not come from the movement= s constituency (Marx and Useem 1971; Ganz 2000). Insiders understand the history and culture of the constituency group and can use this knowledge to mobilize support and choose appropriate tactics. For example, during the campaign of the United Farm Workers movement for a union contract with grape growers, Mexican and Mexican-American leaders chose to hold a crucial meeting in Our Lady of Guadalupe Church in Delano, California, the community= s religious center, and scheduled it for September 16, Mexican Independence Day (Ganz 2000). Insider leaders are more apt to be trusted by their constituency. Leaders with strong ties to the constituency group are especially effective at recruiting participants and locating indigenous resources. They are apt to have a strong commitment to the movement because of their personal ties to the group.

On the other hand, outsiders bring new information and alternative viewpoints to the decision making mix. Better strategy is formulated when the leadership team can draw from a range of potential tactics, and outsiders can bring insights from their experience in other movements and organizations. They are important in linking the movement with outside sources of participants, resources, and ideas. They can be especially effective at building alliances that further the movement= s goals. For example, in a local antiwar group of mostly middle-class people, a leader with ties to the working-class community would be valuable in forming a broader-based organization.

Finally, leaders make better strategic decisions when they are held accountable to the group= s members through decision making procedures that are regular, open, and authoritativeB for instance, regular meetings that are open to all members at which they can take part in making important decisions (Ganz 2000). Accountability of leaders is enhanced also when the organization relies on its constituents for at least some of its resources (Ganz 2000).

3. Groups that seek to displace the opponent are less likely to succeed.

The corollary is that groups whose goals do not include replacing the opponentB for example, that seek to change a law but not replace the lawmakersB are more likely to succeed (Gamson 1975; Frey et al. 1992). For example, revolutionary groups in the United States, which seek to replace the government, have been unsuccessful. Labor organizations that have sought concessions from employers have had some success, while those early labor groups that advocated that workers take over companies and run them themselves were not.

4. Disruption (but not necessarily violence) contributes to success, but only when the target has been unresponsive to less disruptive tactics.

A number of studies have found that movements that use or threaten to use disruptive tactics are more likely to succeed than those that do not. Note that this is not the same as saying that violence succeeds. Disruptive tactics can be violent or nonviolent; the defining feature is that they upset the social order, thereby threatening the power of elites.

Piven and Cloward= s Poor People= s Movements made a major contribution to the literature by demonstrating that when movements of the poor succeedB such as the industrial workers= movement in the 1930s and the civil rights movement of the 1960sB it is by disruption, or specifically by withholding the contributions that they normally make to institutional life in the form of work, or rent, or obeying laws. Cress and Snow= s (2000) study of homeless organizations found that those that were successful either used disruption and had allies on the city council, or used nondisruptive tactics and were situated in cities that already had a government agency that dealt with homelessness. Where an organization both had city council allies and was in a responsive city, tactics were irrelevant. Whether disruptive tactics contributes to success or not is likely dependent on the circumstances under which they are used. When the political environment is favorable toward a movement, disruption can be counterproductiveB it can waste an organization= s resources and alienate real and potential political allies (Cress and Snow 2000; Amenta et al. 1999).

Studies of the effects of using violence have produced equivocal findings. For labor movements, the use of violence was found to be negatively associated with success in the United States (Taft and Ross 1969) and Italy (Snyder and Kelly 1976), but not in France (Shorter and Tilly 1971) (cited in Giugni 1999). A study of the effects of the movement against the war in Vietnam found that demonstrator violence and property destruction increased the number of pro-peace votes in Congress, but decreased the number of times that Congress voted on war-related measures (McAdam and Su 2000).

5. Tactical innovation aids movements.

Disruption depends on the target= s lack of effective response to it. Over time, targets adapt to particular tactics, becoming better at controlling them in order to minimize disruption. For movements whose primary power is the power to disrupt, then, tactical innovationB the periodic introduction of new tacticsB is key to regaining the initiative (McAdam 1983). The civil rights movement used tactical innovation to maintain momentum. The bus boycotts, initiated in the mid-1950s, signaled the beginning of mass participation in the civil rights movement. However, after a time opponents were able to neutralize the impact of bus boycotts through legal means and extra-legal harassment of participants. The movement experienced a lull in activity until a new tactic, the sit-in, was introduced in 1960. Similarly, one result of the widespread protests of the 1960s was that town governments created a permit system for marches and demonstrations. They adapted to the protests by providing a legal channel for them, removing much of their ability to disrupt.

6. The presence of radical groups within the broader movement aids success.

This is termed the "radical flank effect" (Haines 1988) and it has three components. First, the presence of radicals within a particular movement makes moderate groups in the movement more attractive negotiating partners to the target. The target (state officials, for example) would rather sit at the bargaining table with moderates than deal with the radicals. Second, the presence of radical groups tends to prompt an increase in financial support for the movement from outsiders--again, the moderates become the preferred organization for them to support. Third, when radical groups emerge, moderate groups themselves tend to become more radical in their demands. The experience of the NAACP (National Association for the Advancement of Colored People) illustrates these effects (Haines 1988). During the 1950s the NAACP was driven underground in the deep south; it was considered too radical for white power structure to tolerate. When new, more radical civil rights groups were created in the early 1960s that used direct action, like sit-ins, to confront segregation, like SNCC and the revitalized SCLC, the NAACP suddenly looked moderate by comparison, and over the next several years contributions to the NAACP increased tenfold. In addition, the goals of the NAACP became more radical during this time, though not as radical as the other groups.

7. Ideological and strategic/tactical diversity within a movement aids success.

Though there is less systematic research on this issue, some studies have found that the presence of different organizations with a range of ideologies and strategies optimizes a movement= s chances. In the civil rights movement, the SCLC was the central organization for many campaigns. The NAACP fought and won legal battles. SNCC brought young people into the movement who were especially willing to engage in dramatic, risky actions like the sit-ins. CORE was a multiracial group that sponsored the Freedom Rides to integrate interstate transportation in the south (Morris and Clawson 2002). It is unlikely that any one of these organizations alone would have achieved what the civil rights movement as a whole was able to. In the feminist movement, the existence of two branches with different ideologies, organizational styles, and strategies--the older branch, which formed NOW, and the younger branch, which formed consciousness-raising groups and engaged in direct action--meant that diverse women could find a A home@ in the movement. Research on the impact on policymakers of the movement against the Vietnam War found that different tactics were effective at different times (Small 1987).

These findings are particularly interesting given the frequent calls for more unity within the peace and justice movement (broadly defined) in the United States. The diversity among peace and justice organizations is an asset that should be maintained (though this does not preclude greater collaboration among groups).

8. People and strategy can compensate for lack of money.

While material resourcesB money, office space, computers, and the like--can certainly help social movements to meet their goals, they are not necessary for success. For example, civil rights organizations received money from wealthy donors and attracted support from northern whites, but only after they had demonstrated their mass appeal and power (Morris 1984). Movements can compensate for a lack of material resources with committed people and a good strategy. As stated above, Piven and Cloward found that movements of the poor do not win by mobilizing material resources, but by using their powers of disruption. The homeless movements that Cress and Snow (1996) studied that were successful had in common not money, but leadership, outside support and advice, and office and meeting spaces. The United Farm Workers succeeded despite very few material resources (Ganz 2004).

9. The effective framing of issues influences success.

Framing refers to the process of shaping one= s message so that it captures the attention of and appeals to some audience. While there is much research on how social movement organizations frame their messages, there is little analytical research on the relationship between framing and success. Cress and Snow (2000) found that all of the successful homeless organizations that they studied used clear and well-articulated diagnostic frames (which identify a problem and its cause) and prognostic frames (which specify what needs to be done to fix the problem).

External Factors

10. Powerful allies aid social movements, but are not always necessary for success.

Social movement scholars have addressed whether movements can succeed directly, solely through their own efforts, or whether success only comes indirectly, by either influencing public opinion or gaining the support of powerful allies, like political officials. Evidence exists for the direct effects of movements. For example, antiwar protest increased slightly the number of pro-peace votes in the Senate between 1964 and 1973 (Burstein and Freudenberg 1978). In Mississippi, levels of nonviolent protest from 1960 to 1970 were positively related to the number of welfare recipients (Colby 1982). A study of the movement to ratify the Equal Rights Amendment in different states found that the presence of powerful allies, the strength of opponents, and public opinion all influenced success, but the movements themselves had a direct effect regardless of these other factors (Soule and Olzak 2004). The nuclear freeze movement of the 1980s forced President Reagan to soften his rhetoric about fighting and winning a nuclear war and strengthened the arms control caucus in Congress, leading eventually to the first arms reduction agreement between the United States and the Soviet Union (Meyer 1999). At the same time, in most cases powerful allies help. The United Farm Workers succeeded where previous efforts had failed in part because of support from a liberal coalition of religious groups, civic organizations, and government officials (Jenkins and Perrow 1977).

12. Repression can aid or hinder a nonviolent movement, depending on other factors.

Most studies that have investigated the effects of government repression on the success of social movements have looked at non-U.S. movements. Repression tends to lead to more radical--sometimes violent--protest (Goldstein 1983; Wilson 1976; della Porta 1995; Moore 1998). There is some evidence that in the short term, repression inhibits protest, but in the long term it deligitimizes the state and so increases the level of protest (Rasler 1996; Olivier 1991). When used at the beginning and end of a protest cycle, repression may inhibits protest, but it encourages protest when used in the middle of a cycle (Brockett 1993). Repression is more likely to spur increased protest when it is seen as illegitimate and when protesters are encouraged to continue by their social group (Opp and Roehl 1990).

In terms of its ultimate effect on a movement= s success, several studies have found that repression inhibits success (Gamson 1975; Andrews 1997). During the anti-Vietnam War movement, repression of demonstrators by police increased the number of times Congress voted on war-related measures, but decreased the percentage of pro-peace votes (McAdam and Su 2002).

Recommendations and Obstacles

One set of reasons for the dearth of research on the outcomes of nonviolent action is methodological (Earl 2000). First, it is difficult to decide what constitutes success. A movement may experience a temporary setback, then later succeed. A specific campaign conducted by activists may be successful, while the movement as a whole is not. A movement may achieve some of its objectives, but not all. Movements have unintended effects. A second problem is that of causation. When nonviolent activists achieve their goals, it is always possible that something else besides their actions caused the change, or that their actions contributed but were not the sole cause. For example, some sociologists have noted that while Americans= attitudes toward gender changed during the feminist movement of the 1970s, such change may have also been due to the increased participation of women in the paid workforce.

A second obstacle to conducting research on nonviolent action for the use of activists is that most people with the expertise to do so work in colleges and universities. These are institutions which either reward faculty for conducting research for publication in prestigious academic journals, but not for research intended for non-academic audiences; or, if faculty are not pressured to publish in academic journals, then their teaching loads are heavier, leaving less time to do any research at all. In addition, academic writing tends to be highly technical and jargon-laden, leaving what research does exist on nonviolent action relatively inaccessible to non-academics.

If these obstacles can be overcome, what should researchers study? The best recommendations for research topics will come from social movement activists themselves: they are situated to know best what information can help increase their effectiveness. As an activist myself, I can think of a few fruitful avenues of research.

$ the effects of confrontational tactics (property damage; pie throwing; taunting police). Peace and justice movement activists are somewhat divided on the advisability of these tactics; for example, the act of damaging property was hotly debated in the anti-globalization movement following the 1999 Seattle protests. Opponents have argued that such tactics decrease public sympathy for the movement; is this supported by the evidence? Or do militant tactics draw in some people who might not otherwise participate, as claimed by some? Do they keep some groups away?

$ what types of frames are most effective in specific situations. A colleague and I have been planning to erect a freeway banner expressing opposition to the war in Iraq. Would passers by be most receptive to a religious message--A Love thine enemies@ --or a message that highlights the cost, in money and lives, of the war? We live in the rural Midwest. Would the answers to these questions be different if we lived in a different part of the country?

$ the effects of various tactics that are commonly used by peace and justice groups, including silent vigils, leafleting, and mass marches. Most peace and justice groups draw from the same standard repertoire of tactics, few if any of which have been systematically studied as to their actual effects.

$ micro-interactions in protests. Do protesters= communications with police officers during a nonviolent action matter? How likely are the actions of police on the A front lines@ of a protest to be influenced, positively or negatively, by what protesters do and say?

$ insider research into what makes decision makers change their mind. Besides Small= s (1987) research on the influence of anti-Vietnam War protests on political leaders, there is little research into how and why nonviolence is effective in changing the minds of decision makers. Does the size of the movement matter? Is direct communication with leaders desirable? Is it effective to create undesirable media coverage of targets? At what points in a campaign are conciliatory messages more effective, and when is protest more effective?


Amenta, Edwin, Drew Halfmann, and Michael P. Young. 1999. AThe Strategies and Contexts of Social Protest: Political Mediation and the Impact of the Townsend Movement in California.@ Mobilization 4 (1): 1-23.

Andrews, Kenneth T. 1997. AThe Impacts of Social Movements on the Political Process: The Civil Rights Movement and Black Electoral Politics in Mississippi.@ American Sociological Review 62: 800-91.

Brockett, Charles. 1993. AA Protest Cycle Resolution of the Repression/Popular-Protest Paradox.@ Social Science History 17: 457-84.

Burstein, Paul and W. Freudenburg. 1978. A Changing Public Policy: the Impact of Public Opinion, War Costs, and Anti-war Demonstrations on Senate Voting on Vietnam War Motions, 1964-73.@ American Journal of Sociology 84: 99-122.

Colby, David C. 1982. AA Test of the Relative Efficacy of Political Tactics.@ American Journal of Political Science 26, 4 (November): 741-53.

Cress, Dan and David A. Snow. 2000. "The Outcomes of Homeless Mobilization: The Influence of Organization, Disruption, Political Mediation, and Framing." American Journal of Sociology 105: 1063-1104.

Cress D, Snow D. 1996. AMobilization at the Margins: Resources, Benefactors, and the Viability of Homeless Social Movement Organizations.@ American Sociological Review 61: 1089-109.

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

Earl, Jennifer. 2000. AMethods, Movements, and Outcomes: Methodological Difficulties in the Study of Extra-Movement Outcomes.@ Research in Social Movements, Conflicts and Change, Volume 22, pp. 3-25. Stamford, Conn.: JAI Press.

Frey, R. Scott, Thomas Dietz, and Linda Kalof. 1992. "Characteristics of Successful American Protest Groups: Another Look at Gamson's Strategy of Social Protest.@ American Journal of Sociology 98:368-387.

Gamson, William A. 1975, 1990. The Strategy of Social Protest. Wadsworth.

Ganz M. 2000. AThe Paradox of Powerlessness: Strategic Capacity in the Unionization of California Agriculture, 1959-1966.@ American Journal of Sociology 105:1003-62

Goldstein, Robert J. 1978. Political Repression in America. Cambridge, Mass.: Schenkman.

Giugni, Marco. 1999. AHow Social Movements Matter: Past Research, Present Problems, and Future Developments.@ Pp. xiii-xxxiii in How Social Movements Matter, edited by Marco Giugni, Doug McAdam, and Charles Tilly. St. Paul: University of Minnesota Press.

Haines, Herbert H. 1988. Black Radicals and the Civil Rights Mainstream, 1964-1970. Knoxville: University of Tennessee Press.

Herda-Rapp, Ann. 1998. "The Power of Informal Leadership: Women Leaders in the Civil Rights Movement." Sociological Focus 31(4): 341-355.

Jenkins, J. Craig and Charles Perrow. 1977. AInsurgency of the Powerless: Farm Worker Movements (1946-1972).@ American Sociological Review 42, 2 (April): 249-268.

Marx, Gary T. and Michael Useem. 1971. AMajority Involvement in Minority Movements.@ Journal of Social Issues 27: 81-104.

McAdam, Doug. 1983. ATactical Innovation and the Pace of Insurgency.@ American Sociological Review 48: 735-754.

McAdam. Doug and Yang Su. 2002. AThe war at home: antiwar protest and congressional voting, 1965 to 1973.@ American Sociological Review 67: 696-721.

Meyer, David S. 1999. AHow the Cold War Was Really Won: The Effects of the Antinuclear Movements of the 1980s. Pages 182-203 in How Social Movements Matter, edited by Marco Giugni, Doug McAdam, and Charles Tilly. St. Paul: University of Minnesota Press.

Moore, Will H. 1998. A Repression and Dissent: Substitution, Context, and Timing.@ American Journal of Political Science 42, 3 (July): 851-873.

                    Morris, Aldon D. 1984. The Origins of the Civil Rights Movement: Black Communities Organizing for Change. New York: Free Press.

Morris, Aldon D. and Dan Clawson. 2002. ALessons of the Civil Rights Movement for Workers= Rights/Union Organizing.@ Prepared for presentation at the AFL-CIO/Michigan State University Worker Rights Conference, Michigan State University, East Lansing, Michigan, October 11-12, 2002. http://www.cas.northwestern.edu/sociology/faculty/files/morris-clawson.pdf. Accessed June 6, 2005.

Olivier, Johan L. 1991. A State Repression and Collective Action in South Africa, 1970-1984.@ South African Journal of Sociology 22: 109-117.

Opp, Karl-Peter and Wolfgang Roehl. 1997. A Repression, Micromobilization, and Political Protest.@ Social Forces 69, 2 (December): 521-547.

Rasler, Karen. 1996. AConcessions, Repression, and Political Protest in the Iranian Revolution. American Sociological Review 61: 132-52.

Small, Melvin. 1987. AInfluencing the Decision Makers: The Vietnam Experience.@ Journal of Peace Research 24, 2 (June): 185-198.

Snyder, David and William R. Kelly. 1976. AIndustrial Violence in Italy, 1878-1903.@ American Journal of Sociology 82: 131-62.

Soule, Sarah A. and Susan Olzak. 2004. AWhen Do Movements matter? The Politics of Contingency and the Equal Rights Amendment.@ American Sociological Review 69: 473-497.

Taft, Philip and Philip Ross. 1969. AAmerican Labor Violence: Its Causes, Character, and Outcome.@ Pages 281-395 in Violence in America: Historical and Comparative Perspectives, edited by Hugh D. Graham and Ted R. Gurr. New York: Bantam.

                    Wilson, John. 1976. "Social Protest and Social Control." Social Problems 24: 469-81.

Zunes, Stephen. 2000. ANonviolent Action and Human Rights.@ PS: Political Science and Society 33 (2): 181-187.