Coding Definitions

Here we show how the database

  • operationally identifies “leaders,” “partners,” and “allies,”
  • shows the timing of the appearance (and exiting) of allies,
  • rates the success or failure of the campaign,
  • defines violence by campaigners and nonviolent action by opponents


Leaders usually, but not always, initiate a campaign. Otpor in Serbia initiated the campaign to overthrow Milosevic. The mayor of Culebra, Puerto Rico (and some people close to him) initiated the campaign against the U.S. Navy. A key African American women’s group, labor leader E.D. Nixon, and some black clergy initiated the Montgomery Bus Boycott on the occasion of the arrest of Rosa Parks.

The leaders are usually the organizers/spokespeople who get there first. There are exceptions: In Burma in 1988 student leaders initiated the democracy movement then Aung San Suu Kyi came forward to be the primary leader, a role she retained through many years of house arrest.

Leadership can shift in the course of the campaign, also, as Burma reminds us. Aung San Suu Kyi built a political party to run in the election which the dictatorship had been forced to allow, and the building of the party created an additional leadership structure, while the students who had started the campaign were, before the formation of the party, killed, forced into exile, or had to flee into the jungle. So timing is an important variable in identifying leadership. In neighborhood-based campaigns, the initiative may be taken by professional organizers who seek local individuals with leadership potential and support them.

That is also true for some labor campaigns. What’s complicated here for researchers is that the discipline of community organizing in the U.S. makes a distinction between “organizers” and “leaders” – the former are expected to stay in the background and not get publicity; the latter are expected to be the public spokespeople and to grow in leadership skills as they chair meetings and head negotiating delegations. So for researchers working from written secondary sources, it’s sometimes hard to know who were the initiators.  For this database, however, it is sufficient for researchers to name
the leaders indicated by the secondary sources even if a more finely tuned research study would reveal that there are organizers behind the scenes.

Partners usually respond to the invitation of the leadership. Consider Freedom Summer in Mississippi, 1964. SNCC had been working in Mississippi and decided to create the campaign. It invited partners: CORE, SCLC, NAACP, who created a structure through which the partners could meet, communicate, collaborate, be visible, make some decisions. However, SNCC remained the “senior partner”/leader – everyone knew that the leader SNCC had the biggest stake, and the most experience in Mississippi, and the greatest expertise in working with young adults.

Partners have a stake; they can benefit organizationally or in terms of personal prestige from success, and they can also take a loss from failure. Partners have influence on the conduct of the campaign, although less than the leadership.

Another characteristic of partners (in contrast to allies) is stability. The fisher folk of Culebra could expect stable relationships with their partners (Puerto Rican Independence Party, Clergy Committee) for the duration of the campaign. Partners also “own” the campaign and usually see it to the end. One reason why the opponent tries to drive a wedge between one partner and the others, through rumor-mongering, for example, or planting evidence of corruption, is to damage the campaign – precisely because partners are expected to be active for the duration of the campaign.

Allies, on the other hand, are not expected to be so stable. They’re a little like people who come to a party – they may come late, and leave early. They are less likely than the partners to stay after the party to help clean up. They do not “own” the campaign; they support it. Two criteria illuminate this: allies in general are exposed to less risk, and do not expect to have decision-making power.

Consider the role of the National Council of Churches (NCC) in Freedom Summer 1964. As an ally, the NCC offered to organize and fund two weeks of training in the North for the hundreds of students who volunteered to go to Mississippi. That gave the NCC no decision-making power or ongoing influence in the campaign, so it was not a partner. What the NCC did for the campaign was extremely important, but its commitment didn’t make the campaign dependent on it. A partner would be expected to stay in there doing other things to meet new challenges. An ally can more easily put a boundary around the degree of commitment and length of commitment.

The stakes are less for allies: If the Freedom Summer campaign faltered in July, and by August was a shambles, the NCC would have been excused by most of its member religious groups as having made an error of judgment, or their members could just say “things don’t always work out.”



The database intends to support the comparison of cases with regard to how quickly/slowly the allies of the campaign make themselves known, and also when elite involvement happens (if ever). We also want to know when groups exit the campaign, if they do. (Maybe they get scared off by repression, or bought off by the opponent, and even though the campaign continues, they leave.)

Researchers cannot, however, simply write down dates; we need a way of comparing cases that takes into account the varying durations of the campaigns. The procedure follows:

  1. Define the life span of the campaign. (3 months? 1 year? 3 years?)
  2. Mark off the campaign’s length into six equal segments.
  3. For each ally/elite/partner, note the timing of their entrance by placing their name in the field for the segment in which they entered.
  4. If a participant exits, mark this by entering the name in the segment in which they exited, followed by "(exit)"
  5. Note in “Comments” anything that you think might make this case an anomaly with reference to timing.


One standard approach in the world of evaluation is to assess success/failure in terms of the stated goals of the protagonist.  Success is rarely all or nothing, of course, so it helps to have a scale.  The scale (of maximum ten points) will be based on the stated goals of the campaigners/interveners, but also on some process goals that are often unstated but assumed. 

Weighted score: from zero to ten points.  We will take three variables into account: goal attainment, survival, and growth.  The three are weighted differently. While this particular scoring system will no doubt be controversial in the field of social movements – people can always argue about how much weight to attach to what – this procedure can still work for us, because it will provide internal consistency so we can compare campaigns and interventions with each other.

Procedure: 6 points: success in achieving specific demands/goals

Give up to six points for this, calibrated according to degree of success.

Because historical cases are unique, each of these ratings will need to fit the character of the struggle and its outcome.  That means that each rating will to some extent be disputable; evaluation in political matters is almost always controversial.  What enables us to justify such an inherently subjective/interpretative procedure is to make it (a) transparent and (b) open to feedback. 

The reason to have six points available is to be able to make a continuum.  Six points would represent complete success of the campaigners’ original goals.  Zero points would represent total failure to achieve the campaigners’ original goals.  The basic framing is, whenever possible, to find a statement of original goals, outlined by the leadership of the campaign.

For example, the ANC launched its mid-fifties campaign to force withdrawal of the pass laws in South Africa.  It failed to achieve that to any degree, so gets zero points.  The 1920 campaign in Germany to force the reinstatement of the Weimar Republic after the Kapp putsch succeeded.  It removed Wolfgang Kapp and brought back the Social Democrats.  It therefore gets six points.

Another German case is interesting in this regard: the defense of the Ruhr Valley against the French and Belgian invasion after World War I.  Historians often call this a mixed success, because even though the French and Belgians retreated and the Ruhr was restored to German governance, and under the Dawes Commission the Germans actually gained a better position regarding reparations as a result of their nonviolent struggle, they won at a high cost – the Social Democrats paid for the struggle by printing more money than they could back up, causing what became runaway inflation, social turmoil, and sowing the seeds for eventual Nazi rise to power.

We nevertheless give the Ruhrkampf a score of six, because we’re scoring by degree of success in achieving original goals.  I don’t believe we should get into the empirical swamp called “assessment of costs,” let alone the then-necessary question of whether the campaigners could have found ways of reducing the costs and still winning (which in theory they always could) etc. etc.  To keep some kind of stable control of the materials we’re working with, I believe we need to separate the function of “assessment” from the function of “second-guessing.”  The danger of second-guessing is why the field of evaluation evolved into the perspective of looking at achievement in terms of the actor’s own goals!

Sometimes a campaign will have multiple goals, of course.  An anti-nuclear power campaign may seek to stop a specific nuclear power plant from being built AND to de-commission one that is already on line.  Maybe the campaign succeeds in the first and fails on the second, so it would get a score of 3 unless there is reason to think that the campaigners themselves were mainly seeking the first and the second was an add-on.  We could debate that, and maybe end up giving a score of 4 or even 5.

What if there is delayed achievement of goals?  Political scientist Erica Chenoweth takes account of this possibility by using a two-year window after the campaign ceases, and we will do the same.  Some opponents only make concessions after they have “swept the streets clean of protesters,” in order to save face.  We join Erica in calling it a “win” if the change happens within two years. 

What if the opponent concedes some or all of the campaigners’ demands, then takes it back a couple of years later, or much later? 

For this database we use more of a snapshot approach to the achievement of demands than we do a feature film.  At the conclusion of the campaign, we ask, "To what degree did they achieve their stated goal?"   If it is complete achievement at that time or within two years, we give a 6, even though the opponent may later rally its forces and make a come-back.

We do that because history is full of such contests between movements and opponents, in which, over time, first one side seems to be winning and then the other side. History is like a football game and the sides race up and down the field, gaining touchdowns and temporary advantage, but no one knowing how it will finally end up until Time itself runs out.

Because we're not cosmic and willing to wait until the end of history, and we don't have an operational way of knowing when a campaigner goal is secured forever, we use the end of the campaign (plus two years) as the point of measure.

Process Goals

While we mainly rate degree of success/failure in terms of the explicit goals of a campaign or movement, there are also process goals that may be “too obvious to state” but nevertheless are inherent in most groups that are in struggle, and are important in assessing degree of success:

  • Survival. Nearly every group wants to survive, so that even if it loses the current battle, it can fight again at a more opportune time. The ANC lost its ‘50s campaign against the pass laws in S. Africa but survived to fight again and eventually to defeat apartheid.
  • Growth. Every group wants to grow. If it wins its specific objective it wants to be able to consolidate gains. If it loses its specific objective, it wants to be in better shape to win its next campaign.

An advantage of including these process goals is that by doing so we affirm the basically historical nature of most social struggle. A database project needs to take extra care to remain in history. Nonviolent campaigns and interventions, like military struggles, operate within the context of unpredictable lengths of time.

A U.S. presidential candidate maps a campaign with a certain date of conclusion. The campaign goals of social movements are rarely framed in a time period. The very unpredictability of the time frame influences the eventual outcome in complex ways. Inclusion of at least these two process goals, survival and growth, is a way of acknowledging the flow of history in the outcome of winning and losing.

One point: survival

Give one point for survival, by which we mean survival of the infrastructure (which is more measurable). For example, although the African National Congress was at one point banned by the government, it nevertheless retained its organization, re-located offices, budget to support paid staff, etc. So even though its ability to grow was hampered, the infrastructure survived. It gets one point, therefore. (It’s possible to give only half a point if the infrastructure is severely weakened but not destroyed.)

Three points: growth

Measure growth of the campaign by: (a) number of members and/or participants, (b) size of budget and staff/organizers/full-time campaigners, (c) number and power of allies.
The points are awarded in this way:

  • One point for a small amount of growth,
  • Two points for a medium amount of growth,
  • Three points for a large amount of growth.


Largely nonviolent campaigns often include some violence done by the campaigners.  By violence we mean “the threat or use of injurious force directed toward humans” --  toward police, for example, or soldiers, officials, or replacement workers in a strike.  It might be spontaneous, in the heat of the moment, or acted out in planned reprisals.  If the record shows such campaigner violence to have happened, database researchers report it.

Researchers are asked to consider whether the injurious force is in fact being done by campaigners, or by assorted people who spontaneously “join” the action but have no organic ties to the campaign.  A campaign can hardly take responsibility for the latter, whether they are sincere people enraged by what’s happening, or are paid by police to use violence while pretending to be campaign activists. 

When we find a case in which the violence is planned by or encouraged by the campaign’s leadership, the case is excluded from this database.  An example of exclusion is the British woman suffrage campaign, in which the leadership encouraged the horsewhipping of Members of Parliament.

Property destruction may or may not be "injurious force toward humans."  Slashing a painting that an artist spent the last six months on can be considered violence.  Cutting a fence to get into and expose a concentration camp would not be. So, while property destruction is not routinely regarded by this database as violence, the question requires thoughtfulness.  Campaigners may do property destruction that is read by prospective allies as violence (even though database researchers would not define it as such), and in the unfolding of the struggle it is the allies’ interpretation that counts!  If a researcher finds this to be so, we try to include it in the writing.

On some occasions opponents have used nonviolent action against the campaigners.  In 1931 during the Salt Satyagraha police reportedly conducted a sit-down in the street to prevent protesters from completing their march, although this method failed when the marchers handed over food and blankets to the rain-soaked police.  The researchers’ guide on this question is again scholar Gene Sharp’s 198 methods plus additional ones added for this database.

- George Lakey 19/08/2011