Civil resistance and the 3.5% rule

by morganism1 min read2nd Feb 201715 comments


Personal Blog

Interesting, haven't seen anything data-driven like this before...


Civil resistance and the 3.5% rule.

"no campaigns failed once they’d achieved the active and sustained participation of just 3.5% of the population—and lots of them succeeded with far less than that."

"Then I analyzed the data, and the results blew me away. From 1900 to 2006, nonviolent campaigns worldwide were twice as likely to succeed outright as violent insurgencies. And there’s more. This trend has been increasing over time—in the last fifty years civil resistance has become increasingly frequent and effective, whereas violent insurgencies have become increasingly rare and unsuccessful."


Data viz:



Interesting strategic viewpoint

1. Size and diversity of participation.

2. Nonviolent discipline.

3. Flexible & innovative techniques. switching between concentrated methods like demonstrations and dispersed methods like strikes and stay-aways.

4. Loyalty shifts.
if erstwhile elite supporters begin to abandon the opponent, remain silent when they would typically defend him, and refuse to follow orders to repress dissidents, or drag their feet in carrying out day-to-day orders, the incumbent is losing his grip.


(observations from article above)

"The average nonviolent campaign takes about 3 years to run its course (that’s more than three times shorter than the average violent campaign, by the way)."

"The average nonviolent campaign is about eleven times larger as a proportion of the overall population as the average violent campaign.

"Nonviolent resistance campaigns are ten times more likely to usher in democratic institutions than violent ones."




original overview and links article:


and a training site that has some exercises in group cohesion and communication tech, from Guardian.


edit: The article that got me looking, how to strike in a gig economy, and international reach

13 comments, sorted by Highlighting new comments since Today at 7:26 PM
New Comment

From 1900 to 2006, nonviolent campaigns worldwide were twice as likely to succeed outright as violent insurgencies.

Obvious-looking alternative to the interpretation I take it they have in mind: One of the factors that makes a campaign turn to violence is a sense that other means won't get them what they want. Imagine a toy world in which there are causes with success probabilities ranging from 0 to 1, and ones whose success probability is <0.1 turn to violence, and doing so makes them 3x more likely to succeed. And imagine (toy world, remember) that these success probabilities are uniformly distributed. Then the typical nonviolent campaign succeeds 55% of the time, and the typical violent campaign only 15% of the time -- but that doesn't mean the violent campaigns should have stayed nonviolent to maximize their chances of success.

A good point to keep in mind, though it looks like Erica Chenoweth has tried to address it:

In my book with Maria Stephan, we devote an entire quantitative chapter (where we use multiple two-stage models identifying both the choice of NV/V resistance and the link between this choice and the outcome) and four cases studies to this possibility. We find that, while it’s true that people consider the costs before acting, the information environments in which they are operating are highly uncertain and that the selection process isn’t really influencing a majority of the cases. Activists do not know when they go into the streets whether the regime is going to crack down indiscriminately or whether the regime will ignore them. And some of them disagree about which method is most effective, resulting in some people within the campaign choosing nonviolent resistance and others using violent resistance under identical circumstances. Indeed, we found no structural factors (including violent repression) that systematically influenced whether people resorted to nonviolent or violent resistance. In other words, it is not the case the the probability of success influences the choice to use nonviolent resistance.

I guess it's up to the reader to decide whether these qualitative & statistical controls adequately address the selection effect.

Ah, good. (I haven't attempted to assess whether it's addressed well, but it certainly looks like there's a serious attempt.)

Please post an introductory paragraph explaining what this is about and why you are posting your scattered links?

It seems kind of common sense that a small group of people using violence against a very large, well-armed group are going to have a tough time.

It feels like reading something like this may have mislead the Syrian rebels. Nonviolent resistance runs the risk that the other side presses the violence button.

Reminds me of a story where Gandhi tried using nonviolent resistance against Nazis. Didn't work as expected, the Nazis just shrugged and said "makes throwing you into the ovens easier".

Also, if the enemy propaganda is strong enough, it probably doesn't matter that you didn't use violence, because their TV will still report that you did. There needs to be some impartial observer to report on your actions.

Very true, although believing in the idea of an impartial observer feels a little old fashioned.

Nowadays the Resistance Channel would report that everybody is peaceful, regardless of the truth (anything that can't be ignored is false flag), and the Regime Channel would report that the protests are an Orcish Uprising that wants to murder your kids.

Outside observers will just be taken to be Regime loyalists by Rebels, and Rebels by Regimes, since their narrative will disagree with one or the other propoganda storylines.

This describes the case where there's an establishment and a resistance. What happens when there's two establishments and two resistances, each opposing the other? I think this is the more relevant scenario for your interest.

Or a resistance and no definitive establishment

The issue, as I understand gjm, is whether there's a confounding factor.

Treating the correlation between choosing nonviolence (call that N for short) and succeeding in a political campaign (call that S) as accurately representing causality works if one assumes causation runs exclusively as N → S. gjm observes that expected probability of success (P) could be a confounding factor: maybe we have the extra causal arrows N ← P → S, as well as N → S.

According to her blog comment, Chenoweth did try to tackle this by establishing that there's no N ← P causal arrow. Reconstructing her apparent logic, she says she found that external structural factors (X), which activists presumably use to form their beliefs about the probability of success, are uncorrelated with the decision to use nonviolence: i.e., she supposes that X → P and observes that X is uncorrelated with N, so P → N is implausible (because otherwise we'd have X → P → N and X would correlate with N, assuming, uhhh, the causal Markov condition I think?). If we accept that argument, then P isn't a confounder.

Chenoweth also implies that she compared the outcomes within campaigns which split into a nonviolent arm and a violent arm. That would arguably control for P, if one accepted that for any given campaign which splits into two, P is the same for the two new subcampaigns. I don't think I do accept that, but I could be wrong, and if so this alternative analytical approach would address the issue.

It's hard for me to come to a strong final conclusion because I haven't read the book. Based on the blog comment, I wouldn't trust that Chenoweth's isolated the potential confounder, but I might change my mind if I read the book. Along similar lines, you could well be correct about what's typical for campaigns, but I can't conclusively say you are because I haven't systematically studied a wide range of political campaigns from this perspective.

A misnamed idea which assumes there are effectively no costs to accommodating the minority.

This depends heavily on how one defines campaigns. Is the NAACP of 1910 the same campaign as Malcolm X in 1960?

I suspect that each group would say no, but their common opponents would say yes.