The evidence in the last level is impressive, but a fair reading also has to weigh the objections. Some come from scholars who question the data, some from activists who have faced regimes willing to kill, and some from people who believe that armed resistance is sometimes both justified and necessary.
Who chooses nonviolence?
The first objection is about cause and effect. Perhaps nonviolent campaigns succeed more often because movements may choose nonviolence where winning is already easier: against weaker regimes, in more urban and educated societies, or where the army is already wavering. If so, nonviolence would be a symptom of favorable conditions rather than their cause. Chenoweth and Stephan tested for this, comparing campaigns against regimes of similar strength and repressiveness, and found that the nonviolent advantage held. Critics reply that some conditions are hard to measure, and the debate continues.
What counts as nonviolent?
A second objection is about coding. Many campaigns counted as nonviolent included riots, stone throwing or burning of police stations. In Egypt in 2011, crowds attacked and burned police stations in the first days of the uprising. The sociologists Mohammad Ali Kadivar and Neil Ketchley studied this "unarmed collective violence" and found that it was sometimes linked to later democratization, rather than always harming movements. Their finding does not show that violence is good, but it does suggest that the line between violent and nonviolent campaigns is blurrier than the headline numbers imply.
Radical flanks
Movements are rarely of one mind, and their wings affect each other. A positive radical flank effect occurs when militants make moderates look like safe partners, so that rulers prefer to bargain with the moderates before the militants grow stronger. The sociologist Herbert Haines argued that something like this happened in the American civil rights era. A negative radical flank effect is the reverse: militants give rulers an excuse to crack down on everyone, and frightened bystanders rally to the government. The same movement can experience both effects at different moments, which makes the evidence hard to read.
When regimes will kill
Nonviolent resistance is least likely to win when security forces stay united and willing to shoot. In Beijing in June 1989, after weeks of protest, the army cleared the protest by force; hundreds and possibly thousands were killed, and the party's rule was not shaken. In Bahrain in 2011 the monarchy, aided by troops from Saudi Arabia and the United Arab Emirates, crushed a very large movement. Defenders of civil resistance answer that armed revolts against such regimes usually fail too, often at far greater cost, and that a strategy is not refuted by losing against long odds.
The case for armed struggle
Some movements turned to arms after peaceful methods were met with bullets. In March 1960 South African police killed 69 people at a protest against pass laws at Sharpeville, and the government banned the major liberation movements. In 1961 the African National Congress formed an armed wing. Mandela later explained at his trial that peaceful protest had been met with massacres and bans, leaving no lawful path. Many people who honor nonviolence still accept a right to self-defense, and many independence wars are remembered as justified. A fair account admits that this is a genuine moral argument, not a failure of nerve.
Toppling is not building
Finally, removing a ruler is only half the task. Egypt shows that ousting a ruler does not guarantee democracy: the movement that forced Mubarak out in 2011 was united on what it opposed but divided on what to build, and two years later the army was back in charge. Ukraine's Orange Revolution of 2004 overturned a fraudulent election, yet the candidate it defeated won the presidency fairly in 2010. A movement's skill at saying no to a ruler is different from its capacity to agree on what comes next. Movements need plans, organizations and coalitions for the day after.
Worked example: Tunisia and Syria
In December 2010 a Tunisian street vendor, Mohamed Bouazizi, set himself on fire after officials confiscated his goods. Protests spread across the country, and on 14 January 2011 President Zine el-Abidine Ben Ali fled to Saudi Arabia. The army largely stayed out of the repression, though accounts of who refused which orders still differ. When politics deadlocked in 2013, four civil groups, a labor union, an employers' association, a human rights league and the bar association, brokered a compromise and won the Nobel Peace Prize in 2015. Tunisia adopted a democratic constitution in 2014. Yet in 2021 President Kais Saied suspended parliament and began ruling by decree.
Syria's protests began in March 2011. The regime fired on crowds from the start, its core military units stayed loyal, and defecting soldiers formed armed groups. A civil war followed that killed hundreds of thousands and drove millions from their homes, until Bashar al-Assad fled in December 2024. The two cases show how much depends on whether the security forces hold together, and how even a success can later erode.