| (2019) Chenoweth has backed off her previous conclusions in recent years, observing that nonviolent protest strategies have dramatically declined in effectiveness as governments have adjusted their tactics of repression and messaging. See eg https://www.harvardmagazine.com/2025/07/erica-chenoweth-demo... One current example of messaging can be seen in the reflexive dismissal by the current US government and its propagandists of any popular opposition as 'paid protesters'. Large attendance at Democratic political rallies during the 2024 election was dismissed as being paid for by the campaign, any crowd protesting government policy is described as either a rioting or alleged to be financed by George Soros or some other boogeyman of the right. This has been going on for years; the right simply refuses to countenance the possibility of legitimate organic opposition, while also being chronically unable to provide any evidence for their claims. |
In a case of relatively organic or somewhat spontaneous action, 3.5% of people doing something is huge. The reason is because in organic or spontaneous action, those 3.5% probably represent the views of vastly more than 3.5% of people. But as actions become more organized and less spontaneous, you reach a scenario where those 3.5% may represent fewer and fewer people other than themselves. At the extreme example of effective organization (where you get 100% participation rate), those 3.5% of people may represent nobody beside themselves.
I was perusing the dataset they used [1] for the '3.5% rule' and it seems that a more unifying theme is leaders losing the support of their own base. And it's easy to how that could strongly correlate with large organic protest since you've done things to the point of not only pissing off 'the other side' but also your own side.
I think Nixon is a good example of this. There were vastly larger protests against Nixon's involvement in Vietnam than there were for Watergate. Yet the Vietnam protests had no effect whatsoever, while he left office over Watergate. The difference is that he lost the confidence of his own party over Watergate. Had he not resigned, he would likely have been impeached and convicted. Had 3.5% of people protested Watergate, he would even be included on this list, which I think emphasizes that protests (or lack thereof) are mostly a tangential factor.
[1] - https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi...