Experimental Evidence for Tipping Points in Social Convention

Experimental Evidence for Tipping Points in Social Convention

Damon Centola, Joshua Becker, Devon Brackbill, Andrea Baronchelli

Science 361(6397): 1116–1119 (2018) DOI: 10.1126/science.aas8827

Theoretical models of critical mass have shown how minority groups can initiate social change dynamics in the emergence of new social conventions. Here, we study an artificial system of social conventions in which human subjects interact to establish a new coordination equilibrium. The findings provide direct empirical demonstration of the existence of a tipping point in the dynamics of changing social conventions. When minority groups reached the critical mass—that is, the critical group size for initiating social change—they were consistently able to overturn the established behavior. The size of the required critical mass is expected to vary based on theoretically identifiable features of a social setting. Our results show that the theoretically predicted dynamics of critical mass do in fact emerge as expected within an empirical system of social coordination.

Key finding: When a committed minority reached roughly 25% of the group, they consistently overturned the established convention. Below that threshold, the majority generally held. The paper claims this provides the first direct empirical demonstration of a tipping point in social convention change.

Commentary

This paper has been hugely influential — cited in Scientific American, The Atlantic, Harvard Business Review, and widely invoked in activist circles (notably by George Monbiot) as evidence that a committed 25% minority is enough to tip society. That sweeping interpretation is where the controversy lies.

James Özden — "Social tipping points are probably overrated" (2023) Substack(https://jamesozden.substack.com/p/social-tipping-points-are-probably)

Probably the best critical write-up available. Özden's core claim: the paper is methodologically neat but has been wildly over-extrapolated. Some highlights:

"There are several examples of putting a lot of belief in social tipping points in society, whether it's in popular media outlets, social media forums, academia or civil society groups. This paper… finds that 47% of all papers referencing social tipping points were published in 2020-2021, focusing on them as a climate solution."

"Long story short, this paper provides extremely little evidence that 25% is a likely number for social tipping points, for issues that people actually care about."

His three key critiques:

  1. The task is trivial and artificial. Participants matched names to random faces for $0.10 — nothing culturally or societally ingrained. Real issues (meat-eating, flying, car ownership) are deeply embedded. Even for a far simpler behaviour — choosing reusable mugs over single-use cups — research finds tipping points of 50–65%, roughly double Centola's 25%.

  2. People were incentivised to coordinate. Participants won $0.10 every time they coordinated. In reality, pro-social behaviours (EVs, plant-based food) often cost more, so people are disincentivised, not incentivised. You can't tell from this design whether people actually changed their beliefs or just wanted the cash.

"In reality, this is almost the polar opposite of what we might expect for some of the social issues we care about. For example, electric vehicles are often more expensive than internal combustion engine vehicles, and planet-and-animal-friendly meat alternatives are more expensive than animal products. So, people are disincentivised from adopting these pro-social behaviours, rather than incentivised to adopt them."

  1. Life isn't a series of one-shot pairwise interactions. The experiment paired people 1-on-1. Real norm formation happens through family, friends, groups, social media — one-to-many and many-to-many dynamics. (Cites Peter Licari's blog post on the same point.)

Özden is generous to the paper:

"I think this is a very neat experimental design, and it's certainly interesting. The issue I have with this paper is the claims that this will generalise, in any way, to broad social issues people are referring to, whether it's climate change, meat consumption, anti-racism, or something else."

He also notes the paper's own theoretical model says the threshold "is expected to vary based on theoretically identifiable features of a social setting" — but the popular framing treats 25% as a fixed universal rule. That, he argues, is the real problem: not the experiment itself, but the gap between what was shown and what people claim it proves.

"I'm worried that people read the headlines, fail to understand the methodology and develop some ill-placed belief that social tipping points will solve the world's issues. Of course, they might, but this paper certainly doesn't make me feel much more confident about it."

See also: Damon Centola, Change: How to Make Big Things Happen (2019) — Özden notes this book is "a much more reasonable and non-hyperbolic overview" of the same literature.

Edit this page