Threshold Models & Cascades
Mark Granovetter / Thomas Schelling · 1978
"Threshold models show identical groups can tip at very different points even with similar average preferences — because cascades depend on the distribution of thresholds, not just the mean, and a few low-threshold early actors determine whether contagion tips."
Granovetter asked in 1978 why two crowds with the same average willingness to riot behave oppositely — one riots, one doesn't. The answer is not averages but thresholds: who needs one other person to join, who needs ten.
Each person has a threshold — riot if at least k others riot. If low-threshold actors are present, their action pushes the next threshold holder over, triggering a chain. Schelling's segregation model shows the same: mild individual preferences tip into sharp collective segregation without anyone wanting segregation per se. Outcome is distribution-sensitive and path-dependent; seeding the first few actors matters more than persuading the average.
Why can two groups with the same average threshold behave oppositely in threshold cascade models?
Read more about the topic
The explanation above is written with AI assistance. These are the originals — go to them to check it.
- Threshold Models of Collective Behavior (Granovetter 1978; Schelling 1971)Granovetter & Schelling / Wikipedia
The Strength of Weak Ties
"Granovetter showed that infrequent, bridging acquaintances — weak ties — matter more for new information and job finds than close friends, because close friends overlap in what they already know."