The entries in this subject share a single idea with large consequences: what a group does depends on how its members are connected, not only on what they individually believe. The same people wired together differently will spread an idea or smother it, cooperate or fragment, and see themselves accurately or not at all.
How networks form
The structural findings describe the shape most human networks take. Homophily finds ties forming mainly between similar people, and Triadic Closure finds friends of friends becoming friends. Preferential Attachment explains why a few nodes collect most of the links, and Hub Vulnerability warns that networks shaped this way survive random failures and collapse when their hubs are attacked. The Watts–Strogatz Model shows how a few long-range links make even a clustered world small.
How things spread
A second cluster concerns diffusion. Rogers's Innovation Diffusion Curve and the Bass Model describe adoption moving from a few innovators to the majority. Granovetter's Strength of Weak Ties finds new information and job leads arriving through acquaintances rather than close friends. Complex Contagion adds an important distinction: a simple piece of news spreads through a single contact, but a costly behaviour needs reinforcement from several. Critical Mass, the Tipping Point and the Percolation Threshold all describe the moment when slow accumulation becomes sudden change. Taleb's Minority Rule shows a small inflexible group imposing its preference on a flexible majority.
How crowds misread themselves
The most striking entries concern perception. The Friendship Paradox shows that most people's friends have more friends than they do, because popular people are overrepresented in everyone's circle. The Majority Illusion shows a minority view looking like the majority when its holders are well connected. Kuran's Preference Falsification explains why regimes and norms can collapse overnight: when people hide their private views, nobody knows how many others share them, until a small shift reveals it. Information and Availability Cascades show people rationally following others until the crowd's apparent consensus rests on very little evidence.
What happens to communities
The community entries record how groups change as they grow. Eternal September, named for the month Usenet was opened to AOL's users, describes norms overwhelmed by a permanent influx of newcomers. Evaporative Cooling of Group Beliefs shows moderates leaving first, so a group grows more extreme as it shrinks. Griefing and the Geek Social Fallacies describe members who damage a shared space, and the group norms that make removing them hard.
Where it applies
The practical uses are wide. Public-health campaigns use complex contagion to decide whether to seed a behaviour widely or in dense clusters. Job seekers are better served by acquaintances than by close friends. Community managers meet Eternal September and evaporative cooling as soon as a forum grows or a dispute drives members away. Engineers designing roads and power grids meet Braess's Paradox, in which adding capacity makes a network slower, and hub vulnerability, which argues for redundancy at the points everything depends on. Businesses meet the Network Effect and its competing valuation rules, Metcalfe's, Reed's and Odlyzko's.
Where it is argued over
Several famous claims in this subject rest on less evidence than their fame suggests. The Small-World Phenomenon rests on a 1969 study in which only about a fifth of chains completed. Dunbar's Number has a confidence interval so wide that reanalysis finds it compatible with group sizes from a few dozen to several hundred. Critics of Metcalfe's law, Odlyzko among them, argue that it overstates network value, since not all connections are worth the same. Social Capital and Network Closure are argued over against Structural Holes, which values exactly the gaps closure fills. The Condorcet Jury Theorem and the Miracle of Aggregation hold only when errors are independent, and researchers of media environments find that correlated errors are common.
Why it is worth knowing
Most explanations of collective behaviour look at what individuals think. This subject shows how much depends instead on the wiring between them: a well-connected minority can pass for a majority, a hidden majority can stay silent for decades, and a crowd's confidence can rest on a handful of early choices. Knowing that makes surprising collective events, from sudden revolutions to viral products, less mysterious, and makes a person more cautious about reading the mood of a crowd from inside it.
The entries behind this essay: Networks & Collective Behavior. Every subject has its essays on the essays page.