Psychology, Department of


Date of this Version



Published as: Pachur, T, Schooler, L.J., & Stevens, J.R. (2013). When we meet again: Regularities in social contact dynamics reflected in memory and decision making. In R. Hertwig, U. Hoffrage, & the ABC Research Group (Eds.), Simple heuristics in a social world (pp. 199-224). Oxford: Oxford University Press.


Copyright 2013 Ralph Hertwig and Ulrich Hoffrage Published by Oxford University Press. Used by permission.


In 1967, the social psychologist Stanley Milgram reported a fascinating observation. For a study on social connectivity, he randomly picked several individuals in Wichita, Kansas, and asked them to get a message delivered to a particular target person in Cambridge, Massachusetts—a distance of more than 1,300 miles. There were two catches. First, Milgram provided no address but only the name and some basic information about the target person, such as the area where he lived and his profession (the target person was the same for everybody). Second, people could only forward the message to someone they knew on a firstname basis. That is, the task for each person in the chain was to pass on the message to someone she thought would be most likely to know the target person or know someone else who might know him. The astonishing result was not only that many of the messages reached the target, but also that the number of intermediaries needed turned out to be rather small, ranging between 2 and 10 (with a median of 5).

Milgram’s (1967) findings became known as the small-world phenomenon or sixdegrees of separation because it takes six separate journeys to get from the sender to the target via five intermediaries. His results suggest that the social world possesses regularities that allow most people to reach any other person using a relatively small number of steps. Researchers have only recently begun to understand the nature of these regularities in greater detail. Watts and Strogatz (1998) identified conditions under which networks are both “small worlds”, in which the average number of connections between all agents in the network is rather low, and, at the same time, display tight clustering consisting of densely connected subgroups. Specifically, some connections must link members of a cluster to a randomly selected member from the entire network. Numerous investigations have followed, exploring the nature of social environments.