For those interested in open-source systems (everything from public policy to software) should listen to Christopher Allen’s talk (his blog here) on the Dunbar Number.

Dunbar’s number, which is 150, represents a theoretical maximum number of individuals with whom a set of people can maintain a social relationship, the kind of relationship that goes with knowing who each person is and how each person relates socially to every other person.

Malcolm Gladwell brought the Dunbar number into popular discourse when he referenced it in his book The Tipping Point.

However, Allen’s talk tries to nuance the debate. Specifically, he wishes that those who reference the Dunbar number would be more aware that in he research literature, the mean group size of 150 only applies to groups with high incentives to stay together. As examples he cites nomadic tribes, armies, terrorist organizations, mafias, etc… in short, groups in which mutual trust and strong relationships are essential for survival. This is in part due to the fact that there is a cost group members must pay to maintaining this groups of this size: one must spend 40% of ones time engaged in “social grooming.” This means sitting around listening to one another, talking, being engaged, etc… Without this social grooming it is difficult to develop and maintain the unstructured trust that holds the group together.

More interestingly Allen’s research suggests that in modern groups there is a correlation between group satisfaction and the size of the group. Things work well between 3-12 people and from 25-80. But in between there is this hole. Groups in this “chasm” are too be too big to use many of the tools (like meetings) that small groups can use, but too small to successfully rely on the tools (such as hierarchies and reporting mechanisms) that allow larger groups to function.

Open source projects (and really any new project) should find this interesting. There is a group size chasm that must, at some point, be crossed. When I’m less tired I will try to wander over to sourceforge and see if I can plot the size of the projects there to see if they scale up nicely against Allen’s graph.

In addition, I’m curious as to whether some softer skills around facilitation would allow groups to function more effectively, even within this “chasm.”