I’ve long been interested in Dunbar’s Number, an estimated maximum number of people that a person can maintain a stable relationship with. (See Neocortex size as a constraint on group size in primates. Put simply, our minds are only wired to be able to maintain relationships with around 150 people at a time.

Some people have suggested that sites like Twitter, where I currently follow about 2900 people and have over 3,350 followers, Facebook, where I am approaching 2000 friends, and even LinkedIn where I’m approaching 700 connections is a reason to rethink Dunbar’s number. Perhaps technology gives us the ability to maintain broader relationships than our neocortex permits.

However, a recent article, Validation of Dunbar's number in Twitter conversations explores the nature of twitter conversations.

We find that users can entertain a maximum of 100-200 stable relationships in support for Dunbar's prediction. The "economy of attention" is limited in the online world by cognitive and biological constraints as predicted by Dunbar's theory.

Basically, they analyze a mass of tweets and find that as the number of people a person is connected with on twitter, the number of people they regularly communicate with starts dropping off somewhere between when they reach 100 and 200 people. They conclude

In this paper we show that social networks did not change human social capabilities. We analyze a large dataset of Twitter conversations collected across six months involving millions of individuals to test the theoretical cognitive limit on the number of stable social relationships known as Dunbar's number. We found that even in the online world cognitive and biological constraints holds as predicted by Dunbar's theory limiting users social activities.

I think they are significantly overstating things. If anything, their paper shows that for a very large sample of uses of a particular social network the human behavior around conversations did not significantly change as the result of using the technology. The lack of observed change in a behavior in a specific time with a specific tool does not mean that capabilities haven’t or are not changing. It may well be that there is something about Dunbar’s number that is immutable, even with the use of technology. This paper just doesn’t show it.

As I thought about it, I also thought about things like Klout and PeerIndex; tools aimed to measure influence. Online influence varies greatly. This might be because online communications isn’t just conversational, but there is a broadcast element as well. If we are in fact limited by the number of people we can converse with, it is an interesting topic for people interested in social media, especially those focused on the value of conversations over broadcasts.

Another thought that I had was about how many patients a typical doctor sees during a year. Numbers that I read suggest that the best ‘patient panel’ for a primary care doctor is in the range of 1800-2000 patients, way past Dunbar’s number. Should our conversations on social media be more like a broadcast? An informal conversation? A Doctors’ Appointment?

What do you think?