Friday, 8 May 2009

do virtual group members crowd out real ones?

Since Robin Dunbar had put his group size theory forward (an early summary is his Grooming, Gossip book, for instance), we seem to know about the 'natural' group size. Some fifteen years ago, Vili and I had long conversations about the modern media creating 'quasi-archaic virtual group members'. People gossip about politicians, celebrities, soap opera characters as if these were their friends, in fact as if these were members of their own Dunbarian groups.

This week I got to go to a talk of Robin Dunbar, in which he had put forward the latest version of his social brain hypothesis. Which in turn reminded me of an old question that we were wondering about back then: do these virtual groupmembers push out real-life friends from the inner circle? If the group size is limited to, say 150-200 people, then following the life of a large number of public figures perhaps limits our capacity to gossip about and groom with members of our own real group. Or maybe not.

I asked Robin Dunbar about this, he said they never checked.

Anybody with an answer? (Or bother to check it at that?)


Checking would probably be rather straightforward: get a set of people, for each of them measure media consumption, assess virtual group size, assess real group size. Regress.


If this hypothesis was to turn out to be confirmed, then the consequences might be interesting. Obviously: what is the effect on social cohesion? But also: to what extent do these virtual members affect the social identity of people? Probably there is a language barrier. Is that so? Is there a growing set of globally shared virtual group members (how many people on the planet have seen the puppy of the Obama's or know about the Obama date nights)? Is anyone engineering these virtual group members as such, or the celebrity PR merely evolved that way? And many more questions, I presume.


  1. Professionally designed virtual group membership, exhibit A.

  2. I think two things could be distinguished here:

    1. the possible crowding out due to the limited social capacity of each individual that needs be shared between 'virtual' and 'real' -- that sounds straightforward indeed..

    2. the possibility that 'real' group cohesion might become even stronger through a set of 'virtual' members, known by all the 'real' members providing a common reference group -- eventually even integrating previously isolated 'real' groups, and hence increasing the actual size of 'real' groups. I am not sure about this, but if it is so, this effect is clearly different from the first one, I think

    In terms of methodology, I suppose, that would imply controlling for the overlap between the virtual group of each person in your set.