Mapping Associations - Mapping Scientific Frontiers: The Quest for Knowledge Visualization - page 132

Fig. 3.42 A structural hole between groups a , b and c (Reprinted from Burt 2002 )

In the 1950s and 1960s Anatol Rapoport studied social networks as random
graphs (Rapoport and Horvath 1961 ). He showed that if the placement of edges was
not completely random, it could produce a graph with a lower overall connectivity
and a larger diameter. Sociologist Mark Granovetter ( 1973 ) argued that it is through
casual acquaintances, or weak ties, that we obtain new information, rather than
through strong ties, or close personal friends. The weak ties across different groups
are crucial in helping communities mobilize quickly and organize for common
goals easily. In this vein, Ronald Burt ( 1992 ) extended the strength of weak ties
argument and argued that it was not so much the strength or weakness of a tie that
determined its information potential, but rather whether there was a structural hole
between someone's social network. A structural hole can be seen as a person who
has strong between-cluster connections but weak within-cluster connections in a
social network. Figure 3.42 illustrates two persons' connections in a social network
(Burt 2002 ). While both Robert and James have six strong ties and one weak tie,
Robert is in a more informed position than James because much information for
James would be redundant. Robert, on the other hand, is a bridge to cluster A and C.
Therefore, the number of connections in a social network is important, but the value
of each connection depends on how important it is for maintaining the connectivity
of a social network.
The degree of separation between two people is defined as the minimum length
of such chains between them. In a graph, this is equivalent to the diameter of the
graph. You may have heard that everyone on Earth is separated from anyone else
by no more than six degrees of separation. Normally, the social world we know is
confined to a group of our immediate acquaintances; most of them know each other.