Sex differences in social networks
15
F(1,25)=5.52, p< .05, d = .98; Whole-networks T2 - F(1,27)=6.13, p< .05, d = .92). Between the
core and group levels, boys’ networks gained on average between 4 and 5 additional members but
girls’ networks only increased by one member. Male cores, but not those of females, are then
expanded by additional members joining the network at cluster and group levels.
For females, the number of networks increased at each level indicating that some networks
did not have members at the core or even cluster levels. To study this in more depth we examined at
which threshold the social networks of boys and girls at both time points were established. The
majority (82%) of male networks, though only just over half (53%) of female networks, were
established with a core of members (for example, Figure 1 Networks 1, 3 and 4). Twenty percent of
female networks, as opposed to 6% of male networks, consisted of children coming together at the
25% level only (for example, Figure 1 Network 2). Children in these group level only networks
were thus loosely connected since they spent little time with each other. The remaining 12% and
27% of networks, for boys and girls respectively, were established at the cluster level. There are
therefore two different structuring patterns. All of the male social networks had a core of members
to whom others, who spend less time in the network, would join. By contrast some girl networks
spent little time together and female networks expanded only minimally in size across the three
levels indicating that members spend similar amounts of time together in the network. Because of
the limited number of male networks, statistical analyses cannot be used to examine the data further.
Of further interest is the difference in the size of play networks in comparison with the social
networks at the group level (see Table 1). Boys’ play networks were smaller, while those of girls
were larger, than their social networks at the group level. This is significant given that the play
network data could include children from other classes and suggests that boys socialised in play
networks that were smaller than their general social network.