Relationships on collaborative websites are far from being fully reciprocal: between any two people, it is very often the same one who most often initiates the interaction. This 'imbalance' in relationships can be traced back to asymmetry in user profiles.

We tend to think that the relationship between two users of a social networking site such as Facebook is either very close – among family or friends – or virtually non-existent, i.e. with fairly distant acquaintances. However, according to a team of six researchers at Notre DameUniversityin Indiana, USA, two people may be linked, but only one of them will tend to initiate communication with the other, without this causing the relationship to break down.  The researchers carried out a study with a sample of Internet users, analysing the type and number of messages sent by each user to each of their contacts.

Relationship imbalances arise from profile differences 

So, although the usual models predict that all social relationships must be reciprocal if they are not to collapse, it now seems that social networks are in reality full of unequally weighted relationships, where one member of the duo might kick off up to 90% of the interactions. Most of the time this disparity has the same root cause: asymmetric profiles of the two users. Moreover, the researchers came up with a second finding: there is a causal link between the degree of profile similarity of two users and whether or not their relationship is equally balanced. For example, a wide difference in connection habits or the number of contacts each member has may lead to a significant imbalance between the two – the one with fewer contacts will be the more active.

Focussing activity on a small number of contacts

Another notable characteristic is the predominance of some contacts. Ideally, in order to form an equal relationship, there would be homogeneous “distribution” of network activity with each of a user’s contacts. But in reality users focus more than 80% of their activity on a small number of their neighbours - those with whom the "relationship nodes" are the strongest. This is where the disparity of interaction between contacts arises. According to the researchers, the study enables them to set up a predictive model of the workings of a social group. If we are talking about a group where the members value profile homogeneity, there is a strong chance that disparity of interactions will be fairly low. On the other hand, groups which promote diversity between members will tend to favour a high degree of imbalance in their interactions.