Half a network turns over in seven years

Gerald Mollenhorst, Beate Völker and Henk Flap repeatedly interviewed the same people in the Netherlands, collecting at each wave the detailed composition of their personal network, as well as the context in which each relationship had been formed.

That design makes it possible to measure not the size of the network, which varies little, but its content: who leaves, who enters, and at what rate. One-off surveys, which take a snapshot of a network at a given moment, cannot see that movement.

Over a seven-year period, a substantial share of the people named at the first wave are no longer named at the second, while the overall size of the network stays comparable.

In other words, the network maintains its volume through replacement. That result has been found in other longitudinal work, at rates varying by age and context. Barry Wellman and colleagues had already observed substantial turnover in their repeated surveys in East York, a suburb of Toronto.

The authors note that turnover follows changes of context: moving house, changing job, a child arriving, finishing studies. Relationships born in a context disappear with it.

That finding echoes work on proximity: what sustains a relationship is largely the undeliberate repetition of encounters, not a decision to maintain it.

It also extends the focus theory proposed by Scott Feld in 1981, according to which ties form around shared places and activities: a team, a club, a stairwell. When the focus breaks up, the ties it organised lose their support.

A small part of the network is stable across the whole period, and it consists mainly of family ties and a small number of old friendships.

Those durable ties are characterised less by contact frequency than by the capacity to resume after an interruption, a property surveys measure poorly because they ask about frequency. A friendship seen once a year can thus be more stable than a daily office relationship.

It does not describe deterioration. High turnover is also a sign of exposure to new contexts, and periods of complete stability often correspond to a reduction in opportunities. Neither configuration constitutes a norm.

Above all it invites a distinction between two questions public debate conflates: the size of the network, fairly stable across adult life, and its identity, which changes a great deal. A circle can stay the same size while being made up, a few years later, of largely different people.