Relationships born online, twenty years of results

The displacement hypothesis held that time spent online would substitute for existing relationships. The stimulation hypothesis held the opposite: the tools would allow more ties to be maintained.

Early studies, in the late 1990s, seemed to support the first. A 1998 American study led by Robert Kraut and his colleagues, reporting an association between internet use and isolation, was widely taken up, before being partly contradicted by its own authors in a later follow-up.

At the same time, field studies described the opposite. Keith Hampton and Barry Wellman's study of a wired neighbourhood in suburban Toronto found that connected residents knew more of their neighbours than the others did, a result consistent with the second hypothesis.

As use became general, comparing users and non-users ceased to make sense: non-users became a particular population, older and less educated.

The question therefore shifted towards types of use. Studies now distinguish active use, exchanging with identified people, from passive use, consulting content. This is the second change of answer announced at the outset: the question is no longer whether one is online, but what one does there.

Several studies find associations of opposite sign by type of use, active use being associated with a more favourable reported state and passive use with the reverse.

Those results are mostly correlational, and the few available experiments cover short durations and student samples, which limits their reach. More recent critical reviews question the distinction itself, noting that the same actions may fall under either kind of use depending on the person.

Surveys find that relationships born online often migrate to other channels, and that those remaining exclusively online are described as less close on average. Malcolm Parks and Kory Floyd already observed this in 1996 among participants in discussion forums.

That average covers very different situations, notably for people whose offline opportunities to meet are constrained by health, geography or belonging to a locally underrepresented minority. For them, the channel says nothing about the importance of the tie.

Platforms hold precise contact data but do not publish it, and analyses conducted with their cooperation raise questions of independence and consent.

Independent surveys, for their part, rest on reported use whose agreement with recorded use is poor, as the meta-analysis by Douglas Parry and his colleagues established in 2021. The field thus finds itself in an uncomfortable and durable position.