One of the big questions raised by social networking sites is what the heck those “friendships” really are. In this paper, written with my former Ph.D. student and now Ohio University professor Andrew Ledbetter, we examine this in the case of Last.fm. Based on a large survey of users, we pose the question of what predicts how strong or well developed Last.fm friendships are.

The short answer is that the best predictor is not shared taste in music (which has no effect on relational development), but how many different ways people communicate with one another. For each medium added, people’s relationships are a little closer. This means that sites like Last.fm can provide pairs with an additional way to maintain and strengthen their relationship that goes over and beyond what they get through email, instant messaging, phone calls and other means of interaction.

On the other hand, the “friendships” that begin via Last.fm don’t go very far, even if shared taste was important to the relationship’s initiation.

Overall, the “friendships” on Last.fm are pretty weak. The notion that shared taste makes people “musical soulmates” makes for good mythology, it seems, but not strong interpersonal connections.

You can download and read the paper here. For reasons I don’t understand, the tables did not get included in this PDF. If you really want to see them, email me.

If you’re in Copenhagen this week for Internet Research 9.0, drop on by and hear this presented live in person.