How Good a Friend Is a Last.fm Friend?

Tunes that Bind? Predicting Friendship Strength in a Music-Based Social Network
To be presented at Internet Research 9.0; Copenhagen, 2008.

“Friendship” is an inherently ambiguous relational descriptor. In social network sites, where “friend” is often the only word available to label relationships, the ambiguity seems only to be enhanced (e.g. boyd, 2006; Fono & Reynes-Goldie, 2006; Gross & Acquisti, 2005). This paper seeks to shed light on the nature of “friendships” in one social networking site. Founded in London in 2005, Last.fm functions as both a social network site and a music recommendation, streaming and, to a lesser extent, downloading service. In May 2007, when it was bought by CBS Corporation for US$280 million, Last.fm boasted more than 15 million active users in hundreds of countries. To our knowledge, there has not been any academic study of social dimensions of Last.fm.

The data reported here come from an international survey of Last.fm users. The 559 respondents (36.5% female, 63.5% male) from 48 countries were recruited through messages posted to Last.fm’s two general interest site-wide discussion forums. Each time one opens a Last.fm profile page, one’s friends list appears in a random order. Participants were asked to open their profile in another window and think about the first person on that list in answering a series of questions about their relationship. After assessing a number of baseline facts about Last.fm friendships (number of friends, proportion that began on Last.fm, average length of Last.fm friendships), we conducted a 4-step multiple regression analysis to determine the predictive value of four sets of variables on relational strength.

We measured relational strength with the scale used by Chan, Cheng, and Grand (2004), a shorter version of that created by Parks and Floyd (1996). This 18-item scale assesses six of the dimensions Parks (2007, p. 27) argues, “constitute a definition of the relational change process.” These include interdependence, depth or intimacy of interaction, breadth or variety of interaction, commitment, predictability and understanding, and code change and coordination. Because scores on each of these dimensions showed high intercorrelation, we treated the scale as a single measure of relational strength.

We examined four sets of variables’ correlations with relational strength, controlling for each previous set with the introduction of the next set. First we considered demographic factors including age, gender, and geographic location. Second, we looked at partner similarity (homophily) in terms of those demographic variables and musical taste. Third, we addressed the extent to which relational partners use media other than Last.fm (face-to-face, telephone, text messaging, email, chat, instant messaging, communication via other websites, and postal mail) to communicate. Finally, we examined whether communicating via Last.fm itself correlates with relational strength above and beyond communication via other media.

We found that on average, the relationships were of moderately low strength, just below the midpoint on the scale. Last.fm friendships were likely to be stronger when (1) the partner was female, (2) the relationship was between partners of different sexes, (3) the partners did not meet through Last.fm, (4) the partners also communicated face-to-face, on the telephone, through text messaging, via email, via IM, or on another website, and (5) the partners communicated via Last.fm. Homophily, even in musical taste, did not correspond to friendship strength except in the case of sex, where it lessened relational strength. Chat and postal mail did not correlate with relational strength.

These results suggest that Last.fm – and likely other social network sites – serves as just one node in stronger relationships. By itself, Last.fm does not seem to lead to strong relationships. As a relationship-formation site, it fosters weak ties. However, in conjunction with other modes of communication, it may enhance already strong partnerships. The findings lend further support to Haythornthwaite’s (2005) theory of “media multiplexity,” in which she argues that the number of media through which people communicate should be added to the definition of “strong ties.” Our results also demonstrate the importance of considering diverse modes of online interaction separately, as well as examining how their use is interwoven.

References

boyd, d. (2006). Friends, Friendsters, and MySpace Top 8: Writing community into being on social network sites. First Monday, 11 (12). http://www.firstmonday.org/issues/issue11_12/boyd/

Chan, D., Cheng, K.S. & Grand, H.L. (2004). A comparison of offline and online friendship qualities at different stages of relationship development. Journal of Social and Personal Relationships. Vol 21(3), 305-320.

Fono, D., & Raynes-Goldie, K. (2006). Hyperfriendship and beyond: Friends and social norms on LiveJournal. In M. Consalvo & C. Haythornthwaite (Eds.), Internet Research Annual Volume 4: Selected Papers from the AOIR Conference (pp. 91-103). New York: Peter Lang.

Gross, R., & Acquisti, A. (2005). Information revelation and privacy in online social networks. Proceedings of WPES’05 (pp. 71-80). Alexandria, VA: ACM.

Haythornthwaite, C. (2005). Social networks and Internet connectivity effects. Information, Communication, & Society, 8 (2), 125-147.

Parks, M.R. (2007). Personal Relationships and Personal Networks. Mahwah, NJ: Lawrence Erlbaum Associates.

Parks, M. R., & Floyd, K. (1996). Making friends in cyberspace. Journal of Communication, 46(1), 80-97.