Mike Thelwall is Professor of Information Science in the School of Computing & Information Technology at the University of Wolverhampton. He has recently published a fascinating piece of "work in preparation" about MySpace users - Social Networks, Gender and Friending: An Analysis of MySpace Member Profiles [250 kB DOC], which he is happy for people to quote and cite.
The analysis uses two nice big data sets, and a third, much smaller one. The large data sets were obtained in the following interesting way:
"The raw data for this article are three samples of MySpace public user profiles. Each MySpace user has a personal identification number, and these numbers are apparently given out in sequence. We identified approximately the last ID issued by MySpace on July 3, 2007 and for the first collection selected every 10,227th ID starting at 1,939 (a random starting point) to give a large total sample size (20,064). The profile page associated with each of these user IDs was then downloaded via the URL http://profile.myspace.com/index.cfm?fuseaction=user.viewprofile&friendid= followed by the user ID. These profiles were downloaded at a rate of 8,000 per day over three days to avoid overloading the MySpace server. This is the 'all members collection', and encompasses MySpace users from a wide variety of joining dates. "
"The second collection is the 'July 3, 2006 members collection' and consists solely of users who joined on July 3, 2006, as identified by trial and error through MySpace ID ranges, selecting 10,000 IDs (90,306,349 to 90,316,348). These were downloaded over four hours on July 17, 2007 starting at 2.30am central U.S. time in an attempt to access the data when most users were asleep in order to (a) minimise impact upon the MySpace servers and (b) capture comparable data."
The article draws many useful insights, and represents the data visually with great clarity. Here is a small selection:
- ".... (half of the users had not checked their account for at least two and a half months), because they had given up MySpace, had a pattern of infrequent checking, or had switched to another MySpace account ."
- "Excluding users with 0 or 1 friends on the grounds that they are probably mainly inactive users, the median length of time since the last logon was approximately one week (8 days)."
- "In simple terms, it seems that about a third of members give up immediately, a third become regular users and the rest log on occasionally."
- Users lie about their ages: "The surprisingly high number of centenarians could be due to children signing up their oldest living relatives but certainly reflects joke ages for at least some: several claimed to want children 'someday'; and there are examples in the data like 'Kevin' (apparently 102) who appears to be about 13 from his picture and has a typical teen action-packed MySpace profile."
- "A large majority (77%) of users were from the U.S., although – surprisingly - the UK was in second place (5%), followed by Canada (2%), Australia, (2%), the Philippines (1%) and Mexico (1%)."
- "The median number of friends is 1, and the median number for users who have at least 2 friends is 27."
Thelwall concludes:
"If there is a typical MySpace user then she is probably 21, single, with a public profile, interested in online friendship and logging on weekly to engage with a list of mixed but majority female 'friends' that are predominantly acquaintances (including bands). This is a very different profile to those of the teen users frequently discussed in the contemporary press, and so it seems likely that public perceptions of MySpace are at variance with reality."
Note. After writing this piece I came across dana boyd's extensive and up-to-date Research on Social Network page, which contains a link to Thelwall's paper, and a lot of other material besides.
Comments