One of the interesting things I’ve been doing with my Facebook studies (see here, here, and here for more info) is collecting data on the Freshman class over time. I believe that the Freshman can tell us a lot about technology and adoption trends, and from a broader perspective, what social and political trends may be emerging on campus. To both extents, it’s fairly easy to pull out some simple metrics on these from my collected data. I’ll present a little of that data here.
First, who uses the Facebook
Basing my data analysis on these UNC-provided figures, I can say with relative confidence that nearly every Freshman on campus uses TheFacebook. On day 1 of class, I was able to discover 3095 Freshman profiles (!!!). 5 weeks later, I was able to discover 3292 Freshman profiles. 88 percent of Freshmen have an active profile on the Facebook. Lets look at these trends over time:
Click on any graph for a larger version
This graph shows total Freshman profiles found, and the lower line in the graph indicates the 2005 Freshman profiles with no “Privacy” flags set. I’ve found that a little more than 1% of Freshmen use privacy flags. Interestingly, privacy use has declined over time, at a constant rate (Week 1: 3% privacy use; Week 3: 1.5% privacy use; Week 5: 1% privacy use). This feeds into the notion that students that do not share their profiles online are not “participating” as normal social actors.
Interesting in this data is the “Week 4 depression”, which may represent students having buyers remorse about living so publicly. This is just a guess, though.
How often are Freshmen using the facebook?
One of the questions posed to me during a recent talk was “How often are students using the Facebook”, which I roughly translate to “How often are students logging in to the system, etc”. Facebook does not provide “last access” data, though it does provide “last update” information, indicating the last time the user has updated their profile. Accepting that “updating” one’s profile places a slightly higher bar on system use than simply “logging in”, I took a look at how often students are updating their profile. The results amazed me.
Using data collected on September 27, I analyzed last update; as you can see, 2871 profiles were updated in September, with 1819 profiles updated in the last week. I was really blown away by this, so I broke out the September data by day.
We can clearly see a trend that profiles are being updated most frequently in the days preceding the data collection. Assuming that “last update” stands in as a good proxy for system use, I believe we can safely say that students are using TheFacebook extremely frequently.
Friend network expansion
One of the interesting metrics we can pull out of the data is average friend network size; in doing so, we watch the obvious and interesting aggregation of a social network on campus.
The average moves up at a fairly regular interval. For the network geeks out there, here’s data on the Freshman social network size:
Week 1: 3095 nodes, 144,260 edges
Week 2: 3188 nodes, 167,783 edges
Week 3: 3241 nodes, 198,668 edges
Week 4: 3217 nodes, 218,534 edges
Week 5: 3292 nodes, 245,883 edges
Freshmen political orientation
2423 Freshmen report a political orientation in the Facebook; this is 74%, or about 3 out of ever 4 students. Reported here are some well thought out reasons why to take political orientation data in the Facebook with a grain of salt. And to that extent, I agree completely. The notion of labeling in the Facebook may be completely arbitrary; a good amount of the data in the system represents in-jokes and the like. The students use codes in the system to communicate and display in-network group strength and identity. Stepping back from that, though, I still think its interesting to see the data. And of course, it has surprised me.
While Liberal students again outnumber Conservative students, there is suprising parity between the two factions (assuming, of course, that UNC is a liberal institution). Is this what the post-9/11 student body looks like? When political orientation becomes a face-value aspect of identity, how does it force factionization in friend-making behaviors? If we all purport to know each other’s politics, is it not hard to imagine freshmen making decisions based on this highly characteristic identity element? Please see this post for more information on this data, the pooling that goes on, etc.
Now, for some fun
First, I present what I’m calling the Happy Index. This is Freshmen relationship identification over time. I’ve decided to include all fields available in the Facebook, though the graph can more effectively be segregated along Single/Not-Single lines. It is interesting to see how Married and Open Relationship increase over time as the in-joke nature of the claim moves through the network. Since the network size changes each week, this is presented as a ratio of the total network.
Finally, I analyzed what sort of websites students are identifying as their personal websites. As we might expect, the majority of personal websites claimed by students are of the photo, blog, or profile hosting nature.
Interestingly, only six unc.edu sites were returned, with not one being a unc.edu/~ personal site. Not one freshman reports using his or her UNC provided webspace. This is alarming, but not surprising.
For questions about this data, please contact Fred Stutzman