editor's notebook If you took the countless number of Facebook "status updates," or out-loud thoughts, that are plastered on the site every day and put them under a microscope, what patterns would emerge from these many billions of words? What glimpse might we get into the hearts and minds of the social network's millions of users?
Facebook's data-crunching team decided to take a look this week, and it gleaned some insight into the different ways in which older and younger, many-friended and more-intimate members express themselves.
The team collected about 1 million status updates generated by U.S. English speakers, anonymized them, and picked them apart word-by-word with the help of a text-analysis application called Linguistic Inquiry and Word Count. LIWC was originally developed, with the help of the National Institutes of Health, to aid the study of how written and verbal language reflects mental and physical well-being.
The Facebook team used the LIWC dictionary and its many word groupings to dissect the status posts. The dictionary, for example, has a "psychological processes" category, with an "affective processes" subcategory and a "positive emotion" subsub--which includes words such as love, nice, and sweet. There's also a "swear words" subcategory--under the larger "linguistic processes" heading--that includes words I won't repeat here.
The resulting data-dumps provide ample opportunity for musing about the Facebook phenomenon. The data group looked, for instance, at status updates that had triggered comments from readers or inspired them to hit the "Like" button, and it noted the word types most often used in such updates. People seem to "like" updates that include sexual words. Words related to death--not so much. No real surprise there. The team, however, also found that members with a lot of friends tend to use different types of words than members with fewer friends.
My cohorts in the media have made much of this, with headlines like "Facebook Unveils Secrets to Being Popular Via One's Updates." No real harm in that; still, I couldn't help but look at it the other way. For example, according to the data team's results, the Facebookers with many friends tend to use fewer "emotional words" than do members with less friends. I'm not sure this means that people flock to those who are unemotional; it could just as easily mean that people who tend to form deeper, more-emotional relationships use Facebook in a different way (or not at all)--i.e., that "popular" Facebookers, with more "friends," form shallow connections, or indeed, that the Facebook platform itself, as Zadie Smith recently suggested in The New York Review of Books, encourages shallowness.
At any rate, whether Smith is right or wrong--or whether my own little musing makes any sense--the Facebook Data Team's results are fun to look at and use as fuel for your own ideas about Zuckerberg's monster (and people in general). I'm appending them here, and inviting you to offer up your own heady theories--serious or not--in the comments section.
(A list of LIWC word groupings can be found here. The Facebook Data Team noted other interesting tidbits as well--negative emotional updates receive more comments than do positive emotional updates, for instance. The team's full post can be found here.)