In recent years, and with the explosion of sites like Facebook or Twitter, the idea that contagion is the main factor in the formation of networks has gained popularity. However, as I mentioned in a previous post, there have also been jobs that seem to put in doubt the mechanism of contagion (or, at least, limit their effects to be phenomena very specific).
A new article, written by Kevin Lewis, Marco Gonzalez and Jason Kaufman, offers us more material to the controversy homofilia – contagion. In particular, Lewis, Gonzalez and Kaufman we show that lto spread of certain tastes cultural in Facebook (music, movies, and books is due to the homofilia and not to contagion.
The conclusions of the article can be summarized in other words: while it is more likely that people with similar tastes about music and movies end up being friends, few interests that passed between them.
The article is from the past month of December, but I’ve recently discovered thanks to a press note of the University of Harvard.
The researchers followed a cohort of students of a university anonymous during 4 years, from march 2006 to march 2009. Each year, collected information on the profiles of Facebook of the individuals, to examine the coevolution of friendships and tastes in music, movies and books.
The data obtained in these snapshots were analyzed to develop clusters more specific, thus obtaining areas ámplias of interest for the study of its evolution and propagation in the network.
The modeling statistics of these clusters, and their evolution in time, yielded the following results:
In the case of music, people who listened to classical or jazz tended to be friends. The same thing happened with other genres, such as rock. But of these genera, only classical music and the jazz continued a pattern of contagion. Why? According to the researchers:
the one type of preference that does “spread” among Facebook friends—classical/jazz music— may be especially “contagious” due to its unique value as a highstatus cultural signal; whereas students whose friends like “indie” or alternative bands may try to symbolically distance themselves from these peers
In short: the genre classical and jazz are an indicator of cultural status, and therefore social recognition, so that it is more likely to spread to other genres that are looking for exactly the opposite effect, the social differentiation.
In the case of the movies, fans of genres such as the comedy, the gore or the sàtira also tended to interact among themselves, showing little spread among friends.
There is a result that is surprising: the study found no connection between the preference for certain books and friendship. As we said in the article, this could be explained by the factors that determine the social impact of a taste: not only imports the content itself, but also the environment, so that the taste in music might be perceived as more consequential, and therefore easier to categorize, tastes in books.
These results could have important implications for the business model of Facebook. And is that Facebook assumes that users influence each other through the expression of their tastes (“like”): the expression of these tastes would make friends were more likely to adopt those same tastes, podruciendo a viral effect (contagion). Lewis, Gonzalez and Kaufman seem to show the opposite: that only in very certain cases it might work this strategy viral.
Even so, as he says a review of the study of Lewis et al on the blog Social Commerce Today, these conclusions seems to enter in contradiction with the results of studies carried out by the company’s own Facebook (and by experimenters independent), in which it is shown that lI “like” generated dramatic increases in the influence between the same.
So, does it work the strategy is viral or not? In my opinion, the key lies precisely in the expression “among equals”. And is that the study of Lewis et al not totally rule out the influence of viral transmission of taste: what it does is limit it seriously.
As I said above, Lewis is of the opinion that the contagion takes place only in very specific instances (see press release). And it could not be otherwise: those instances are the ones that correspond to the cases in which the tastes are propagated within a network of friends created previously on those same tastes:
Our findings suggest that friends tends to share some tastes not because they influence one another, but because this similarity was part of the reason they became and remained friends in the first place
Worth to pick up the final conclusion of the authors:
our findings would support a view of contemporary online interaction as having less to do with influencing our neighbors and more to do with strengthening social ties among those whom we already resemble