Long post coming up:

One of the web pages I frequent is Edge. According to its founders, it is supposed to be a meeting place for “the third culture”. This stems from CP Snows classic “The two cultures”, where he bemoans the gulf between the social and natural sciences. Edge attempts to bridge the gap. The site is great, even if what they are championing is one culture’s invasion of the other. Just look at all the geneticists “proving” the genetical cause of all sorts of social phenomena [1].

Today, their main article is one on the dispersion of happiness in social networks. The whole article, which is published in the British Medical Journal, can be found here. Basically, it states that people are affected by others’ mood, and goes on to map out one network and quantify to what degree this is true. Supposedly, a close friend who becomes happier increases the chance of a similar development in you with 25 %. The same number for a coresident spouse is 8 %, siblings living nearby 14 % and next door neighbours 34 %, all with 95 % confidence interval.

Some thoughts and questions arise from this. First, for my own part, I realise I have to look into network analysis as method. Some of the results here seem somewhat ridiculous to me (Coresident spouse has the least effect on happiness? One fourth that of your neighbour?), and the whole idea of putting this in terms of probabilistic measures is very foreign to me. The question is how this method works, and what it is safe to use it for. I’ll ask Thomas Berker here at the institute, since he’s working on network analysis of real estate advertisement (a much safer venue, I think) [2].

Second, there is a central point in the article that I have trouble swallowing. According to the authors, happiness can be distributed by up to three degrees’ distance. That means that your happiness can be afflicted by the friends of your friends’ friends’ happiness, even if you don’t know them at all. Now, it’s obvious that this happiness effect must be distributed through the friends you actually know, but there is no discussion of this in the paper. Is there a linear distribution? Does the happiness increase disperse equally through the network? Do the causes of happiness increase matter for the dispersion? In fact, I don’t see how it is possible to measure that sort of distributed effect at all.

I’m not the only one thinking about these issues. According to this New York Times article (NYT again? That’s two in a row, mister),

A study also to be published Friday in BMJ, by Ethan Cohen-Cole, an economist at the Federal Reserve Bank of Boston, and Jason M. Fletcher, an assistant professor at the Yale School of Public Health, criticizes the methodology of the Christakis-Fowler team, saying that it is possible to find what look like social contagion effects with conditions like acne, headaches and height, but that those effects disappear when other factors are considered.

A more speculative method than the actually quite sober longitudinal study the authors use as main source (even if it is based on self reporting of emotional issues) is their treatment of social networking online. They want to see how internet affects the effect proximity on happiness distribution. How do they do it? By counting smiling and frowning people on Facebook. Here is a graphical representation of the network of smiling and frowning people:


Blue nodes: frowning | Green nodes: neutral | Yellow nodes: smiling

Smiles in a photo, for example while drunk at a party, as measure of general happiness? What about people who don’t like having pictures of themselves on these sites? They must be unhappy.

Anyway, I still think this is interesting. I’m looking forward to checking out this methodology, but it seems difficult to say anything of substance that isn’t riddled with problematic assumptions.

Did you make it all the way down here? Good boy/girl. Here’s your treat, a ridiculously overblown analysis of Star Wars aestethics. I love it.

[1] Bruno Latour warns us of using the term “social” as an explanatory variable for anything, and I see his point, but I’ll do it anyway. Sue me.

[2] By the way, check out Thomas’ book on Internet use. My German is a little rusty, but I’m not sure whether I would consider the Internet itself in a media analysis perspective at all. I’m sure he discusses the difference between the Internet in general and Internet media. I would have chosen a better cover, however…


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s