If those riding intellectual fads are sometimes guilty of sloppy reasoning, imagine what happens when two fads collide.
That’s what happened when the British Medical Journal elected to publish a study analyzing 1) happiness in 2) social networks. The study, by James Fowler and Nicholas Christakis, concludes that happiness is contagious within social networks.
According to the authors, your happiness depends on the happiness of your friends, and their friends, and their friends. It’s a fascinating finding, and it was duly reported by hundreds of newspapers. Indeed, according to Fowler, “if your friend’s friend’s friend becomes happy, that has a bigger impact on you being happy than putting an extra $5,000 in your pocket.”
Unfortunately, it’s probably not true. Here’s the crux of the research: the authors show that your happiness is positively related to the happiness of your friends, and that this holds even after accounting for a number of other variables, including how happy you and your friends were a few years back. That’s correlation; what about causation?
There are (at least) three reasons why happiness is correlated within social networks. It may be that — as the authors posit — happiness is contagious. Or perhaps people with similar dispositions are more likely to be friends. Economists call this the confounder “selection effects,” while medical journals call it “homophily.” The authors partly account for this by adding statistical controls for the past happiness of both you and your friends.
The third reason is perhaps the most likely: if you and I are friends, we are often subject to similar influences. If a buddy of ours dies, we’ll both be less happy. Or, less dramatically, if our favorite football team wins, we’ll both be happier. But this isn’t contagious happiness — it is simply a natural outcome of the shared experiences of people in the same social circles. Unfortunately, observational data cannot distinguish the headline-grabbing conclusion — that happiness is contagious — from my more mundane alternative: friends have shared emotional influences.
Interestingly, the same issue of the BMJ contained a very careful article by Ethan Cohen-Cole and Jason Fletcher making precisely this point. They employ a pretty cheeky research strategy: if you want to show that a research design is silly, show that it leads to silly conclusions.
They use Fowler and Christakis’s approach on another dataset, and show that it leads to the unlikely conclusion that height, headaches, and acne are also contagious. The more likely explanation, of course, is that all are subject to similar environmental influences. For instance, the same jackhammer causing your headache is likely causing mine.
I bet that a similar analysis would show that stories about happiness being contagious are, well, contagious. After all, what else explains last week’s epidemic, with stories in The New York Times; The Boston Globe, and The Washington Post? Of course, it may just be that this “epidemic” reflects a shared environmental influence, like each newspaper receiving the same press release.
So we have two studies drawing two conclusions. The first finds that happiness is contagious; the second finds that researchers can too easily draw false conclusions about contagion. Guess which one grabbed the attention of headline writers.
and 
. You can also watch it on 
Entirely unscientific and anecdotal support for the study’s conclusion: I am an introvert and a curmudgeon. I enjoy being around happy-go-lucky people. They make me feel less Eeyore-like. Indeed, their happiness is, well, contagious.
That said, I’m sure there are many curmudgeons that would rather be alone all the time.
GREAT article. That made me happy.
Now that my skeptic hat is on, though, it should be pointed out that Cohen-Cole and Fletcher measure reported heights, headaches, and acne. The former is likely to be fairly objective, and by itself is probably dispositive. The latter two could, however, be plausibly “contagious,” inasmuch as both the psychological perception of, and reports to doctors of, headaches and acne might indeed be affected by others around you complaining of the same thing. I’ve certainly known people who were great complainers about headaches and caused those around them to notice aches they might otherwise have ignored. As as for acne, the self-consciousness felt by those with mild acne is highly affected by what those around them are saying — just go to an acne self-help website and see. So though I agree with their conclusions, the debunkers should have chosen two additional variables more like height to fill out their study, since otherwise the true skeptics might dismiss both Fowler’s and their studies and say, well, that’s the nature of statistics — I still don’t know anything one way or the other.
I’ll admit to not having looked at the details of the study.
I a few things to mention:
Merely thinking of most of the people on my Facebook friends list tend to make me happy. Most of them are from college, and we had some great times.
Could those associations have cued a responce in those taking part in the survey?
Also, is there a positive correlation between college graduates and happiness, perhaps even a causal relationship?
Again, I don’t know how the study was administered. You’d think they’d control for those types of things, but given the “explanations” for co-workers, spouses and neighbors it would not surprise me if these things were not properly considered.
I heard a scientific proof and have experienced myself that the meal one has is not only determined by the quality of it (product used), the presentation of the food, the taste is actually greatly affected by the environment – that is, whom s/he is having the meal with. Others saying how the food is delicious, which make the individual’s neuron to fire – “yummy” signals (in addition to the smell and other stimulus) to the brain – thereby making the person enjoy the food more than the actual response. Hence, surrounding people – environment – affects others. This is particularly true about happiness – happiness IS contagious! Everybody likes and wants to be happy, rather than miserable, and angry. People who are “jolly” are popular or liked by the majority. They would usually say nice things about other which make the person happy which gives him an extra “margin” to be kind to others, do them a favour. This is a cycle that keeps going on. Hence “happiness” is contagious as it keeps repeating.
Economists dismiss social influence as a cause because they ignore causal studies in psychology. The issue isn’t if social influence is a cause, because decades of behavioral experiments show it is. The real issue is if the effect size is great enough that happiness spreads in a major way outside the lab. Clearly, the stochastic model in the C&F study has flaws, but so do the vast majority of observational studies that use panel data instead of experiments.
The author cites interdependent sampling as an alternative explanation to social influence, when I see it as a type of social influence.
My question is, why don’t economists use their considerable modeling skills to address social influence head on, controlling for it with models, instead of irrationally pretending it doesn’t exist!
Why don’t you let your readers know that the authors of the original study (Fowler and Christakis) have responded quite ably to the concerns raised by Cohen-Cole and Fletcher? They have a supplement to the happiness study that discusses some of these issues available here:
http://tinyurl.com/5v2cdg
Also, this isn’t the first time that Cohen-Cole and Fletcher have been critical of Fowler and Christakis; the authors put together a response to that previous criticism as well, which can be found here:
http://tinyurl.com/5lkhks
I’m surprised that t one of the authors contact this blog and ask that you make your readers aware of these responses… or did they?
Regards,
NCProsecutor
As outlined in this response, the Fowler and Christakis paper in the JHE has several problems:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1262249