
How your friends' friends can affect your mood
* 30 December 2008 by Michael Bond * Magazine issue 2689.
IF YOU live in the northern hemisphere, this is probably not your favourite
month. January tends to dispirit people more than any other. We all know
why: foul weather, post-Christmas debt, the long wait before your next
holiday, quarterly bills, dark evenings and dark mornings. At least, that is
the way it seems. For while all these things might contribute to the way you
feel, there is one crucial factor you probably have not accounted for: the
state of mind of your friends and relatives. Recent research shows that our
moods are far more strongly influenced by those around us than we tend to
think. Not only that, we are also beholden to the moods of friends of
friends, and of friends of friends of friends - people three degrees of
separation away from us who we have never met, but whose disposition can
pass through our social network like a virus.
Indeed, it is becoming clear that a whole range of phenomena are transmitted
through networks of friends in ways that are not entirely understood:
happiness and depression, obesity, drinking and smoking habits, ill-health,
the inclination to turn out and vote in elections, a taste for certain music
or food, a preference for online privacy, even the tendency to attempt or
think about suicide. They ripple through networks "like pebbles thrown into
a pond", says Nicholas Christakis, a medical sociologist at Harvard Medical
School in Boston, who has pioneered much of the new work.
At first sight, the idea that we can catch the moods, habits and state of
health not only of those around us, but also those we do not even know seems
alarming. It implies that rather than being in charge of where we are going
in life, we are little more than back-seat drivers, since most social
influence operates at a subconscious level.
But we need not be alarmed, says Duncan Watts, a sociologist at Columbia
University, New York. "Social influence is mostly a good thing. We should
embrace the fact that we're inherently social creatures and that much of who
we are and what we do is determined by forces that are outside the little
circle we draw around ourselves." What's more, by being aware of the effects
of social contagion we may be able to find ways to counter it, or use it to
our own benefit. "There's no doubt people can have some control over their
networks and that this in turn can affect their lives," says Christakis.
To get an idea of what is going on, take Christakis's findings on the spread
of happiness, which were published last month. His team looked at a network
of several thousand friends, relatives, neighbours and work colleagues who
form part of the Framingham Heart Study, an ongoing multi-generational
epidemiological survey that has tracked risk factors in cardiovascular
disease among residents of Framingham, Massachusetts, since 1948. They found
that happy people tend to be clustered together, not because they naturally
orientate towards each other, but because of the way happiness spreads
through social contact over time, regardless of people's conscious choice of
friends (BMJ, DOI: 10.1136/bmj.a2338).
Christakis also found that a person's happiness is dependent not only on the
happiness of an immediate friend but - to a lesser degree - on the happiness
of their friend's friend, and their friend's friend's friend. Furthermore,
someone's chances of being happy increase the better connected they are to
happy people, and for that matter the better connected their friends and
family. "Most people will not be surprised that people with more friends are
happier, but what really matters is whether those friends are happy," says
Christakis. Happiness is near
They also discovered that the effect is not the same with everyone you know.
How susceptible you are to someone else's happiness depends on the nature of
your relationship with them. For example, if a good friend who lives within
a couple of kilometres of you suddenly becomes happy, that increases the
chances of you becoming happy by more than 60 per cent. In contrast, for a
next-door neighbour the figure drops to about half that, and for a nearby
sibling about half again. Surprisingly, a cohabiting partner makes a
difference of less than 10 per cent, which coincides with another peculiar
observation about some social epidemics: that they spread far more
effectively via friends of the same gender.
All this poses a key question: how can something like happiness be
contagious? Some researchers think one of the most likely mechanisms is
empathetic mimicry. Psychologists have shown that people unconsciously copy
the facial expressions, manner of speech, posture, body language and other
behaviours of those around them, often with remarkable speed and accuracy.
This then causes them, through a kind of neural feedback, to actually
experience the emotions associated with the particular behaviour they are
mimicking. Actions and feelings can be as contagious as a virus
Barbara Wild and her colleagues at the University of Tübingen, Germany, have
found that the stronger the facial expression, the stronger the emotion
experienced by the person observing it (Psychiatry Research, vol 102, p
109). She believes this process is hard-wired, since it acts so rapidly and
automatically.
Others have suggested it works through the action of mirror neurons, a type
of brain cell thought to fire both when we perform an action and when we
watch someone else doing it, though it is not clear whether the mimicking
would cause the neurons to fire or whether their firing would trigger the
mimicry. What is clear is that unconscious imitation allows people to "feel
a pale reflection of their companions' actual emotions" and even "feel
themselves into the emotional lives of others", says Elaine Hatfield at the
University of Hawaii, Honolulu, whose review of the latest research will
appear next April in The Social Neuroscience of Empathy.
There is plenty of evidence for emotional contagion outside the lab. In
2000, Peter Totterdell at the University of Sheffield, UK, found a
significant association between the happiness of professional cricketers
during a match and the average happiness of their teammates, regardless of
other factors such as whether the match was going in the team's favour
(Journal of Applied Psychology, vol 85, p 848). He found a similar effect
among nurses and office workers. It has also been shown that if a college
student suffers from mild depression their roommate will become
progressively more depressed the longer they live with them, and that
emotional displays by bank employees have a direct impact on the moods of
their customers.
We can see, then, how a phenomenon such as happiness might pass quickly
through a social network and infect clusters of friends and relatives. What
none of these studies explains, however, is why the strength of the
infection varies according to who is passing it to whom. Why are we so much
more strongly affected by the happiness of a nearby friend than a nearby
sibling? Why does a next-door neighbour have a significant impact, yet
someone living a few tens of metres away on the same block have none? The
power of strangers
Two factors appear crucial: the frequency of social contact, and the
strength of the relationship. This is not too surprising: we know that
emotional contagion requires physical proximity. It is also likely that the
closer we feel to someone, the more empathetic we are towards them, and the
more likely we are to catch their emotional state. However, how these two
factors play out in day-to-day interactions is uncertain. What is also
unclear - because it has never been properly tested - is the extent to which
emotions can propagate through virtual networks, where the opportunity for
physiological mimicry is much reduced.
So much for emotions - what about other phenomena that we unwittingly pick
up, and pass on, through our social networks? In 2007, Christakis's team,
again tracking members of the Framingham Heart Study, found that obesity is
transmitted in a similar way to happiness. Your risk of gaining weight
increases significantly when your friends gain weight, and it is also
affected by the weight of people beyond your social horizon. "Obesity
appears to spread through social ties," Christakis says. Again, how likely
you are to catch it depends on who you are interacting with: after
controlling for factors such as difference in socioeconomic status, the
researchers found that an individual's chances of becoming obese increased
by 57 per cent if one of their friends became obese, 40 per cent if a
sibling did and 37 per cent if their spouse did, irrespective of age (The
New England Journal of Medicine, vol 357, p 370).
However, neighbours have no influence, and how far away you live from a
friend counts for little, which implies that obesity spreads via a different
mechanism to happiness. Rather than behavioural mimicry, the key appears to
be the adoption of social norms. In other words, as I see my friends gain
weight, this changes my idea of what an acceptable weight is. One similarity
with happiness is that friends and relatives have a far greater influence if
they are of the same gender. While it is not evident why that should matter
for emotional contagion, norms of body size are clearly gender-specific:
"Women look at other women, men look at other men," says Christakis. This
could also help explain the epidemics of eating disorders reported among
groups of schoolgirls in recent decades.
The spread of a social norm appears to account for another of Christakis's
findings: that when people stop smoking, they usually do so along with whole
clusters of friends, relatives and social contacts. As more people quit, it
becomes the socially acceptable thing to do, and those who choose to
continue smoking are pushed to the periphery of the network. In this case,
people are most strongly influenced by those closest to them - if your
spouse quits, it is 67 per cent more likely that you will too. Your work
colleagues can also have an effect, particularly if you are in a small,
close-knit workplace; and more highly educated friends influence one another
more than less educated (The New England Journal of Medicine, vol 358, p
2249).
Happiness, obesity, smoking habits - activities that we traditionally think
of as shaped by individual circumstances, turn out to be ruled to a large
degree by social forces. Many other day-to-day phenomena fit a similar
pattern, often counter-intuitively. Take autism: Peter Bearman at the
Institute for Social and Economic Research and Policy at Columbia University
who in 2004 uncovered a link between suicidal behaviour and certain
friendship patterns (American Journal of Public Health, vol 94, p 89), is
looking at whether the recent rise in the diagnosis of autism is in any way
socially determined. His study is ongoing, but he says his findings could be
"explosive". "It is likely that if you have an autistic child in your
community the probability of your child being diagnosed with autism is
significantly higher." Happiness, obesity, smoking habits - all turn out to
be ruled to a large degree by social forces Why three degrees?
While the mechanism of social contagion varies depending on the phenomenon
being spread, in many cases the dynamics are very similar. For example,
Christakis has found that with happiness, obesity and smoking habits, the
effect of other people's behaviour carries to three degrees of separation
and no further. He speculates that this could be the case with most or
perhaps all transmissible traits. Why three degrees? One theory is that
friendship networks are inherently unstable because peripheral friends tend
to drop away. "While your friends are likely to be the same a year from now,
your friends of friends of friends of friends are likely to be entirely
different people," says Christakis.
This poses the question: what shapes the architecture of our social networks
and our position in them? Clearly, many factors contribute: where we live,
where we work, family size, education, religion, income, our interests, and
our tendency to gravitate towards people similar to us. New research by
Christakis's team, due to be published in the next few weeks, suggests there
is also a strong genetic component. The study compared the social networks
of identical and fraternal twins, and found that identical twins had
significantly more similar social networks than fraternal twins, suggesting
the structure of your social network is influenced by your genes. That may
not sound remarkable, since personality traits such as gregariousness and
shyness clearly play a role in determining how connected we are. But there
is much more to it, says Christakis. "It's not just about having a genetic
predilection to be friends with a lot of people, it's about having a genetic
predilection to be friends with a lot of popular people. That's mysterious:
how could our genes determine our actual location in this socio-topological
space?"
Answering that should help us understand more about the "collective
intelligence" of social networks, which some researchers liken to the
flocking of birds - the decision to quit smoking, for example, is no more an
isolated move than the decision by a bird in a flock to fly to the left.
Sociologists and others are using mathematical models to test these dynamics
to try to understand better what triggers the spread of behaviours. Duncan
Watts at Columbia University has shown that seeding localised social groups
with certain ideas or behaviours can lead to the ideas cascading across
entire global networks. This contradicts the notion - promoted by the author
Malcolm Gladwell in The Tipping Point and others - that social epidemics
depend on a few key influential individuals from whom everyone else takes
their cue. It doesn't ring true, argues Watts, because such "influentials"
typically interact with only a few people. The key for the spread of
anything, he says, from happiness to the preference for a particular song,
is a critical mass of interconnected individuals who influence one another.
Is there any way to mitigate the effects of such powerful and pervasive
social forces? It is unlikely we can ever escape social influence entirely,
even if we wanted to. "Even when you're aware of it, you're probably
susceptible," says Watts. Still, being aware can help, especially when we
are seeking to avoid undesirable behaviours or adopt positive habits. We can
be choosy about new friends, seeking out people whose lifestyles we aspire
to: if you want to lose weight, for example, join a running club and - most
importantly - socialise with its members.
Actually cutting ties with old friends might be a bit drastic, though
perhaps spending less time with those whose traits we do not wish to share
would be a good idea - lazy people, perhaps, or those inclined to negative
thinking. And beware those who hang out with such people even if they do not
display their views or behaviours - remember the three degrees of contagion
rule. Finally, if you really cannot avoid spending time with certain people
whose behaviours or emotional state you would rather not take on board
(certain relatives at family gatherings, perhaps), you could always try
repressing your natural inclination to mimic their body language and facial
expressions, and so limit the contagion effect - though be prepared for them
to instinctively cool towards you as a result.
What this game plan amounts to is a kind of subtle social reorientation. We
will always be vulnerable to what those around us are doing, so as far as
possible make sure you are with the right people. Remember the new adage: we
are who we hang out with. Five tips for a healthier social network
1. Choose your friends carefully.
2. Choose which of your existing friends you spend the most time with. For
example, hang out with people who are upbeat, or avoid couch potatoes.
3. Join a club whose members you would like to emulate (running, healthy
cooking), and socialise with them.
4. If you are with people whose emotional state or behaviours you could do
without, try to avoid the natural inclination to mimic their facial
expressions and postures.
5. Be aware at all times of your susceptibility to social influence - and
remember that being a social animal is mostly a good thing. Issue 2689 of
New Scientist magazine
* From issue 2689 of New Scientist magazine, page 24-27.