January 22, 2016

Online or offline: Connecting with close friends improves well-being

By: Moira Burke, Robert Kraut

Does social technology draw us closer to our friends or isolate us? Understanding how technology affects our well-being is difficult, in part because it’s very hard to measure. And there’s no shortage of debate about the issue in the popular press and academic literature [1].

In a new paper, we’ve reviewed two studies, fifteen years apart, and what we found is fairly intuitive: The Internet’s effect on your well-being depends on how you use it. Talking online with friends is associated with improvements in well-being, while talking with strangers or simply reading about other people is not. The studies combine monthly surveys measuring social support, depression, and other aspects of psychological well-being with log data of participants’ Internet use.

Kraut, R. and Burke, M.
Internet use and psychological well-being: Effects of activity and audience
. Communications of the ACM. 58(12), December 2015, p.94-100.

Here’s what we found:

    1. About 15 years ago (long before Facebook), our research lab at Carnegie Mellon found that the more people used the Internet, the more depressed they became, as measured with standard psychological survey questions.
    2. However, back then, the Internet was generally comprised of universities; people’s closest friends weren’t online yet, so our participants were mostly talking with strangers.
    3. When our lab replicated the study a few years later, once participants’ friends were online, too, we found that the more they used the Internet to talk to friends, the less depressed they became. On the other hand, talking to strangers was again linked to increases in depression.
    4. Fast forward to today. In contrast to 15 years ago, most people don’t seek out strangers to talk to on Facebook [2]; instead they communicate with a range of acquaintances and close friends. To conduct this analysis, we relied on data from three opt-in surveys administered one month apart measuring social support, depression, and other aspects of well-being combined with de-identified and aggregated counts of Facebook activity from participants who agreed to allow us to access this information. The Facebook activity that was measured consisted of information such as the number of wall posts and comments posted and read, likes delivered and received, stories read, and photos viewed; no text was analyzed.
    5. We found that the effects really depended on how people used the site: The more people talked one-on-one, such as writing wall posts or comments, especially with close friends, the more their well-being improved (see Figure 1).post00064_image0002
      Figure 1. Satisfaction with life, positive mood, social support, and loneliness all improved 1-3% among people who received approximately 50 more comments than average from close friends. Stress, depression, and negative mood were unchanged. These effects control for age, gender, and the participant’s reported well-being the prior month.
    6. But not all activities are equal. Receiving comments from acquaintances (rather than close friends) was linked to smaller improvements (0.5-1.5%), and spending time reading about acquaintances without talking to them was linked to a small but statistically significant increase in negative mood (about 1%). Other well-being outcomes were unchanged after reading about acquaintances (see Figure 2).


Figure 2. Negative mood increased about 1% among people who read more stories about acquaintances on Facebook. Other well-being outcomes were unchanged.

These effects are all pretty small, because well-being is relatively stable month-to-month, and a lot of things influence your well-being more than your experiences online, such as whether you have a good job, how much you exercise, and how much sleep you get.

To understand how these effects compare to major events like getting married, having a baby, or losing a loved one, we asked people whether they’d experienced major events in the month between surveys, and measured how much their well-being changed from those events. The effects of one-on-one communication from strong ties was roughly comparable (though in the opposite direction) to the effect of an illness, and half to a third of the size of divorce or losing a job. Said another way, receiving more one-on-one communication from close ties was linked to a boost in well-being that was about the same magnitude as well-being changes caused by major events in people’s lives.

Some challenges to studying well-being and the Internet

This study is one small piece of the puzzle, and measuring both Facebook use and well-being is methodologically difficult. Here are some challenges that researchers face that are worth keeping in mind when evaluating research about the Internet.

Challenge #1: Studying people at a single point in time is misleading. Many studies are cross-sectional, but cross-sectional studies don’t allow you to determine causation. For example, in a cross-sectional 2010 study, we found that people who spent a lot of time “passively consuming” feed stories and photos on Facebook were lonelier. We couldn’t tell if reading stories makes you lonely, or if lonely people spent more time reading online. Later studies have shown that people sometimes choose to go to Facebook when they’re feeling blue, and that reflecting on close relationships on the site helps them feel better [3,4].

So, be cautious of claims made from cross-sectional studies. A better approach uses a longitudinal design, where you take into account how lonely someone feels at the start of the study, and how much their loneliness *changes* over time. The present study uses that technique and finds that loneliness decreases about 3% over time when people have interactions with close friends.

Challenge #2: It’s difficult to report how much time you spend online. Many studies are survey-based, asking participants how much time they spend on Facebook. But that’s very difficult to answer. Figure 3 shows that there’s only a modest correlation between self-reported time spent on Facebook and actual time spent (see also [5]). And the common perception that the Internet causes loneliness may also jointly bias self-reports of time online and loneliness [6]. Instead, look for studies that collect data from multiple sources, such as a combination of surveys, server logs, a lab study, nanny software, and/or heart-rate monitors (all of which have been used by technology researchers). Our studies combine surveys with server data to get more accurate measurements of time online.

Figure 3: Surveys measuring how much time people spend on Facebook bear little relationship to how much time they actually spend on Facebook.

Challenge #3: Facebook isn’t a monolith and not all people are the same. As the present study shows, the effects depend on how you’re using the site, and whom you’re talking to. Furthermore, people vary in extraversion, optimism, social skills, and technology access. Results can be very different, depending on what’s happening in a person’s life. In a 2013 study [7], we found that for most people, talking with close friends was linked to reductions in stress, but people who said they’d recently lost a job felt *more* stress among close friends. In a 2011 study, we found that well-being didn’t change after reading feed stories about others. But for people who said they were uncomfortable communicating face-to-face, their feelings of social capital increased the more they read. It’s worth taking into account differences in people, differences in how people use the site, and differences in whom they’re talking to.

It is becoming clearer: a great prescription for dealing with life’s travails is social interaction with family and close friends. Facebook is one way – given the distance that often separates us from our friends and family – for people to engage in interactions with people they care about.


[1] For example:
Hampton, K., Sessions, L., Her, E., & Rainie, L. (2009). Social isolation and new technology. Pew Internet & American Life Project.
Burke, M., Marlow, C., & Lento, T. (2010). Social network activity and social well-being. Proceedings of the SIGCHI conference on human factors in computing systems (CHI 2010), 1909-1912.
Sleeper, M., Acquisti, A., Cranor, L. F., Kelley, P. G., Munson, S. A., & Sadeh, N. (2015). I Would Like To…, I Shouldn’t…, I Wish I…: Exploring Behavior-Change Goals for Social Networking Sites. Proceedings of the 18th ACM Conference on Computer Supported Cooperative Work & Social Computing (CSCW 2015), 1058-1069.
[2] Lampe, C., Ellison, N., & Steinfield, C. (2006). A Face (book) in the crowd: Social searching vs. social browsing. Proceedings of the 2006 20th anniversary conference on Computer supported cooperative work (CSCW 2006), 167-170.
[3] Toma, C. L., & Hancock, J. T. (2013). Self-affirmation underlies Facebook use. Personality and Social Psychology Bulletin, 39(3), 321-331.
[4] Sheldon, K. M., Abad, N., & Hinsch, C. (2011). A two-process view of Facebook use and relatedness need-satisfaction: disconnection drives use, and connection rewards it. Journal of Personality and Social Psychology, 100(4), 766-775.
[5] Junco, R. (2013). Comparing actual and self-reported measures of Facebook use. Computers in Human Behavior, 29(3), 626–631.
[6] http://en.wikipedia.org/wiki/Common-method_variance
[7] Burke, M., & Kraut, R. (2013). Using Facebook after losing a job: Differential benefits of strong and weak ties. Proceedings of the 2013 conference on Computer supported cooperative work (CSCW 2013), 1419-1430.