Nonprofit, nonpartisan journalism. Supported by readers.


UCare generously supports MinnPost’s Second Opinion coverage; learn why.

Twitter accounts provide a linguistic window into people’s loneliness, study suggests

These findings suggest, say the researchers, that social media may offer a “loneliness prediction system.”

Twitter users
Findings suggest that social media may offer a “loneliness prediction system.”
REUTERS/Kacper Pempel/Illustration

The words people use when posting on their Twitter accounts may serve as a kind of “loneliness prediction system,” helping to identify people with high levels of loneliness, according to a study published this month in the journal BMJ Open.

As the study’s authors point out, finding better ways of identifying such people is important because loneliness — defined in the study as “the discrepancy between a person’s desired and actual social relationships” — is a risk factor for several chronic physical and mental conditions, including depression, anxiety, heart disease, stroke and cognitive decline.

Loneliness is also associated with early death.

“Loneliness can be a slow killer, as some of the medical problems associated with it can take decades to manifest,” said Sharath Chandra Guntuku, the study’s lead author and a research scientist at the University of Pennsylvania, in a released statement. “If we are able to identify lonely individuals and intervene before the health conditions associated with the themes we found begin to unfold, we have a chance to help those much earlier in their lives. This could be very powerful and have long-lasting effects on public health.”

Article continues after advertisement

Indeed, loneliness is now recognized as a major — and growing — public health problem. The United Kingdom has even appointed a “minister for loneliness” to tackle the issue. In the United States, public health officials have also been trying to raise awareness about the health consequences — for individuals and for society — of loneliness.  In 2017, for example, Dr. Vivek Murthy, who served as United States Surgeon General in the Obama administration, wrote a widely publicized article for the Harvard Business Review that argued that loneliness needed addressing in the workplace.

Estimates of the proportion of American adults who are lonely range from 17 to 57 percent, although a study published in 2018 found that three-quarters of Americans reported moderately high levels of loneliness.

How the study was done

Past research has shown that social media platforms, such as Twitter, Facebook and Instagram, can serve as linguistic windows on an individual’s health and well-being. Studies have found, for example, that people who are stressed and depressed use more first-person singular pronouns.

Guntuku and his colleagues wanted to see if the words individuals use on social media platforms could “confirm the existing understanding of loneliness and give new insights into the daily lives of those who express being lonely.”

For their study, the researchers searched through about 400 million public tweets collected from Twitter users in Pennsylvania between 2012 and 2016. They found 6,202 individuals whose posts included the words “lonely” or “alone” more than five times during that period. The average age of these individuals was 21, and almost 70 percent were women.

The researchers then selected a matching (by age and gender) “control” group of Twitter users who had no posts with the words “lonely” or “alone.”

The tweets of both groups were then analyzed for keywords associated with psychological well-being.  That analysis revealed that the “lonely” individuals tweeted almost twice as much as the control group. They were also significantly more likely to do so in the evening or at night.

In addition, the “lonely” tweeters posted much more frequently about relationship issues (using phrases such as “want somebody” and “no one to”), needs and feelings (“I just wanna” and “I need”), substance use (“smoke,” “weed,” and “drunk”) and insomnia (“sleep,” “tired,” “bed”). They also tended to use words associated with anger, depression and anxiety — and they used more expletives.

Article continues after advertisement

The control group, on the other hand, posted more often about games (“season,” “coach,” “tea”) and positivity “(awesome”). They were also more likely to post the names of other Twitter users in their tweets — a sign, perhaps, of strong social connections.

These findings suggest, say the researchers, that social media may offer a “loneliness prediction system.”

A promising possibility

The study comes with caveats. Most notably, the people whose Twitter accounts were analyzed weren’t actually screened for loneliness. In addition (as the study’s authors note), they weren’t representative of the general population.

Still, the study corroborates other research that has shown that people who are lonely express more negative sentiments about relationships, more feelings about helplessness and have a greater difficulty self-regulating their emotions.

Guntuku and his colleagues plan to conduct further studies to see if people express loneliness similarly on other social media platforms, such as Facebook, Instagram and Reddit.

“As loneliness is a public health challenge, a better understanding of how loneliness is expressed online can inform passive assessment of loneliness and interventions targeted at addressing it in regards to the behaviour of lonely individual who may be a risk of developing a severe mental health condition,” they write.

“From our studies, people experiencing loneliness seem to be using social media as a venue to express themselves, but they’re not having a lot of interpersonal actions,” said Guntuku in an interview with the Philadelphia Inquirer. “But it’s promising because they’re open to sharing things from their electronics, and we can target some of these interventions to them.”

FMI:  The study can be read in full on the BMJ Open website. For strategies on overcoming feelings of loneliness, go to the National Institute of Aging’s website. Although that site is aimed at older adults, its suggestions are applicable to people of any age.