We live so much of our lives on social media, and that’s not always a good thing. But there may be some benefits, as well, when it comes to diagnosing depression.

A new study from the University of Pennsylvania and SUNY Stony Brook explores the link between the language we use in Facebook posts and depression; its results, published in the Proceedings of the National Academy of Sciences, point to some striking correlations.

The researchers analyzed language in Facebook data through machine learning techniques, and were able to identify which study participants had depression diagnoses. They could make those identifications using language that the participants had posted as far back as three months before their diagnoses.

The findings are significant because they indicate that by screening language that people use online, it may be possible to identify and treat more cases of depression, an illness that often goes unreported and undiagnosed, according to the study authors. They also cite the startling projection, from a study by the World Health Organization and published in the journal PLOS, that by 2030, depression will be the leading cause of disability in the United States. Depression can have serious consequences, and a tool that helps identify it is worth examining. Social media may actually be a useful tool in addressing depression — even if it could also be one contributing factor.

In the meantime, as we work towards better systems of diagnosing and treating depression, it’s useful to understand the language cited in the study. If you or a loved one are concerned about a risk of depression, take a look to see if this kind of language feels familiar to you:

Mentions of hostility and loneliness

The context in which you use this kind of language actually didn’t matter, at least in this study. Researchers found that high numbers of references to hostility and loneliness in posts indicated possible signs of depression.

Words like “tears” or “feelings”

Another linguistic correlate with depression was language relating to sadness. Again, the study wasn’t looking at how the language was used. What mattered is simply whether people were posting these words frequently.

Lots of personal pronouns

Personal pronouns are a good linguistic indicator of preoccupation with the self and rumination — processes associated with depression, according to the study authors. The study found that high rates of their usage online also correlated with depression. We all certainly use these terms (along with the other ones listed here), and if you notice them in your posts, you aren’t necessarily suffering from depression — the study explains that those eligible for clinical depression diagnoses statistically use these words more than others, a comparative rubric that’s impossible to analyze looking just at your own (or even at a few) profiles. But if you see them popping up in high numbers, especially in combination with some of the other linguistic indicators identified by the study, it may be worth considering turning to a professional for an expert opinion.

Follow us on Facebook for all the latest news on how you can keep Thriving.

More from Thrive Global:

8 Things You Should Do After 8 P.M. If You Want to Be Happy and Successful

The One Relationship You’re Probably Ignoring

The One Word That Can Hurt Your Reputation at Work


  • Nora Battelle

    Multimedia Staff Writer at Thrive

    Nora Battelle is a writer from New York City. Her work has been published on the Awl, the Hairpin, and the LARB blog, and she's written for podcast and film. At Swarthmore College, she studied English and French literature and graduated with Highest Honors. She's fascinated by language, culture, the internet, and all the small choices that can help us thrive.