Instagram activity linked to two distinct forms of loneliness

By Amara Osei

A UC San Diego study analyzed the complete Instagram histories of 415 undergraduates and found that posting behavior predicted loneliness better than it measured real-world support networks.

The findings concerned self-reported loneliness, meaning how students described their own feelings, and real-world support networks, meaning the people they said they could contact. The research appeared in npj Mental Health Research.

What did the study examine?

The researchers reviewed full account histories, including information Instagram users normally keep private. All 415 participants gave free and informed consent, and the researchers said they applied privacy protections during analysis.

More than 60% of participants had private accounts. The authors said this made the study unusual. Instead of focusing on influencers whose stylized content aimed to build a personal brand, the team could examine ordinary users interacting with friends and sharing everyday experiences, including activity outside observers normally cannot see.

The team used machine learning, meaning algorithms trained to identify patterns in data, to detect signals in the records. The researchers compared those signals with students’ responses to established loneliness measures. Students also reported how many friends and relatives they contacted regularly and whom they could turn to for support.

The study produced measurable associations, not a standalone loneliness test.

How did the two types of loneliness differ?

The researchers separated emotional loneliness from social loneliness. Emotional loneliness means lacking a close, meaningful connection, even when surrounded by people; it can also involve emptiness and emotional distress. Social loneliness means feeling disconnected from a broader circle of friends or community despite having a best friend or partner.

The authors said emotional loneliness warrants particular concern because previous research associated it more strongly with depression and anxiety.

Piotr Winkielman, a psychology professor at UC San Diego, said: “Loneliness is not an experience that applies to everyone in the same way, and neither are the solutions. Someone who feels disconnected from a broader community may benefit from more opportunities for socialization, while someone without a close emotional connection may need a very different kind of support.”

The study also challenges the popular assumption that Instagram activity mainly presents a happy, curated side of life. The researchers found measurable behavior differences associated with two forms of loneliness.

The researchers reported distinct patterns for each type, with some overlap:

Patterns reported for emotional loneliness:

• Fewer photos containing human faces.
• More past-focused language, including “was” and “had,” which may reflect a tendency to dwell on the past.
• More late-night posts, especially between 23:00 and 05:00.
• More negative language in captions.
• Profile photos showing sadder or angrier facial expressions.
• More references to family and friends, which may reflect a longing for closer bonds.

Patterns reported for social loneliness:

• Fewer photos containing human faces.
• Profile photos with the head tilted upward instead of facing the camera, a pattern also associated with other indicators of loneliness.
• More past-focused language, including “was” and “had,” which may reflect a tendency to dwell on the past.
• Less frequent likes and comments on other people’s posts.
• Fewer weekend posts.
• Fewer words expressing affection or appreciation, such as “thank you” and “take care.”
• Fewer words related to communication, such as “said” and “to say,” possibly reflecting fewer references to interactions with other people.

Both profiles showed fewer human faces and more past-focused language. Emotional loneliness added late-night posts, negative captions and family references; social loneliness added lower interaction and fewer weekend posts.

Can Instagram reveal a person’s support network?

The researchers found Instagram behavior useful for predicting the level of self-reported loneliness. They found it much less effective for predicting the size of real-world support networks.

Sherry (Yueyi) Jiang, the lead author, said: “Instagram really reflects how lonely people feel, but not necessarily how many people they can contact in real life.”

Digital behavior may eventually help identify people experiencing loneliness, including those reluctant to discuss their feelings, the authors said. They proposed that universities, counselors and health professionals could use early warning signs to offer support tailored to each form of loneliness.

Instagram reflected group-level loneliness more clearly than it measured the size of a person’s support network.

What remains unproven?

The researchers stressed that their findings represent patterns across groups, not reliable indicators for diagnosing individuals. They also said the study cannot establish whether loneliness changes how people use Instagram, whether Instagram use contributes to loneliness, or whether both occur.

For marketers, the defensible reading is narrow: the study identifies group-level associations, not a basis for inferring an individual’s mental state.

The next test is whether researchers can use those signals for early, tailored support without turning social activity into an individual diagnosis.