Parasocial Relationships

What the Research Actually Says

In November 2025, Cambridge Dictionary named "parasocial" its word of the year, a strange fate for a term that spent almost seventy years gathering dust in communication journals. The word arrived in the mainstream carrying a faint accusation. To call a feeling parasocial is usually to hint that something has gone slightly wrong, that a person has mistaken a screen for a friend. The behavior it describes is far older than the panic around it. People have always formed attachments to figures who cannot see them back, and the research on parasocial relationships tells a quieter, more human story than the headlines do. What changed recently is the medium. The figures now read our names out loud on live streams, and the one-sided parasocial interaction of the television age has started to answer back.

The idea itself comes from a 1956 paper in the journal Psychiatry, written by two University of Chicago scholars, Donald Horton and Richard Wohl. They were watching the young medium of television do something peculiar. A host would look straight into the camera, greet the audience as if it were a single person, ask how everyone at home was keeping, and behave like a friend who had let himself in through the back door. Horton and Wohl called this intimacy at a distance, and they named the bond it produced a parasocial relationship. The detail that usually falls out of the retelling is their calm. They were not sounding an alarm. For the ordinary viewer, they wrote, this kind of attachment tended to sit alongside real friendships rather than crowd them out. Where it curdled into something isolating, that was the unusual case, and they said so plainly. The moral panic came later, and it came without them.

The lonely-fan myth

The first serious attempt to put a number on the feeling arrived in 1985. Alan Rubin, Elizabeth Perse, and Robert Powell built a parasocial interaction scale and aimed it, of all places, at local television news anchors, which is about as mundane a bond as media can produce. Their survey asked people whether they felt their favorite newscaster was almost a friend, whether they missed the person when the broadcast was off, whether the anchor's warmth felt like it was meant for them. The instrument held up well, and versions of it have been used in hundreds of studies since. Later researchers pointed the same measure at soap opera characters, talk radio hosts, home shopping presenters, and eventually online influencers, and the pattern kept repeating, which told the field it was measuring something ordinary and widespread rather than a quirk of the evening news. Sitting inside that first dataset, though, was a result that should have quieted a lot of noise and somehow did not. Loneliness and parasocial interaction were not correlated. The lonely viewers were not the ones reaching for these bonds, at least not more than everyone else was. Scandinavian researchers had found much the same thing a few years earlier, and David Giles, reviewing the entire field in 2002, concluded that parasocial experience was better treated as a normal part of media life than as a symptom of anything at all.

That finding sits crosswise to the assumption most of us carry without examining it. Across the 1970s and 1980s a tidy theory took root: that these bonds were compensation, a stand-in connection for people running short on the real kind. Gayle Stever, a media psychologist who has studied celebrity and fan attachment for decades, is direct about where that theory ended up. Lonely people do form parasocial relationships. People who are not lonely form them just as readily, and just as strongly. The picture only gets less tidy from there. Extraverts, who are not short of company, form parasocial bonds more often than introverts, partly because they simply use more media and reach toward more people, on screen and off. The one thread that holds is attachment style. People with an anxious, ambivalent way of relating, the ones who crave closeness and fear its withdrawal, tend to form the most intense parasocial bonds and to hurt the most when those bonds break, which is precisely how they behave with the people in their actual lives. A parasocial bond is also, as the psychologist Jaye Derrick has pointed out, safe. The person on the screen cannot reject you, cannot grow bored of you, cannot leave. For someone bruised by ordinary rejection, that safety is not a delusion. It is a feature they have correctly identified. A 2024 review led by Holger Schramm, pulling together 281 studies from a single five-year window, found the same broad shape across cultures and formats: parasocial bonds are common, mostly benign, and poorly explained by any simple story about social failure.

A 2022 experiment by Jan-Philipp Stein and colleagues drew a fine line under all this. They asked people to spend a few minutes dwelling on a real friend, a beloved media figure, or nothing social at all, then measured mood and loneliness on either side of the task. Thinking about the media figure lifted people's mood roughly as much as thinking about a real friend did. It did nothing at all for their loneliness. Parasocial bonds, then, are a way to feel better, not a way to feel less alone, and the confusion between those two jobs is where a good deal of the worry goes wrong. Ask them to brighten an afternoon and they deliver. Ask them to replace human company and they quietly fail, because that was never the work they were built for. The distinction matters more than it sounds, because a person who leans on a favorite streamer to lift a bad evening is doing something sensible, while a person who expects that streamer to fill the place of a friend is setting up a disappointment neither of them can fix.

If any doubt remains that these attachments are real, watch what happens when one is taken away. Jonathan Cohen has spent much of his career on parasocial breakups, the grief people report when a series ends or a favorite character is written out. After the last episode of Friends aired in 2004, he and Keren Eyal surveyed viewers and found that the strength of the parasocial bond predicted the pain of the ending better than anything else they measured, better than loneliness, better than sheer hours watched. When the 2007 writers' strike froze scores of shows mid-run, other researchers caught the same distress in a smaller key. Cohen's earlier work had already shown that people with anxious attachment styles brace for the worst when a beloved character leaves, exactly as they would before a real goodbye. The brain, in Stever's framing, does not keep one ledger for people it has met and another for people it has only watched. It files the streamer, the sitcom character, and the friend across the table in roughly the same drawer. That is why the loss can feel like a loss, and why there is rarely a funeral or a sympathy card for it, which is part of what makes the grief so hard to put down.

When the screen talks back

Everything described so far grew out of a one-way medium. Horton and Wohl's television host could greet you but could never hear you answer. Live streaming tore that wall down. On Twitch, on YouTube, on the cam platforms, the persona answers. Your message scrolls up the chat and the streamer reads it aloud, sometimes laughs at it, sometimes argues with it. Your subscription sets off an alert with your name spelled across the screen, a small public moment engineered to feel like recognition. Your tip buys a thank you, occasionally a whole minute of undivided attention. In a 2021 experiment, Tim Wulf and colleagues sat viewers in front of a simulated stream and changed a single variable: whether the streamer noticed them, addressed them by name, reacted to their messages. The viewers who were noticed felt distinctly closer to the streamer than the ones left in the scroll. Caitlin McLaughlin and Donghee Yvette Wohn, studying real audiences rather than a lab setup, found that the streamer's own conduct, how much they acknowledged the room, was the strongest thing shaping how deep these bonds ran. Attention is the raw material, and live platforms have turned its supply into a craft.

This is the point where the old vocabulary begins to buckle. A bond with zero reciprocity, a viewer and a recorded character who will never know they exist, is parasocial in the pure sense. A bond with full reciprocity is just a friendship. Live streaming lives in the space between, and Stever now argues that parasocial and social are better pictured as the two ends of a single line than as a pair of separate boxes. Some researchers describe the streamer bond as one-and-a-half sided, which catches the lopsidedness without pretending that nothing travels back the other way. The viewer knows the streamer with real texture, the in-jokes, the moods, the dog's name. The streamer knows the viewer as a handle that lights up now and then, and as a number at the bottom of the screen.

Nowhere is this middle ground bought and sold as openly as on erotic cam sites, which is exactly what makes them such a clean case to study. In 2024, Ellen Kaufman and Amanda Gesselman of the Kinsey Institute surveyed 2,047 clients of one large webcam platform. Sixty-five percent said they had felt an emotional bond with a model at some point. More than half, 51.6 percent, believed the model actually cared about their life away from the screen. These are not people muddled about the transaction. They are describing something the sociologist Elizabeth Bernstein named bounded authenticity: a connection that feels genuine and is also shaped, and fenced, by the fact that it is a paid service with a clock running. The performer is doing what Arlie Hochschild called emotional labor, the trained work of producing and displaying feeling as part of a job, the same labor a flight attendant or a nurse performs from a different script. The warmth a viewer receives can be sincere in the moment and professionally manufactured at the same instant. Both are true, and refusing to hold them together is how people end up either sneering at the client or lying to him. The filmmaker Sean Dunne, after time inside these rooms, reported that the men often described not a strip club but a community, a place where the feeling in the chat mattered more than the nudity. Viviana Zelizer, who studies the economics of intimacy, calls it a new kind of closeness, not quite a relationship and not quite a commodity, something we do not yet have a settled word for. Camming only crystallizes what streaming does everywhere, with the price tag made plain. On Twitch the same exchange happens in softer currency, a follow, a gifted subscription, a raid from a bigger channel, but the underlying trade of attention for the feeling of being known runs the same way underneath.

It is worth seeing that the intimacy is not produced by people alone. It is built into the platforms. Every alert chime, every animated banner that throws a subscriber's name across the screen, every gift that makes the streamer stop and say thank you, is a design choice, and the choices all lean the same direction: turn attention into feeling, and feeling into payment. Superchats, tokens, tiered subscriptions, timed private shows, all of them reward the moments when the bond feels most personal and most exclusive. Recommendation systems push the creators who keep people watching, which tends to mean the ones best at manufacturing that sense of being seen. Large studies of streaming income find the money is steeply concentrated, with a thin sliver of creators taking the bulk of it, which turns the screw tighter still: the incentive is to make every regular in the room feel like the only person in it. None of this is an accusation aimed at the models or the viewers. It is the shape of the machine they are both standing inside, and the viewers who can see that shape tend to hold the feeling more lightly, and enjoy it more honestly.

The real hazards live in the gap between how the bond feels and what it is. When a viewer forgets that the closeness runs one way, ordinary things sour. A streamer who stops sharing personal stories can register, to a devoted follower, as a friend going cold. The musician Chappell Roan spent much of 2024 pushing back on fans who read her fame as a standing invitation, treating familiarity as a claim on her time, her family, her body. The streamer Emiru was grabbed by a fan at a meet-and-greet, a hard reminder that a one-sided feeling can reach into a two-sided world. At TwitchCon in 2025, creators sat on panels about the strain of being everyone's imaginary friend at once, of absorbing thousands of attachments while owing nothing back and somehow feeling that they do. The evidence is steady that most parasocial bonds are harmless, and often better than harmless. They offer comfort, a template for working out who you are, a sense of belonging to a crowd of fellow fans who feel the same pull. Work by Cynthia Hoffner and Bradley Bond ties them to real gains in well-being and identity. A smaller, measurable minority slides into obsession, and that deserves attention and care rather than a shrug. The distinction the research keeps drawing is between the bond itself, which is normal, and a particular kind of fixation that runs over every boundary the other person sets, which is not. But the person worth worrying about is almost never the woman crying when a character dies. She is doing what a social animal does when it is shown a face and a voice and a name that seem to be turned toward her.

For most of a century, the story we told about parasocial attachment was a story about lack. The lonely, the shy, the socially fumbling, reaching into a screen for the closeness they could not manage in a room. The data never really backed it, and the sharper version of the panic, the one aimed at teenage girls with crushes on pop stars, has it close to upside down. Feeling a bond with a voice that seems to speak to you is not a fault in the human design. It is the design running as intended, on inputs it never evolved to meet. The caution the research genuinely supports is quieter, and less comfortable for the people building these products. It is not that audiences feel too much. It is that a large and growing industry has learned, in granular detail, how to be felt for, and how to put that feeling on a meter.