Body Doubling: Why You Work Better With Someone in the Room

It is eleven at night and you have a tax form open in one window and, in the other, a stranger in a city you will never visit, writing in a notebook under a desk lamp. Forty thousand people are watching her. Nobody is speaking. Every so often she sips tea and turns a page, and somehow, with her on the screen, the form gets filled in.

There is also the co-working app that pairs you with a stranger for fifty minutes, cameras on, microphones off. The "clean with me" video that gets a kitchen done. The friend who works at your table for two hours, says nothing, and leaves you with a draft you had been circling for a week.

None of these people helped. They did not check your work or set a deadline. They were just there, and it worked anyway, which is a question with a real answer, one older than video.

A New Name for an Old Habit

The practice is now called body doubling, and the name has a traceable origin. It came out of neurodivergent communities, where people with ADHD had been describing for years that a task they could not start alone became possible with someone else in the room. Services built around the idea followed. When Tessa Eagle and colleagues surveyed 220 people about it in 2023, they described their paper as the first academic exploration of the topic.

Two things in that survey are worth holding onto. Most respondents said they had used the technique long before they learned its name. And the range was wide: body doubling at a café and with YouTube, for studying, dishes and exercising, with motivations like generating momentum and staying on task. The authors proposed a model of body doubling as a continuum along two dimensions they call space/time and mutuality: how co-located and simultaneous the other person is, and how much you are doing it for each other.

If you have ADHD and want to know whether it works better for you than for everyone else, the studies in this article do not say. The Eagle survey recorded how neurodivergent people use body doubling and why. It did not compare them with anyone. None of the experiments that follow set out to compare people with and without ADHD either. What they do show is that mere presence shifts performance for people in general, so the thing your community named is real. Whether it is stronger for you is a question this evidence leaves open.

A livestream of a stranger sits at one corner of that map and a friend beside you at the other. The research on why either corner does anything predates the name by more than a century.

The Ten Children Who Got Worse

In 1898 Norman Triplett noticed that cyclists' official records were faster when paced or in competition than when riding alone against the clock. He took it into the lab and had children wind fishing reels as fast as they could, alone and side by side with another child.

Of the forty children in his main tables, twenty were, in his words, stimulated positively by the competitor. Ten were overstimulated and did worse. Ten showed slight evidence of being stimulated at all. His conclusion was that the bodily presence of another contestant serves to liberate latent energy not ordinarily available.

That paper is usually counted as the first social facilitation experiment, and it already contained the complication: a quarter of the children got worse with company.

Robert Zajonc's 1965 paper in Science organised decades of scattered audience and co-action findings around one idea: the presence of others strengthens whatever response is dominant for you. On a well-learned task the dominant response is the right one, so you improve. On a task you are still learning, the dominant response is often wrong, so you get worse. We covered the version that applies to an accountability partner elsewhere. What Zajonc left open was sharper: does the other person have to be watching you, or is it enough that they exist?

Getting Dressed in Front of a Stranger

Hazel Markus answered that in 1978 with an experiment designed to remove every reason to care what the other person thought.

The task was getting dressed. Participants took off their own shoes and socks and put them back on, which is about as well-learned as a task gets, and they also dressed in unfamiliar clothing, which is not. The task was chosen because it has no clear performance criteria and is very unlikely to make anyone worry about being judged. Nobody is competitive about their socks.

They did this alone, with an attentive spectator, or with a passive, inattentive person who was paying them no attention at all.

That is the body doubling result, thirty years before the term. A person who is not watching you, cannot evaluate you, and has nothing to say about how fast you tie a shoelace still moves the needle. The stranger on the livestream does not have to be looking at your tax form. She has to be there.

Easy Goes Up, Hard Goes Down

Five years later Charles Bond and Linda Titus pooled 241 studies of the presence of others, covering nearly 24,000 people, and the shape held across the whole literature.

Two more findings matter for anyone trying to use this on purpose. The effects were small, accounting for between .3% and 3% of the variance in a typical experiment. Consistent in direction, not a miracle. And social facilitation effects were, in the authors' words, surprisingly unrelated to the performer's evaluation apprehension. Fear of being judged was not what was driving it.

Worth separating this from a neighbouring effect. Choking under pressure is about stakes and evaluation dismantling a skill that had become automatic, and it sits in a different literature from mere presence. Here nothing is at stake and nobody is grading you. The room is simply occupied, and that alone sorts your work into the kind that speeds up and the kind that does not.

A Face on a Screen Counts

The obvious objection is that a video is not a person. Sung Park and Richard Catrambone tested something close in 2007. People solved anagrams, mazes and modular arithmetic alone, next to another person, or next to a virtual human on a computer screen, an animated character with no ability to see them.

With the virtual human on screen, people did better than alone on the easy tasks and worse than alone on the difficult ones. The same split, produced by pixels. The authors concluded that, as with a human, virtual humans can produce social facilitation, and they added a warning that reads differently once you have a study stream playing on your second monitor.

A recording of a real person is not an animated character, so the step from their study to your livestream is an inference, not a finding. But if a cartoon on a monitor can produce the effect, the stranger with the tea almost certainly can, and that includes the half of the effect you do not want.

What to Double, and What to Do Alone

Put the findings together and body doubling stops being a vibe and becomes a tool with a specification.

Use it for the work you already know how to do. Starting, which is a well-learned action you are avoiding, not a hard one. Admin. Reps. The dishes. The thirty emails. Anything where the problem is doing it rather than working out how. Every study above says this is where the effect pushes in your favour.

Take the genuinely new thinking somewhere quiet. The first time you structure an argument, learn a notation, or make a decision you have never made before, the research says company costs you speed and accuracy. The fix is usually not to abandon the double but to split the task. Outline alone, then type with the stream on. Decide alone, then execute in company.

Let it be small. A few percent of variance will not rescue a task you fundamentally do not want to do. What it reliably does is lower the cost of beginning, and beginning is where most goals actually die.

Move along the mutuality line when a stream stops working. A stranger who cannot see you is the thinnest version of presence. A friend who is also working, who knows what you said you would do today, is a thicker one, and it borrows from a separate effect: knowing someone shares your goal intensifies your pursuit of it. The stream gives you a body in the room. A person who notices when you show up gives you that plus a reason to come back tomorrow, which is roughly the bet Future You makes, free on iOS and Android.

Setting one up is less formal than it sounds. The options are the ones from the start of this article: a study or clean with me video, a co-working app that pairs you with a stranger for a fixed block, or a friend who brings their own work to your table. A stream is the easiest to start at eleven at night. Cameras on and microphones off is a sensible default, since Markus showed the other person does not need to watch you or speak for the effect to appear. Choose the task before the session begins, so the first minutes go on doing it.

The Room Was Always Part of the Work

The folk version of productivity treats focus as something you generate alone, by wanting it enough. More than a century of research says something plainer. The presence of another person is an input, it works without their attention, it works through a screen, and it helps the practised and hinders the new.

Which means the stranger with the tea was never a distraction you got away with. She was doing a job, and now that you know what the job is, you can hire her for the right tasks and send her home for the rest.