The Twenty-Second Hug

You have met this instruction. It arrives in a caption, or from a friend who read it in a caption: hold the hug for twenty seconds, because that is how long it takes for the good chemistry to arrive. Anything less doesn’t count. People stand in kitchens counting.

It is a strangely precise number for something nobody has ever timed you doing. And precision, in this line of work, is the thing worth pulling on. A round twenty is the sort of figure that gets repeated because it sounds measured, not because anyone measured it.

The belief, stated kindly

Underneath the stopwatch there is a real observation: a quick, distracted hug feels like nothing, and an unhurried one does not. That much is true and the research supports it. The instruction is people trying to describe the difference between being touched and being processed, and reaching for a number because a number is easier to pass on than a feeling.

Where the number came from

The trail runs back to a study of 183 adults in cohabiting couples. Half the couples spent ten minutes holding hands while a romantic film played, and then embraced for twenty seconds. The other half sat quietly, not touching. Everyone then faced a public-speaking task. The couples who had touched showed lower blood-pressure and heart-rate responses to the stressor than the couples who had not.¹

Read that protocol again, because it is where the myth is made. The twenty seconds is one component of a ten-minute warm-contact procedure. It was a design decision, not a finding. Nothing in the study compared twenty seconds against ten, or forty, or none, and nothing in it isolates the hug from the film, the hand-holding, or the fact that the couples in that condition had just spent ten quiet minutes together. The number escaped the paper and lost the paragraph it came in.

What happened when somebody finally varied it

In 2021 a group in London ran the experiment the folklore assumes exists. Participants received hugs of one, five and ten seconds. One-second hugs were rated as the least pleasant and left people feeling least in control. Five- and ten-second hugs were rated better — and, crucially, were not different from each other.²

So the only direct dose-response we have on hug length flattens out somewhere around five seconds and never gets tested as far as twenty. It is a laboratory hug between people who may not know each other well, and rated pleasantness is not the same as anything happening in your blood. But it is the closest thing to an answer anyone has produced, and it does not say twenty.

What the pooled evidence says

The largest synthesis of touch research to date pooled 137 studies, reviewed 75 more, and covered 12,966 people. Touch interventions came out ahead across physical and mental outcomes, with the clearest signal on cortisol (Hedges’ g = 0.78, 95% CI 0.24 to 1.31).³

And then the line that ends the stopwatch argument. Delivering more touch sessions increased the benefit in adults. Increasing the duration of touch per session did not — in adults or in newborns. The variable that carried the effect was how often, not how long.

Out of the lab

The strangest study in this literature is also the most persuasive. Four hundred and four healthy adults were interviewed by telephone every evening for a fortnight about conflict in their day and whether anyone had hugged them. They were then deliberately exposed to a virus that causes the common cold and kept in quarantine while researchers recorded who became infected and who showed symptoms.⁴

People who felt supported were protected against the rise in infection risk that came with more frequent conflict, and hugs accounted for 32 per cent of that buffering effect. Note what was counted: the number of days someone was hugged. Nobody asked how long the hugs lasted. Frequency was enough to show up in whether a person caught a cold.

Where it gets complicated

The oxytocin half of the story is thinner than the sharing suggests. The work most often cited for it found that women who reported more frequent partner hugs had higher oxytocin and lower blood pressure and heart rate — but it is observational, rests on self-reported hug frequency, and was conducted in premenopausal women only.⁵ It cannot tell you which way the arrow points. People in warmer relationships hug more and may differ in a hundred other ways.

And a randomised trial of 159 people put through a standardised social stressor found something worth keeping. Those who were hugged, and those who simply placed a hand on their own chest, had lower cortisol afterwards than the control conditions. But their heart rates did not differ, and neither did what they said about how stressed they felt.⁶

That is the honest shape of this whole field: the measurable change and the felt change are not the same size, and the felt one is often absent. Which is worth knowing before you go looking for a rush at second nineteen.

One thing worth sitting with

Twenty seconds is long. Try it and you will find out exactly how long — there is a moment, somewhere past the fifth second, where an ordinary hug would have ended and one of you would have moved to let go. The instruction’s real function may be nothing to do with chemistry. It may simply be a rule that stops you being the one who ends it first.

That is not nothing. But it is a social device, not a dose.

The hug survives the evidence. The stopwatch does not — the only experiment that varied the length found the gain had arrived by about five seconds, and the biggest review of touch research found that hugging more often mattered while hugging for longer did not.

What this might mean at your front door

1. Stop counting. There is no threshold to cross and no evidence of one. Counting also puts your attention on the count, which is its own small problem.
2. If you want a number, five seconds is the honest one. It is where the only real test found the difference, and it is short enough that you will actually do it.
3. More often beats longer. That is the clearest instruction the pooled data gives, and it is the opposite of the one going round.
4. Expect not to feel it. The reliable changes in this research show up in cortisol and blood pressure, not in how people rate their own state. Nothing has gone wrong if the hug just felt like a hug.

If the wider question interests you — what touch is actually doing, and how much of the oxytocin story survives contact with the papers — we went through that separately in Touch, Closeness, and the Quiet Hormone. Our closeness work lives in the same territory: it is company for an unhurried hour, and it claims none of what is described above.

← Back to The Lab