Put a pen between your teeth. Not your lips — your teeth, so your cheeks pull back and your mouth stretches into something that's technically a smile. Now read a comic strip. According to one of the most famous experiments in psychology, you'll find it funnier than you would have otherwise, because your face just told your brain it was having a good time.
That experiment is the grandfather of every "just smile, you'll feel better!" you've ever been handed. And here's where the science stands now, so you don't have to scroll for it: the pen trick is on genuinely shaky ground. When seventeen labs tried to repeat the famous study, they got nothing. Then, in 2022, a global team partly rescued the idea. A deliberate, voluntary smile really does seem to lift your mood, by roughly as much as looking at a mildly pleasant photo. A nudge. Not a fix, and definitely not a treatment plan. The myth was never that smiling does nothing. The myth is that it does much.
The pen study
In 1988, the German psychologist Fritz Strack and his colleagues published an elegant little experiment. Participants held a pen either in their teeth, which forces the smile muscles to fire, or in their lips, which blocks them, and then rated how funny some cartoons were. The teeth group found the cartoons funnier.
The clever part was in the subtitle: a nonobtrusive test. Nobody in the room knew the pen was making them smile, so nobody could be accused of playing along. The face moved, the mood followed, and the participants had no idea why. It looked like clean proof of the facial feedback hypothesis: the old idea that expressions don't just broadcast emotions, they help manufacture them.
The finding escaped the lab almost immediately. It landed in textbooks, in corporate trainings, in the general cultural water supply that produces advice like "fake it till you make it" and mugs that say choose happy. For nearly thirty years, if you asked why forcing a smile should work, the pen study was the answer.
Seventeen labs and one camera
Then psychology's replication crisis came for it. In 2016, a Registered Replication Report, the format where many independent labs run the exact same vetted protocol and publish whatever they find, put the pen study to the test. Seventeen labs found no support for the effect. Pooled together, the difference between the smile condition and the pout condition was 0.03 units on the rating scale, with a confidence interval sitting comfortably on both sides of zero.
Three hundredths of a point. Essentially nothing.
But there's a twist, and it's my favorite part of the whole saga. To verify that participants held the pen correctly, the replication labs had pointed a video camera at them. The original study hadn't. In 2018, Tom Noah, Yaacov Schul, and Ruth Mayo at the Hebrew University of Jerusalem reran the experiment both ways: without a camera, the facial feedback effect showed up; with a camera present, it vanished. Feeling observed, they argued, was enough to erase it.
You can read that as a rescue or as an indictment. Either way, it tells you something useful: an effect that a webcam can kill is not an effect to build a philosophy of happiness on.
Meanwhile the broader literature was getting its own audit. In 2019, Nicholas Coles, Jeff Larsen, and Heather Lench pooled 286 effect sizes from 138 studies covering more than 11,000 participants and concluded that facial feedback effects on emotional experience are real on average, but small and all over the map. Not dead. Not impressive either.
The comeback, with an asterisk
Coles didn't leave it there. He organized the Many Smiles Collaboration, a deliberately global, deliberately careful study published in Nature Human Behaviour, with 3,878 participants across 19 countries. Participants tried one of three things: mimicking photos of smiling actors, deliberately pulling the corners of their mouths toward their ears, or the classic pen-in-the-mouth maneuver.
The first two worked. Mimicking a smile or consciously posing one produced a real, detectable increase in reported happiness. The pen trick, the one that started the whole thing, stayed inconclusive. So the science changed its mind twice: first the effect collapsed, then half of it stood back up. The half that survived is the boring half — a smile you know you're making, whole face involved. The magic half, where your face secretly manipulates your brain without your consent, is still in intensive care.
And how big is the surviving effect? The Many Smiles team put it at roughly the size of the mood bump you get from looking at mildly positive images. Detectable across 3,878 people. In your actual Tuesday? About the lift of a nice photo of a dog.
What a nudge is actually for
So here's the honest, slightly complicated scorecard. A voluntary smile you're aware of making: small, real lift. A covert smile smuggled in by a pen: unproven, possibly nothing, possibly just camera-shy. A reason to tell someone who's struggling to "just smile": still zero.
Psychology is full of advice like this: a modest, conditional finding that got rounded up into a law. The thought-suppression rule turned out to be half wrong in a similar way, and the meditation apps have their own fine print. The pattern is always the same: the nuance dies in transit, and what arrives is a slogan.
The real problem with "just smile" isn't that it's useless. It's that it treats a low mood as a display error, something to be patched at the screen. Sometimes that's fine. Some moods really are weather, and a posed smile in the elevator costs you nothing and might buy you a flicker. Take the free flicker.
But when the mood is low because something in your life is actually unresolved — the conversation you're avoiding, the thing you haven't said to anyone — your face is the wrong organ to work with. The research above gives you a dog-photo-sized bump. Naming the actual problem, out loud, in your own words, is playing a different sport entirely. That's more or less why we built Joice around talking instead of posing: you say what's actually going on, something listens and asks the next question, and the mood gets addressed at the source instead of at the surface.
Run the experiment yourself tonight. Pull your mouth into a real, deliberate smile and hold it for ten seconds; notice the small flicker, or notice its absence. Both are honest data. Then say one true sentence, out loud, about whatever's actually sitting on you.
Notice which one moved the needle.