It's 11:47 on a Tuesday night and someone you sat next to in tenth-grade chemistry just got engaged on a clifftop in Lisbon. You are eating cereal over the sink. The engagement is not the problem. The problem is the quiet arithmetic that follows: her, promoted-guy-from-college, the friend training for the marathon, the roommate who "just felt like buying a house." Data point by data point, the conclusion assembles itself. Everyone else is doing fine. You are the only one white-knuckling it.
That conclusion has been studied, carefully, and it's wrong in a specific and fixable way.
the evidence is rigged
In 2011, Alexander Jordan and colleagues at Stanford published a series of studies with a title that gives the ending away: Misery Has More Company Than People Think. Across four studies, people consistently underestimated how often their peers felt lonely, rejected, homesick, or overwhelmed. Not by a little. And the underestimate had a mechanism: negative emotions mostly happen in private, and when they do surface in public, people actively suppress them. You cry in your car and argue in your kitchen. Then you go get brunch.
So everyone's dataset is corrupted the same way. You have complete, unedited footage of your own interior, every 3am doubt included, and only the trailer of everyone else's. Comparing the two isn't pessimism. It's a sampling error.
The Stanford team found the error has teeth, too. The more people underestimated others' dark moments, the lonelier they felt and the more they ruminated. Believing you're the only one struggling is itself a way of struggling.
a very old bug
None of this started with smartphones. Psychologists call the pattern pluralistic ignorance, and the classic demonstration is from 1993, when Deborah Prentice and Dale Miller surveyed Princeton students about drinking on campus. Privately, most students were uncomfortable with how much everyone drank. Publicly, nobody said so. The result was a campus where the average student believed the average student was fine with it, and some students drank their way toward a norm that, strictly speaking, no one held.
Miller spent the next three decades on the phenomenon, and his 2023 retrospective in Frontiers in Social Psychology traces a century of versions of the same mistake: we read each other's public performances as private truth, forgetting that everyone else is performing too. The lecture hall where nobody asks a question because nobody else looks confused. The friend group where everyone is exhausted by the group chat and no one leaves.
Growing up online just industrialized the bug. The feed is pluralistic ignorance with a recommendation algorithm: a machine built to show you other people's public selves at unprecedented volume, at the exact hours you're most alone with your private one.
the comparison you're actually running
There's decent experimental evidence on what that does. In a 2023 study in Behavioral Sciences, Phillip Ozimek and colleagues showed 409 people Instagram content built around ability comparisons, the achievement-shaped stuff: fitness progress, career wins, enviable trips. Compared to people shown opinion-based content, they came away with lower self-esteem, more negative feelings, and lower life satisfaction, from minutes of exposure. The clifftop engagement isn't neutral information. It's an input your mood does math with.
The maddening part is that knowing all this barely helps on its own. You can recite "it's just a highlight reel" like a catechism and still feel the drop when the photo lands, because the comparison doesn't run in the part of you that knows things. It runs in the part that counts them. Any real fix has to change what gets counted.
What helps isn't logging off forever, since nobody does, or deciding everyone's secretly miserable, which is just the same error flipped. What helps is keeping honest books on your own side of the ledger. The researchers' word for the missing data is suppression: the true version of people exists mostly in private, spoken to no one. So say your version somewhere, out loud, on the record. That's roughly why we built Joice, a voice journal that listens like a friend: a few minutes of saying how the day actually went builds you an archive of your real interior, and it gets harder to believe in everyone else's highlight reel once you can hear how different your own truth sounds from your own posts.
And once in a while, run the experiment socially. Ask a friend how they're actually doing, and answer the same question honestly when it comes back. Prentice and Miller's students needed someone to say the quiet thing before the norm could crack.
The person on the clifftop in Lisbon posted the photo. She did not post the argument about the guest list, the job she's afraid to quit, or her own 11:47pm arithmetic about someone else's feed. That footage exists. It just never ships. Yours doesn't either, which is worth remembering the next time you mistake somebody's trailer for their life.