Why Can't You Stop Scrolling?

You pick up your phone to check one thing.

Fifteen minutes later you are watching a video of someone’s vacation in a country you’ve never thought about, and you cannot identify the exact sequence of steps that brought you here.

You were not trying to do this. You do not feel particularly good. You are going to put the phone down in a second.

You do not put it down.

Skinner’s Most Powerful Schedule

In the 1930s, behavioral psychologist B.F. Skinner discovered that not all rewards are created equal — that the schedule on which a reward is delivered has a dramatic effect on how hard organisms will work for it and how resistant the behavior becomes to extinction.

He identified four basic schedules. Fixed ratio (reward after every N responses) and fixed interval (reward after N time has passed) both produce predictable, moderate behavior. Variable interval (reward at unpredictable times) produces steady, patient behavior. But variable ratio — reward after an unpredictable number of responses — produced something different: the highest rate of responding, the most persistent behavior, and the greatest resistance to stopping once the rewards ceased.

Variable ratio is the schedule of slot machines. You pull the lever. You don’t know if this pull will pay out, but you know it might. The not-knowing is the engine. When the reward is unpredictable, the response pattern never settles — you keep going because you have no clear signal that this time won’t be the one.

The social media feed is a variable ratio schedule.

Each scroll is a lever pull. You don’t know if the next post will be funny, beautiful, infuriating, or completely uninteresting. The uncertainty is the feature, not the bug. The designers built a slot machine and put it in your pocket.

What Dopamine Is Actually Doing

The common explanation — “social media triggers dopamine” — is correct but imprecise. It implies the reward itself causes the dopamine. The mechanism is more specific and more troubling.

Wolfram Schultz at the University of Cambridge spent decades recording from dopamine neurons in primates. In a now-classic finding published in 1997 in Science, he demonstrated that dopamine neurons don’t fire at the reward — they fire at the prediction of the reward. When an animal has learned that a light predicts food, the dopamine response shifts from the food itself to the light that precedes it.

And crucially: if the reward is uncertain, dopamine neurons fire even more intensely during the anticipation than they do under certainty.

Uncertainty amplifies dopamine. The “maybe” is more potent than the “yes.”

This is why checking your phone feels compelling even when you don’t expect anything interesting. The act of checking is the cue that might — might — predict a reward. The dopamine fires before the outcome. The pull is in the opening of the app, not the content inside.

The People Who Built It Said So

In 2017, Sean Parker — the founding president of Facebook — gave an interview in which he described the explicit psychological model behind the platform’s design.

“How do we consume as much of your time and conscious attention as possible? And that means that we need to give you a little dopamine hit every once in a while, because someone liked or commented on a photo or a post or whatever. And that’s going to get you to contribute more content, and that’s going to get you… It’s a social-validation feedback loop… exactly the kind of thing that a hacker like [Zuckerberg] would come up with, because you’re exploiting a vulnerability in human psychology.”

He added: “God only knows what it’s doing to our children’s brains.”

This was not a criticism from outside the system. This was a description of the intended design, from the person who helped build it, in retrospect.

Tristan Harris, a former design ethicist at Google, spent years documenting the same mechanisms. His central argument: a small number of people at a small number of companies make choices about interface design that collectively determine how billions of people spend their attention. The people making those choices are optimizing for engagement — time spent, not wellbeing.

The Removal of Stopping Points

Aza Raskin designed infinite scroll — the feature that eliminates the natural pause between pages of content, replacing discrete pagination with an endless, frictionless feed.

He later calculated that infinite scroll wastes approximately 200,000 hours of human attention per day, just on Twitter alone. He publicly expressed regret for having built it.

The mechanism is straightforward. Before infinite scroll, loading a new page created a natural stopping point — a gap in the experience where the choice to continue was explicit. You had to click. With infinite scroll, there is no gap. The content continues. The choice to stop requires generating momentum against inertia rather than simply pausing.

Slot machines operate on the same principle. The most important design feature of a slot machine is the speed of the next pull. Eliminate the wait, and you eliminate the moment of reflection where a person might decide to stop.

The Paradox: It Doesn’t Even Feel Good

Research on social media use and mood consistently shows a paradox: passive consumption of social media — scrolling, watching, reading, without posting or interacting — tends to make people feel worse, not better.

A 2015 study by Phillipe Verduyn and colleagues at the University of Leuven, published in the Journal of Experimental Psychology: General, found that Facebook use predicted lower life satisfaction and more negative affect — and that the effect was mediated by social comparison and envy. Seeing curated highlights of other people’s lives produced unfavorable comparisons.

Yet people continue. Use does not decline proportionally with negative experience, because the variable ratio schedule doesn’t require positive affect to maintain behavior. You don’t need to feel good to keep pulling the lever. You need to feel the possibility that the next pull might be good.

The behavior has been reinforced thousands of times across years. The grooves are deep. The “just one more” is not a decision. It’s a pattern that has been built into the response repertoire by intermittent reinforcement across thousands of sessions.

What Would Actually Help

The behavioral literature suggests several things.

Removing the variable element reduces the compulsiveness. Turning off notifications — which are delivery mechanisms for variable-ratio reward signals — reduces the frequency of checking. Batch-checking (email or messages at scheduled times) imposes a fixed-interval schedule, which is substantially less compelling than variable ratio.

Creating physical friction helps: putting your phone in another room is more effective than trying not to look at it while it’s in your pocket. The choice not to look requires ongoing willpower. The phone in the other room doesn’t.

Understanding the mechanism helps, but not as much as you’d hope. Knowing that you’re being subjected to a variable ratio schedule does not fully neutralize it — the dopaminergic anticipation doesn’t care what you know. It responds to cue, not understanding.

The phone is not the problem. The variable ratio schedule is not new — gambling has used it for centuries. What changed is that the most sophisticated behavioral engineering in human history is now distributed to everyone, for free, and optimized continuously by algorithms that measure engagement in milliseconds.

You are not weak. You are human.

The lever is very well designed.