Skip to content
MyFreud

The Psychology of Videos You Cannot Stop Watching

A video ending is not what keeps you watching, the next one starting is. The psychology of anticipation, reward, and why "one more" is so hard to refuse.

5 min read

Pop-art illustration of a man in profile looking down at his glowing phone screen at night, a lit city skyline reflected in water behind him.

Key takeaways

  • The urge to keep watching comes from anticipation, not from the reward itself. Dopamine neurons fire most strongly to the prediction of what happens next, which is why a video ending is not the point where the pull stops.
  • Unpredictable rewards are the most persistent kind. Skinner and Ferster showed that a variable schedule, where you cannot tell in advance whether the next thing pays off, produces far more persistent behaviour than a reliable one does.
  • This is not a personal failing. Formats that escalate stakes and withhold the outcome until the last moment are built around exactly this mechanism, and they work on a well-functioning brain, not a broken one.
  • Deciding to stop and actually stopping run on different systems. Research on binge-watching found viewers who intended to stop often kept going anyway, especially when stopping meant confronting an unrelated task they had been avoiding.
  • The reliable lever is friction, not willpower. Removing autoplay, closing the tab, or deciding the cutoff before you start work better than resolving to want it less in the moment.

A video ending is not what makes the pull to watch another one stop, because the pull was never really about the video you just finished. It comes from anticipating the next one, and that anticipation runs on its own machinery, largely separate from whatever satisfaction the last video actually delivered.

Why the next thing matters more than this one

The clearest evidence for this comes from research on dopamine, and it says something narrower and more useful than the popular version of “dopamine hit.” Recording from dopamine neurons directly, researchers found they fire most strongly to the prediction of a reward, and specifically to the gap between predicting it and knowing whether it actually arrives. [schultz-1997-prediction] Once the outcome is certain, the signal drops away. The system is built to track what might happen, not to linger on what did.

That has a direct consequence for anything designed to be watched. A video that resolves its own question quickly, a joke that lands, a fact that gets stated and explained, gives the anticipation system nothing further to do. A video that keeps a question open, who wins, what happens at the end, whether the attempt succeeds, keeps that system engaged for as long as the answer is withheld. The pull to keep watching is strongest exactly where the outcome is least certain, which is a property of how the content is built rather than a comment on the viewer watching it.

The reward schedule almost nobody notices

The second half of the mechanism is older and comes from a different field entirely. In the 1950s, B.F. Skinner and Charles Ferster ran extensive experiments on how the timing of a reward changes behaviour, and one finding has held up ever since: a reward delivered on an unpredictable schedule produces far more persistent behaviour than one delivered reliably. [ferster-skinner-1957] Animals rewarded on a variable schedule kept responding for far longer after the reward stopped altogether than animals on a fixed schedule ever did, because a fixed schedule teaches you exactly when to expect the payoff and exactly when to stop bothering. A variable one never teaches you that, so you keep checking.

A high-production YouTube video built around an escalating, uncertain contest is a modern, voluntary version of that schedule. Jimmy Donaldson, the creator known as MrBeast, built one of his most-watched uploads, “$456,000 Squid Game In Real Life!”, entirely around this shape: 456 competitors, a series of elimination games, and one outcome withheld until the closing minutes. Nobody sits through a 25-minute video for the premise; the pull is the same one Skinner described, an outcome that stays genuinely uncertain for as long as the format can sustain it.

$456,000 Squid Game In Real Life!

Recommendation feeds add a second layer on top of the same mechanism. Where one video ends on an uncertain note, the next one starts before the feeling has anywhere to go, so the schedule that used to reset between episodes now barely resets at all.

What viewers report keeps them watching Illustrative
Why people keep going
  • Wanting to know how it ends 40% of reasons given
  • Enjoying the content itself 25% of reasons given
  • Avoiding a task or feeling 20% of reasons given
  • Habit, not a specific reason 15% of reasons given

Illustrative breakdown, not a measured statistic. Drawn from the categories of motivation described in binge-watching research; no single study reports these exact proportions.

Deciding to stop and actually stopping are different things

The gap between meaning to stop and stopping is not a failure of character, and it has been studied directly. A study using event-reconstruction methods, where participants reported on real recent viewing sessions rather than describing their habits in the abstract, found that people who intended to stop watching frequently continued anyway, and that this was especially likely when stopping meant switching to a task they had been putting off. [lades-2022-bingewatching] The same research found that viewers with higher self-control did not necessarily watch less, but reported less guilt and tiredness afterward, largely because their viewing interfered less with whatever else they were supposed to be doing. The problem, where there is one, is usually about what watching displaces rather than about the watching itself.

That reframes what actually helps. Deciding in the moment to want it less is competing against a system built to keep predicting what happens next, and it tends to lose. Removing the decision in advance works better: turning off autoplay, setting a stopping point before starting, or moving the habit somewhere it takes more effort to reach. None of that requires more willpower. It just removes the point at which willpower would otherwise have to do the work.

This is a mechanism, not the whole picture. Our guide to screen time and mental health covers the broader question of what all this adds up to, including why total hours turns out to be a weak measure and what the displacement it causes actually costs.

When to speak to someone

Speak to a doctor if low mood, anxiety or sleep problems have lasted more than a couple of weeks, whatever you think is causing them, or if screen use is being used specifically to avoid a feeling that keeps coming back once the screen is off. Watching a lot of video is common and usually harmless. It is worth a closer look when it is consistently taking sleep, replacing contact with people, or is the only way a difficult feeling gets managed.

There is no published screener for screen or video habits on this site, and none is linked here for that reason. If low mood, anxiety or sleep disruption is the concern sitting underneath the viewing, the self-assessment hub has instruments for those instead.

How MyFreud can help

The easiest way to see whether video habits are actually costing you sleep, rather than just filling the evening, is to track both for a couple of weeks and look at them side by side. A pattern is much easier to act on than an impression of it.

Download MyFreud and start today: App Store or Google Play.

Frequently asked questions

Why can I not stop watching YouTube videos?

Because the pull is built around anticipation rather than payoff. Dopamine neurons respond most strongly to the prediction of what happens next, not to the moment of resolution, so a video ending does not switch the urge off the way you would expect. Recommendation systems and escalating formats are built around exactly this gap between predicting and knowing, which is why closing one video rarely feels like a natural stopping point.

Is binge-watching a sign of a bigger problem?

Not on its own. Watching several episodes or videos in a row is common and, for most people, is a source of enjoyment rather than a symptom of anything. It is worth a closer look when it is consistently displacing sleep, when it is the only way you can tolerate a difficult feeling, or when you regularly watch for longer than you intended and feel bad about it afterward. Those are questions about cost, not about hours.

Why do some videos feel impossible to look away from?

They are usually built around an unresolved question held open as long as possible: who wins, what happens, how it ends. Uncertainty of that kind is what a variable reward schedule is, in psychology terms, and it is a far stronger pull on attention than a predictable outcome delivered on schedule. That is a property of the format, not a comment on the viewer.

Why is it so hard to actually stop once I decide to?

Deciding to stop and stopping draw on different processes, and a study of binge-watchers found that people who intended to stop often kept going regardless, particularly when the alternative was returning to a task they were avoiding. Meaning to stop is not the same skill as stopping, which is why a rule decided in advance, like closing the laptop at a set time, tends to work better than relying on willpower in the moment.

What actually helps if I want to watch less?

Change what happens automatically rather than trying to want it less. Turning off autoplay, moving the app off the home screen, or deciding the stopping point before you start all remove a decision you would otherwise have to make while already absorbed, which is the point at which willpower is weakest. Protecting sleep specifically matters most, since that is usually the first and largest cost.

References

  1. 1.Schultz W, Dayan P, Montague RR ( 1997). A neural substrate of prediction and reward. Science. doi:10.1126/science.275.5306.1593
  2. 2.Ferster CB, Skinner BF ( 1957). Schedules of Reinforcement. Appleton-Century-Crofts.
  3. 3.Lades LK, Barbett L, Daly M, Dombrowski SU ( 2022). Self-control, goal interference, and the binge-watching experience: An event reconstruction study. Computers in Human Behavior Reports. doi:10.1016/j.chbr.2022.100220