Most people check their phones without ever deciding to. The screen lights up, or it doesn’t even need to — the phone is just there, and a hand reaches for it before the thought catches up. Two separate studies published in 2023 asked what’s actually happening in that reach, and found it isn’t one thing. One describes a pattern that looks compulsive but mostly isn’t. The other describes a pattern that is compulsive, and comes with measurable costs attached.
The habit that looks worse than it is
A group of Chinese scientists ran two experiments with 62 college students, published in Frontiers in Psychology in September 2023. They tracked how often participants unlocked their phones under different conditions: phone visible versus hidden, with or without instructions to exercise self-control. Visible phones and no self-control instruction produced significantly more prolonged use, a result the authors describe as consistent with cue-reactivity, a phone in sight simply invites a glance, largely outside conscious decision-making.
The paper’s own framing draws a line here. Habitual behaviour, in their description, “occurs outside of awareness” and “does not necessarily lead to negative consequences.” That’s a meaningfully different category from what researchers term problematic smartphone use, which does carry documented harm. Picking up a phone because it’s sitting there isn’t the same act as picking it up because something uncomfortable needs managing.
The pattern that isn’t harmless
A second 2023 study, led by Ana Nikolić and published in Frontiers in Public Health, surveyed 761 medical students across Belgrade and Niš. Just under 22 percent screened as having smartphone addiction by the study’s measure, and those students carried real statistical costs alongside it: 1.65 times higher odds of poor sleep quality, 2.04 times higher odds of anxiety, 2.29 to 2.51 times higher odds of depression, and 1.75 times higher odds of elevated stress, compared with students who didn’t screen as addicted.
These are two different studies with two different designs, not one continuous piece of research. But put together, they describe something coherent: a version of phone use driven by habit, cued by the object simply being visible, that carries little apparent cost, and a version driven by something closer to compulsion, associated with worse sleep and more anxiety in the people who report it. The same physical act, on a screen that looks identical either way.
The third complication: thinking you’re addicted may be its own problem
A more recent study adds a layer that neither of the 2023 papers anticipated, and it changes how the habitual-versus-compulsive distinction should probably be used. Ian Anderson, at the California Institute of Technology, and Wendy Wood, at the University of Southern California, published research in Scientific Reports in November 2025 built around a simple mismatch: across 1,204 participants and two studies, only 2 percent of active Instagram users met clinical criteria for addiction on the Bergen Instagram Addiction Scale, a validated clinical measure, while 18 percent of the same population self-reported being “at least somewhat addicted.” That’s roughly a ninefold gap between how many people actually show clinical symptoms and how many believe they do.
The more striking part of the study wasn’t the gap itself. It was what believing the label did to people. In a second experiment, Anderson and Wood had participants reflect on their use through language borrowed from the U.S. Surgeon General’s advisory on social media and addiction. Simply prompting people to think about their scrolling in those terms reduced their sense of control over their own use, increased how much they blamed themselves for past lapses, and did all of this without any actual change in how much they were using or how compulsively. The addiction framing itself produced some of the effects usually attributed to addiction.
What’s actually driving the cue in the first place
None of the three studies above explain why a phone lying on a table pulls attention in the first place, and it’s worth being careful here about how much the underlying mechanism is actually understood.
The most-cited explanation in consumer technology writing borrows the language of variable-ratio reinforcement, the same reward structure behind slot machines, where a notification, a like, or a new post arrives on an unpredictable schedule rather than a fixed one. Unpredictable rewards are well established in behavioural psychology as producing more persistent checking behaviour than a reward that always arrives on schedule. What the research on smartphones specifically supports is narrower than the popular “slot machine in your pocket” framing suggests: notification design clearly shapes how often people check their phones, but the leap from that to a specific dopamine mechanism is more often asserted by app critics and product writers than demonstrated in the underlying neuroscience.
The honest version of this claim is behavioural, not chemical: unpredictable, socially relevant interruptions get checked more than predictable ones, for reasons psychology has studied in other contexts for decades, and phones happen to be an extremely efficient delivery mechanism for that kind of interruption.
The direct evidence on what happens when the interruption is actually removed complicates the cue-based story rather than confirming it. Dekker, Baumgartner, Sumter and Ohme ran a field experiment with more than 200 Android users aged 18 to 30, published in Media Psychology, in which one group disabled non-essential notifications for a set period. The result was not the clean reduction in checking a simple cue-reactivity model would predict: the researchers “found no difference” between groups in either time spent on the phone or how often it was checked. Removing the buzz didn’t remove the behaviour. There was also a genuine drawback, not just a null result: participants in the no-notifications group reported more fear of missing out, which the authors read as evidence that disabling notifications carries its own cost rather than being a straightforward improvement.
The one benefit that did show up was smaller and more interesting than “checked less”: people in that group described their phone use as more intentional, and reported a stronger sense of control over it, even though the amount of use hadn’t changed. Whatever is driving habitual checking, it survives the removal of the specific cue that’s supposed to trigger it, which suggests the phone itself, not any single notification, has become the cue.
Why this matters more than the addiction framing
Most public conversation about screen time treats all of it as roughly the same problem at different intensities, more scrolling equals more concern. This research suggests intensity isn’t actually the variable that matters, and that the language used to describe the behaviour may matter almost as much as the behaviour itself. What distinguishes the costly pattern from the harmless one is closer to function: is the phone being picked up because it’s there, or because something else, boredom, distress, the need to not feel a particular way, needs somewhere to go. Two people with identical daily screen-time totals could be doing entirely different things with that time, for entirely different reasons, with entirely different outcomes attached. And a third person who is doing the harmless, habit-driven version of it may talk themselves into the costly one simply by deciding they’re “addicted.”
Neither the Li nor the Nikolić study supports the idea that less screen time is automatically better, or that any given hour of scrolling should be read as a warning sign. The Anderson and Wood findings add a specific caution to that: labelling ordinary, habitual phone use as an addiction, whether in a public health campaign, a workplace wellness program, or a person’s own self-assessment, is not a neutral description. It appears to carry its own cost, separate from whatever the underlying behaviour was actually doing.
Put together, what the three studies support is a narrower, more useful distinction than the standard screen-time debate offers: the scrolling worth paying attention to is the scrolling that happens because something needed managing, and a phone was the fastest way to manage it — not the scrolling that happens because a phone is sitting on the table, and not the scrolling someone talks themselves into worrying about by deciding they have a problem.
