You open the app to check one thing. Forty minutes later you surface, slightly dazed, unable to say what you watched. The suspicion that this habit is doing something to your attention has become one of the most common self-diagnoses of the decade — usually delivered as a joke, usually not entirely a joke.
A research team at Griffith University in Australia pulled together one of the most comprehensive syntheses of evidence to date on the question. Their meta-analysis, published in Psychological Bulletin (Nguyen and colleagues, 2025), reviewed 71 studies covering 98,299 people, pooling 70 of them, and found that heavier short-form video use was associated with measurably poorer cognition — a moderate overall correlation, with attention and self-control showing the strongest links. But the detail that deserves more airtime than the headline is this: it mattered far less how many minutes people spent watching than how compulsively they watched. Time on the apps showed only a weak link to cognition and a still weaker one to mental health. Scoring high on “addiction-style” measures — the habitual, hard-to-stop kind of use — carried markedly stronger associations than any time-based measure.
We should say plainly what we are: writers reading a body of research, not psychologists rendering judgment on anyone’s scrolling. What follows is what the evidence says, including the parts that cut against the scarier headlines.
What the pooled numbers actually show
The team, led by Lan Nguyen at Griffith’s School of Applied Psychology, gathered every quantitative study they could find on short-form video — TikTok, Douyin, and “general” short-form use where no specific app was named — and its relationship to cognitive and mental health measures. No study in the pool looked directly at Instagram Reels, Facebook Reels or YouTube Shorts. The protocol was preregistered, and the data sit openly on the Open Science Framework.
Across the 14 pooled studies that measured cognition, most of them relying on self-rated attention and self-control rather than lab-tested performance, heavier use correlated with poorer cognition at r = −.34 — a moderate effect, in the convention where .10 is weak and .50 is strong. Attention (−.38) and inhibitory control (−.41) — the psychologist’s term for the ability to stop yourself from acting on an impulse — showed the strongest associations. Memory, language, and working memory showed weak links; reasoning showed none at all — though only one study measured it.
Across the 61 studies measuring mental health, the overall association was weaker (−.21) but consistent: stress (−.34) and anxiety (−.33) led, with depression, loneliness, and sleep quality trailing behind. The pattern held for young people and adults alike — age made no statistical difference. Platform did matter, but only on the mental-health side, where general short-form use showed a somewhat stronger association than TikTok-specific use; for cognition, the platform made no difference.
The finding hiding under the headline
The more interesting result is what moderated those associations. The studies measured short-form video use in different ways: some counted hours, some counted how often people opened the apps, and about half used addiction-style scales — questionnaires about loss of control, preoccupation, use that interferes with daily life.
Those measurement choices changed the answer. Duration — sheer time spent — was weakly linked to cognition (−.20) and barely linked to mental health (−.10). How often people opened the apps showed no significant link to mental health at all. But addiction-style scores correlated with poorer cognition at −.37 and poorer mental health at −.32.
In other words, the person who watches a long documentary’s worth of cooking clips on a Sunday afternoon and puts the phone down is, on these numbers, the weaker signal, though the authors are explicit that the high-duration, low-compulsion viewer is a profile nobody has yet studied directly. The person who keeps flicking upward through content they are not enjoying, because stopping somehow doesn’t happen, is the one the numbers are worried about. The authors’ own framing is that not all engagement is problematic, and that the function and feel of use — not the quantity alone — is what tracks with health.
Two more results deserve mention because they cut against expectation. Short-form video use showed no significant association with body image or self-esteem — a surprise given a decade of commentary, which the authors suggest may reflect the diversity of content and creators on these platforms. And reasoning ability, in the single study that measured it, showed no link either. Whatever is happening here, the current evidence does not support “these apps make you stupid.”
Where the evidence runs out
Now the load-bearing caveat: 87 percent of this research is correlational rather than group-comparison, and most of it is cross-sectional — a snapshot, not a film. The data cannot say that scrolling causes weaker attention. The arrow could run the other way: people who already struggle with attention or anxiety may be drawn harder into an interface built from rapid rewards, or find it harder to leave. The authors say this directly, noting that people experiencing distress may turn to short-form video for relief rather than being harmed by it, and that both processes may operate at once.
The samples skew particular, too: nearly three quarters of the studies were conducted in Asia, participants averaged about 23 years old, and 60 percent were women. The addiction-style scales that carried the strongest associations measure problematic use on a continuum — scoring high on one is not a clinical diagnosis, and the authors caution that the pooled result may reflect the more problematic end of the engagement spectrum rather than the typical user. Heterogeneity between studies was considerable, which is statistician’s language for “the studies disagreed with each other quite a lot.” And a moderate correlation, in psychology, still leaves most of the variation in people’s attention explained by something other than their video habit. Two structural limits matter as much as the samples. Only about one study in ten controlled for general social media use, so this evidence cannot cleanly separate short-form video from time online in general. And the database search closed in October 2024, which on a subject moving this fast is a while ago.
There is suggestive experimental work — the meta-analysis cites a three-week trial by Hunt and colleagues in which people who restricted their social media use reported reductions in depressive symptoms while a control group showed no change — but for short-form video specifically, the longitudinal and experimental record is thin. That is the honest state of the science: a consistent pattern, a plausible mechanism, and an unproven arrow.
What to do with this on a Tuesday night
The practical takeaway is not “delete the apps,” which the evidence doesn’t demand. It is that the useful question to ask yourself is not how long but how it feels to stop. If the answer is “fine, when I’m done,” this research has little to say to you. If the answer is “stopping doesn’t really happen — the feed decides,” that is the pattern the strongest associations attach to, and it may be worth treating the endless scroll the way you’d treat any designed environment: something built by professionals to be hard to leave.
And a note of care: the mental-health findings here — the links to anxiety, stress, and low mood — are associations in aggregate data, not verdicts on any one person’s evenings. If scrolling has become the main way of managing feelings that are getting heavier, a therapist or counselor is a better next step than a screen-time app. The research describes a pattern; it cannot see you.
What the feed knows is how to keep you; what it cannot know is what it costs you. On current evidence, neither do we — not yet, not for certain. The count that matters, for now, is not the minutes. It’s who decides when they end.
