Scrolling and Memory: What a Stanford Preprint Found

Analysis. A preprint from Stanford University and Georgia Tech tested 502 U.S. adults aged 18 to 24. Heavier TikTok and Instagram use, read from phone screen time reports, was not linked to sustained attention. It was linked to poorer memory for pictures seen minutes earlier, especially the rare ones.

What the researchers did

Participants were recruited online through Prolific and did a task on a computer. They sorted 500 scene photos into indoor or outdoor with a key press. Indoor scenes were frequent (450) and outdoor scenes rare (50). Without warning, they then took a recognition test on 100 of those pictures mixed with 100 new ones.

Use was measured in minutes per week of TikTok plus Instagram, copied from phone screen time reports. Among the 398 people with some use, the average was 862 minutes a week, about two hours a day. The other 104 had zero minutes on both apps. Many of them still reported using YouTube Shorts or Facebook Reels.

What they found

For attention, there was no association with use (standardized coefficient 0.004, p = 0.94). For memory, there was one: more use went with a lower overall memory score (coefficient minus 0.174, 95% confidence interval minus 0.275 to minus 0.073). Splitting users into thirds, the light group (2 to 521 minutes a week) scored 0.79, and the medium and heavy groups both scored 0.69.

What the headlines leave out

The authors call the relationship small. Medium and heavy users did not differ from each other, so this is not a clean dose and response pattern. Users and non users did not differ either, which the authors tie to non users watching video on other platforms. The attention test could only rule out correlations of about 0.14 or larger. The frequent pictures were remembered poorly overall, which may hide a real effect there.

This is a preprint, so it has not been peer reviewed. It is correlational: scrolling might change memory, people with weaker memory might scroll more, or both. A pilot with 338 mostly Stanford students showed a similar pattern. The study did not test a ban, a limit or any habit change.

The questions that remain

Would a randomized cut in scrolling change recognition scores? Does the pattern hold for YouTube Shorts and other apps? And if attention holds up while memory slips, where in the chain from watching to keeping does the difference appear? This article is not medical advice.