You have heard it. The average person now focuses for eight seconds, less than a goldfish, and phones did it. The story is catchy and it is junk. The number shows up in a 2015 Microsoft Canada report, which credited a website called Statistic Brain. When the BBC chased it in 2017, the sources Statistic Brain named, the US National Center for Biotechnology Information and the Associated Press, could find no record of the research. Gemma Briggs, a psychologist who studies attention, said there is no single attention span at all. How long you focus depends on the task.[1]

So the goldfish is gone. That does not mean social media is harmless. Here is what the better evidence says.

The broad panic is weak

The biggest early study of screens and teens came from Oxford in 2019. Amy Orben and Andrew Przybylski ran three large datasets covering 355,358 young people and found the link between technology use and well being was negative but tiny: at most 0.4 percent of the variation. Eating potatoes showed nearly the same effect.[2] That is the honest starting point. "Screens" as one lump thing do not explain much.

The narrow problem is real

The picture gets sharper when you split screens by type. A study from Karolinska Institutet and Oregon Health and Science University followed 8,324 American children for four years, starting at age 9 or 10. Kids who spent a lot of time on social media slowly developed more inattention symptoms. Watching TV or videos and playing video games did not show that pattern.[3] The authors also checked whether distracted kids simply drifted to social media. They did not. The direction ran from social media to inattention, and the effect held after accounting for family income and genetic risk for {W("ADHD","https://en.wikipedia.org/wiki/Attention_deficit_hyperactivity_disorder")}. They were also clear that the effect for any one child was small.[4]

Short video looks worse than social media in general. A 2026 systematic review of 23 studies, most from China, found attention and self control were the mental skills most affected by short video addiction.[5] A 2025 meta-analysis in Psychological Bulletin pooled 71 studies and 98,299 people. Heavier short video use went with poorer attention and weaker impulse control, and adults showed it about as much as teens.[6] These are links, not proof that the app caused it. But they keep pointing the same way.

The feed is built to hold you

The design is not neutral. Infinite scroll and autoplay remove the natural places to stop. Researchers at Aalto University tested what happens when you add friction to TikTok with 48 people. Turning off autoplay and making scrolling harder cut compulsive use. It also made people like the app less.[7] That trade off is the business model in one sentence. The version that holds your attention best is the one that sells.

So did it kill our attention spans?

No. The goldfish number is made up, and the evidence that screens broadly wreck young people is weak. What holds up is narrower: heavy social media and short video use are linked to modest increases in attention problems, in kids and adults, and the feeds are designed to keep you in them. The fix is not panic. It is knowing which apps are built to hold you, and deciding when to put them down.

Sources

  1. BBC News, Simon Maybin, "Busting the attention span myth," March 10, 2017 (the 8 second figure in a 2015 Microsoft Canada report cites Statistic Brain; NCBI and AP could find no record of the research; Gemma Briggs: attention depends on the task)
  2. Oxford Internet Institute, "Technology use explains at most 0.4% of adolescent wellbeing," Jan 2019, on Orben and Przybylski, Nature Human Behaviour 3, 173 to 182 (three datasets, 355,358 young people)
  3. Karolinska Institutet, "Using social media may impair children's attention," Dec 8, 2025, on Nivins et al., Pediatrics Open Science (8,324 US children followed four years from age 9 to 10; social media, not TV, video or games, linked to rising inattention)
  4. Down To Earth, Torkel Klingberg and Samson Nivins (the study's authors), "Social media, not gaming, tied to rising attention problems in teens," Dec 2025 (direction ran from social media to inattention; effect not large for one child)
  5. Mona et al., "Short video addiction and its impact on cognitive functioning in adolescents and youth: a systematic review," International Journal of Adolescence and Youth, 2026 (23 studies, most from China; attention and self control most affected)
  6. Business Standard, on "Feeds, feelings, and focus," a meta-analysis in Psychological Bulletin, Nov 2025 (71 studies, 98,299 people; heavier short video use linked to poorer attention and impulse control in teens and adults)
  7. Aalto University research portal, "Effects of Adding Friction to Attention-Capture Patterns in TikTok," CHI 2026 (N=48; disabling autoplay and adding scroll friction cut compulsive engagement but users liked the app less)

Inspired by Happy. Prepared by DeepSeek (self reported, version unknown). Reviewed and edited by Claude.

Scan to open this post on your phone, or share it: kwm59.io/attention-spans-social-media/
Next relatedInstagram Links Stay in the App
Revision history (R1 current)
RevDateWhat changedPage
D1Nov 20, 2026, 3:45:21 AM CST (logged on arrival; chat message 3:45 am, minute only)What DeepSeek returned, unedited (self report, version UNSURE, search on). Batch J prompt handed out Nov 20, 3:31:56 AM CST. About 13 minutes round trip. Its citation numbers ran past its own source list.view
R0Nov 20, 2026, 10:30:30 AM CST (reviewed and posted)First publish, reviewed and edited by Claude (Opus). Remapped citations; replaced the Zenodo review with Orben and Przybylski 2019 and added the 2025 Psychological Bulletin meta-analysis; added the TikTok study's user experience trade off and the Karolinska authors' own account; cut the prefrontal claim, completion rates, "faster evaluation" and the mindfulness and detox lines.view
R1Nov 22, 2026, 1:14:22 AM CSTAdded a small QR code after the credit line that opens this post, so a photo of a screen can share it (Kenny asked, Nov 22). No text changed. Done by Claude (Opus) in one batch.view