Focus
Is our attention span really shrinking?
In plain English
The famous claim that human attention span has fallen to eight seconds, shorter than a goldfish, is not backed by any real study. Nobody has measured the same people’s attention reliably before and after to prove a decline happened. What genuinely gets studied instead is narrower: switching between several apps at once, and how much screen time children get, and both show only small, mixed effects rather than a dramatic collapse in the ability to concentrate.
The evidence
- Evidence
- ?Insufficient
- Human evidence
- No · No study measured attention span decline directly; related studies ask narrower questions.
- Studies reviewed
- 4
- Verdict
- Misleading
- How sure are we?
- Very uncertain
- Last reviewed
- Grade status
- Provisional — awaiting editorial validation
Short answer
Attention span is not shown to be “shrinking” in any study that measured the same people over time with a standardised task — the widely repeated “8 seconds, shorter than a goldfish” figure traces back to an uncredited marketing statistic, not a peer-reviewed study, and no comparable measurement exists from decades ago to compare against today. What is genuinely measured, and mixed, is media multitasking and its relationship to distractibility and screen time’s small correlation with children’s attention-related behaviour.
- Verdict
- Misleading
- Evidence
- ?Evidence currently insufficient
- Why this grade
- No study has measured human attention span longitudinally with a consistent method to show a population-level decline, so the core claim cannot be graded against direct evidence; the related, testable question of whether media multitasking impairs attention has produced a weak, inconsistent effect that becomes non-significant after correcting for small-study bias 12.
- Last reviewed
- 24 September 2026
The quick explanation
The “goldfish” statistic usually cited is: humans now have an attention span of 8 seconds, one second less than a goldfish, supposedly “since the rise of mobile technology”. This number circulated in a 2015 marketing report and has been repeated by media outlets ever since, but no underlying scientific study measuring attention span this way, in this population, over this timeframe, has ever been produced, and the “goldfish” comparison itself is not a documented finding about goldfish either. What researchers do measure is narrower: whether people who report multitasking across several media score differently on lab attention tests, and whether children’s screen time correlates with attention-related behaviour problems. Both are modest, contested findings, not evidence of a shrinking attention span.
What the research actually shows
Start with the number itself, because it is the part most people actually mean when they ask this question. The claim usually goes: the average human attention span is now 8 seconds, down from 12 seconds in 2000, and shorter than a goldfish’s 9 seconds. This travelled through news articles and marketing decks for years. It is not a scientific finding. There is no published, peer-reviewed study that measured a representative sample of people’s attention span in the year 2000 and again years later using the same method and found this decline. The figure traces to an uncredited statistic in a 2015 marketing report, and the “goldfish” comparison is not a documented measurement of goldfish attention either. Repeating a number this specific, sourced this thinly, is a textbook case of a statistic becoming true by repetition rather than by evidence.
Why this particular claim cannot be graded on the usual A-to-E scale
Grading evidence normally means weighing what the studies found. Here there are no studies measuring the thing being claimed — a real, method-consistent decline in human attention span over time — so the honest answer is that the evidence is insufficient to support or refute the number, not merely weak. That itself is useful information: a specific, widely quoted statistic with no traceable source should be treated as unverified, not as a rounded-down truth.
What is actually measured: media multitasking and lab distractibility
Underneath the soundbite is a real research question that has been studied properly: does using several media at once change how well people filter distraction? The influential 2009 study that started this line of research reported that people who scored higher on a media-use questionnaire performed worse on tests of ignoring irrelevant information. Independent labs then tried to replicate it.
What the replication effort found
Two pre-registered replication studies (14 tests in total, well-powered) found a statistically significant effect in the expected direction in only five of the fourteen tests, and only two of those held up under a more conservative statistical approach.1
Combining these replications with every other published study into a single meta-analysis (39 effect sizes) produced a weak but significant overall association between media multitasking and distractibility — which became non-significant once the analysis corrected for the tendency of small studies to report larger effects.1
A newer study of 214 young adults tested four cognitive domains at once — attention control, working memory, associative thinking, fluid reasoning — rather than one narrow lab task, and found no overall significant relationship between media-multitasking scores and combined cognitive performance, with only a small effect on one specific creativity-style measure.2
Put plainly: the most thoroughly studied, most direct scientific cousin of the “attention span is shrinking” claim is an effect that keeps shrinking itself every time it is tested more carefully. That is a meaningfully different story from the one in circulation.
Where a real, if small, signal does exist
Two other bodies of evidence are genuinely relevant, and both show something, just something much smaller and more qualified than “attention span has collapsed”.
In children, a large meta-analysis covering 87 studies and more than 159,000 participants under 12 found a small but statistically significant correlation between the amount of screen time a child gets and externalizing behaviour problems, a category that includes attention-deficit-type symptoms, alongside a smaller correlation with internalizing problems such as anxiety and low mood.3 Small and significant is the accurate description: the correlation coefficients were around 0.07–0.11, which means screen time is one contributor among many, not a dominant explanation, and the review notes substantial differences between studies in how screen time and behaviour were measured.
Separately, a 2026 systematic review pulled together 43 studies specifically about the psychological effects of fast-paced online interaction — constant notifications, rapid replies, always-on connectivity. It found consistent associations with what the authors call attentional fragmentation, along with higher cognitive load and technostress.4 This is a real and growing research area, but it is built almost entirely from cross-sectional, self-reported studies, so it shows an association between fast digital environments and how fragmented people feel their attention is, not a measured, before-and-after decline in attention span itself.
| Claim | Evidence status |
|---|---|
| “Attention span has fallen to 8 seconds” | No traceable scientific source; not a measured finding |
| Media multitasking impairs lab attention tests | Weak effect that disappears after correcting for small-study bias |
| Screen time correlates with children’s attention-related behaviour | Small but real correlation (r ≈ 0.1) in a very large meta-analysis |
| Fast digital interaction feels attention-fragmenting | Consistent self-reported association across 43 studies, not an objective attention-span measurement |
Why a claim like this spreads even without evidence
The 8-second figure has stayed in circulation for a decade because it is simple, quotable and fits an intuition many people already have — that phones make concentration harder. That intuition is not unreasonable on its own, and some of the evidence above (fragmentation, small screen-time correlations in children) is consistent with a milder version of it. The problem is specific to the number: a precise-sounding statistic (“8 seconds”, “one second less than a goldfish”) creates an impression of scientific precision that the underlying research cannot support, because the underlying research measuring exactly that thing does not exist. That gap — between a real, fuzzy, well-hedged research finding and a crisp, alarming headline number — is common across health and psychology claims generally, and worth watching for whenever a single number is repeated without a named study attached to it.
What this means in real life
It is reasonable to notice that your own concentration feels different around a phone than it does without one — that is a real, personal observation, and it is consistent with the fragmentation and cognitive-load findings above. What is not supported is treating that feeling as proof of a measured, population-wide biological decline, or repeating the 8-second goldfish figure as if it were a scientific fact. If concentration genuinely feels worse day to day, the more useful and better-evidenced question is the practical one covered on why you can’t concentrate: sleep, mood, and how many things are competing for attention at once tend to explain far more than a single “attention span” number ever could.
For the specific question of whether phone habits meet the bar for “addiction”, and whether compulsive checking is really the same thing as impaired attention, see is phone addiction real. The honest, connecting thread across all of these pages is the same: several individually modest, genuinely measured effects exist, and the popular narrative tends to fuse them into one dramatic and unsupported claim.
A practical test worth applying whenever a striking attention-span statistic appears is to ask three questions: who was measured, how was attention actually operationalised (a lab task, a questionnaire, a device log), and is there a named, checkable study behind the number, or only a chain of articles citing each other. Applied to the goldfish claim, none of the three checks out. Applied to genuine findings, like the screen-time meta-analysis in children or the media-multitasking replication effort, all three do — the population and method are stated, and the effect size, however small, is reported honestly rather than rounded up into a slogan.
What we know
- A meta-analysis combining the original 2009 media-multitasking study with all published replications (39 effect sizes across multiple labs) found only a weak association between self-reported media multitasking and laboratory distractibility, one that became non-significant once corrected for small-study publication bias 1.
- A 2026 study of 214 young adults testing attention control, working memory, associative thinking and reasoning together found no significant overall association between media multitasking and combined cognitive performance 2.
- A meta-analysis of 87 studies and over 159,000 children found screen time had a small but significant correlation with externalizing behaviour problems, which include attention-deficit-type symptoms (r = 0.11) 3.
- A 2026 systematic review of 43 studies on “digital speed” (rapid online interaction) found consistent associations with attentional fragmentation and cognitive load, but the review is built from cross-sectional, self-report studies rather than measured attention span over time 4.
- The original media-multitasking study this whole research area grew from used a single lab-based questionnaire and cognitive-test sample rather than a population survey tracked over time, and its findings have not consistently replicated 1.
What we don’t know
- Whether population-level sustained attention has changed at all over the past two decades, since no consistent measurement exists from before and after the relevant period to compare.
- Whether any observed differences in attention-related tasks are caused by device use, or whether people who are already more distractible simply multitask media more (the direction of causation is not settled) 1.
- Whether the same pattern applies to adults as convincingly as it does to the child screen-time literature, where the correlation is smallest and most studied.
How sure are we?
- Very confident
- Reasonably confident
- Uncertain
- Very uncertain
Very uncertain. The specific claim has no traceable study behind it at all, and even the closest related evidence 1 shows a weak effect that disappears after correcting for small-study bias.
Confidence describes how settled the answer on this page is, not how important the topic is. It can change as new research is published.
What would change the answer?
The evidence that is currently missing:
- Longitudinal studies tracking the same people’s attention with one consistent method over decades.
- Clear evidence on whether device use causes distractibility or distractible people use devices more.
- Larger adult-focused studies matching the scale of the child screen-time research.
Common misconception
The human attention span has fallen to 8 seconds, shorter than a goldfish’s, because of smartphones.
That specific number has no traceable scientific study behind it, and the actual research on media multitasking and attention finds a weak, inconsistent effect rather than a documented population-wide collapse 1.
Relevant studies
- Two replication studies plus a meta-analysis of 39 published effect sizes (following on from Ophir, Nass & Wagner 2009) found media multitasking weakly associated with lab distractibility, an effect that turned non-significant after correcting for small-study effects.1
- Study of 214 healthy young adults (18-30) tested attention control, working memory, associative thinking and fluid reasoning together; found no significant overall link between media-multitasking scores and combined cognitive performance, with only a small effect on one sub-measure.2
- Systematic review and meta-analysis of 87 studies and 159,425 children (age 12 or under) found a small but significant correlation between screen time duration and externalizing behaviour problems (r = 0.11) and internalizing problems (r = 0.07).3
- Systematic review of 43 peer-reviewed studies on the psychological effects of fast-paced online interaction found consistent associations with attentional fragmentation, cognitive load and technostress, based on cross-sectional and self-report data.4
The bottom line
The eight-second goldfish statistic has no traceable scientific source and should not be repeated as fact. Real research on media multitasking and children’s screen time shows only small, inconsistent effects, not a proven decline in attention span. If concentration feels worse day to day, sleep, mood and competing demands are more likely explanations.
Related questions
The Brief
Understand your brain a little better each week.
Join The Brief from dopamine.co.uk. Free. No spam, no supplements, no selling your address. Unsubscribe in one click.
References
- 1. Wiradhany W and Nieuwenstein MR Cognitive control in media multitaskers: Two replication studies and a meta-Analysis. Attention, perception & psychophysics. 2017. PMID 28840547. doi:10.3758/s13414-017-1408-4
- 2. Caprì T Media Multitasking and Cognitive Abilities: A Multi-Domain Assessment of Attention, Working Memory, Convergent Associative Thinking, and Fluid Reasoning. Journal of Intelligence. 2026. PMID 42646021. doi:10.3390/jintelligence14080175
- 3. Eirich R, et al. Association of Screen Time With Internalizing and Externalizing Behavior Problems in Children 12 Years or Younger: A Systematic Review and Meta-analysis. JAMA psychiatry. 2022. PMID 35293954. doi:10.1001/jamapsychiatry.2022.0155
- 4. Tripathy S, et al. Psychology at digital speed: A systematic review of how rapid online interaction alters cognitive and affective processes. Acta psychologica. 2026. PMID 42508353. doi:10.1016/j.actpsy.2026.107536