GelişimYaşam

Is Your Attention Span Really Shrinking? What the Data Shows About Focus in 2026

Reader absorbed in a printed book in a teal armchair by a sunny window, phone set aside on the side table

TL;DR. The famous “8-second attention span, shorter than a goldfish” comes from one slide in a 2015 Microsoft Canada marketing report, and that slide credits “Statistic brain”, a statistics website, rather than any study; the report’s own research (a 2,000-person online survey and a 112-person EEG study) never measured an 8-second span. Workplace logging studies did record shorter screen focus: a mean of 75.5 seconds per window in 2012 and 47.0 seconds in 2016 (median 40.2). Yet in the 2012 study, cutting email off raised it to 131.9 seconds for the same 13 people, a 74.7% jump within days. Twitter hashtags spent 32.0% less time in the global top 50 between 2013 and 2016, but the authors define that as collective attention, distinct from individual cognition. The one 43-year series on span capacity shows small, mixed trends, not a collapse. The evidence points to a changed environment more than a broken brain. CEO move: treat fragmented focus as a systems design problem before treating it as a personal defect. Student move: run the one-week self-test below to separate capacity from environment.

This piece is the fact-check layer of our attention cluster. For how to train focus once you know the problem, see attention as a trainable skill; for the broader evidence on screens and cognition, see screen time and cognitive performance beyond the moral panic; and for why the “dopamine hijack” story is weaker than it sounds, see the dopamine myth.

This article is general information about published research. It does not diagnose or give guidance on attention disorders such as ADHD; anyone worried about their own attention should speak to a qualified clinician.

Why does the shrinking-attention-span claim need a verdict?

Because most people already believe it. In a survey of 2,093 UK adults run by Savanta ComRes for King’s College London’s Policy Institute in September 2021, 49% agreed that their attention span is shorter than it used to be, against 23% (a ratio of about 2.1 to 1), and 50% thought it true that the average adult attention span is only 8 seconds, against 25%. The report calls that belief a myth and adds: “A lack of long-term studies means we can’t tell whether attention spans have actually declined.”

The way to close the gap between a strong feeling and thin evidence is to take each popular number apart: where it first appeared, what unit was measured, on whom and how, and whether it describes a change in capacity (what a person can do when conditions allow) or in environment and behaviour (what people do in the conditions they are in). The last question matters most. A leader who misreads an environment problem as a capacity problem buys brain-training apps; one who reads it correctly redesigns the inbox.

What is an “attention span”, and why do the numbers disagree?

Psychology does not use a single “attention span” measured in seconds. Even the 2015 Microsoft report split attention into three types (sustained, selective and alternating) and stated that attention cannot “be simply characterized as how long people can concentrate”. The King’s College report makes the same point: the ability to sustain attention depends on the task, whether listening to a lecture, reading a book or driving.

Direct measures of sustained attention capacity use controlled tasks. The largest lifespan study, by Fortenbaugh and colleagues (2015), tested 10,430 people aged 10 to 70 on a 4-minute continuous performance test and found that sustained attention ability peaks in the early forties, at about 43, and then declines gradually. That is a capacity measure: same task, same rules, compared across ages. How long a person stays on one browser tab at work is a behaviour measure, shaped by notifications, open windows and job demands. Most viral attention statistics quietly swap one for the other.

Where did the 8-second goldfish figure come from?

The figure was printed, not measured. The Microsoft Canada Consumer Insights report “Attention spans” (Spring 2015) contains a slide headed “We know human attention is dwindling” with three numbers: 12 seconds for the average human attention span in 2000, 8 seconds in 2013, and 9 seconds for a goldfish. The only attribution is “Source: Statistic brain”. No study, sample or method is given, and no peer-reviewed paper reporting a 12-second or 8-second human attention span was found for this article.

The report’s own research was a “gameified” online survey of 2,000 Canadians fielded in the fourth quarter of 2014 and an EEG study of 112 Canadians fielded in December 2014. Both were single snapshots; neither can show a change since 2000, and neither reports an 8-second span. The report does make its own claim, in its executive summary, that digital lifestyles “deplete the ability to remain focused on a single task”, but that rests on comparing heavier and lighter users at one point in time, not on a measured decline. The same summary adds that connected consumers are becoming better at “doing more with less via shorter bursts of high attention”, and the foreword opens with “Think digital is killing attention spans? Think again”. The headline that travelled the world was the one number the authors did not produce. The report is commercial research written for marketers, and the goldfish comparison has no stated basis at all.

Verdict: untraceable. The 8-second figure should not be cited as data on human attention.

Did screen focus really fall from minutes to 47 seconds?

This is the strongest real evidence behind the story, from in-situ workplace studies led by Gloria Mark at the University of California, Irvine. The often-repeated sequence (about two and a half minutes in 2004, 75 seconds in 2012, 47 seconds recently) is popularised in Mark’s 2023 book. The papers themselves give a more precise picture.

Study Who Method What was timed Figure as printed
González and Mark, CHI 2004 14 information workers at one firm Human observer with a stopwatch, about 26 hours per person Continuous time on an “event” using a PC before switching 2 min 53 sec per PC event; 2 min 11 sec with any device or paper
González and Mark, CHI 2004 Same 14 workers Same Time in a “working sphere” (a project-level unit of work) About 12 minutes
Mark, González and Harris, CHI 2005 Larger sample, over 700 observation hours Observation Time in a working sphere 11 min 4 sec
Mark, Voida and Cardello, CHI 2012 13 workers, one research organisation Window-logging software, normal email Mean time a window stayed active 75.5 sec; 37.1 switches per hour
Same study Same 13 workers Same, email cut off for five workdays Mean time a window stayed active 131.9 sec; 18.2 switches per hour
Mark, Iqbal, Czerwinski, Johns and Sano, CHI 2016 40 information workers, one high-tech workplace, two work weeks Window-logging software Mean focus on any computer screen Mean 47.0 sec, median 40.2 sec; 272.7 switches between apps per day

Source: the five papers named in the first column, as published on the first author’s university pages. Figures as printed.

First, the 47-second figure is real, and it is a mean. The 2016 abstract leads with the median: “The median duration of online screen focus was 40 seconds”; the 47.0-second mean sits in Table 1. The “two and a half minutes” for 2004 does not appear in the 2004 paper, so it is treated here as a book-only number and not used.

Second, the methods changed. The 2004 figure came from a human observer coding “events”; the 2012 and 2016 figures came from software that records every change of foreground window. Software catches switches an observer would merge into one event, so part of the fall is built into the change of instrument. The samples were also small and drawn from different single workplaces.

Third, task focus and screen focus differ. In 2004 the workers who switched PC events every few minutes stayed in a working sphere for about 12 minutes, roughly 4.2 times longer. A window switch is not necessarily a switch of task. For what each switch costs, see our single-tasking vs multitasking breakdown.

How much of the fall is capacity, and how much is environment?

This CEOtudent analysis of González and Mark (2004), Mark, Voida and Cardello (2012), Mark and colleagues (2016) and Lorenz-Spreen and colleagues (2019) recomputes the changes from the printed figures; every cell is reproduced in our calculation script.

Comparison From To Change Same people and method? What it most likely reflects
PC event (2004) to window focus (2012, with email) 173 sec 75.5 sec -56.4% No: different firms, observer vs software Mixed: behaviour plus change of instrument
Window focus 2012 (with email) to 2016 75.5 sec 47.0 sec -37.7% No: different firms, same type of software Behaviour in different workplaces
PC event (2004) to window focus (2016) 173 sec 47.0 sec -72.8% No The headline comparison; least controlled
With email to without email, 2012 75.5 sec 131.9 sec +74.7% Yes: same 13 people, same software Environment
Window switches per hour, with vs without email, 2012 37.1 18.2 -50.9% Yes Environment
Hashtag time in global Twitter top 50, 2013 to 2016 17.5 hours 11.9 hours -32.0% Same platform and rule Collective turnover of topics
Distinct hashtags reaching the hourly top 50 per year 25,031 36,703 +46.6% Same platform and rule More content competing for the same slots

Read the bold rows first. The only comparison holding people and instrument constant is the 2012 email experiment, and there an environment change moved focus duration further in five working days (+74.7%) than the whole 2012-to-2016 decline (-37.7%). The cross-study decline describes how logged screen work looked in those offices, but it cannot separate weaker capacity from a busier environment. Caution: the email study had only 13 participants, durations varied widely (standard deviation 568.1 seconds without email), and a five-day cut-off shows what is possible, not what every role can sustain.

Is collective attention speeding up?

Yes, by one well-defined measure of the information environment. Lorenz-Spreen and colleagues (2019, Nature Communications) analysed the global top 50 Twitter hashtags, sampled hourly from about 43 billion tweets, plus Google Books (100 years), movie ticket sales (40 years), scientific citations (25 years), Google Trends, Reddit and Wikipedia. A hashtag stayed in the Twitter top 50 for 17.5 hours on average in 2013 and 11.9 hours in 2016, as the number of distinct hashtags passing through rose from 25,031 to 36,703 a year. The other datasets showed similar steepening.

The authors state that collective attention “is distinct from the definition of attention used in cognitive psychology”, and their model attributes the acceleration to “increasing production and consumption of content, resulting in a more rapid exhaustion of limited attention resources”. Topics burn out faster because more of them compete for the same public, not because each person holds a thought for less time.

Verdict: true for topics, not evidence about individuals. Collective attention cycles have shortened; that is an environment finding.

Can students really focus for only 10 minutes?

Two literature reviews found the popular “10 to 15 minute” lecture limit poorly supported. Wilson and Korn (2007, Teaching of Psychology) reviewed note-taking studies, classroom observations, self-reports and physiological measures and concluded that the research “provides little support for the belief that students’ attention declines after 10 to 15 min”, adding that most studies failed to account for individual differences. Bradbury (2016, Advances in Physiology Education) reached the same conclusion: some institutions had cut lectures to 15 minutes on “common knowledge”, yet “the available primary data do not support the concept of a 10- to 15-min attention limit”, and the most often cited source “barely discusses student attention at all”. The most consistent finding was that attention varies most between teachers, not formats.

Verdict: not supported. Attention in a lecture varies with the teacher and the material more than with a fixed clock.

What does classroom data show about devices?

The OECD’s PISA 2022 assessment, covering 15-year-olds in 81 countries and economies, asked how often distraction happens in mathematics lessons. On average across OECD countries, 30% of students (described in the report as one in three) said they get distracted by using digital devices in most or every lesson, and around one in four said they get distracted by other students using devices. Students who reported being distracted by others’ devices in at least some lessons scored 15 points lower in mathematics than those for whom it never or almost never happens, after accounting for socio-economic profile. Where phones are banned on school premises, students are less likely to report device distraction.

All of this is correlational, self-reported and cross-sectional, so it cannot show that capacity has fallen. It does show that the classroom environment, including other people’s screens, is linked to learning outcomes, and that school rules about it are linked to reported distraction.

Verdict: environment effect, well documented; not a capacity measure.

Are heavy media multitaskers’ minds actually different?

This is the closest the field gets to a capacity question. Uncapher and Wagner’s 2018 review in PNAS found that “the balance of evidence suggests that heavier media multitaskers exhibit poorer performance in a number of cognitive domains”, while noting that “many studies find no performance differences between groups”. For working memory, about half the adult tests showed heavier multitaskers underperforming, the rest null; none showed them doing better. Differences tend to involve sustained goal-directed attention, which the authors frame as a possible tendency towards attentional lapses. The direction of causality is listed as open: heavy multitasking might weaken attention, or lapse-prone people might multitask more.

Verdict: unproven as a cause. A link exists in some tasks; whether screens cause it has not been shown. Our notification audit protocol is a practical response that works under either explanation.

Has measured span capacity changed across generations?

The one long-run dataset on span-type capacity points away from collapse. Wongupparaj and colleagues (2017, Intelligence) ran a cross-temporal meta-analysis of 1,754 independent samples (n = 139,677) covering 43 years, 1973 to 2016. Forward digit span, the classic test of how many items a person can hold in mind, rose slightly with year of publication (r = 0.12), as did forward Corsi block span (r = 0.10). Backward spans, which add a manipulation demand, fell slightly (r = -0.06 for digits, r = -0.17 for Corsi blocks).

Squaring those correlations shows how small the trends are: publication year accounts for about 1.4% of the variance in forward digit span and 0.4% in backward digit span. Digit span measures short-term and working memory, not sustained attention, and the series ends in 2016. No comparable multi-decade series for sustained attention was found, matching the King’s College conclusion that long-term studies are missing.

Verdict: no evidence of a capacity collapse; small, mixed shifts.

The claim-by-claim verdict

Unlike the origin table in our attention-skill piece, this one names the sample and method and classifies each claim on the capacity-versus-environment axis. It is a CEOtudent editorial verdict based on the primary sources above.

Claim Origin What was measured, on whom, how Capacity or environment? Verdict
“Human attention span fell from 12 to 8 seconds, below a goldfish’s 9” Microsoft Canada consumer insights report, 2015, slide credited to “Statistic brain” Nothing: the report’s own research was two single snapshots (2,000-person survey, 112-person EEG study) Neither Untraceable; do not cite
“Attention spans are shorter than ever” Public perception, e.g. 49% of 2,093 UK adults (2021 survey) Self-report of a feeling Perception Widely felt; no long-term data confirms it
“People switch screens every 47 seconds” Mark et al., CHI 2016 Mean window duration, 40 workers, one workplace, software logs Behaviour in an environment Accurate as a mean (median 40.2 sec); not task focus or capacity
“Screen focus fell from minutes to seconds since 2004” Mark’s studies 2004-2016 Different small samples, observer then software Environment plus change of instrument Partly supported; the 2.5-minute start is not in the 2004 paper
“Students can only focus for 10 to 15 minutes” Lecture folklore, reviewed by Wilson and Korn (2007), Bradbury (2016) Two literature reviews Neither established Not supported by primary data
“Collective attention is speeding up” Lorenz-Spreen et al. 2019 Turnover of top topics, Twitter and six other datasets Environment (collective) Supported for topics; not about individuals
“Screens distract students in class” OECD PISA 2022 Self-reported distraction, 15-year-olds, 81 systems Environment Supported as an association
“Heavy multitasking damages attention” Uncapher and Wagner review (2018) Lab tasks, heavier vs lighter media multitaskers Possible capacity link Mixed; causality unknown

Self-test: is it your capacity or your environment?

The test below is a CEOtudent editorial framework: a self-observation exercise, not a psychological assessment, with scoring bands that are design choices rather than research findings.

Part 1: answer yes or no.

  1. With notifications fully off for two hours, does your focus noticeably improve?
  2. Do you focus well in some settings (a film, a game, a hobby) but not at your desk?
  3. Did the problem grow when your tools, role or team changed, rather than gradually?
  4. Do most switches start from an outside cue (a ping, a badge, a person)?
  5. Is your work split across more projects or channels than a year ago?
  6. Reading on paper somewhere quiet, can you finish a long chapter?
  7. Do you often open a tab or app without deciding to?
  8. Is the difficulty present in every setting, including quiet ones with no device, and has it lasted most of your life?

Scoring. Count “yes” answers to questions 1-7; score question 8 separately.

Pattern Reading First move
5 to 7 “yes” on 1-7, “no” on 8 Mainly an environment problem Redesign inputs: batch email, remove badges, close idle windows; use the micro-environment design checklist
2 to 4 “yes” on 1-7, “no” on 8 Environment plus habit Fix the environment, then add practice from the attention-skill programme
0 or 1 “yes” on 1-7 Environment is not the main driver Check sleep, workload and task design; measure uninterrupted time with the Deep Work Deficit Index
“Yes” on 8 A long-standing pattern across all settings This is beyond a self-test; speak to a qualified professional

Part 2: the one-week contrast, a small copy of the 2012 email study. For three days, work as usual and estimate each hour how many times you changed window or app. For the next three days, open email and chat only at two fixed times and log the same count. If switches drop by anything like the 50.9% seen in 2012, the cause sits largely in your environment, and the fix is a system, not willpower.

The CEO move: own the environment. Decide which channels may interrupt you and when, and write the rule down for your team like a meeting policy. If switching between people and AI tools drives the fragmentation, read our analysis of attention residue in the AI era. The student move: treat your attention as a research question; compare the two halves of the week before deciding whether you need training, redesign or both.

What this does not mean

  • It does not mean attention is fine. The screen-focus decline is real behaviour with documented switch costs; the question is where the cause sits.
  • It does not prove capacity is unchanged. The best series ends in 2016 and covers memory span; no evidence of collapse is not proof of no change.
  • It does not mean phones are harmless in class. A 15-point PISA gap matters, even if correlational.
  • It does not make the workplace studies representative. They used 13 to 40 people in single organisations: careful measurements, not national statistics.

Frequently asked questions

Is the 8-second attention span a real scientific finding?
No. It appears on one slide of a 2015 Microsoft Canada marketing report, credited to “Statistic brain”, with no study behind it.

Do people really switch screens every 47 seconds?
In one 2016 study of 40 information workers, the mean window focus was 47.0 seconds (median 40.2). That measures window switching at work, not how long a person can concentrate.

Has the average attention span declined since 2000?
No long-run study has measured it. Many people feel it has, but the longest capacity series, on memory span from 1973 to 2016, shows small mixed trends.

Why did focus time almost double when email was switched off?
Removing one input changed behaviour within days: 131.9 versus 75.5 seconds per window for the same 13 people, with half as many switches. That points to the environment, not ability.

Should lectures be cut to 10 minutes?
Two reviews found no solid primary evidence for a 10 to 15 minute limit; teachers mattered more than lecture length.

Does this mean screens have no effect on attention?
No. Heavier media multitaskers do worse on some tasks, and classroom distraction is linked to lower scores. What is not established is that screens have permanently reduced capacity.

Sources

  1. Microsoft Canada Consumer Insights (2015). Attention spans. Spring 2015 research report (commercial research for advertisers; archived copy of the original PDF).
  2. King’s College London Policy Institute and Centre for Attention Studies (2022). Do we have your attention? How people focus and live in the modern information environment. Survey by Savanta ComRes of 2,093 UK adults, 24-26 September 2021.
  3. Fortenbaugh, F. C., DeGutis, J., Germine, L., et al. (2015). Sustained attention across the life span in a sample of 10,000: Dissociating ability and strategy. Psychological Science, 26(9), 1497-1510.
  4. González, V. M., and Mark, G. (2004). “Constant, constant, multi-tasking craziness”: Managing multiple working spheres. Proceedings of ACM CHI 2004.
  5. Mark, G., González, V. M., and Harris, J. (2005). No task left behind? Examining the nature of fragmented work. Proceedings of ACM CHI 2005.
  6. Mark, G. J., Voida, S., and Cardello, A. V. (2012). “A pace not dictated by electrons”: An empirical study of work without email. Proceedings of ACM CHI 2012.
  7. Mark, G., Iqbal, S. T., Czerwinski, M., Johns, P., and Sano, A. (2016). Neurotics can’t focus: An in situ study of online multitasking in the workplace. Proceedings of ACM CHI 2016.
  8. Lorenz-Spreen, P., Mønsted, B. M., Hövel, P., and Lehmann, S. (2019). Accelerating dynamics of collective attention. Nature Communications, 10, 1759.
  9. Wilson, K., and Korn, J. H. (2007). Attention during lectures: Beyond ten minutes. Teaching of Psychology, 34(2), 85-89 (abstract only).
  10. Bradbury, N. A. (2016). Attention span during lectures: 8 seconds, 10 minutes, or more? Advances in Physiology Education, 40(4), 509-513 (abstract only).
  11. OECD (2023). PISA 2022 Results (Volume II): Learning During and From Disruption. OECD Publishing, Paris.
  12. Uncapher, M. R., and Wagner, A. D. (2018). Minds and brains of media multitaskers: Current findings and future directions. Proceedings of the National Academy of Sciences, 115(40).
  13. Wongupparaj, P., Wongupparaj, R., Kumari, V., and Morris, R. G. (2017). The Flynn effect for verbal and visuospatial short-term and working memory: A cross-temporal meta-analysis. Intelligence, 64, 71-80 (abstract only).
  14. Wongupparaj, P., Wongupparaj, R., Kumari, V., and Morris, R. G. (2020). Verbal and visuospatial short-term and working memory data across a 43-year period (1973-2016) worldwide: Flynn and anti-Flynn effects. Data in Brief, 29, 105231.

Notes on method. The capacity-versus-environment table, the 4.2-times working-sphere ratio and the variance-explained (r squared) figures are CEOtudent analyses of the cited sources; the claim-by-claim verdict table and the self-test are a CEOtudent editorial framework. Deliberately not used: the “two and a half minutes in 2004” figure, which does not appear in the 2004 paper and is known here only from secondary accounts of a 2023 book; the 12-, 8- and 9-second goldfish figures as data, because the report credits them to a statistics website and no study behind them was found; and anything from Bradbury or Wilson and Korn beyond their abstracts.


This content was compiled with the support of AI following in-depth research, then written and prepared for publication by the CEOtudent editorial team.

This post is also available in: Türkçe Français Español Deutsch

Benzer içerikler