Mark ở vũ trụ khácVIMark học AI
Mark ở vũ trụ khác · Another Universe Common myths · Attention · Source tracing

Common myth #7: An attention span shorter than a goldfish

8 sec? 9 sec?

"People can now focus for only 8 seconds; a goldfish manages 9" · spread from May 2015

The claimhuman attention: 8 seconds, goldfish: 9 seconds
Spread bya Microsoft Canada marketing report, then Time magazine
In factthe number cannot be traced to any study
Main sourcesBBC More or Less (2017), Bradbury (2016)
One-minute recap

Essentials

A number nobody measured"8 seconds" does not come from any experiment anyone can find.
Borrowed three timesStatistic Brain → Microsoft Canada → Time. Each hand looked more credible than the last.
The goldfish is innocentGoldfish learn feeding times and can estimate how far they have swum.
A grain of truthWe switch screens faster than before. That is a habit, not a limit of the brain.
Three questionsMeasured what? Measured whom? Measured by whom? Ask before a number goes on a slide.

Sources: BBC News, More or Less, "Busting the attention span myth" (S. Maybin, 10 March 2017); Microsoft Canada, "Attention spans" (2015); Gee, Stephenson and Wright (1994); Sibeaux et al. (2022); G. Mark, Attention Span (2023).

Part 1

The question behind the myth

A number printed in a Microsoft report, put in a Time headline, then repeated on thousands of slides. Why did nobody ask how it was measured? Because it travelled under credible names, and because it matched what we already felt.

~2013A statistics siteStatistic Brain lists 12 seconds (2000), 8 seconds (2013), goldfish 9 seconds.
→
2015A report, then a headlineMicrosoft Canada reprints it with "Source: Statistic Brain". Time writes "shorter than a goldfish".
→
2017Traced backThe BBC asks the sources Statistic Brain named. Nobody can find a study.

Sources: BBC More or Less (10 March 2017); Microsoft Canada, "Attention spans" (2015); Time, K. McSpadden (14 May 2015).

Part 2

Where the myth came from

Following each hand that passed the number on. Arrows show who cited whom.

The citation chain. Thousands of slides and articles cite Time magazine, 2015. Time cites the Microsoft Canada report. Microsoft Canada credits its chart to the Statistic Brain website. Statistic Brain vaguely credits the US National Library of Medicine and the Associated Press. In 2017 the BBC asked both; neither found any study. The chain ends in an empty box. thousands ofslides, articles2015 to now Time magazine14 May 2015"than a goldfish" Microsoft Canada2015 report"Source: StatisticBrain" Statistic Brainstats website"Source: NCBI, AP" original study?not found 2017: BBC More or Less traces the whole chain NCBI and AP both "can't find any record of research" behind the numbers
The further down, the bigger the name and the fainter the source. Time readers saw "Microsoft research"; Microsoft readers saw a small source line; anyone opening Statistic Brain found two big names and no paper.Sources: BBC More or Less, S. Maybin (10 March 2017); Microsoft Canada, "Attention spans" (2015), attention span chart credited to Statistic Brain; Time, K. McSpadden (14 May 2015).
~2012–2013 · Statistic BrainIts "Attention span statistics" page gives 12 seconds (2000), 8 seconds (2013), goldfish 9 seconds. No paper, author or method. (Description of the archived page, secondary source.)
Dec 2014 · Microsoft Canada runs a real surveyAbout 2,000 Canadians played online attention tests; another 112 had their brain activity recorded (EEG).
Spring 2015 · The "Attention spans" reportThe 12 → 8 second chart and the 9-second goldfish note are printed with "Source: Statistic Brain". They are not results of Microsoft's own survey.
14 May 2015 · Time magazineHeadline: "You now have a shorter attention span than a goldfish". Many outlets copied it within days.
2015 to now · Into slides, books, even law reviewsA piece on Temple University's law school site counted more than a dozen law review articles using the number as fact. (Secondary source.)
10 March 2017 · The BBC traces it backMore or Less contacts NCBI and AP. Neither finds any study. Psychologist Gemma Briggs: "average attention span" is not something psychologists measure.

Sources: BBC More or Less (2017); Phys.org (15 May 2015) on the Microsoft report's method; Time (14 May 2015); law.temple.edu (law reviews, secondary); Medium, "The attention span stat everyone keeps getting wrong" (archived Statistic Brain page, secondary).

Why is this number so easy to believe?

It matches a feeling. Everyone has scrolled a phone and forgotten what they just read. A number that "confirms" that feeling is rarely questioned.

It carries big names. "Microsoft" and "Time" stand in front of it. Few people read the small source line under the chart.

It makes a funny comparison. Losing to a goldfish is memorable and shareable. The older "goldfish have a 3-second memory" saying, of unknown origin, makes the comparison feel plausible.

It is precise to the second. "8 seconds" sounds measured, though nobody measured it.

Part 1

Where it came from

8 secondsStatistic BrainMicrosoftTime
Part 2

What the evidence says

goldfishtask-dependent attention47 secondslectures
Part 3

Work lesson

trace to the sourcewhat, whom, by whomclean slides
Part 3

What the evidence says

There is no single number for attention

attention is task-dependentBBC 2017; Fortenbaugh 2015
Four attention numbers compared on three questions: measured what, measured whom, measured by whom. The 8-second figure: unknown on all three. Microsoft survey 2015: game-like online tests, about 2,000 Canadians, Microsoft's marketing team. Gloria Mark: time on one screen before switching, office workers, UC Irvine scientists. Fortenbaugh 2015: steadiness on a continuous attention test, 10,430 people aged 10 to 70, a university team. measured what?whom?by whom? "8 seconds"unknownunknownunknown Microsoft 2015game-like onlinetests + EEG~2,000 Canadians+ 112 with EEGmarketing teamnot peer reviewed Gloria Markhow long beforeswitching screensoffice workersdoing real workUC Irvinemany studies Fortenbaugh 2015steadiness on acontinuous attention test10,430 peopleaged 10 to 70university teamPsychological Science
Four numbers, four different things. Psychologist Gemma Briggs told the BBC that attention is "very much task-dependent", and that "average attention span" is not something psychologists measure. Fortenbaugh's large study measured how steadily people perform an attention test, and found this ability peaks around age 43.Sources: BBC More or Less (2017); Phys.org (May 2015) on Microsoft's method; G. Mark, Attention Span (2023); Fortenbaugh et al., Psychological Science 26(9), 2015 (10,430 people; peak ~43 per the APS press release).

Goldfish remember more than we think

goldfish cognitionGee 1994; Sibeaux 2022
Two experiments. Left: in 1994, goldfish learned to press a lever for food when food was available in only a one-hour window each day; they concentrated their presses around feeding time. Right: in 2022, nine goldfish were trained to swim 70 cm in a long striped tank; without the cue, eight turned on their own after 74 cm on average. food for 1 hour a day fish learn to press around that time Gee, Stephenson and Wright (1994) trained: 70 cm on their own: 74 cm on average (8 of 9) Sibeaux et al. (2022)
Goldfish are a standard animal in memory research. Briggs called their "goldfish brain" reputation "an interesting irony". No study has measured a "9-second attention span" or a "3-second memory" in goldfish.Sources: Gee, Stephenson and Wright, J. Exp. Anal. Behav. 62(1), 1994; Sibeaux, Karlsson, Newport and Burt de Perera, Proc. R. Soc. B, 2022 (70 cm, 74 cm and 9 fish via Oxford's summary and science press, secondary); BBC More or Less (2017).
More: how smart are fish in general?

Culum Brown's review (Macquarie University, Animal Cognition, 2015) concludes that fish cognition often matches that of other vertebrates. In an interview he said fish can learn to escape a trap and still remember a year later. That is fish in general, not goldfish specifically.

In 2008, Adelaide schoolboy Rory Stokes trained goldfish to swim to a beacon for food and found they still remembered days later. It was a school project reported only in the news, so treat it as a fun anecdote.

Sources: C. Brown, "Fish intelligence, sentience and ethics", Animal Cognition 18(1), 2015, and an interview on Macquarie University's site (secondary); news reports on Rory Stokes, Feb 2008 (secondary).

The grain of truth: we switch screens faster

screen switchingGloria Mark, 2023
Average time on one screen before switching to another, from Gloria Mark's studies: about 150 seconds in 2004, about 75 seconds some years later, 47 seconds recently. 2004~150 s years later~75 s recently47 s measured: how long office workers stay on one window or screen before switching
This is a work habit, not a brain limit. This number has a researcher and a method, and it really has gone down. But it measures window switching, not how long we can focus when we need to. Do not put "47 seconds" on the slide where "8 seconds" used to be.Source: G. Mark, Attention Span (2023) and CNN interview, Jan 2023, reposted by UC Irvine. The year of the 75-second figure could not be verified, so it is not given.

Collective attention to a topic is getting shorter

collective attentionLorenz-Spreen 2019
Illustration: a curve of public interest in one topic. In the past the curve was wide and gentle; now it is narrow and steep, rising and falling fast. before: longer now: up fast, down fast time → (illustration, not data)
Crowds move on faster. Lorenz-Spreen and colleagues analysed decades of data (Twitter, books, films, searches) and found each topic holds public attention for a shorter time, because more content competes for it. This is the attention of a crowd, not of one brain.Source: Lorenz-Spreen, Mønsted, Hövel and Lehmann, "Accelerating dynamics of collective attention", Nature Communications 10, 2019.

Even "10 minutes in a lecture" is weak

lecture attentionBradbury 2016

The claim

Students only focus for 10–15 minutes, so some institutions cut lectures to 15 minutes.

Rereading the sources

Many papers cite each other, very few primary studies. The most cited source barely discusses attention.

What varies most

The teacher. Differences between teachers are bigger than differences between formats.

Source: N. A. Bradbury, "Attention span during lectures: 8 seconds, 10 minutes, or more?", Advances in Physiology Education 40(4), 2016, pp. 509–513.

Part 4

Fact check: Solid, Caution, False

11 common claims about "goldfish attention", checked against the original sources. Click each line for the reasoning.

SOLIDMicrosoft Canada did run a real study in 2015

A survey of about 2,000 Canadians plus EEG on 112 people (Phys.org May 2015; BBC 2017). But it is a marketing report, not peer reviewed, and the 8-second figure did not come from it.

SOLIDOffice workers switch screens faster than before

In Gloria Mark's studies: about 2.5 minutes (2004) down to 47 seconds recently. That measures behaviour, not capacity.

SOLIDThe public stays on any one topic for less time

Lorenz-Spreen et al. (Nature Communications, 2019), on decades of data. This is collective attention and says nothing about individuals.

CAUTIONSocial media ruins our focus

The Microsoft report itself found heavy social media users did worse at prolonged attention but had more short "bursts of high attention". That is a correlation in one sample, not proof that social media is the cause. (Report described via Phys.org and CSO Online, secondary.)

CAUTIONStudents only focus for 10–15 minutes in a lecture

Bradbury (2016) reread the literature and found the primary data do not support this limit. Listeners do drift, but there is no "10-minute clock".

CAUTIONYoung people focus worse than older people

Fortenbaugh et al. (2015, 10,430 people) found sustained attention ability rises to about age 43, then declines. But that compares age groups at one moment; it does not show "this generation is worse than the last".

FALSEThe human attention span is 8 seconds

No study can be traced. The BBC (2017) asked the two sources Statistic Brain named (NCBI and AP); neither found anything. Psychologists do not measure "average attention span" as one number.

FALSEAttention fell from 12 seconds (2000) to 8 seconds (2013)

Same empty source. No published 2000 measurement exists to compare with.

FALSE"Microsoft proved we are worse than goldfish"

The number was printed in Microsoft's report, but credited to Statistic Brain. Microsoft's own survey results are a different matter.

FALSEGoldfish have a 9-second attention span

No study has measured this. Statistic Brain gave no source for it.

FALSEGoldfish only remember for 3 seconds

Goldfish learn feeding times (Gee 1994) and estimate distance swum (Sibeaux 2022). Nobody knows who first said "3 seconds".

Part 5

Verdict on the myth

Scores are Mark's assessment based on the sources in Part 8, not data from any study.

How wrong
5/5
How persistent
5/5
How harmful
3/5
Grain of truth
2/5
What makes sense
  • We really do switch screens faster
  • More content competes for attention
  • Writing tightly for busy people is good
What is wrong and harmful
  • An unmeasured number used as fact
  • Audiences treated as "goldfish brains", content chopped up
  • Goldfish wrongly maligned

A number nobody measured passed through three credible names and became "fact" on thousands of slides. It hurts nobody's health, but it teaches us to look down on our audience.

More: why it survives being debunked

The BBC debunked the number in 2017, and many researchers and educators have written about it since. But "shorter than a goldfish" is shorter and funnier than any correction. A correction must explain "measured what, measured whom"; the myth needs one sentence.

Beware the opposite overstatement: "attention has not changed at all". Screen-switching habits and the rhythm of crowd attention really have changed. The accurate sentence is: we have no number for "the human attention span", and the way we spend attention is changing.

Part 6

Work lesson: check a viral number before it goes on a slide

For sales decks, reports to your manager, company posts and training material.

Four steps to check a number: trace each link back to the origin, ask what was measured, ask who was measured, ask who measured it and who paid. If you cannot reach the origin in 10 minutes, drop it or mark it unverified. 1 · trace to the originclick each source linkuntil you hit a study 2 · measured what?what "attention" meanswhich test 3 · measured whom?how many, wherelike your audience? 4 · by whom?who did it, who paidpeer reviewed? No origin after 10 minutes: drop the number, or mark it "unverified".
For "8 seconds", step 1 is where it stops. Through Time, Microsoft and Statistic Brain, the trail ends at two big names and no paper.
Read the small source line"Source: Statistic Brain" sits right under Microsoft's chart. One second of reading is enough to doubt.
A big name is not a source"According to Microsoft" or "Harvard" only says where it was printed. Ask: did they measure it or copy it?
Beware numbers that are too neat, too funPrecise to the second, compared with an animal, matching a common feeling: three signs to check harder.
Replace it with a true sentenceSwap "people only focus for 8 seconds" for "clients are busy; put the main point first". You do not need a fake number to write tightly.
WhenWhat to doTime
When you get a great numberSave the link to the origin right away, not a screenshot1 minute
Before sending the deckPick the 3 most prominent numbers and run the 4 steps15 minutes
When you cannot find the originDrop it, or write "unverified" right next to it1 minute
When a colleague forwards oneAsk gently: "Do you have the link to the original study?"1 sentence

The schedule above is Mark's suggestion, not a research finding.

Average attention span is not a metric that psychologists would try to measure and quantify.Gemma Briggs, psychologist, Open University, speaking to BBC More or Less (2017).
Part 7

Further reading

BBC News, "Busting the attention span myth" (Simon Maybin, 2017)The piece that traced the 8-second number back. Short and readable.
Neil Bradbury, "Attention span during lectures: 8 seconds, 10 minutes, or more?" (2016)A lecturer rereads every source on the "10-minute focus". Free to read.
Gloria Mark, Attention Span (2023)By the researcher who actually measures how people switch screens at work.
Common myths #08: Group brainstorming produces more ideasDoes a group "thinking big" really beat people thinking alone and then pooling ideas?
Part 8

Sources

Primary sources first. Many original pages could not be opened directly from where this was written; anything checked only through secondary sources is marked "secondary".

  1. S. Maybin, "Busting the attention span myth", BBC News / More or Less, 10 March 2017. bbc.com
  2. Microsoft Canada, Consumer Insights, "Attention spans", 2015 (survey of ~2,000 people, EEG on 112, chart credited to Statistic Brain). Method described by Phys.org, 15 May 2015. phys.org
  3. K. McSpadden, "You Now Have a Shorter Attention Span Than a Goldfish", Time, 14 May 2015. time.com
  4. N. A. Bradbury, "Attention span during lectures: 8 seconds, 10 minutes, or more?", Advances in Physiology Education 40(4), 509–513, 2016. pubmed
  5. P. Gee, D. Stephenson, D. E. Wright, "Temporal discrimination learning of operant feeding in goldfish", J. Exp. Anal. Behav. 62(1), 1–13, 1994. pmc
  6. A. Sibeaux, C. Karlsson, C. Newport, T. Burt de Perera, "Distance estimation in the goldfish (Carassius auratus)", Proc. R. Soc. B, 2022. ora.ox.ac.uk
  7. C. Brown, "Fish intelligence, sentience and ethics", Animal Cognition 18(1), 1–17, 2015. mq.edu.au
  8. F. C. Fortenbaugh et al., "Sustained attention across the life span in a sample of 10,000", Psychological Science 26(9), 1497–1510, 2015. pmc
  9. G. Mark, Attention Span, Hanover Square Press, 2023; CNN interview, Jan 2023. ics.uci.edu
  10. P. Lorenz-Spreen, B. M. Mønsted, P. Hövel, S. Lehmann, "Accelerating dynamics of collective attention", Nature Communications 10, 1759, 2019. pmc
  11. Secondary: a piece on Temple University's law school site about law reviews using the number (law.temple.edu); Medium, "The attention span stat everyone keeps getting wrong" (describes the archived Statistic Brain page).