attck-research

Evidence review: How Attention Works Online

Article: How Attention Works Online
Reviewed: November 2026 · Method version 2 · How these reviews work

How this review works

Each factual claim in the article was checked against the source it cites, then challenged by reviewers whose job was to find the strongest evidence against it: replication attempts, meta-analyses, later papers that cite the original work, and research from other fields that study the same mechanism. A separate reviewer weighed both sides, and a script re-checked every quoted source. Where a claim was overstated or misattributed, the article was corrected. The research below supports the article as it now reads.

Background research

Introduction

In 2015 a statistic that the average human attention span fell from 12 seconds in 2000 to eight, one second shorter than a goldfish’s, appeared in Time, the Telegraph, the Guardian, USA Today and the New York Times.

Every reference led back to a 2015 Microsoft Canada consumer insights report, but the figure was not from Microsoft’s research; the report credited Statistic Brain, and the sources that website listed could find no record of the study.

Open University psychologist Gemma Briggs said the idea of an average attention span is close to meaningless: “It’s very much task-dependent.”

Quote confirmed verbatim.

The research points to something closer to a budget, moved around constantly and spent generously when the content repays it.

Framing is an editorial synthesis; supporting dwell-time research added.

A Scarce Budget, Spent All Day

Every page, alert and message draws on the same limited supply, and people tend to move it toward whatever looks most worth the cost.

Uncited; information foraging theory is the closest established source.

Focus on one screen rarely lasts a minute

In an earlier experiment on an email task, Mark’s team found that interrupted people finished faster, with no loss of quality, but reported more stress, frustration, time pressure and effort.

Boundary condition: in procedural and reading tasks, interruptions do cost accuracy (Altmann et al. 2014; Foroughi et al. 2016).

Long reads still earn long attention

Readers who followed a link from elsewhere on the same site stayed longest, and those arriving from social media stayed the shortest time.

Pew also notes social media drives roughly 40% of cellphone visits.

Conclusion

The eight-second figure was never measured, and there is no good evidence that people have lost the capacity to focus. What has changed is how often they switch: in Mark’s studies, time on one screen fell from about two and a half minutes in 2004 to 47 seconds by the late 2010s, a drop she ties to the habits people have built around their devices. People spend a scarce resource across many competing claims, and they put it where the return looks highest.

The 2004 figure comes from observer-timed fieldwork that counted paper documents as well as screens (Gonzalez and Mark, CHI 2004: ‘somewhat more than two minutes using any electronic tool or paper document’), and the later figures from computer logging of different samples, so the trend is suggestive rather than exact. On capacity, meta-analyses of media multitasking and attention (Wiradhany and Nieuwenstein 2017; Parry and le Roux 2021) find small or non-significant links, while a 2018 review reads the balance of evidence as showing poorer sustained attention among heavy multitaskers with causality unknown. A 2025 three-level meta-analysis on the same question exists (Current Psychology, doi 10.1007/s12144-025-07624-2) but its abstract could not be retrieved, so it was not used.

What changed in the article

These edits were made on November 12, 2026 after the review.

Introduction: supported with caveats

Claim reviewed: The research describes attention as something closer to a budget, moved around constantly and spent generously when the content repays it, rather than a timer running out.

Now reads: The research points to something closer to a budget, moved around constantly and spent generously when the content repays it.

Why: The ‘budget, not timer’ framing fits the evidence the article goes on to cite and web dwell-time research showing people stay much longer on pages that pass an early check. But ‘the research describes’ implies a source the article never names at this point, and it is the article’s own synthesis. Now cites Nielsen, How Long Do Users Stay on Web Pages? NN/g (2011), summarizing Liu, White and Dumais (2010).

A Scarce Budget, Spent All Day: supported with caveats

Claim reviewed: Every page, alert and message draws on the same limited supply, and people keep reallocating it to whatever looks most worth the cost.

Now reads: Every page, alert and message draws on the same limited supply, and people tend to move it toward whatever looks most worth the cost.

Why: The idea that people spend attention where the expected return looks best matches information foraging theory and web dwell-time data, but the article cites neither. Mark’s own 2016 study also finds that switching partly reflects personality and stress, not only a judgment of value. Now cites Pirolli and Card, Information Foraging, Psychological Review (1999).

Focus on one screen rarely lasts a minute: supported with caveats

Claim reviewed: People with shorter focus rated their own productivity lower by the evening.

Now reads: Shorter focus went with lower self-rated productivity by the end of the day.

Why: The paper reports that shorter focus went with lower self-rated productivity at day’s end, and an earlier independent sample found the same pattern. It is a correlation in 40 people, and the abstract does not say whether it compares people or days, so the phrasing should not imply a fixed trait.

Focus on one screen rarely lasts a minute: supported with caveats

Claim reviewed: In an earlier experiment, Mark’s team found that interrupted people finished tasks faster, with no loss of quality, but reported more stress, frustration, time pressure and effort.

Now reads: In an earlier experiment on an email task, Mark’s team found that interrupted people finished faster, with no loss of quality, but reported more stress, frustration, time pressure and effort.

Why: The study found exactly this, and a 2022 meta-analysis supports the stress result. The ‘no loss of quality’ result comes from one lab email task, though; other experiments find that even brief interruptions raise errors in step-by-step tasks and hurt reading comprehension.

Focus on one screen rarely lasts a minute: supported with caveats

Claim reviewed: People make up for interruptions by working harder, and they feel it.

Now reads: People make up for interruptions by working faster, and they feel it.

Why: The authors say people compensate by working faster, at a price; the article says ‘harder’, which shifts the mechanism slightly. Other experiments show the compensation does not always protect accuracy.

Long reads still earn long attention: not supported

Claim reviewed: Read together, the studies show attention moves every minute or so, and it roughly doubles when the content repays it.

Now reads: Read together, the studies fit each other. At a desk, attention moves every minute or so; on a phone, readers gave long articles about twice the time they gave short ones, roughly what the difference in length would predict. Neither study measured why people stayed, but the pattern fits a budget rather than a timer: people keep spending as long as the page looks worth it.

Why: Mark’s 40-second median supports ‘every minute or so’, and Pew’s 123 versus 57 seconds supports ‘roughly doubles’. But Pew sorted articles by word count, not by whether they repaid attention, and at an average reading speed of about 238 words a minute, twice the time for pieces at least twice as long is what length alone would predict; dwell-time research confirms that time on a page rises with its length regardless of satisfaction. The budget idea itself has support in foraging models of web use, so the thesis is reasonable, but these two studies do not show it. Now cites Pirolli and Fu, Rational analyses of information foraging on the Web, Cognitive Science 29(3) (2005).

Conclusion: supported with caveats

Claim reviewed: There is no evidence that people are losing the ability to focus.

Now reads: The eight-second figure was never measured, and there is no good evidence that people have lost the capacity to focus. What has changed is how often they switch: in Mark’s studies, time on one screen fell from about two and a half minutes in 2004 to 47 seconds by the late 2010s, a drop she ties to the habits people have built around their devices. People spend a scarce resource across many competing claims, and they put it where the return looks highest.

Why: The BBC says there is no evidence attention spans are shrinking, and the strongest evidence on capacity agrees: two meta-analyses of media multitasking and attention find the link small, or gone after correcting for small-study effects, and no cohort data show the ability to focus declining. What has clearly changed is behavior. Mark’s logging shows time on one screen falling from about two and a half minutes in 2004 to 47 seconds, and she ties that to habits built around devices. ‘No evidence’ still overstates, because reviews report heavier media multitaskers doing worse on sustained-attention tasks, with causality unresolved. Now cites Our 47-second attention span, with Gloria Mark, Steelcase podcast transcript (2022).