Is Your Attention Span Really Shrinking? What Reading Research Shows

Where the goldfish comparison comes from, what attention span research says about reading, and which focus practices the evidence actually supports.

You open a long article. Three paragraphs in, your eyes land on the end of a sentence and nothing has gone in. You read it again. Somewhere in that moment, a half-remembered comparison floats up: attention spans are shrinking, and the goldfish now has the better one.

The line stings because it sounds like confirmation. If you cannot finish a chapter without reaching for your phone, the goldfish comparison feels less like a headline and more like a diagnosis of you.

Here is the first useful correction. Fading focus while reading is a mechanics problem, not a character flaw, and the comparison at the center of the panic is the shrinking attention span myth at its most confident: widely repeated, personally painful, and not backed by any published peer-reviewed study. What follows traces the claim to its source, reviews what controlled reading and interruption research actually finds, separates the real findings from the overhyped ones, and closes with the focus practices the evidence supports. If you want a calmer setup for long text while you read on, SReader does focal-point reading with no account, no ads, and no feed.

The shrinking attention span myth: where the goldfish comparison comes from

The claim itself is a comparison: human attention is collapsing, the story goes, and the goldfish comes out ahead. There is no paper behind the comparison.

Attention researchers reject the premise before the arithmetic even starts. There is no single average attention span to shrink. The concept is "pretty meaningless," in the words of Dr Gemma Briggs, a psychology lecturer at the Open University who studies attention in drivers and crime witnesses, because attention is "very much task-dependent." How much attention we apply, she explained, "will vary depending on what the task demand is." It also varies with our expectations and our prior experience.

A related belief has been debunked as well: the idea that people can focus on a single task, such as a lecture, for one fixed typical length of time. Anyone who has sat through both a gripping talk and a dull one has the intuition already. Duration follows the task, the stakes, and the material.

The goldfish half of the comparison fails too, and it fails in the opposite direction.

Goldfish deserve an apology

There is no evidence that goldfish have short attention spans or short memories. The scientific record points the other way.

Professor Felicity Huntingford, who has studied fish behavior for nearly half a century, told the BBC that goldfish "can perform all of the kinds of learning that have been described for mammals and birds." That is exactly why researchers use goldfish as a model system for studying memory formation: they learn, and they remember.

So the comparison insults both species. A comparison wrong in both directions was never a measurement. It was marketing.

What attention span research says about reading

Debunking the comparison is the easy half. A harder question remains: has human attention capacity itself declined? Reading is a good test case, because researchers can measure sustained attention under controlled conditions, and because the conditions around reading have transformed over the same years the claim has circulated.

The scientists who measure it say the capacity has not moved. "I've been measuring college students for the past 20 years," said Edward Vogel, a professor of psychology and neuroscience at the University of Chicago, and the results are "remarkably stable across decades." Michael Posner, the psychologist who identified the brain networks underlying attention, puts the horizon further back still: "There is no real evidence that it's changed since it was first reported in the late 1800s."

Twenty years of college students. More than a century of reports. The measured capacity to sustain attention looks stable even as reading moved from paper to screens and from quiet desks to pocket-sized feeds. That contrast carries the rest of this article. If capacity is stable while reading feels harder, the explanation sits in the environment, not in your head.

The real numbers: two decades of screen-switching data

Something real has been measured, and it deserves an honest reading. Gloria Mark, Chancellor's Professor of Informatics at UC Irvine, has tracked knowledge workers since 2003, first with stopwatches and later with computer-logging software. Her subject is not attention capacity but a closely related behavior: how long people stay on a single screen before switching away.

The trend in that behavior is steep. In 2004, people averaged two and a half minutes on a screen before switching. By 2012, the average was 75 seconds. From 2014 through 2020, about five studies found durations between 44 and 50 seconds, and these were "not just my studies, but also replicated by others," as Mark describes the work. In her 2016 study, the average was 47 seconds and the median was 40, "which means half of the observations we found were 40 seconds or less."

Take those numbers seriously, then read them precisely. They measure screen-switching behavior in interruption-rich digital environments: people moving between windows, apps, and devices many times an hour. They do not measure the underlying capacity to attend, which is what the goldfish claim is supposedly about. The shortening of screen durations is real as behavior. It is evidence about how we now work and read, not about brains that have shrunk.

Why reading feels harder now

None of this lets the modern reading environment off the hook. Several findings about interruption and screens are solid, and together they explain most of what the goldfish myth tries, wrongly, to explain.

Interruptions carry real costs. Returning to a page after a switch takes effort that the switch itself conceals, so a brief glance can cost more reading time than the glance took. Research on screen inferiority finds that deep reading is often harder on screens than on paper. And correlational work links heavier media multitasking with attention measures, though correlation is not cause and the direction of effect is not settled.

Mind-wandering research completes the picture. Work by Smallwood and Schooler, extended by Unsworth and McMillan, shows that lapses during reading are normal and frequent: attention drifts, the eyes keep moving, and comprehension quietly drops, even among skilled readers. That is the baseline for every reader who ever lived. What is new is the environment stacked on top of that baseline, with feeds engineered for the next tap, notifications, and a well-rehearsed habit of self-interruption that Mark's sub-minute numbers document.

Put the pieces together and the felt experience makes sense. Reading long text today often means deliberately defying the default design of the device you are reading on. When a chapter will not stay in your head, you are not broken. The conditions around reading changed, and they changed recently.

One part of the mind-wandering findings is especially practical: lapses during reading are partly reduced by pacing and task structure. Attention drifts less when the task gives it an anchor. That points directly at what helps.

Speed-reading claims are overhyped too

The skepticism that sank the goldfish comparison should also be aimed at the products selling the cure. Speed-reading apps promise enormous words-per-minute counts, usually built on RSVP, or rapid serial visual presentation, in which words appear in quick succession at a fixed point instead of scrolling past.

A 2016 review by Rayner and colleagues in Psychological Science in the Public Interest, "So Much to Read, So Little Time," tested those promises against the evidence. RSVP-style presentations can reach high speeds. Comprehension drops sharply at those speeds. Pushing pace past comprehension trades real understanding for the feeling of speed. That is the same shape of error as the goldfish story: a striking claim that does not survive contact with the evidence.

We have gone through this trade-off in detail in Why Speed Reading Apps Drain You and What Actually Works, including the pace a normal reader can realistically sustain. The short version for any pacing tool is one rule: comprehension first. Speed up when the material is easy; slow down when the ideas need more room.

What actually helps: focus practices the evidence supports

The research above points to three levers, and willpower is not one of them.

Anchor the task. Because mind-wandering during reading is partly reduced by pacing and task structure, give attention something to hold onto. Focal-point reading does this directly: a highlighted line acts as a reading rail, staying centered while the surrounding text keeps context visible. The line stays put; the page moves. When drift happens, it has an address, so you notice it sooner and return faster. If retention is your main worry, we covered the connection in How Focused Reading Helps You Retain What You Read.

Match pace to material. Faster for easy passages, slower for dense ones, with comprehension as the benchmark instead of words per minute. Before tuning anything, test your real reading speed and comprehension to get a baseline. Changing one variable against a known baseline is the only honest way to tell whether a change helped.

Design the environment, not the person. Every feed, badge, and notification you remove is one less switch to recover from. Local-first, no-account reading tools take attention-hostile design out of the session by default: no ads, no feed, no reason for the app to want your eyes anywhere else. SReader is one such option. It imports TXT, Markdown, EPUB, and PDF text; keeps your library on your device; and offers line, word, and book views with an adjustable pace.

Two limits, stated plainly. Pacing helps anchor attention; it does not eliminate lapses, because lapses are normal. And no setting diagnoses, treats, or universally solves anything about how your brain works. This is a reading plan, not a race.

If you want to test all of this, keep it small: one long text, two weeks, one change at a time, judged by what you understood rather than how fast you finished.

The verdict: your attention is not shrinking, your reading environment changed

The clear answer the goldfish headline never gives: there is no credible evidence that human attention capacity has declined. What is real is behavioral. People now switch screens in under a minute on average, and those switches carry costs. Deep reading is harder in distraction-rich environments, and it is helped by task structure, pacing anchors, and tools built without a feed.

So stop blaming your brain and start redesigning the conditions. The goldfish was never the problem. Neither are you.

If the calmer setup described here sounds worth a try, Try SReader. Your library stays on your device, and your pace stays yours.