WHAT THE RESEARCH FOUND

It was there
all along.

Two pictures, side by side. One obvious change. Where is it?

Two nearly identical printed photographs of a café street scene laid side by side on a table, with a finger pointing between them — AI-generated image, not a photograph of a real person or place.

We look, but miss
Attention finds changes
Interest guides attention
Why the obvious hides

Spot-the-difference feels like a test of sharp eyes. Psychology research says it is really a test of attention: we only notice a change when we happen to be attending to the place where it happens.

Updated 16 October 2026

THE SHORT ANSWER

We see less
than we think.

Researchers call it change blindness.
It is a limit of attention, not of eyesight.

FLICKER

A blink hides the change

When an original and a changed photo alternate with a brief blank between them, large changes become very hard to spot, even when people know one is coming.

Check a change

INTEREST

Central things get noticed

Changes to what people describe as the main point of a scene were found far faster than changes to details nobody mentioned.

See the numbers

REAL LIFE

Even a new person

In a 1998 street study, nearly half of people giving directions did not notice that the stranger they were talking to had been swapped.

The door study

HOW LONG TO SPOT IT?

Same change,
very different times.

Pick a condition from a classic 1997 experiment to see how long people needed to spot a large change between two photos. The bar shows how hard it was.

Average times from 48 colour photographs of real-world scenes, with one large change in each.

Change check

Choose a conditionNo blank between picturesFlicker, change to the main subjectFlicker, change to a background detailThe hardest scenes

No blank between picturesAbout 0.9 s

Without the flicker, the change shows up as movement and people found it in about 1.4 alternations on average.

Flicker, change to the main subjectAbout 4.7 s

With a blank between images, changes to objects of central interest took about 7.3 alternations to find.

Flicker, change to a background detailAbout 10.9 s

Changes to things nobody mentioned when describing the scene needed about 17.1 alternations on average.

The hardest scenesOver 50 s

For some images, observers needed more than 80 alternations, about 50 seconds, to identify a change that was large and repeating.

Source: Rensink, O'Regan and Clark (1997), 'To see or not to see', Psychological Science.

TWO NUMBERS WORTH KNOWING

Confident,
and often wrong.

How often people miss a real-world change,
and how sure they are that they would not.

≈ 50%

Pedestrians who missed a swapped stranger

Nearly 50 percent of pedestrians giving directions failed to notice they were talking to a different person after a door was carried between them.

THE DOOR STUDY · 1998

≈ 83%

People sure they would spot film changes

Nearly 83 percent said they would notice changes in short film scenes, yet only around 11 percent of people in earlier studies with the same clips actually did.

OVERCONFIDENCE · 2000

Researchers call this mismatch 'change blindness blindness': people do not believe they are prone to change blindness, even after being told about it.

Two workers carrying a large wooden door across a park path past a person holding a map, seen from a distance — AI-generated image, not a photograph of a real person or place.

FOUR WAYS TO SPOT MORE

Beat the puzzle,
with attention.

These tips follow from what the experiments found about attention. They may also help whenever you compare two versions of a picture.

Search, don't stare
Changes are seen only when attention is on that spot, so move through the picture area by area.
Check the boring bits
Background details were missed far longer than the main subject, so look where you would not normally look.
Name what you see
In the 1996 study, verbal cues improved performance considerably, and mentioned objects were easy to check.
Avoid the blink
Flipping straight between two versions lets the change show as motion; a gap or a glance away hides it.

THE DOOR STUDY

Directions,
a door, a new face.

In a study published in 1998, an experimenter holding a map asked a pedestrian for directions. Two people then carried a wooden door between them, and the experimenter was replaced by someone who looked and sounded markedly different.

Nearly half the pedestrians carried on without noticing. The researcher behind it argues that we focus on the task in hand, and that noticing everything would simply be too distracting.

A busy city square seen from above with many people, bikes and market stalls, full of small details — AI-generated image, not a photograph of a real person or place.

Questions

Questions people ask

What is change blindness?

A perceptual effect in which a change to something we are looking at goes unnoticed. For example, observers often fail to notice major differences in an image that flickers off and on again.

Why is spot-the-difference so hard?

Research suggests that to perceive a change, attention must be directed to it while it is happening. Without that, even large changes can go unseen for many seconds.

Is change blindness a problem with eyesight?

No. In the 1996 flicker study, verbal cues improved performance considerably, showing the effect was not due to reduced visibility.

Has this been known for long?

Film editors noticed that audiences missed changes in backgrounds, and William James mentioned the difficulty of detecting change in 1890. Laboratory research grew in the 1990s.

Does it happen without a blink or flicker?

Yes. In 'mudsplash' experiments, small high-contrast shapes scattered over a picture, away from the change, are enough to stop people noticing it.

It was there
all along.