Change Blindness and Inattentional Blindness

In brief

In 1999, Daniel Simons and Christopher Chabris asked people to watch a video of two teams passing basketballs and count the passes made by the team in white shirts. Partway through, a person in a full gorilla suit walked into the court, faced the camera, thumped her chest, and walked off, clearly visible. Roughly half the viewers, absorbed in counting, never saw her. This is inattentional blindness: failing to perceive a fully visible, unexpected object because attention is engaged elsewhere. Its sibling, change blindness, is failing to notice that a scene has changed at all, even a large difference, once the change is masked by a brief interruption such as a flicker or a cut. Both converge on an uncomfortable conclusion: the rich, continuous visual world you feel you are seeing is not what your visual system delivers moment to moment. What you consciously see tracks what you attend to, not what light falls on your retina.

The full treatment

The problem it answers

Everyday introspection treats vision like a camera running continuously, registering everything in view whether or not anyone is looking for it. If that were true, a gorilla crossing a basketball court, or a stranger's face changing mid-conversation, would be nearly impossible to miss. Cognitive psychologists from the 1970s through the 1990s tested that assumption directly, building paradigms to isolate the gap between what hits the eye and what reaches awareness.

The two paradigms

Inattentional blindness is tested by giving a subject a demanding primary task, then introducing an unexpected, fully visible stimulus and asking afterward, cold, whether they saw anything unusual. No comparison across time is needed; the stimulus is simply present and still frequently goes unreported. Change blindness is tested differently: two versions of a scene, identical except for one alteration (an object present or absent, a color shifted, an object relocated), are shown in alternation, with a brief blank field or cut inserted between them. That interruption erases the motion signal that would otherwise pull the eye to the change. Without it, observers can stare for many seconds, or many cycles, without spotting a difference that looks glaring once pointed out. The two paradigms differ in design, one compares two states, the other has only one, but share a diagnosis: seeing a stimulus is not the same as perceiving it, and the difference is attention.

What the studies claim

The claim is not that people are careless. It is that focused attention binds a stimulus into a stable, reportable, conscious representation, and attention is a scarce resource a demanding task consumes. Divert it, and a stimulus can register on the retina and still never become something the person can say they saw. The finding generalizes past vision: Polly Dalton and colleagues have documented an auditory analog, "inattentional deafness," missing unexpected sounds while focused on a competing listening task.

The key study or demonstration

Simons and Chabris published "Gorillas in Our Midst" in the journal Perception in 1999. Their central number, that roughly half of attentive, motivated observers miss a fully visible gorilla walking through a video they are watching to track objects on screen, has been replicated many times and became one of the best known results in modern psychology, earning the pair the 2004 Ig Nobel Prize. The effect is not confined to novices: Trafton Drew, Melissa Vo, and Jeremy Wolfe (2013) had experienced radiologists search chest CT scans for lung nodules, then quietly inserted a gorilla image, forty-eight times the size of an average nodule, into the scan. Eighty-three percent never noticed it, though eye-tracking showed many had looked straight at it.

The parallel change blindness demonstration came two years earlier. Rensink, O'Regan, and Clark (1997) alternated pairs of photographs of real scenes, an aircraft engine present or missing, a helicopter's color shifted, separated by a brief blank flicker, and found observers often took many seconds and many alternations to detect changes central to the scene's meaning, sometimes failing entirely within the time given. Simons and Daniel Levin (1998) took the idea into the street: an experimenter stopped pedestrians for directions, and while two confederates carrying a door passed between them, the experimenter was swapped for a different person entirely. About half the pedestrians finished giving directions to the stranger without noticing the switch.

The applied stakes are not hypothetical. Strayer, Drews, and Johnston (2003) found simulator drivers on a hands-free phone showed sharply reduced recognition memory for roadside objects they had visually fixated on, looking but not seeing, evidence hands-free laws target the wrong bottleneck. Chabris, Adam Weinberger, Matthew Fontaine, and Simons (2011) modeled the case of Kenny Conley, a Boston officer convicted of perjury for testifying he had not seen fellow officers beating a man during a foot chase he ran at the same time: subjects jogging behind an assistant at night, some tracking a counting task, noticed a staged assault far less often when attention was loaded, real grounds for treating testimony like Conley's as plausible rather than dishonest.

Related distinctions

These are often confused with bad eyesight; they are not. Subjects have normal vision and are frequently looking straight at the missed stimulus. A further, distinct finding is change blindness blindness, documented by Daniel Levin, Nausheen Momen, Sarah Drivdahl, and Simons (2000): people confidently, wrongly predict they personally would notice such a change, a metacognitive error layered on the perceptual one. The missed events are also large and often bizarre, a gorilla, a swapped conversation partner, not small or peripheral, which separates the phenomena from ordinary distraction.

Lineage

The direct ancestor is Ulric Neisser and Robert Becklen's "selective looking" experiments (1975), tracking passes in one of two superimposed videotaped games while a person carrying an umbrella walked through the other, unseen by most viewers, an outgrowth of Neisser's ecological view of perception. In reading research, George McConkie and Keith Rayner's work on saccade-contingent display changes (1975) showed alterations made while the eye is in motion go unnoticed, an early cousin of the flicker paradigm, later extended to scene photographs by John Grimes (1996). Arien Mack and Irvin Rock systematized the static-display version through the 1990s and named the phenomenon in Inattentional Blindness (1998). Anne Treisman's feature integration theory (1980) supplied the scaffolding: focused attention binds separate visual features into an object available to consciousness. Downstream, the work fed research on distracted driving, forensic treatment of disputed eyewitness accounts, and a philosophical argument, from J. Kevin O'Regan and Alva Noë's sensorimotor contingency theory (2001), that visual consciousness is not a stored, detailed internal picture.

The strongest case for it

The evidence is hard to dismiss because it keeps reappearing in settings with real stakes, not only contrived lab videos. Drew, Vo, and Wolfe's radiologists show domain expertise does not immunize a trained searcher against missing something enormous. Strayer's driving research predicts a measurable, policy-relevant failure: hands-free phone laws, aimed at manual distraction, do not address the cognitive bottleneck that actually causes missed hazards. Chabris and colleagues' assault study shows the phenomenon transfers from a basketball video to a staged nighttime street, lending real support to a specific, previously incredible piece of eyewitness testimony. Across labs, materials, and decades, the core result, that a demanding task can make a large, unexpected stimulus vanish from conscious report, keeps replicating.

The strongest case against it

What the finding actually means is genuinely contested. Ronald Rensink's "coherence theory" (2000) argues no detailed representation of the unattended scene exists at all, so there is nothing to compare: the change was never encoded, rather than simply inaccessible. Simons's own 2000 review, "Current Approaches to Change Blindness," notes the data cannot rule out a rival account: a representation forms but cannot be retrieved without attention, a failure of memory and comparison rather than of perception itself. The distinction remains unresolved.

A second challenge complicates the headline that attention blocks everything unattended equally. Steven Most, Brian Scholl, Erin Clifford, and Simons (2005) showed that noticing an unexpected object depends heavily on how closely it resembles the category the observer is already monitoring, an effect they called attentional set: an object sharing a defining feature with the attended set is caught far more reliably than one resembling only the ignored set. Inattentional blindness is graded and category-sensitive, not the blanket blindness the gorilla anecdote suggests.

A third caution: "satisfaction of search," an older, distinct radiology finding dating to Tuddenham's work in the early 1960s, in which finding one abnormality lowers the odds of finding a second nearby one, is sometimes conflated with inattentional blindness in discussions of missed diagnoses, though the two likely involve different mechanisms.

Where it stands now

Both paradigms are standard, heavily replicated tools in attention research and are taught in nearly every introductory psychology course, most often through the gorilla video itself. The applied consequences are taken seriously in road-safety research and, more cautiously, in the eyewitness-testimony literature that courts and legal scholars consult, though experts warn against treating any one lab paradigm as a courtroom-ready explanation for a specific missed observation. The deeper theoretical question, whether visual experience is a rich internal picture or something thinner and attention-dependent, remains an active dispute in the philosophy and science of consciousness.

Test yourself

Think of a time you were sure you would have noticed something, a text while driving, a changed detail in a familiar room, a person's new haircut, and did not. Before you reach for an excuse about a bad angle or poor lighting, consider the less comfortable explanation the research points to: your attention was elsewhere, and elsewhere is where most of the world goes to become invisible.