The study
In 1999 Daniel Simons and Christopher Chabris, then at Harvard, published “Gorillas in our midst” in the journal Perception. Observers watched a video of two teams, one in white and one in black, passing basketballs. Their task was to count the passes made by one team. In the easy version they kept one count; in the hard version they kept separate counts of bounce passes and passes through the air.
Partway through, an unexpected figure crossed the scene for five seconds: either a woman carrying an open umbrella or a woman in a full gorilla costume. Of the 192 observers whose results were analysed, 46 per cent did not notice either. The gorilla was missed more often than the umbrella, and the harder counting task made missing more likely.
The authors then filmed a longer version in which the gorilla stopped among the players, turned to the camera, thumped its chest and walked on, nine seconds in all. Twelve new observers watched it. Half noticed.
An older idea
The study was an update rather than an invention. In the 1970s Neisser and colleagues had superimposed films of ball games and shown that people following one game often missed events in the other. In one demonstration, only six of twenty-eight observers reported the woman with the umbrella. Simons and Chabris borrowed the umbrella and added the gorilla, and found the effect even when the players were filmed together in one ordinary shot.
The name for the effect is inattentional blindness: failing to see something clearly visible because attention is committed elsewhere. It is not a problem with the eyes. The paper’s conclusion was that what we notice depends on what we are attending to, and on how similar the unexpected thing is to what we are looking for. The dark gorilla was noticed by 58 per cent of people counting the black team’s passes and 27 per cent of those counting the white team’s.
Experts too
In 2013 Drew, Võ and Wolfe of the Visual Attention Lab at Harvard Medical School asked 24 radiologists to search a set of lung scans for nodules, a task they do every working day. In the last case they had inserted a gorilla 48 times the size of an average nodule. Eighty-three per cent of the radiologists did not see it. Eye tracking showed that most of those who missed it had looked directly at the spot where it was.
Knowing about it
The gorilla is now famous, which creates a problem for anyone showing the video. In 2010 Simons made a new one. People who already knew the original saw the gorilla, as expected. But the new video also contained two other unexpected events, a curtain changing colour and a player leaving, and these viewers were slightly less likely to notice them than people who had never heard of the gorilla.
What it shows
The 1999 paper was itself a replication of Neisser’s findings, and the effect has since been shown in experts. But the “half” is not a constant. It changes with the difficulty of the task and how much the unexpected object resembles what people are watching for. The broader point holds: looking at something is not the same as seeing it, and expertise in a search does not protect against missing what the search is not for.