Do pictures of eyes make people behave better?

The coffee-room experiment

In 2006, Melissa Bateson, Daniel Nettle and Gilbert Roberts ran a simple test in a staff coffee room at Newcastle University. Staff paid for tea, coffee and milk by dropping coins in an honesty box. Nobody checked.

Each week the researchers changed a small picture above the price list. Some weeks it showed flowers. Other weeks it showed a pair of eyes looking straight at the reader. They measured how much money went in for each litre of milk used.

In the eyes weeks, people paid nearly three times as much. Nobody was watching, and everyone knew it was a picture. The researchers called it the first evidence from a real setting that cues of being watched change how people cooperate.

The idea spread fast. People care about their reputations, and our brains respond quickly to faces and eyes, so a picture of eyes might switch on "someone could see this" without anyone noticing.

What happened next

The same team kept testing it outside the lab.

• Litter. In a university cafeteria, posters with eyes made people more likely to clear their own tables (Ernest-Jones and colleagues, 2011).
• Bike theft. Signs with eyes and the words "Cycle thieves, we are watching you" went up over three bike racks on a campus. Thefts at those racks fell by 62 per cent over a year. At racks without signs, thefts rose by 65 per cent (Nettle and colleagues, 2012). The signs seem to have moved theft somewhere else rather than stopped it.

That second result matters. A sign protects the spot it hangs over, not the whole campus.

Then the reviews came in

One eye-catching study proves little on its own. The real test is what happens when you pool every study that tried the same thing.

Review	What it pooled	What it found
Northover and colleagues, 2017	Lab studies where people chose how much money to give away	Pictures of eyes did not make people more generous
Bradley, Lawrence and Ferguson, 2018	117 papers on being observed and helping others	Being observed has a small positive effect, and it's stronger when a real person watches than when people only feel watched
Dear, Dutton and Fox, 2019	15 experiments with 2,035 people on antisocial acts such as littering and theft	Pictures of eyes cut antisocial behaviour by about 35 per cent

Put together, the picture is clearer than the headlines. Eye images don't reliably make people kinder or more generous. They do seem to make people less willing to do something they know others disapprove of, such as dropping rubbish or taking a bike. And actual people watching beat pictures of people.

Why the difference?

Researchers are still arguing about it, but one reading fits the evidence. Giving more than you must is optional, and nobody expects it, so a picture changes little. Littering and theft break a clear rule, and a reminder that someone could see you is enough to tip some people away from it.

The coffee-room study also fell between the two. Paying for your milk was the rule, so skipping payment was a small act of cheating, not a failure to be generous.

What it means in everyday life

The broader lesson is older than the eye pictures. In 1961 the writer Jane Jacobs argued that city streets are safest with "eyes upon the street": windows facing the footpath, shops open, people around. The reviews point the same way. Being seen by real people shapes behaviour more than a symbol of being seen.

So:

• A sign with eyes is cheap and may help against littering or theft at one spot. Expect some of the problem to move next door.
• Don't expect it to raise donations. The pooled evidence says it doesn't.
• Places where people can see each other do more. Shared spaces, windows onto the street and neighbours who know each other are the real version of the effect.

If you read a claim that a simple trick changes behaviour, look for the review that pooled every attempt, not the first study. Here, the first study was real, and the effect turned out smaller and narrower than it first looked.