a/justice· 31 October 2026 · 5 min read

The bikes didn't vanish. The thieves walked to the next rack.

Art: Omxus Press

Three bike racks at a British university kept losing bikes. The university hung a sign over each: a pair of eyes and the words "Cycle thieves: we are watching you." Thefts there fell from 39 to 15 in a year. Then someone added up the whole campus.

70 and 68

At the racks with signs, theft fell by 62%. At the rest of the campus it rose from 31 to 51. Add the two together and the campus lost 70 bicycles in the year before the signs and 68 in the year after (Nettle et al., 2012). The thieves had not stopped. They had moved out of sight of the eyes.

That isn't a failure of the signs. It's the lesson. People respond to the place they are standing in. Make a few spots feel watched and behaviour changes in those spots and nowhere else.

The coffee room that started it

The signs came from a smaller experiment. In 2006 a university psychology department ran its tea and coffee on an honour system: 30 pence for tea, 50 for coffee, coins in a box. Researchers swapped the picture above the price list each week, flowers one week and a pair of eyes the next. In the flower weeks people paid 15 pence per litre of milk used. In the eyes weeks they paid 42 pence, 2.76 times as much (Bateson et al., 2006). The same 48 people, buying the same coffee, with a photograph above them.

We usually treat honesty as something a person has or lacks, and we build justice around finding the person who lacks it. The coffee room says the people did not change. The wall did.

How far the effect goes

Pooling 15 experiments with 2,035 people, a meta-analysis found eye images cut antisocial behaviour by about 35% (Dear et al., 2019). A review of 44 evaluations found real CCTV cut crime by 16%, most of it in car parks (Welsh & Farrington, 2009). The effect has a clear edge: in laboratory games where people give money away in secret, eye pictures mostly do nothing (Northover et al., 2017). Eyes work where there is a rule to keep and a reputation to lose by breaking it.

Seen is not watched

A camera records you for someone who is not there. A neighbour sees you because you share a place. The coffee-room eyes worked through the second: they stood for the group, not for a guard.

Jane Jacobs tells a story about this in her 1961 book. From her window on Hudson Street in New York she saw a man trying to coax a little girl to go with him, and the girl stiffening against it. Jacobs wondered whether to step in. Then she noticed she didn't need to. The butcher's wife had come out and was standing with her arms folded. The deli owner and his son-in-law had appeared, men had come to the door of the bar, and the fruit seller and the locksmith were at their doorways. The man had no idea he was surrounded. He turned out to be the girl's father. Nothing happened, and that's the point: nobody was paid to watch, and everyone was watching (Jacobs, 1961). She called it "eyes on the street".

The research on whole neighbourhoods says the same. In 343 Chicago neighbourhoods, places where neighbours trust each other and will act together had less violence (Sampson et al., 1997). In Philadelphia, 541 vacant lots were randomly cleaned and greened; in neighbourhoods below the poverty line, gun violence fell 29% (Branas et al., 2018). Nobody was watched more closely. People came back outside and saw each other.

What we build instead

Much of modern life removes that visibility: garages that swallow the driver and close the door, blocks with long blank corridors, workplaces where nobody knows a colleague, prisons that take people out of every community that knew them. Australia spent $26.5 billion on police, courts and corrective services in 2024–25 (Productivity Commission, 2026), most of it after the harm.

The eyes did not make anyone a better person. They showed what a place can do. Build places where people can see each other and know each other, everywhere, and the coffee gets paid for.

How sure are we?

The coffee-room test is ten weeks in one room: a large effect, in a small test. The bike study has one site and no randomisation, and some of the move could reflect where bikes were parked. The step from eye pictures to neighbourhood design is an argument from mechanism, supported by the neighbourhood studies, not tested directly. And nobody should want a world of being watched. What Hudson Street had was people who knew each other.

What you can do

Sit out the front. Learn the names of the people either side of you, then the people opposite. Leave the garage door open while you potter. If a house nearby stands dark, or a lot has gone to weeds, ask your council to fix it or offer to weed it yourselves: Philadelphia shows that isn't just tidiness. And if you run a coffee room or an honour box, stop checking on people and let them see each other.

Nobody on Hudson Street was a better person than you. They were just out the front.

References

  1. Bateson, M., Nettle, D., & Roberts, G. (2006). Cues of being watched enhance cooperation in a real-world setting. Biology Letters, 2(3), 412–414. https://doi.org/10.1098/rsbl.2006.0509
  2. Branas, C. C., South, E., Kondo, M. C., Hohl, B. C., Bourgois, P., Wiebe, D. J., & MacDonald, J. M. (2018). Citywide cluster randomized trial to restore blighted vacant land and its effects on violence, crime, and fear. Proceedings of the National Academy of Sciences, 115(12), 2946–2951. https://doi.org/10.1073/pnas.1718503115
  3. Dear, K., Dutton, K., & Fox, E. (2019). Do 'watching eyes' influence antisocial behavior? A systematic review & meta-analysis. Evolution and Human Behavior, 40(3), 269–280. https://doi.org/10.1016/j.evolhumbehav.2019.01.006
  4. Jacobs, J. (1961). The death and life of great American cities. Random House.
  5. Nettle, D., Nott, K., & Bateson, M. (2012). 'Cycle thieves, we are watching you': Impact of a simple signage intervention against bicycle theft. PLoS ONE, 7(12), e51738. https://doi.org/10.1371/journal.pone.0051738
  6. Northover, S. B., Pedersen, W. C., Cohen, A. B., & Andrews, P. W. (2017). Artificial surveillance cues do not increase generosity: Two meta-analyses. Evolution and Human Behavior, 38(1), 144–153. https://doi.org/10.1016/j.evolhumbehav.2016.07.001
  7. Productivity Commission. (2026). Report on government services 2026, Part C: Justice. Australian Government. https://www.pc.gov.au/ongoing/report-on-government-services
  8. Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277(5328), 918–924. https://doi.org/10.1126/science.277.5328.918
  9. Welsh, B. C., & Farrington, D. P. (2009). Public area CCTV and crime prevention: An updated systematic review and meta-analysis. Justice Quarterly, 26(4), 716–745. https://doi.org/10.1080/07418820802506206

Keep reading