Safety in numbers

Safety in numbers is the hypothesis that, by being part of a large physical group or mass, an individual is less likely to be the victim of a mishap, accident, attack, or other bad event. Some related theories also argue (and can show statistically) that mass behaviour (by becoming more predictable and "known" to other people) can reduce accident risks, such as in traffic safety – in this case, the safety effect creates an actual reduction of danger, rather than just a redistribution over a larger group.

The mathematical biologist W.D. Hamilton proposed his selfish herd theory in 1971 to explain why animals seek central positions in a group. Each individual can reduce its own domain of danger by situating itself with neighbours all around, so it moves towards the centre of the group. The effect was tested in brown fur seal predation by great white sharks. Using decoy seals, the distance between decoys was varied to produce different domains of danger. The seals with a greater domain of danger had as predicted an increased risk of shark attack.Antipredator adaptations such as the flocking of birds, herding of sheep, and schooling of fish. Similarly, Adelie penguins wait to jump into the water until a large enough group has assembled, reducing each individual's risk of seal predation.

In 1949 R. J. Smeed reported that per capita road fatality rates tended to be lower in countries with higher rates of motor vehicle ownership. This observation led to Smeed's Law.

In 2003 P. L. Jacobsen compared rates of walking and cycling, in a range of countries, with rates of collisions between motorists and cyclists or walkers. He found an inverse correlation. This inverse correlation can be explained by theories including (1) safer walking and cycling conditions cause more people to walk and cycle, (2) more people walking and cycling cause walking and cycling to be safer, and (3) external factors cause walking and cycling to increase while simultaneously causing walking and cycling to be safer. For further discussion, see for example Correlation does not imply causation.

...
Wikipedia