When the first Ebola case was diagnosed in the United States this led Twitter posts mentioning Ebola to jump from 100 posts per minute to 6,000 per minute and rapidly produced inaccurate claims that Ebola could be transmitted through food, water, and air.

It makes sense that people are motivated to identify threats. This is a cognitive bias that seems perfectly reasonable. Threats can hurt you and your family. 

However, such a bias can rapidly lead to public hysteria when information starts to be communicated from one person to the next. When we receive and communicate information we are naturally biased to identify and re-communicate risks at the expense of a more balanced understanding of events. This is called the social amplification of risk (Kasperson et al., 1988). 

Research by Moussaid et al. (2015) showed that in social diffusion chains information becomes more distorted and risk-focused as information passes from one individual to another. Mirroring the classic children's game "Telephone," individuals heard information and then shared it with the next person in a chain, with the information moving in this way from one person to the next.  In Moussaid et al.’s study, the first individual in the chain read unbiased neutral information about triclosan, a controversial antibacterial agent. As this information was shared, the chain rapidly lost key facts but tended to preserve facts associated with risks. This included the physiological effects of exposure to triclosan and the means for potential exposure, such as cosmetics and household cleaning products. The messages also became more distorted over time, including the addition of new information, amplifying the apparent risk. 

Similar evidence was found in two studies that Robert Jagiello and I recently conducted (Jagiello & Hills, 2018). These studies explored social diffusion chains in response to food additives and nuclear energy. They were inspired by the observation that people tend to disproportionately fear dread risks, or risks that are hard to predict, can kill many people, and do so indiscriminately (Slovic, 1987). Classic dread risks include plane crashes and terrorist attacks.

The studies found that a dread risk (nuclear power) tended to experience social risk amplification more readily than did a non-dread risk (food additives).  But we went one step further by introducing another manipulation. At the sixth spot in the chain, half of the individuals were re-exposed to the original information. This had almost no effect on the upward trend in social risk amplification. 

In other words, biased information shared through social channels strongly outweighed the balanced information provided in the original journalistic articles. The figure below shows how repeatedly moving information through people leads to increasingly risk-focused information.

Source: Courtesy of Thomas Hils

This may not be too surprising. A study by Mickes et al. (2013) found that compared with sentences from books and newspaper headlines, information shared through Facebook status updates was easily the most memorable. 

People know how to make information interesting and they appear to do so adaptively. This may come down to language's evolutionary roots as a social adaptation, which appears to trump its utility as a mechanism for communicating ecological information (Redhead & Dunbar, 2013). Communication may be first about providing social cues, second about communicating risks ('Fire!', ‘Run!’, 'Aaaagh') and only third about providing an accurate picture. So it is in brains built by the sausage-fingered hands of evolution.

References

Kasperson, R. E., Renn, O., Slovic, P., Brown, H. S., Emel, J., Goble, R., ... & Ratick, S. (1988). The social amplification of risk: A conceptual framework. Risk analysis, 8(2), 177-187.

Mickes, L., Darby, R. S., Hwe, V., Bajic, D., Warker, J. A., Harris, C. R., & Christenfeld, N. J. (2013). Major memory for microblogs. Memory & cognition, 41(4), 481-489.

Moussaïd, M., Brighton, H., & Gaissmaier, W. (2015). The amplification of risk in experimental diffusion chains. Proceedings of the National Academy of Sciences, 112(18), 5631-5636.

Redhead, G., & Dunbar, R. I. M. (2013). The functions of language: an experimental study. Evolutionary Psychology, 11(4), 147470491301100409.

Slovic, P. (1987). Perception of risk. Science, 236, 280–285.

Jagiello, R. D., & Hills, T. T. (2018). Bad News Has Wings: Dread Risk Mediates Social Amplification in Risk Communication. Risk Analysis. Advance online publication: https://doi.org/10.1111/risa.13117