During serious global crisis, such as pandemics, one of the most important emotions involved in decision-making is fear. Responses to fear are usually presented as “fight-or-flight”, but it is not obvious which attitude is chosen, under which circumstances, and by what fraction of the population. Therefore, monitoring collective emotions is fundamental to devise appropriate responses, anticipate risky practices, and compliance levels. However, measuring fear(s) is not trivial, neither at the individual nor at the collective levels and, for this reason, few studies have directly analysed its impact.
The response during an infectious disease outbreak is a clear example of when apparently rational individual actions (such as going to an emergency room or buying medication) can lead to collectively risky behaviours (such as overcrowding hospitals) and even panic (if medical supplies run out). Using the 2009 influenza pandemic as a case study we will present preliminary work showing it is possible to use searches on Google as a proxy for collective concern. This tool can be very promising, being considerably cheaper and faster than traditional methods.
First identified in late March 2009, in America, a new highly infectious flu virus spread to Europe and Asia within weeks; by June the WHO had declared a pandemic. This new strain was eventually found to be less virulent than feared and the overall death toll was relatively low. However, media reports were constant during this period and there were several manifestations of fear and anxiety. Our hypothesis, in line with sociological research, is that during the pandemic period, the levels of fear in the population were relatively high and that these levels positively correlated with trust in the authorities and with willingness to comply with imposed measures. The pandemic offered a unique opportunity to test this hypothesis: it originated at least 60 surveys, covering 20 different countries, from January 2009 to December 2013, including two conducted by this research group. We collected these surveys and analysed anxiety levels, perceived-risk, intention to comply with and attitudes towards health authorities. We then tested whether searches for flu-related terms on Google could be a good proxy for fear levels by comparing weekly search-volumes with the results of the 60-survey analysis, actual numbers of flu cases and media reporting.
The results of this work can prove to be very effective in anticipating (and eventually helping to prevent) pandemics and other risk-prone situations, in a fast and effective manner.