Modeling the Dynamics of Risk Perception and Fear: Examining Amplifying Mechanisms and Their Consequences and Examining the Potential of Using Twitter Data to Study Public Response to Terrorist Threats

Principal Investigator: 
Other Researchers: 
Paul Slovic, Daniel Salazar
Performance Period: 
July 2014 to June 2015
Commercialization Status: 
Project Keywords: 
Twitter data
Boston attacks
risk perception
risk communication
The purpose of this research is to improve our understanding of how perceived risk and negative emotions (e.g., fear, anger, sadness) might spread throughout a community and ripple to other parts of the nation following events differing in their risk signal. It specifically seeks to understand the public’s comparative response to a wide array of hazards. It also seeks to examine how the effects of risk-communication messaging spread throughout a population offering protection from counterproductive attitudes and behaviors.   Study 1. Findings from a risk communication experiment were examined with the objective of learning whether communicating with the public in advance of a crisis might mitigate overreaction to an adverse event in terms of perceived risk, emotional response and collapse in confidence that terrorist events could be prevented and or harm mitigated. Two groups are compared across three national surveys, one receiving a carefully designed risk communication message and other not receiving a message. The first survey collected baseline measures, the second survey communicated the risk communication message to the risk message group and the third exposed both groups to a hypothetical airline attack scenario. Findings clearly demonstrated that the group receiving the risk communication message showed less loss of confidence in being protected from terrorism attacks following a simulated attack than the group not receiving the message.   Study 2. Findings from a second risk communication experiment were examined with the objective of corroborating the results from study 1 and now having the risk communication message delivered from video interviews with risk experts. A simulated video newscast was used to portray the attack scenario on a commercial aircraft. Two groups were compared across 4 nationwide surveys. Similar to study 1 the first survey collected baseline measures of confidence in DHS, risk perceptions and emotional reactions to terrorism. The second survey introduced the video risk communication message to half the respondents. The third survey exposed all respondents to the simulated newscast of the attack. The fourth survey followed up on respondents’ reaction to the attack scenario. The findings indicated that the group receiving the risk communication message was significantly more confident in DHS’s ability to prevent attacks on commercial aviation, minimize the harm from such an attack and handle national crisis. While both groups experienced a marked decrease in all these measures following exposure to the attack scenario the group receiving the risk communication message recovered their confidence more quickly than those receiving no risk communication message in advance of the attack scenario. As a result of these findings, respondents receiving a risk communication message were estimated to cancel 6.35 percent of planned airline trips in the first year following an attack on a commercial airline in comparison to 6.93 percent for the group receiving no risk communication message. While these percentage differences appear small an economic analysis of these findings indicated that the risk communication message properly implemented could potentially reduce loses to the airline industry in excess of $100 million in the first year.   Study 3. Findings from a third risk communication study were examined to see if encouraging respondents to talk with others about the risk message they heard would increase its effectiveness. This study used essentially the same risk message and measures as in study 1 and respondents were again part of a nationwide panel. The difference here was that there were now three groups: no risk communication message group; a risk communication message group that was not encouraged to share the message and a risk communication group encouraged to share the message. Findings indicated that perceptions of confidence in DHS to prevent an attack or minimize harm from an attack as well as handle a national crisis were slightly better for the risk communication group encouraged to share the message but only minimizing harm was statistically significant.   The important take away from the findings of these three studies is that well designed risk communications used in advance of an adverse event may mitigate public overreaction and potentially reduce the economic impacts of the event.   Study 4. Findings from longitudinal survey and Twitter data regarding the Boston Marathon bombings attack are compared with two objectives in mind: 1) to compare levels of expressed anger, fear and sadness regarding the attack and 2) to learn about how quickly the public recovered in terms of perceived risk of attending public sporting events, anger, fear, sadness and worry. Three nationwide surveys were conducted from April 2013 to July 2013 following the Boston Marathon bombing attack. Respondents were asked about their perceptions of risk and emotional reactions to the attack. The central findings were that immediately following the attacks a large percentage of those surveyed had moderate to high perceptions of risk, anger and sadness but less fear and worry. Over the next three months perceived risk and emotional reaction to the attack decreased markedly. Twitter data was also examined surrounding the Boston attack beginning in March 2013 and extending through August 2013. Twitter data was examined and counted for emotional references to terrorism during this time period. This was done not only to determine how social media was treating the attack but, to compare this finding with the survey data over a comparable time period. Results from the Twitter data indicated that fear was by far the primary emotion expressed followed by very modest amounts of sadness and anger. This differs significantly from the survey data. Consistent with the survey data, fear, sadness and anger decreased markedly shortly after the attacks.   The important take away from these findings is that the public is likely to recover quickly from adverse events and that government and corporate public relation officers, risk managers and policy makers should not overreact to the immediate response of the public to adverse events because the public is likely to recover quickly.