Social dynamics in hazard preparedness: A multi-method investigation

Bob Meyer, Kenny Broad, Ben Orlove

The goal of this research is to better understand the dynamics of individual and community decisions to invest in short and long-term prevention against losses from natural hazards. Central to this work is the use of simulations that will allow researchers to study in a laboratory setting how individuals make decisions in light of information gathered from a range of real-world information sources such as word-of-mouth (social influences), the web, and television broadcasts.

Previously, a prototype simulation called STORMVIEW was developed to study hurricane preparedness. This prototype is based on modeling ideas developed with support from a companion NSF/NOAA grant on hurricane warning communication and allows us to study how individuals make decisions over time to prepare for storm threats as a function of a variety of social and contextual factors, including the valence of word of mouth provided by neighbors and the visual formats of forecast graphics

The majority of efforts have included running STORMVIEW simulations with over 400 subjects and conducting real-time surveys of over 2000 coastal residents as several hurricanes approached the East coast of the U.S. Primary objectives include identifying key sources of information, objective vs. subjective risk assessment in relation to forecast products, key actions taken in response to information, role of past experience in mediating forecast response.


Major Findings

  • 2011 Survey Results Showed:
    • Participants who viewed forecast graphics, which contained track lines depicting the most likely path of the storm, had higher levels of preparation than those who saw graphics, which showed only uncertainty cones–even among those living far from the predicted center path.
    • Additionally, the participants who were most likely to express worry about an approaching storm and fastest to undertake preparatory action were those who, ironically, had never experienced one.
    • The patterns of information gathering revealed by participants in the simulation closely mirrored those observed in a separately-conducted field study of information gathering in advance of real-world storm threat, Hurricane Earl in North Carolina and Massachusetts in 2010. For example, STORMVIEW participants replicated the rank-order frequency of use of information sources reported in this study: television, the web, radio, and neighbors.
  • 2012 Survey Results Showed:
    • Notable differences in individual’s subjective assessment of risk vs. objective forecast;
    • Consistent misinterpretation of key vulnerabilities in respect to storm characteristics;
    • Confusion over evacuation orders;
    • Confusion over insurance coverage;
    • Experiences with previous storms effected current storm preparation;
    • Identification of challenges of managing the forecast triangle in which the National Hurricane Center releases forecasts, television reworks and disseminates the forecasts and public agencies make decisions on both.


Broader Impacts

  • Initial results have the potential to inform product development and dissemination strategies. These results have been presented to the National Hurricane Center’s headquarters in Miami.
  • STORMVIEW simulation can be used to study a range of behavioral phenomenon and its use can be extended to group settings such as schools, museums, cities, county, and state agencies. It is also amenable to physiometric studies.
  • This project has also been on the cutting edge of the use of computers for behavioral research. This has necessitated the collaboration of academic researchers with computer programmers and online survey experts.  These collaborations have resulted in cross-learning about both the characteristics of natural hazards, hazard warning information, computer simulations/programming, and online computing.


CRED2 Award (2010-2015): Funding was provided under the cooperative agreement NSF SES-0951516 awarded to the Center for Research on Environmental Decisions.

This research was also supported by research funds provided by the Wharton Risk Management and Decision Processes Center, the Florida Catastrophic Storm Risk Management Center, and Grants NSF-SES 0838650 and NSF-SES 0951516 from the National Science Foundation.