• Allaire, M. C., 2016: Using practical and social information to influence flood adaptation behavior. Water Resour. Res., 52, 60786093, https:/doi.org/10.1002/2015WR018258.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bargh, J. A., Ed, 2007: Social Psychology and the Unconscious: The Automaticity of Higher Mental Processes. Psychology Press, 352 pp.

  • Brehm, J. W., 1966: A Theory of Psychological Reactance. Academic Press, 135 pp.

  • Bubeck, P., W. J. W. Botzen, and J. C. J. H. Aerts, 2012: A review of risk perceptions and other factors that influence flood mitigation behavior. Risk Anal., 32, 14811495, https://doi.org/10.1111/j.1539-6924.2011.01783.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Crossett, K., B. Ache, P. Pacheco, and K. Haber, 2013: National coastal population report: Population trends from 1970 to 2020. NOAA, accessed 27 February 2018, https://coast.noaa.gov/digitalcoast/training/population-report.html.

  • de Boer, J., W. J. Wouter Botzen, and T. Terpstra, 2014: Improving flood risk communication by focusing on prevention-focused motivation. Risk Anal., 34, 309322, https://doi.org/10.1111/risa.12091.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Doocy, S., A. Daniels, S. Murray, and T. D. Kirsch, 2013: The human impact of floods: A historical review of events 1980-2009 and systematic literature review. PLoS Curr. Disasters, 1, https://doi.org/10.1371/currents.dis.f4deb457904936b07c09daa98ee8171a.

    • Search Google Scholar
    • Export Citation
  • Eiser, J. R., A. Bostrom, I. Burton, D. M. Johnston, J. McClure, D. Paton, J. van der Pligt, and M. P. White, 2012: Risk interpretation and action: A conceptual framework for responses to natural hazards. Int. J. Disaster Risk Reduct., 1, 516, https://doi.org/10.1016/j.ijdrr.2012.05.002.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fischhoff, B., 2013: The sciences of science communication. Proc. Natl. Acad. Sci. USA, 110 (Suppl. 3), 1403314039, https://doi.org/10.1073/pnas.1213273110.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fox, S., 2014: The social life of health information. Pew Research Center, accessed 21 January 2018, http://www.pewresearch.org/fact-tank/2014/01/15/the-social-life-of-health-information/.

  • GfK, 2017: KnowledgePanel recruitment and sample survey methodologies. GfK, accessed 22 June 2017, https://www.gfk.com/fileadmin/user_upload/dyna_content/US/documents/KnowledgePanel_Methodology.pdf.

  • Grinsted, A., J. C. Moore, and S. Jevrejeva, 2013: Projected Atlantic hurricane surge threat from rising temperatures. Proc. Natl. Acad. Sci. USA, 110, 53695373, https://doi.org/10.1073/pnas.1209980110.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hine, D. W., W. J. Phillips, R. Cooksey, J. P. Reser, P. Nunn, A. D. G. Marks, N. M. Loi, and S. E. Watt, 2016: Preaching to different choirs: How to motivate dismissive, uncommitted, and alarmed audiences to adapt to climate change? Global Environ. Change, 36, 111, https://doi.org/10.1016/j.gloenvcha.2015.11.002.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Holland, G., and C. Bruyère, 2014: Recent intense hurricane response to global climate change. Climate Dyn., 42, 617627, https://doi.org/10.1007/s00382-013-1713-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Horrigan, J. B., 2008: The internet and consumer choice: Online Americans use different search and purchase strategies for different goods. Pew Internet and American Life Project Rep., 42 pp., http://www.pewinternet.org/2008/05/18/the-internet-and-consumer-choice/.

  • IPCC, 2013: Climate Change 2013: The Physical Science Basis. Cambridge University Press, 1535 pp., https://doi.org/10.1017/CBO9781107415324.

    • Crossref
    • Export Citation
  • Janis, I. L., and L. Mann, 1976: Coping with decisional conflict. Amer. Sci., 64, 657667, http://www.jstor.org/stable/27847557.

  • Jonkman, S. N., and J. K. Vrijling, 2008: Loss of life due to floods. J. Flood Risk Manage., 1, 4356, https://doi.org/10.1111/j.1753-318X.2008.00006.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kahneman, D., and A. Tversky, 2013: Prospect theory: An analysis of decision under risk. Handbook of the Fundamentals of Financial Decision Making, World Scientific, 99–127, https://doi.org/10.1142/9789814417358_0006.

    • Crossref
    • Export Citation
  • Kellens, W., T. Terpstra, and P. De Maeyer, 2013: Perception and communication of flood risks: A systematic review of empirical research. Risk Anal., 33, 2449, https://doi.org/10.1111/j.1539-6924.2012.01844.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kopp, R. E., R. M. Horton, C. M. Little, J. X. Mitrovica, M. Oppenheimer, D. J. Rasmussen, B. H. Strauss, and C. Tebaldi, 2014: Probabilistic 21st and 22nd century sea-level projections at a global network of tide-gauge sites. Earth’s Future, 2, 383406, https://doi.org/10.1002/2014EF000239.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kreibich, H., A. H. Thieken, T. Petrow, M. Müller, and B. Merz, 2005: Flood loss reduction of private households due to building precautionary measures–lessons learned from the Elbe flood in August 2002. Nat. Hazards Earth Syst. Sci., 5, 117126, https://doi.org/10.5194/nhess-5-117-2005.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Krueger, N., and P. R. Dickson, 1994: How believing in ourselves increases risk taking: Perceived self-efficacy and opportunity recognition. Decis. Sci., 25, 385400, https://doi.org/10.1111/j.1540-5915.1994.tb00810.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kunreuther, H., 1996: Mitigating disaster losses through insurance. J. Risk Uncertainty, 12, 171187, https://doi.org/10.1007/BF00055792.

  • Leventhal, H., R. Singer, and S. Jones, 1965: Effects of fear and specifity of recommendation upon attitudes and behavior. J. Pers. Soc. Psychol., 2, 2029, https://doi.org/10.1037/h0022089.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lindell, M. K., and R. W. Perry, 1992: Behavioral Foundations of Community Emergency Planning. Taylor and Francis, 630 pp.

  • Lindell, M. K., and R. W. Perry, 2004: Communicating Environmental Risk in Multiethnic Communities. Sage Publishing, 272 pp.

  • Lindell, M. K., and R. W. Perry, 2012: The Protective Action Decision Model: Theoretical modifications and additional evidence. Risk Anal., 32, 616632, https://doi.org/10.1111/j.1539-6924.2011.01647.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Morgan, G. M., B. Fischhoff, A. Bostrom, and C. J. Atman, 2002: Risk Communication: A Mental Models Approach. Cambridge University Press, 351 pp.

    • Crossref
    • Export Citation
  • Paton, D., and D. Johnston, 2001: Disasters and communities: Vulnerability, resilience and preparedness. Disaster Prev. Manage., 10, 270277, https://doi.org/10.1108/EUM0000000005930.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Paton, D., and D. Johnston, 2017: Disaster Resilience: An Integrated Approach. Charles C Thomas, 413 pp.

  • Pew Research Center, 2017: Demographics of internet and home broadband usage in the United States, Pew Research Center, accessed 21 January 2018, http://www.pewinternet.org/fact-sheet/internet-broadband/.

  • Pidgeon, N., and B. Fischhoff, 2011: The role of social and decision sciences in communicating uncertain climate risks. Nat. Climate Change, 1, 3541, https:/doi.org/10.1038/nclimate1080.

    • Search Google Scholar
    • Export Citation
  • Poortinga, W., A. Spence, L. Whitmarsh, S. Capstick, and N. Pidgeon, 2011: Uncertain climate: An investigation into public scepticism about anthropogenic climate change. Global Environ. Change, 21, 10151024, https://doi.org/10.1016/j.gloenvcha.2011.03.001.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Richardson, G. E., B. L. Neiger, S. Jensen, and K. L. Kumpfer, 1990: The Resiliency Model. Health Educ., 21, 3339.

  • Ronan, K. R., K. Crellin, and D. M. Johnston, 2012: Community readiness for a new tsunami warning system: Quasi-experimental and benchmarking evaluation of a school education component. Nat. Hazards, 61, 14111425, https://doi.org/10.1007/s11069-011-0070-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sanquini, A. M., S. M. Thapaliya, and M. M. Wood, 2016: A communications intervention to motivate disaster risk reduction. Disaster Prev. Manage., 25, 345359, https://doi.org/10.1108/DPM-11-2015-0256.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schanze, J., and et al. , 2008: Systematisation, evaluation and context conditions of structural and non-structural measures for flood risk reduction. CRUE Research Rep. I-1, 206 pp., http://www.flood-era.ioer.de/files/FLOOD-ERA_Final_report.pdf.

  • Tierney, K. J., M. K. Lindell, and R. W. Perry, 2001: Facing the Unexpected: Disaster Preparedness and Response in the United States. Joseph Henry Press, 306 pp.

  • Witte, K., and M. Allen, 2000: A meta-analysis of fear appeals: Implications for effective public health campaigns. Health Educ. Behav., 27, 591615, https://doi.org/10.1177/109019810002700506.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wong-Parodi, G., and B. H. Strauss, 2014: Team science for science communication. Proc. Natl. Acad. Sci. USA, 111, 13 65813 663, https://doi.org/10.1073/pnas.1320021111.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wong-Parodi, G., B. Fischhoff, and B. Strauss, 2014: A method to evaluate the usability of interactive climate change impact decision aids. Climatic Change, 126, 485493, https://doi.org/10.1007/s10584-014-1226-9.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wong-Parodi, G., B. Fischhoff, and B. Strauss, 2017: Plans and prospects for coastal flooding in four communities affected by Sandy. Wea. Climate Soc., 9, https://doi.org/10.1175/WCAS-D-16-0042.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    Climate Central’s Surging Seas Risk Finder tool for Connecticut showing the (top left) landing page, (top right) risk zone map, (bottom left) flooding risk, and (bottom right) coastal flood days.

  • View in gallery

    Climate Central’s Surging Seas protective action decision aid showing the (top left) landing page and (top right) small, (bottom left) medium, and (bottom right) large steps that people can take to reduce their risk.

  • View in gallery

    Predicted probabilities of intention to prepare for future flood events (0 = no intention, 1 = intention to prepare) by previous action (gray = had not previously prepared, white = had previously prepared) and condition (Control, Risk, Protective, and Risk+Protective). Error bars represent the 95% confidence interval.

  • View in gallery

    Perceived future flood risk estimates (%) by condition (Control, Risk, Protective, Risk+Protective). Error bars represent ±1 standard error.

  • View in gallery

    Schematic of study design.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 157 157 36
PDF Downloads 102 102 16

Effect of Risk and Protective Decision Aids on Flood Preparation in Vulnerable Communities

View More View Less
  • 1 Department of Engineering and Public Policy, Carnegie Mellon University, Pittsburgh, Pennsylvania
  • | 2 Climate Central, Princeton, New Jersey
© Get Permissions
Full access

Abstract

Although the risks of flooding demand responses by communities and societies, there are also many cost-effective actions that individuals can take. The authors examine two potential determinants of such adoption: individual predisposition to act and the impact of decision aids that emphasize the risk, the actions, both, or neither (control). Respondents are a representative sample (N = 1201) of individuals in the areas most heavily affected by Superstorm Sandy in 2012. The authors find that, in the overall sample, seeing protective actions coupled with risk information or alone produced higher rates of individuals reporting that they intended to take action preparing for future storms, compared to a control group receiving no additional information. Moreover, that occurred despite the aids reducing their perceptions of risk. The authors find that individuals who reported having taken previous action are more responsive to decision aid messages with the exception of the combined message (risk and protective actions)—which had a positive effect on those who had not acted previously, but a negative effect on those who had. These results suggest that, in communities that already are aware of their flood risks, the critical need is for authoritative, comprehensible information regarding the most feasible and cost-effective protective actions that they can take. Providing such information requires analysis to determine which actions qualify and a design process that incorporates user feedback to ensure that recommendations are easily understood and credible.

© 2018 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Gabrielle Wong-Parodi, gwongpar@cmu.edu

Abstract

Although the risks of flooding demand responses by communities and societies, there are also many cost-effective actions that individuals can take. The authors examine two potential determinants of such adoption: individual predisposition to act and the impact of decision aids that emphasize the risk, the actions, both, or neither (control). Respondents are a representative sample (N = 1201) of individuals in the areas most heavily affected by Superstorm Sandy in 2012. The authors find that, in the overall sample, seeing protective actions coupled with risk information or alone produced higher rates of individuals reporting that they intended to take action preparing for future storms, compared to a control group receiving no additional information. Moreover, that occurred despite the aids reducing their perceptions of risk. The authors find that individuals who reported having taken previous action are more responsive to decision aid messages with the exception of the combined message (risk and protective actions)—which had a positive effect on those who had not acted previously, but a negative effect on those who had. These results suggest that, in communities that already are aware of their flood risks, the critical need is for authoritative, comprehensible information regarding the most feasible and cost-effective protective actions that they can take. Providing such information requires analysis to determine which actions qualify and a design process that incorporates user feedback to ensure that recommendations are easily understood and credible.

© 2018 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Gabrielle Wong-Parodi, gwongpar@cmu.edu

1. Introduction

Flooding is the leading cause of natural disaster–related deaths worldwide (Doocy et al. 2013), with coastal flooding alone causing nearly 170 000 deaths and affecting 2.27 billion people between 1975 and 2002 (Jonkman and Vrijling 2008). The risk of coastal flooding is increasing due to more frequent and intense high-impact storm events, rising sea levels, and more people living in flood-prone areas (Grinsted et al. 2013; Holland and Bruyère 2014; Kopp et al. 2014; IPCC 2013; Crossett et al. 2013). Although the risks of flooding demand responses by communities (e.g., zoning, levees, wetlands) and societies (e.g., mitigating climate change), there are also many actions that individuals can take to reduce their personal risk. Low-cost, effective measures include creating an evacuation plan, moving vehicles to higher ground, and making copies of important documents. Higher-cost measures include raising a home on pilings, purchasing flood insurance, and moving. However, even though such measures can be cost-effective for individual households and significantly reduce flood-related damages (Kreibich et al. 2005; Schanze et al. 2008), few people take them voluntarily (Kunreuther 1996).

In their systematic review of research on flood risk communication and perception, Kellens et al. (2013) identified both cognitive processes (e.g., risk perceptions, knowledge) and affective ones (e.g., worry, concern) that can influence relevant attitudes, intentions, and behaviors (Kellens et al. 2013). These processes reflect both individuals’ circumstances (e.g., proximity to risk, previous experience, home ownership) and their personal characteristics (e.g., age, gender, education). In their review of protective behaviors related to flood risks, Bubeck et al. (2012) identified both relatively immutable factors and ones that might be addressed by appropriate programs (Bubeck et al. 2012). As examples of the latter, expecting government to provide protection or compensation for flood damage was associated with less individual action; seeing effective measures was associated with more. Both Kellens et al. (2013) and Bubeck et al. (2012) call for experimental research to clarify the mechanisms underlying these largely correlational results. We offer such a study, drawing on risk communication and disaster planning research.

According to Lindell and Perry’s (1992, 2004, 2012) well-known Protective Action Decision Model, social context, environmental cues, and social information interact with one another and personal experience to elicit perceptions, such as whether a threat is real, what additional information is needed, and which actions could be taken (Lindell and Perry 2012, 1992, 2004). These assessments are essential inputs to the decision-making processes determining, for example, whether individuals create evacuation plans, well in advance of a threat, or move vehicles to higher ground, given a perceived imminent danger.

Thus, messages that make threats seem larger and protective actions more effective (and cost effective) could increase the likelihood of recipients taking those actions. Whether that happens for any specific threat, action, recipient audience, and message is an empirical question. We offer and demonstrate a message testing procedure, using a representative sample of residents of an area recently affected by severe flooding (Superstorm Sandy), a state-of-the art website for communicating flooding risks (Climate Central’s Surging Seas), and intensively pretested messages regarding protective measures.

The importance of linking risk perception and action can be seen in studies in other domains, notably health, finding that messages that evoke negative affect, especially fear and worry, may have little, or even negative effect, unless coupled with opportunities to mitigate those risks (Witte and Allen 2000). For example, Leventhal et al. (1965) found that arousing fear of tetanus produced more favorable attitudes toward vaccination and stronger reported intentions to get a shot (Leventhal et al. 1965). However, actual behavior changed only when that message was paired with a plan of action. Many other studies have found the same pattern of results. In a review of studies from 1953 to 2000, Witte and Allen (2000) concluded that fear appeals alone tend to produce maladaptive responses, such as defensive avoidance (e.g., procrastination, shifting responsibility, taking another path; Janis and Mann 1976) or reactance, taking a contrary position, so as to deny a communication’s legitimacy (Brehm 1966). The most effective fear appeals are those coupled with messages showing effective measures that individuals can take. Drawing on this literature, researchers have found that cognitive and affective appeals coupled with practical advice can promote endorsement of measures for climate change adaptation (Hine et al. 2016) and earthquake-resistant construction (Sanquini et al. 2016).

Some recent studies have begun to examine individual differences in responses to such information. De Boer et al. (2014) identified a subset of individuals whom they described as “chronic prevention focused,” in the sense of focusing on “security, safety, stability, and obeying rules.” Those individuals acted as though guided by a decision-making heuristic that led to taking recommended protective actions, if feasible, without analyzing the options in any detail. De Boer et al. (2014) found that information that increased the fit between a prevention-framed risk message and a person’s chronic prevention focus enhanced adaptation responses.

Most research on preparedness is cross sectional in design, producing correlational evidence that makes causal inferences difficult (Ronan et al. 2012; Kellens et al. 2013; Bubeck et al. 2012). However, a few studies have been done employing quasi-experimental or experimental designs. Using a quasi-experimental design, Ronan et al. (2012) found that a children’s educational program on tsunami warnings increased preparedness indicators for 213 primary and intermediate school students in New Zealand (Ronan et al. 2012). Those indicators included knowledge, intended protective behaviors, intended risk mitigation, hazard awareness, risk perceptions, and perceived ability to cope with disaster. In a randomized field experiment, Allaire (2016) found that providing practical information about neighbors’ purchase of flood insurance increased uptake by 5% among 364 flood-prone households in Bangkok, Thailand; however, providing information about home retrofits had no effect. Such studies suggest both the potential for information-based interventions and the need for testing.

Thus, creating effective appeals requires offering actions that appear effective and feasible (Paton and Johnston 2001, 2017; Eiser et al. 2012), given recipients’ circumstances (Poortinga et al. 2011), responsibilities (Bubeck et al. 2012), and predispositions (de Boer et al. 2014). Here, we test such an intervention drawing on existing research and taking advantage of an Internet-based platform providing authoritative information and developed through intensive user testing. It addresses preventive measures in the context of coastal flooding, with a representative sample of individuals living in vulnerable communities in New York, New Jersey, and Connecticut. We compare two interventions, presented singly and in combination, and a control condition. One intervention emphasizes fear, reiterating the risk of flooding that these individuals likely know already (Risk). The second emphasizes efficacy, showing actions that recipients can take, but without mentioning risk (Protective). The control condition provides no additional information. We know that how messages are communicated matters (Tierney et al. 2001). Indeed, Lindell and Perry propose that proactive decision-making comprises five stages: attention, comprehension, acceptance, retention, and action (Lindell and Perry 1992). Hence, we created interventions designed to capture attention and enhance comprehension, hoping to encourage acceptance, retention, and action. We expected stronger effects for recipients who 1) have a personal predisposition to take preventive action about flooding and 2) are exposed to both interventions, thereby pairing fear and efficacy. We sought to make each intervention as effective as possible by providing a high-fidelity decision aid that showed flooding risk for recipients’ own locale and by offering behaviorally realistic protective strategies.

2. Methods

a. Sampling

Respondents were drawn from the GfK KnowledgePanel, which uses address-based random sampling methods to recruit individuals in U.S. households (GfK 2017). Panelists complete web-based surveys in return for a modest incentive to encourage participation, which includes the chance to enter special raffles or special sweepstakes with both cash rewards and other prizes. GfK also offers a web-enabled device (e.g., a tablet) and free Internet service to panelists recruited from households without Internet access. The target population consisted of adults (age 18 or older) residing in coastal areas below the 100-yr flood height in New York, New Jersey, and Connecticut. Between 26 May 2015 and 4 June 2015, 2200 people were invited to participate in a study about coastal flooding. Among them, 1201 completed the study, for a completion rate of 54.6%. E-mail reminders were sent to nonresponders on day 3 and day 9 of the field period.

b. Protection of human subjects

The Institutional Review Board of Carnegie Mellon University approved all procedures. All participants provided informed consent.

c. Risk decision aid

Climate Central, an independent organization studying and reporting on climate change, has developed Surging Seas (riskfinder.climatecentral.org), a decision aid with authoritative forecasts of coastal sea level rise. Figure 1 shows a screenshot of its website. This aid is intended to inform a range of decision-makers, from homeowners deciding whether to buy flood insurance (or move) to city planners developing recovery and resilience programs. Its Risk Finder tool provides information about the likelihood and potential consequences (e.g., exposure of different populations, property) of flooding for specific coastal locations. Its scientific content was developed by a team of subject matter experts. Its user interface was developed in conjunction with behavioral researchers and subjected to iterative testing (Wong-Parodi et al. 2014; Wong-Parodi and Strauss 2014).

Fig. 1.
Fig. 1.

Climate Central’s Surging Seas Risk Finder tool for Connecticut showing the (top left) landing page, (top right) risk zone map, (bottom left) flooding risk, and (bottom right) coastal flood days.

Citation: Weather, Climate, and Society 10, 3; 10.1175/WCAS-D-17-0069.1

d. Development of protective action decision aid

We designed and developed a decision aid for evaluating protective actions, following the protocol of risk communication research (Fischhoff 2013; Morgan et al. 2002; Pidgeon and Fischhoff 2011). That process began with (normative) analysis of the options proposed in natural disaster preparation guides from the Federal Emergency Management Agency (FEMA) and interviews with nine disaster preparedness experts (e.g., from the American Red Cross). It proceeded to (descriptive) research evaluating the intuitive appeal of measures emerging from this analysis, with both interviews with 14 coastal residents affected by Superstorm Sandy (Wong-Parodi et al. 2017) and a survey in which 346 U.S. adults evaluated measures emerging from this analysis, in terms of whether they could do them, how helpful each would be, and how much each would cost to implement (see Tables A1A10 in appendix A). The present study reports the results of a (prescriptive) intervention, whose design included intensive user testing for comprehension and ease of use (Wong-Parodi et al. 2014).

As shown in Fig. 2, the resulting protective action decision aid allows users to select among three levels of preparation, ranging from small steps (quicker, less expensive, less effective for long-term preparation) to larger ones (slower, more expensive, more effective for long-term preparation). Examples of smaller steps are moving vehicles to higher ground, obtaining and stocking a supply kit, and developing an emergency plan. Examples of medium steps include raising wiring, obtaining an alternative source of power (such as a generator), and getting a landline phone. Examples of larger steps include purchasing flood insurance and raising a home on pilings. See http://sealevel.climatecentral.org/flood-preparation for more details and information.

Fig. 2.
Fig. 2.

Climate Central’s Surging Seas protective action decision aid showing the (top left) landing page and (top right) small, (bottom left) medium, and (bottom right) large steps that people can take to reduce their risk.

Citation: Weather, Climate, and Society 10, 3; 10.1175/WCAS-D-17-0069.1

e. Experimental protocol

All participants were reminded that Sandy made landfall in New Jersey on 29 October 2012. To deepen their engagement with the experiment, they were asked to think about the time before and after that day and then to describe their experience in one or two sentences. To characterize their predisposition to prepare, participants were then asked about previous actions that they had taken in response to Sandy, both before landfall or afterward. They were scored 1 if they reported any action; otherwise, they were scored 0. In responses not analyzed here, respondents were also asked about their exposure to Sandy, views on their community’s efforts to meet the future risk, regrets, and perceptions of the efficacy of past responses.

Participants were then randomly assigned to one of four interventions:

  1. Control: Participants went directly to the response measures (described below).
  2. Risk: Participants were instructed to “please take some time to learn about major flooding risk in your area.” They were told that they had 90 seconds to look at the website, and that a bell would ring when it was time for them to return to the survey. “Once you’ve opened the website, go to the map and find your city or town. When you hear the bell, please return to the survey window to complete the survey. You can return to the flood risk webpage after you are done with the survey to learn more if you’d like. Please turn up your volume so that you will be able to hear the bell when it alerts you that your time is up. If you are unable to hear the bell and feel that you have spent 90 seconds reviewing the website, please return to the survey window to complete the survey.” Depending on their state of residence, participants were taken to Climate Central’s Surging Seas Risk Finder tool (riskfinder.climatecentral.org) for New York, New Jersey, or Connecticut.
  3. Protective: Participants were instructed to “please take some time to learn about how to prepare for major flooding in your area.” Their procedure followed that of the Risk group, except that they were taken to a website describing the time, cost, and effort to prepare for future coastal flooding events (sealevel.climatecentral.org/flood-preparation).
  4. Risk+Protective: Participants were taken first to Risk Finder and then to the protective aid. They had 90 seconds to explore each.

Next, participants were asked about their 1) intentions to take future actions, 2) risk estimates, 3) risk tolerance, 4) perceptions of changing risk, and 5) perceived drivers of changing risk. They then answered demographic questions. Finally, participants were asked if they would be willing to share their home addresses so that we could calculate their flood risk; 662 did so (55.1%). See appendix B for more information about study design.

f. Response variables

1) Protective future action intentions

Participants indicated their intention to undertake protective measures by answering the question “Do you think you’ll take new or additional actions in the next three months to prepare your home for flooding?”, coded as 1 = yes and 0 = no. [Research suggests that the more concrete and specific a plan of action, the greater likelihood of follow-through. Therefore, we chose a short window of time, forcing respondents to consider the likelihood of doing something in the near term (see, e.g., Bargh 2007)].

2) Risk estimates

Participants were asked to answer the question “What do you think the chances are of major flooding happening here in the next 30 years?” on a scale from 0% to 100%.

3) Risk tolerance

Participants first answered the question “Does the risk of major flooding make you think about moving?”, coded as 1 = yes, 2 = no, and 3 = I do not know. Those who responded “no” were then asked about their risk tolerance with the question “How large would the chances have to be before you would move?”, on a scale anchored at 0% = no chance and 100% = certainty.

4) Changing risk

Participants were asked the question “How do you think the chances of another Sandy are changing over time?”, with the options 1 = going up, 2 = going down, 3 = staying about the same, and 4 = I do not know.

5) Drivers

Participants answered the question “When thinking about future major flooding risk in your area, which of the following worries you?” They could check as many of the following as they chose: climate change, sea level rise, the intensity of storms, the frequency of storms, coastal development, and building standards. Each response was coded as 1 if checked and 0 if not.

g. Statistical analysis

1) Weighting

Poststratification weights were iteratively constructed from respondents’ design weights (which adjust for factors from GfK’s initial sampling strategy and various forms of nonresponse and noncoverage), using probability estimates based on multiple demographic characteristics, region of residence, and Internet access. These weights then adjust for sample attrition as well as discrepancies between the final obtained sample and U.S. census benchmarks, allowing for population inferences. The final weighted sample closely matches December 2013 U.S. census data by sampling area.

2) Statistical significance

Unless otherwise indicated, all analyses were specified before examining the data.

3) Analytical approach

Statistical analyses were conducted using Stata (version 14; Stata Corp, College Station, Texas). One-way analyses of variance examined the mean height of each participant’s home and exposure to flood water during Sandy (objective flood risk) by condition (Control, Risk, Protective, Risk+Protective). Fisher’s exact test compared previous protective actions, as reported by participants in each condition, to assess the success of our randomization. We conducted a logistic regression with intentions for future protective actions as dependent variables, and previous actions and condition as independent variables, to assess the effect of the interventions. We conducted a linear regression, with risk estimate as the dependent variable and previous actions and condition as independent variables. Finally, we conducted planned analyses, predicting the effect of previous actions and condition on flood tolerance, changing risk, and perceived drivers of that risk. Because the risk information shown varied by state, we controlled for location in our analyses. All statistical tests were two-tailed.

4) Participants

Our participants reported being on average 47.4 years old [standard deviation (SD) = 17.0], with 46.4% male, 63.2% having at least some college or higher, 55.9% having a household income of $50 000 or greater, 43.4% living in a household with three or more people, and 75.2% in households with no children under the age of 18. Most reported living in a building with two or more apartments (47.9%), followed by those living in a single-family detached home (42.8%), single-family attached (8.3%), mobile home (1.0%), or vehicle [boat, van, recreational vehicle (RV), etc.] (0.02%). Participants most often reported owning their home (54.5%), followed by renting (43.6%) and occupying without paying rent (1.9%). On average, participants reported being politically independent (M = 3.2, SD = 1.9,) where 1 = strongly Democratic and 7 = strongly Republican, and ideologically moderate (M = 3.7, SD = 1.5), where 1 = extremely liberal and 7 = extremely conservative. Finally, participants reported being moderately religious (M = 3.5, SD = 1.6), where 1 = very religious and 6 = not at all religious.

3. Results

a. Previous protective action

About 29% (345 out of 1189) of respondents reported having taken actions to prepare for extreme weather. Typical measures taken before Sandy were “moved beds away from windows” or “filled up the car up with gas and went food shopping.” Typical measures taken after Sandy were often stronger ones, such as “had our backyard regraded so that water would not pool and [now] run[s] away from the house” and “got flood insurance through FEMA.” [This individual further noted that “it will not reduce damage in the event of another natural disaster, but will at least cover some of the cost of cleanup/repair/replacement.”]

b. Randomization check

One-way analyses of variance revealed no difference in the flooding risk of participants in the four conditions, in terms of either the mean flood height for their homes, F(3, 658) = 0.84, p = 0.47, or their exposure to flood waters during Sandy, F(3, 658) = 1.64, p = 0.17. A Fisher’s exact test (p = 0.59) found no difference in reported protective actions. These results indicate successful randomization.

c. Protective intentions

A logistic regression predicting reported intention to take protective measures as a function of reported previous actions and condition revealed a significant main effect for previous actions. Pooling across groups, 31.5% of participants who reported having taken actions in response to Sandy reported intending to take future actions, compared with only 12.2% of those who had not, OR = 4.32, z = 2.66, p < 0.01 [odds ratio (OR), z-score (z), p-value (p); Table 1]. In the baseline (control) group, 23.8% of those who reported previous actions expected to take future actions, compared with 6.7% of those who had not (Fig. 3), suggesting the size of the subgroup of “chronic preparers.”

Table 1.

Logistic regressions with the dependent variables (protective future action intentions, perceived changing chances of future flooding, and climate change perceived as a driver of future flood risk) and independent variables (intervention, previous action, and intervention × previous action).

Table 1.
Fig. 3.
Fig. 3.

Predicted probabilities of intention to prepare for future flood events (0 = no intention, 1 = intention to prepare) by previous action (gray = had not previously prepared, white = had previously prepared) and condition (Control, Risk, Protective, and Risk+Protective). Error bars represent the 95% confidence interval.

Citation: Weather, Climate, and Society 10, 3; 10.1175/WCAS-D-17-0069.1

We also found a significant main effect for condition, with those who used the Protective decision aid, alone or with the Risk decision aid, being more likely to report intending to take future actions than those receiving no information (OR = 3.03, 2.66, respectively; Table 1). However, we observed no difference between those using the Risk decision aid and receiving no information (control). We additionally observed a significant interaction between reported previous protective action and condition (Fig. 3). Among the 71% who had not previously prepared, the pattern was the same: reported future intentions were higher for those who used the Protective decision aid, χ2(1) = 5.11, p = 0.02 or both interventions, χ2(1) = 4.54, p = 0.03 (compared to those who received no information); however, the Risk alone intervention had no significant effects compared to those who received no information. Among the 29% who had previously prepared, Risk or Protective interventions increased intentions relative to the control condition; however, engaging with both interventions (Risk+Protective) resulted in lower intentions to take protective action compared to those who saw only the Protective decision aid χ2(1) = 8.22, p < 0.01, or the Risk decision aid χ2(1) = 5.36, p = 0.02. See Table C1 in appendix C for more details.

d. Risk estimates

A linear regression found a significant main effect for condition on risk estimates (Table 2, Fig. 4). Planned post hoc comparisons found that participants who used the Protective decision aid (M = 51.6, SE = 2.37) or the Risk+Protective decision aid (M = 51.2, SE = 2.56) estimated lower chances of major flooding in the next 30 years than did those in the control condition (M = 59.5, SE = 2.61) or those who saw only the Risk decision aid (M = 55.6, SE = 2.05) (with no difference between those two conditions). See Table C2 for more details. Risk estimates were unrelated to participants’ reports of their previous action or any interaction between reported action and intervention. Thus, information about protective actions reduced estimates of the risks that might motivate them, whereas information about the risks had no effect.

Table 2.

Linear regression with the dependent variable of risk perceptions and independent variables of intervention, previous action, and intervention × previous action.

Table 2.
Fig. 4.
Fig. 4.

Perceived future flood risk estimates (%) by condition (Control, Risk, Protective, Risk+Protective). Error bars represent ±1 standard error.

Citation: Weather, Climate, and Society 10, 3; 10.1175/WCAS-D-17-0069.1

e. Additional planned analysis

1) Risk tolerance

Most respondents (937/1201 = 78.0%) said that they had not considered moving because of coastal flooding risk. When asked what risk it would take for them to move, these individuals give a mean probability of 67.8% (SE = 1.56%). Most people provided risk estimates lower than the risk they would tolerate before moving (73.5%). However, some people provided estimates that exceeded (23.8%) or met their tolerance for risk (2.7%). Neither the answer to the yes/no question nor the probability was related to condition or reported previous actions.

2) Changing risk

Pooling over all four groups, 39.5% of respondents thought that the “chances of another Sandy” were going up, 3.6% thought that they were going down, 36.5% thought that they were staying the same, and 20.4% did not know. A logistic regression was performed comparing belief in an increasing trend (going up) to everything else (going down, staying about the same, I do not know) (Table 1). We found no relationship between acknowledgment of the increasing chances of an event and condition or reported previous actions.

3) Drivers

A series of logistic regressions examining different potential drivers of changing chance found that those who had prepared for flooding events are almost 2.5 times more likely to worry about climate change than those who had not prepared (Table 1). No significant relationships were observed for any of the other proposed possible drivers (worry about sea level rise, the intensity of storms, the frequency of storms, coastal development, building standards).

4. Discussion

Motivating people to take steps to mitigate their risk to hazards such as coastal flooding can reduce damage and even deaths. However, it is difficult to motivate people, especially if they are not naturally inclined toward prevention. Coupling fear appeals about flooding risk with plausible options to reduce that risk is a promising approach toward the design of communication interventions. Here we experimentally evaluated whether such messages enhance motivation among those who habitually prepare for flood risk and those who do not, in a representative sample of individuals drawn from communities with recent flooding experience and future vulnerability. We contrast responses to decision aids providing information about risks, protective actions, both, or nothing

Consistent with recent research on individual differences in predisposition to prepare, we found that participants who reported having taken any act in preparation for flood events (29% of the sample) were significantly more likely to indicate intent to take future actions, in all four experimental conditions.

Consistent with the Protective Action Decision Model and related approaches, we found that a decision aid providing information about protective actions or protective actions and risk increased reported intent to take such actions. However, presenting information about risk alone had no such effect, perhaps because all three aids reduced estimates of risks (compared to the control condition). Thus, information about feasible, effective protective actions increased intended actions despite a perception of reduced threat. This pattern was also observed among the (71%) subset of participants who reported no previous action. However, among participants who had taken previous action, providing information about both protective action and risk was less effective than presenting action about either alone. One possible interpretation of this difference (albeit observed post hoc) is that any guidance encourages those who do not usually prepare to report future intentions. However, those who had already acted may have felt that they had done enough, after reviewing the full picture described in the combined aid. This surprising finding warrants further study into both the underlying mechanism that may explain this phenomenon.

Another surprising finding is that seeing the protective display singly or in combination with the risk display lowered estimates of a major flooding event in the next 30 years. One possible reason is that these aids provided the first quantitative estimates of risk that most participants had ever seen, which turned out to be a lower number than they imagined (as provided by the control group), but not to affect their sense of risk materially (consistent with the relative insensitivity to probabilities in the middle range, as postulated by prospect theory; Kahneman and Tversky 2013).

Research on resilience has found that when people feel empowered, they are more willing to treat risks as learning experiences, which enables them to fight through hard times and feeds a virtuous cycle of resilience (Richardson et al. 1990). Moreover, those with higher levels of perceived self-efficacy are also more likely to take such risks (Krueger and Dickson 1994). Natural hazards research has also found that seeing protective actions as effective increases the likelihood of their adoption. By presenting curated lists of protective actions, the Protective and Risk+Protective decision aids focused users on actions chosen to be feasible and effective (Paton and Johnston 2001, 2017; Eiser et al. 2012). If the risk was already over their threshold of action, then that guidance might have been what they needed.

A final interesting finding is the relatively high tolerance for the risk of another Sandy-like flood. Most respondents had not previously considered moving because of flood risk. They reported that the risk would have to be about 68% before they would consider moving in the future. However, these findings are echoed in previous research. Wong-Parodi et al. (2017) found a similarly high level of risk tolerance (~62%) among 224 New Jersey residents surveyed in August 2014, who had been directly affected by Superstorm Sandy (Wong-Parodi et al. 2017). They found that individuals with greater social support also expressed greater tolerance for flood risk. Possibly, respondents in the present study with higher risk tolerance were also those with more of the social connections that tie them to their community and help them bounce back from adversity.

Limitations

Our study has the strengths of a large, representative sample; an experimental design; and systematically developed decision aids. One of its limitations is providing participants just 90 seconds to review their assigned website (or 180 seconds in the Risk+Protective group) without guidance or instruction in how use the tool. Given the complexity of the tools, this may have created a sense of racing, confusion, and incompleteness. However, all participants with the exception of the control group, who received no information, were exposed to this potential bias. Furthermore, since we did see observable differences between the conditions, and compared to the control, we believe that participants were given sufficient time to get a sense of the risk and protective information to inform their responses. Future studies could look at whether deeper engagement with the tools affects responses. Second, by focusing on one particular flood-prone area, our study leaves open the question of what would be observed in other areas, with their own unique histories, climates, geographies, and populations. Third, only 29% of our participants reported having taken previous action. Although the percentage might reflect actual behavior, it does reduce the statistical power of comparisons between those who have and have not prepared.

5. Conclusions

These findings suggest practical pathways for improving public safety in coastal areas vulnerable to flooding. Providing clearly presented, authoritative recommendations for protective action may encourage action even when it reduces the perceived probability of the threat. These results are consistent with prior research, while showing some of the detail needed to translate its general predictions into specific interventions. The insensitivity to risk information observed here might be characteristic of individuals who already have a sense of risk, but lack the guidance needed to create the feeling of self-efficacy needed to translate their disposition into action. Furthermore, coupling this information with information about the risks appears to enhance motivation among those who do not usually take action. The present interventions took advantage of the analytically and behaviorally informed Surging Seas website, as well as intensive pretesting of the protective action website. Those considerable investments are suited to online decision aids that can reach wide audiences, including those who might otherwise not see actions that they could and should take. Nearly 9 out of 10 American adults are currently online (Pew Research Center 2017), with most using the Internet to inform their decisions (Horrigan 2008; Fox 2014). These results suggest the potential for providing people with critical information in systematically developed form, to help them make informed decisions to reduce flood-related damages and deaths.

Acknowledgments

This work was supported by a research grant from the Connecticut Sea Grant Program (R/CSAP-9-CT) and from the Center for Climate and Energy Decision Making (National Science Foundation 09-544 Award 1463492). The authors declare no competing financial interests.

APPENDIX A

Summary Statistics

Summary statistics of the amount of perceived time, cost, and helpfulness of preparation actions individuals can take to reduce their risk are presented in Tables A1A10.

Table A1.

Mean (and SD) responses to agreement that protective measures for the home will take a long time, cost a lot of money, and will be helpful to get or do (1 = completely disagree, 7 = completely agree).

Table A1.
Table A2.

Mean (and SD) responses to agreement that planning for a natural disaster will take a long time, cost a lot of money, and will be helpful to get or do (1 = completely disagree, 7 = completely agree).

Table A2.
Table A3.

Mean (and SD) responses to agreement that a first aid kit for a natural disaster supplies kit will take a long time, cost a lot of money, and will be helpful to get or do (1 = completely disagree, 7 = completely agree).

Table A3.
Table A4.

Mean (and SD) responses to agreement that medical items for a natural disaster supplies kit will take a long time, cost a lot of money, and will be helpful to get or do (1 = completely disagree, 7 = completely agree).

Table A4.
Table A5.

Mean (and SD) responses to agreement that sanitary and hygiene items for a natural disaster supplies kit will take a long time, cost a lot of money, and be helpful to get or do (1 = completely disagree, 7 = completely agree).

Table A5.
Table A6.

Mean (and SD) responses to agreement that equipment and tool items for a natural disaster supplies kit will take a long time, cost a lot of money, and will be helpful to get or do (1 = completely disagree, 7 = completely agree).

Table A6.
Table A7.

Mean (and SD) responses to agreement that food and water items for a natural disaster supplies kit will take a long time, cost a lot of money, and will be helpful to get or do (1 = completely disagree, 7 = completely agree).

Table A7.
Table A8.

Mean (and SD) responses to agreement that bed and clothing items for a natural disaster supplies kit will take a long time, cost a lot of money, and will be helpful to get or do (1 = completely disagree, 7 = completely agree).

Table A8.
Table A9.

Mean (and SD) responses to agreement that document and key items for a natural disaster supplies kit will take a long time, cost a lot of money, and will be helpful to get or do (1 = completely disagree, 7 = completely agree).

Table A9.
Table A10.

Mean (and SD) responses to agreement that items for a natural disaster supplies kit for car will take a long time, cost a lot of money, and will be helpful to get or do (1 = completely disagree, 7 = completely agree).

Table A10.

APPENDIX B

Study Design

A schematic of the overall study design is shown in Fig. B1.

Fig. B1.
Fig. B1.

Schematic of study design.

Citation: Weather, Climate, and Society 10, 3; 10.1175/WCAS-D-17-0069.1

APPENDIX C

Post Hoc Analyses

Post hoc analyses examining differences between conditions are shown in Tables C1 and C2.

Table C1.

Post hoc analyses examining differences between conditions by self-reported previous preparation on intentions to prepare for future risk.

Table C1.
Table C2.

Post hoc analyses examining differences between conditions on risk estimates.

Table C2.

REFERENCES

  • Allaire, M. C., 2016: Using practical and social information to influence flood adaptation behavior. Water Resour. Res., 52, 60786093, https:/doi.org/10.1002/2015WR018258.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bargh, J. A., Ed, 2007: Social Psychology and the Unconscious: The Automaticity of Higher Mental Processes. Psychology Press, 352 pp.

  • Brehm, J. W., 1966: A Theory of Psychological Reactance. Academic Press, 135 pp.

  • Bubeck, P., W. J. W. Botzen, and J. C. J. H. Aerts, 2012: A review of risk perceptions and other factors that influence flood mitigation behavior. Risk Anal., 32, 14811495, https://doi.org/10.1111/j.1539-6924.2011.01783.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Crossett, K., B. Ache, P. Pacheco, and K. Haber, 2013: National coastal population report: Population trends from 1970 to 2020. NOAA, accessed 27 February 2018, https://coast.noaa.gov/digitalcoast/training/population-report.html.

  • de Boer, J., W. J. Wouter Botzen, and T. Terpstra, 2014: Improving flood risk communication by focusing on prevention-focused motivation. Risk Anal., 34, 309322, https://doi.org/10.1111/risa.12091.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Doocy, S., A. Daniels, S. Murray, and T. D. Kirsch, 2013: The human impact of floods: A historical review of events 1980-2009 and systematic literature review. PLoS Curr. Disasters, 1, https://doi.org/10.1371/currents.dis.f4deb457904936b07c09daa98ee8171a.

    • Search Google Scholar
    • Export Citation
  • Eiser, J. R., A. Bostrom, I. Burton, D. M. Johnston, J. McClure, D. Paton, J. van der Pligt, and M. P. White, 2012: Risk interpretation and action: A conceptual framework for responses to natural hazards. Int. J. Disaster Risk Reduct., 1, 516, https://doi.org/10.1016/j.ijdrr.2012.05.002.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fischhoff, B., 2013: The sciences of science communication. Proc. Natl. Acad. Sci. USA, 110 (Suppl. 3), 1403314039, https://doi.org/10.1073/pnas.1213273110.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fox, S., 2014: The social life of health information. Pew Research Center, accessed 21 January 2018, http://www.pewresearch.org/fact-tank/2014/01/15/the-social-life-of-health-information/.

  • GfK, 2017: KnowledgePanel recruitment and sample survey methodologies. GfK, accessed 22 June 2017, https://www.gfk.com/fileadmin/user_upload/dyna_content/US/documents/KnowledgePanel_Methodology.pdf.

  • Grinsted, A., J. C. Moore, and S. Jevrejeva, 2013: Projected Atlantic hurricane surge threat from rising temperatures. Proc. Natl. Acad. Sci. USA, 110, 53695373, https://doi.org/10.1073/pnas.1209980110.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hine, D. W., W. J. Phillips, R. Cooksey, J. P. Reser, P. Nunn, A. D. G. Marks, N. M. Loi, and S. E. Watt, 2016: Preaching to different choirs: How to motivate dismissive, uncommitted, and alarmed audiences to adapt to climate change? Global Environ. Change, 36, 111, https://doi.org/10.1016/j.gloenvcha.2015.11.002.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Holland, G., and C. Bruyère, 2014: Recent intense hurricane response to global climate change. Climate Dyn., 42, 617627, https://doi.org/10.1007/s00382-013-1713-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Horrigan, J. B., 2008: The internet and consumer choice: Online Americans use different search and purchase strategies for different goods. Pew Internet and American Life Project Rep., 42 pp., http://www.pewinternet.org/2008/05/18/the-internet-and-consumer-choice/.

  • IPCC, 2013: Climate Change 2013: The Physical Science Basis. Cambridge University Press, 1535 pp., https://doi.org/10.1017/CBO9781107415324.

    • Crossref
    • Export Citation
  • Janis, I. L., and L. Mann, 1976: Coping with decisional conflict. Amer. Sci., 64, 657667, http://www.jstor.org/stable/27847557.

  • Jonkman, S. N., and J. K. Vrijling, 2008: Loss of life due to floods. J. Flood Risk Manage., 1, 4356, https://doi.org/10.1111/j.1753-318X.2008.00006.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kahneman, D., and A. Tversky, 2013: Prospect theory: An analysis of decision under risk. Handbook of the Fundamentals of Financial Decision Making, World Scientific, 99–127, https://doi.org/10.1142/9789814417358_0006.

    • Crossref
    • Export Citation
  • Kellens, W., T. Terpstra, and P. De Maeyer, 2013: Perception and communication of flood risks: A systematic review of empirical research. Risk Anal., 33, 2449, https://doi.org/10.1111/j.1539-6924.2012.01844.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kopp, R. E., R. M. Horton, C. M. Little, J. X. Mitrovica, M. Oppenheimer, D. J. Rasmussen, B. H. Strauss, and C. Tebaldi, 2014: Probabilistic 21st and 22nd century sea-level projections at a global network of tide-gauge sites. Earth’s Future, 2, 383406, https://doi.org/10.1002/2014EF000239.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kreibich, H., A. H. Thieken, T. Petrow, M. Müller, and B. Merz, 2005: Flood loss reduction of private households due to building precautionary measures–lessons learned from the Elbe flood in August 2002. Nat. Hazards Earth Syst. Sci., 5, 117126, https://doi.org/10.5194/nhess-5-117-2005.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Krueger, N., and P. R. Dickson, 1994: How believing in ourselves increases risk taking: Perceived self-efficacy and opportunity recognition. Decis. Sci., 25, 385400, https://doi.org/10.1111/j.1540-5915.1994.tb00810.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kunreuther, H., 1996: Mitigating disaster losses through insurance. J. Risk Uncertainty, 12, 171187, https://doi.org/10.1007/BF00055792.

  • Leventhal, H., R. Singer, and S. Jones, 1965: Effects of fear and specifity of recommendation upon attitudes and behavior. J. Pers. Soc. Psychol., 2, 2029, https://doi.org/10.1037/h0022089.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lindell, M. K., and R. W. Perry, 1992: Behavioral Foundations of Community Emergency Planning. Taylor and Francis, 630 pp.

  • Lindell, M. K., and R. W. Perry, 2004: Communicating Environmental Risk in Multiethnic Communities. Sage Publishing, 272 pp.

  • Lindell, M. K., and R. W. Perry, 2012: The Protective Action Decision Model: Theoretical modifications and additional evidence. Risk Anal., 32, 616632, https://doi.org/10.1111/j.1539-6924.2011.01647.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Morgan, G. M., B. Fischhoff, A. Bostrom, and C. J. Atman, 2002: Risk Communication: A Mental Models Approach. Cambridge University Press, 351 pp.

    • Crossref
    • Export Citation
  • Paton, D., and D. Johnston, 2001: Disasters and communities: Vulnerability, resilience and preparedness. Disaster Prev. Manage., 10, 270277, https://doi.org/10.1108/EUM0000000005930.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Paton, D., and D. Johnston, 2017: Disaster Resilience: An Integrated Approach. Charles C Thomas, 413 pp.

  • Pew Research Center, 2017: Demographics of internet and home broadband usage in the United States, Pew Research Center, accessed 21 January 2018, http://www.pewinternet.org/fact-sheet/internet-broadband/.

  • Pidgeon, N., and B. Fischhoff, 2011: The role of social and decision sciences in communicating uncertain climate risks. Nat. Climate Change, 1, 3541, https:/doi.org/10.1038/nclimate1080.

    • Search Google Scholar
    • Export Citation
  • Poortinga, W., A. Spence, L. Whitmarsh, S. Capstick, and N. Pidgeon, 2011: Uncertain climate: An investigation into public scepticism about anthropogenic climate change. Global Environ. Change, 21, 10151024, https://doi.org/10.1016/j.gloenvcha.2011.03.001.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Richardson, G. E., B. L. Neiger, S. Jensen, and K. L. Kumpfer, 1990: The Resiliency Model. Health Educ., 21, 3339.

  • Ronan, K. R., K. Crellin, and D. M. Johnston, 2012: Community readiness for a new tsunami warning system: Quasi-experimental and benchmarking evaluation of a school education component. Nat. Hazards, 61, 14111425, https://doi.org/10.1007/s11069-011-0070-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sanquini, A. M., S. M. Thapaliya, and M. M. Wood, 2016: A communications intervention to motivate disaster risk reduction. Disaster Prev. Manage., 25, 345359, https://doi.org/10.1108/DPM-11-2015-0256.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schanze, J., and et al. , 2008: Systematisation, evaluation and context conditions of structural and non-structural measures for flood risk reduction. CRUE Research Rep. I-1, 206 pp., http://www.flood-era.ioer.de/files/FLOOD-ERA_Final_report.pdf.

  • Tierney, K. J., M. K. Lindell, and R. W. Perry, 2001: Facing the Unexpected: Disaster Preparedness and Response in the United States. Joseph Henry Press, 306 pp.

  • Witte, K., and M. Allen, 2000: A meta-analysis of fear appeals: Implications for effective public health campaigns. Health Educ. Behav., 27, 591615, https://doi.org/10.1177/109019810002700506.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wong-Parodi, G., and B. H. Strauss, 2014: Team science for science communication. Proc. Natl. Acad. Sci. USA, 111, 13 65813 663, https://doi.org/10.1073/pnas.1320021111.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wong-Parodi, G., B. Fischhoff, and B. Strauss, 2014: A method to evaluate the usability of interactive climate change impact decision aids. Climatic Change, 126, 485493, https://doi.org/10.1007/s10584-014-1226-9.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wong-Parodi, G., B. Fischhoff, and B. Strauss, 2017: Plans and prospects for coastal flooding in four communities affected by Sandy. Wea. Climate Soc., 9, https://doi.org/10.1175/WCAS-D-16-0042.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
Save