Abstract

This study examines how individuals’ trust levels and patience are affected by a tornado event. Affected and unaffected people were surveyed after a 2013 tornado in Moore, Oklahoma, that resulted in 24 fatalities. Findings suggest that those who self-identified as affected became more trusting in general as well as more trusting of police and fire authorities. Affected homeowners also exhibited less patience than their unaffected counterparts. The evaluation of differences in trust and patience enables us to learn about how underlying propensities to invest (or reinvest) in critical private and public infrastructure may be influenced by extreme events. Disasters alter trust levels and patience of affected residents, and documenting the direction and magnitude of these changes may help agencies involved in the recovery process.

1. Introduction

Public involvement in weather-related disasters can be characterized as taking place over several phases: preparedness, warning, immediate response, and recovery. Our main focus in this article is the final phase. In recovery, people rebuild homes and workplaces, heal from physical trauma, and adapt emotionally. Whereas it is natural that most public recovery efforts focus on the rebuilding and physical health aspects, changes in mental perspectives should also be considered. In this article, we explore how a disaster experience may affect individual attitudes of trust and patience in the context of a highly developed economy.

Trust is an important dimension of disaster recovery because lack of trust may cause people to withdraw or fail to respond to socially provided signals (e.g., tornado warnings). On the other hand, becoming too trusting may cause people to become overly reliant on others to take care of their needs. Similarly, if disaster-affected people become overly patient, they may not work to address their needs in a timely way. Conversely, people who become impatient may make decisions that negatively impact them further into the future. Altered patience may influence choices such as how and when to rebuild private homes and businesses as well as how to rebuild and invest in critical public infrastructure. For example, an impatient household may accept price gouging from contractors in exchange for rapid reconstruction, thereby affecting their future economic resilience. Tracking changes in trust and patience in the wake of weather disaster shocks increases scholarly understanding of recovery and resilience.

Tornadoes are especially relevant for study of recovery and resilience. The long-term trend in tornado deaths seems to be increasing (National Weather Service 2016,1), even with improvements in the geographic specificity of the warnings starting in 2007 (Nagele and Trainor 2012). In contrast the trend for other weather disasters is relatively flat despite increasing incidence. Kahn (2005) posits that weather-related death rates may be mitigated by increased disaster resistance that comes with economic development. The difference for tornado incidence is therefore even more notable. Several researchers shed light on potential reasons for the increase in tornado-related deaths (Ashley et al. 2014; Paul and Stimers 2014; Strader et al. 2017), such as the need to match forecast polygons with regions more familiar to public understanding (e.g., city limits), population expansions, and more dense development in certain tornado-prone areas. Tippett et al. (2016) note an increase in the number of extreme tornado events in recent years, which might also explain the increase in deaths.

This examination contributes to the study of recovery and resilience by considering how trust levels and patience among affected citizens are influenced by the EF5 tornado that struck Moore, Oklahoma, in May 2013, which resulted in billions of dollars of damage and 24 fatalities (National Weather Service 2017). Increased understanding of how trust levels and patience are affected by experiencing a disaster could inform relief agencies regarding strategic posttrauma recovery activities and counseling. The study compares responses from people living nearby who were not affected by the tornado with the affected population. The unaffected population serves as the untreated (baseline) population. As a prelude to the full analysis, results show that individuals affected by the tornado exhibit more trust of 1) other people, 2) police/firefighters, and 3) friends, but less patience for financial payments, relative to unaffected individuals from the same and nearby communities.

2. Literature review

The research focusing on attitudes regarding disaster preparedness and response is expansive. For brevity, this discussion highlights some of the most germane research. A volume published by the National Research Council (Mileti 1999) addresses a wide range of challenges related to natural hazard mitigation. The book’s premise is that growing human and economic impacts of natural disasters result from “shortsighted and narrow conceptions of the human relationship to the natural environment” (p.1). Emphasizing “sustainable hazard mitigation” (p. 17), hazard management is linked with local economic and social resiliency. The impacts of natural disasters include interactions among 1) the physical environment including hazardous events, 2) the social and demographic characteristics of communities, and 3) the physical infrastructure—buildings, roads, bridges, and other components of the built environment such as basements in the case of tornadoes (Paul and Stimers 2014). Taking a holistic approach to local government policies concerning land use planning, warning and communication systems, building codes, and community hazard mitigation networks is key for improving safety.

Research has examined the effectiveness of local government early warning systems (Collins and Kapucu 2008) and how it might be improved (Drost et al. 2016); the importance of social interactions in disaster preparation, recovery, and resiliency (Masten and Obradović 2008); the role of public participation in local decision-making (Pearce 2003); and the relationship between disaster preparedness and prior experiences with disasters (Sattler et al. 2000). Silver and Andrey (2014) find increases in protection-seeking behavior in response to a warning in the days after a tornado.

Research acknowledges human limitations in processing information associated with disaster preparation behavior. For example, Meyer (2006) shows that underpreparation for disasters is partially the result of cognitive biases. Humans have a tendency to focus on short-term feedback, extrapolate directly from the present conditions to the future, and discount ambiguous future rewards too highly relative to short-term costs, all of which lead to underinvestment in hazard preparations.

Research investigating trust and patience in the context of low-probability, high-consequence events is somewhat limited, particularly in the natural disaster literature. It is difficult to draw definitive conclusions from the few relevant studies because of the different contexts and nature of the disasters studied. Working with the public on high-risk, low-frequency events is complicated by the manner in which people form judgements about probabilities. Under the observed availability heuristic, people will assign greater probability to an event occurring if they can remember a similar event that was attached to a special date, such as Independence Day (Marx et al. 2007). Under the recency heuristic, people will assign too much or too little weight to the probability of an event depending on the length of time since they experienced the event, which is a posited explanation for federal underinvestment in flood control (Marx et al. 2007).

An expanding literature addresses how weather patterns and related anomalies affect individuals’ attitudes. For example, studies explore the relationship between weather and belief in climate change (Hamilton and Keim 2009; Li et al. 2011; Marquart-Pyatt et al. 2014; McCright et al. 2014; Borick and Rabe 2014), support for farm subsidies (Lee et al. 2016), and approval of the U.S. president (Potoski et al. 2015). This research demonstrates real, if temporary, weather influences on individual responses, sometimes in unexpected directions. To date, most of the attention has focused on temperature or deviations from normal temperature. Less studied is the influence of weather-related disasters.

The research concerning the role that disaster shocks, such as a tornado, may play in trust and patience attitudes is scarce and somewhat conflicting. In post-tsunami experiments in Thai villages in 2004, Cassar et al. (2017) find that those affected by the disaster tend to be more trusting than those who were not directly affected. Toya and Skidmore (2014) also conclude that societal trust increases following disasters. In particular, they use data on 3486 major disaster events for many countries over the 1985–2004 period to examine changes in generalized trust in the years following a major disaster. The authors suggest that increased generalized trust could occur if disasters provide opportunities for individuals to work together to address collective challenges. In this sense, social capital appreciates as it is used in the recovery process (Ostrom 1999). As one example, consider the important changes that occurred in New Orleans public education following Hurricane Katrina. Prior to the storm, leaders in New Orleans had been in conflict regarding the implementation of needed reforms in the troubled New Orleans public school system. However, in the wake of Katrina, leaders put aside their differences and implemented a plan to replace the failing school system with an entirely new structure.2 The early evidence suggests that educational outcomes have substantially improved. In contrast to the above studies, Fleming et al. (2011) find that those affected by the 2010 Chilean earthquake were less trustworthy relative to those who were not affected and, using trust games in Bangladesh, Ahsan (2014) concludes that a cyclone event had no influence on trust. The contrast in findings regarding the impact of weather disasters on trust suggest that further study is warranted.

With respect to patience, the economics literature focuses on two concerns: 1) generating appropriate valuations of future benefits for use in benefit–cost computations and 2) the debate around whether the U.S. bond rate is the appropriate choice for a social discount rate (Feldstein 1964). Additional work on related ideas stemmed from seminal experiments in child psychology (Mischel et al. 1972), where patience is characterized as delayed gratification. With the exception of Smith (1759) and few others, economists traditionally assume that preferences regarding patience are an unchanging individual characteristic (Stigler and Becker 1977). That is, the choices people make are influenced by the circumstances in which they find themselves, but the underlying level of patience is not affected by such circumstances. Insights from behavioral economics and psychology, however, suggest that individual experiences affect preferences regarding patience (e.g., Loewenstein and Angner 2003).

More recently, the focus has shifted toward factors that determine individual patience. This reorientation stems from policy-maker’s interest in encouraging savings and employer’s interest in predicting uptake of retirement investment matching plans or buyouts (Warner and Pleeter 2001). However, how the broader environment influences an individual’s patience has received little attention despite evidence that an individual’s patience may depend on context (Frederick et al. 2002). For example, Fleming et al. (2011) and Cassar et al. (2017) both conclude that exposure to disasters makes individuals more impatient in a developing country context. Similarly, Loveridge et al. (2010) found that community conditions as well as personal-level characteristics influence individuals’ patience, which vary across public and private goods, but do not consider postdisaster influences. The present study adds to this limited literature by evaluating changes in patience in the wake of disasters in a U.S. context.

To summarize, a substantial amount of behavioral research considers how external forces may influence preferences. There is also a growing literature on disaster impacts and resiliency. However, the research concerning the role that disaster shocks may play in preference formation is limited, often based in developing countries, and sometimes contradictory. The present study advances the literature by considering impacts in a highly advanced economy using a with/without treatment design, and we empirically test whether trust increases and patience declines after a tornado.

3. Research design and data

This investigation examines tornado impacts on trust and patience via survey. The survey combined mail and online methods and was fielded about 3.5 months after the 2013 Moore, Oklahoma, tornado. The analysis compares revealed trust and patience of individuals impacted by the tornado with those who were not. The nature of tornado events precludes comparing individuals before and after the event because the timing and location of tornadoes are not possible to predict with enough forewarning to make a pretreatment survey feasible. However, the unpredictable nature of tornado events does mitigate endogeneity concerns. Specifically, in contrast to other types of disasters, such as flooding, it is unlikely that those impacted by tornadoes are somehow self-selected or have traits that put them at a higher likelihood of being directly affected by the tornado (Fothergill et al. 1999). Accordingly, causal relationships are based on the counterfactual behavior of nonimpacted individuals.

The primary sample was taken from Moore, Oklahoma. A priori, it is possible that the sample could fail to include individuals who were unaffected by the tornado to serve as the counterfactual. Accordingly, a demographically similar nearby community (Yukon) was also included to ensure that the sample had a sufficient number of respondents who were not impacted (the survey asked respondents from both communities to self-report if they were affected by the tornado to control for potential commuter or extended family impacts). Both cities are suburbs of Oklahoma City (OKC): Moore is due south of OKC and Yukon is to the west of OKC and is about 20 miles (as the crow flies) from Moore. We purchased a listing for a random 50% draw of the homes in census tracts in the path of the Moore tornado as well as homes in Yukon from Infogroup (a market research firm). The Infogroup data included names and mailing addresses as well as other household information (3902 from affected tracts in Moore and 7558 from Yukon). It is important to note that the name and address data are from one month before the tornado event. From our Infogroup draw, 1800 households were selected using Statistical Package for the Social Sciences (SPSS) code for purely random selection—1200 from the Moore pool and 600 from the Yukon pool. Letters were sent to the sample households with a request to participate in the survey. Recipients were instructed to log on at the provided web address using the unique participant ID included in the letter. The unique participant ID allowed us to make sure that the individual taking the survey was the same as the letter recipient. To ensure that participants could only take the survey once, a valid and unused participant ID was required to log on to the survey site.

Given that an estimated 1150 homes were destroyed and many were badly damaged, having contact data from before the tornado helped to ensure that households that relocated were not omitted. In addition, the U.S. Postal Service forwards first-class mail for 12 months free of charge, which helps to mitigate migration bias inherent in much of the natural disaster research. Sampling errors can result if sampling is done after a natural disaster (e.g., researchers often survey rural villages in developing countries years after major events) and people often have moved out of the event area. Whereas sampling bias cannot be completely ruled out, the relative ease of mail forwarding reduces the potential for migration bias relative to other similar studies.

An online survey was implemented instead of a paper-based one because the survey included an experiment designed to elicit measurements of patience. Whereas all survey methods have drawbacks, the obvious shortcomings of online surveys are the incomplete nature of Internet access and differences in use across age categories. The suburban nature of our target population mitigates the lack of access concern. However, it is possible that some of the most affected respondents were not able to reply because of massive losses or serious injury. On the positive side, the online survey approach provides instant learning (feedback on outcomes and earnings associated with choices) and promotes higher response rates compared with paper-based surveys. Performing the survey and experiment online also reduces the potential for variation across enumerators and may lead to less suspicion by participants that the task was rigged or that they would never receive payment. To this end, the invitation letter specifically stated that participants would be paid via a gift card within 24 h of completing the survey. Overall, the survey achieved a 19.9% response rate. More details on the survey contact methods are found in the  appendix.

The survey was designed to elicit measures of trust and patience as well as information on self-reported tornado impact measures and demographics. The trust and patience portion of the survey is discussed in the next section. Table 1 lists individual demographic characteristics of the sample pools from Moore and Yukon. The overall similarity of the characteristic of the initial survey respondents relative to area averages suggests that the survey invitation and implementation method—particularly the fact that it was an online survey—produces a representative (nonbiased) sample. Survey respondents were first asked if they had ever been directly affected by a tornado and, if the answer was affirmative, were then asked if they were directly affected by the 2013 Moore, Oklahoma, tornado. Persons affected by the 2013 Moore event were considered to be in the treatment group.

4. Trust and patience elicitation tasks

The degree of trust is measured using traditional survey methods and includes several dimensions: overall societal trust as well as trust of family, neighbors, police and fire authorities, local and national governments, and relief agencies. To elicit information on the level of trust, standardized attitudinal trust survey questions based on those found in the Generalized Social Survey (GSS) were used. Conducted yearly since 1972, the GSS has provided source data on societal trends for many studies and its questions have been used extensively in other trust-related studies (see, e.g., Capra et al. 2008; Gächter et al. 2004; Glaeser et al. 2000).

In contrast to trust survey questions, experimental methods were used to elicit patience. A well-established literature focuses on how to develop patience elicitation mechanisms that deliver realistic and reasonable measures (see, e.g., Thaler 1981; Loewenstein 1988; Andersen et al. 2008). Coller and Williams (1999) find that both the information about market rate and real payments induce more realistic patience measures. The effort to inform participants about current rates, while important, often increases the complexity of the task that the respondent must complete. Instead of using absolute measures, the present study focuses only on relative measures across affected and unaffected respondents. Thus, calculating exact or unbiased measures of patience is unnecessary as long as the imprecision or bias is the same across all respondents. Consequently, the measurement tasks are relatively simple, both in terms of respondent task comprehension and experimenter implementation.

To investigate how experiencing the Moore, Oklahoma, tornado might impact individual patience, the survey measures individual discount rates using a variation of the common experimental choice task first seen in Thaler (1981). This type of elicitation task is widely used and was found to be the best among three methods in “predicting real-world behavior and outcomes” (Hardisty et al. 2013, p. 247). Before beginning the task, respondents were reminded that the choices would be used to determine how much they would be paid (in the form of an instant online gift card) above the base incentive. The respondents were told that the task would involve 15 decisions and that each decision would involve choosing between two payment options labeled option A or option B. It was noted that for this task, most people start by choosing option A at first and then switch to option B as they progress down the list of decisions. Finally, they were told that only one line from the task would be (randomly) chosen for payment, but they would not learn which line was chosen or the outcome until after they had made all decisions.

As shown in Table 2, the patience task presented participants with 15 decisions. Each decision involved choosing between the “now” option (option A), where they would receive a given payment now (guaranteed to be paid within 24 h) or a “later” option (option B) that was always a higher payment of $15.00 later (paid in 30 days). The now option payment starts at $14.50 in decision 1 and monotonically decreases to $7.50 in decision 15. The lower an individual’s discount rate, the more patience that individual exhibits. In the task that was presented, a very patient person would always choose the later option. The fewer now options a person chooses before switching over to the later options, the more patient a person is inferred to be. While measuring patience is subject to issues of framing or other considerations such as sunk costs, our objective here is to compare treatment and control group individuals, so any bias arising from the structure of the instrument should be net out in cross-group comparisons.

5. Results

Given that the trust and patience data are not distributed normally, the evaluation begins with nonparametric comparisons before analyzing the data using regression analysis. Table 3 shows the results of Mann–Whitney–Wilcoxon tests where the preference data are split based on whether the individual was directly impacted by the 2013 Moore tornado or not. With regard to trust, those who were affected by the tornado expressed greater generalized trust as well as trust in local government officials, police and fire departments, and neighbors and friends. Patience is represented as the sum of now choices, where a now choice is choosing the “guaranteed in 24 hours” or option A choice. As such, an individual with a lower number of now choices relative to another individual is said to have a lower discount rate or be more patient. As shown in Table 3, individuals that were directly impacted by the tornado were less patient than those that were not directly impacted, though the difference is not statistically significant. Of course, there may be other reasons why preferences vary, so the evaluation continues with regression analysis.

Given that the dependent variables of interest are not continuous and are truncated at 0 and 15 (for patience), ordered logit regressions that include demographic variables as well as various indicators of the degree of impact are estimated for both the trust and patience indicators.3 Table 4 reports estimation results for generalized trust, trust in police and fire departments, and trust in friends. Only these estimates are reported because other trust differences identified in the nonparametric analysis (Table 3) become statistically insignificant once demographics are controlled for.4 As shown in Table 4, those who were affected by the tornado indicated significantly greater generalized trust, trust in police and fire authorities, and trust in friends. Several of the control variables were also significantly related with levels of trust. Specifically, older, more highly educated, and higher-income respondents indicated greater levels of trust. However, marital status, gender, voting status, and home ownership were not significant factors in the trust estimations while education and income provided mixed results depending on the type of trust.

Table 5 reports the estimation results for patience. Two specifications are reported. The second includes an interaction term accounting for individuals who owned a house and were affected by the tornado. Without this interaction term, specification 1 finds no significant relationship between being affected by the tornado and the number of now choices selected. The only demographic variable that has a significant coefficient is HouseOwn, indicating that owning a home is associated with more patient behavior. To investigate whether owning a residence and being impacted alters preferences in a specific way, the interaction term mentioned above is included in the regression analysis. Using this specification, it is clear that owning a home is strongly associated with more patient behavior as long as one has not been impacted. However, if an individual has been impacted and is a homeowner, she tends to exhibit the opposite behavior, that is, more impatience (see coefficient of Affected_HouseOwn). The following possible explanation of this finding is offered. If you are a homeowner who lost your house or have to make repairs, you are likely in crisis mode and under immediate financial pressure: insurance may not cover the full costs of repairs and renting while repairs are underway. Alternatively, if one’s business or workplace was affected, one still has mortgage payments and less ability to move to a cheaper house in the short run than a renter. In this circumstance, one might prefer resources now relative to later, leading to less patience. However, those nonhomeowners who were impacted by the tornado can more easily move and relatively quickly return to a semblance of normalcy. In fact, a renter might exhibit more patience because needs have been met, resulting in the weakly significant negative coefficient on “affected” when the affected homeowner control is introduced.

6. Concluding comments

In this study, survey and experimental methods are used to examine the degree to which a significant tornado event may have altered levels of trust and patience. Affected homeowners exhibited less patience than their unaffected counterparts, but affected nonhomeowners were more patient. Findings also suggest that those who were affected became more trusting in general as well as more trusting of police and fire authorities. Overall, these findings are generally consistent with previous research on other types of disasters, though this study reveals more nuanced behavioral changes in the wake of the tornado.

The present work adds to both the research on weather disaster impacts and behavioral economics. Specifically, it offers additional evidence that underlying preferences are influenced by environmental shocks. Importantly, this study provides useful information that may inform the process of private and public rebuilding and recovery. For example, the analysis shows that homeowners affected by the tornado tend to discount the future in favor of benefits now. Future studies may be useful in determining whether this impatience carries forward into major reinvestment and other life decisions, potentially improving long-run welfare and safety. On the other hand, increased trust in local authorities may compel those affected to support increased public funding to improve safety. Suggested policies include increased public education about how to respond to tornado warnings (Drost 2013) or targeting populations who tend to lack shelter-seeking plans (Senkbeil et al. 2012). The finding that trust in government increases is supported by the city of Moore’s 2014 decision to adopt more wind-resistant building codes. In Moore’s case, cost–benefit analysis of the safety investments concludes that they improve welfare (Simmons et al. 2015), but there may be occasions where transitory changes in preferences may result in the overallocation of public funds to disaster safety. In the context of trust, as noted by Toya and Skidmore (2014), disasters present opportunity for both conflict and cooperation; the finding regarding an increase in trust in the wake of the tornado suggests a potential offsetting benefit of increased social cohesion in addition to all the negative consequences from disasters.

Overall, this research offers new insights on how natural disaster shocks can alter preferences and perceptions. Awareness of these effects may help authorities identify appropriate interventions to help those impacted by the disaster events. More generally, whereas the present study focuses on the behavior of individuals experiencing a suburban lifestyle prior to an unpredictable disaster shock, the methods used in this paper could be applied to other situations, for example, military personnel suffering from PTSD or domestic violence victims, to determine the extent to which the results carry across to other highly traumatic situations.

Acknowledgments

Financial support from the National Science Foundation (Grant 1347968) is gratefully acknowledged. The National Science Foundation exerted no influence on the study design or implementation.

APPENDIX

Survey Mailing Details

Following standard mail-solicited survey protocol, a series of reminder letters and postcards were sent out to those that had not yet responded. Specifically, with approximately one week between mailings the initial invitation letter, a color reminder postcard, a reminder letter, and final color postcard were sent out. All of the mailings contained the website, unique participant ID, and simple instructions. In addition, all the mailings were hand stamped to make sure that they were regarded as first-class mail. It should be noted that the first 900 invitation letters did not contain a monetary incentive. In an effort to increase response rates, the second 900 included a $2 incentive. As shown in Table A1, this approach worked and the response rate increased from 17% to just over 23%. For both approaches, 17.8% of invitation mailings were returned as undeliverable, yielding 1479 total number of contacts.5 Of those contacted, 295 individuals took the survey, giving a 19.9% response rate, which is similar to that experienced in a survey of New York City residents after the 9/11 attack (Wagner et al. 2005).

REFERENCES

REFERENCES
Ahsan
,
D.
,
2014
:
Does natural disaster influence people’s risk preference and trust? An experiment from cyclone prone coast of Bangladesh
.
Int. J. Risk Reduct.
,
9
,
48
57
, doi:.
Andersen
,
S.
,
G. W.
Harrison
,
M. I.
Lau
, and
E. E.
Rutström
,
2008
:
Eliciting risk and time preferences
.
Econometrica
,
76
,
583
618
, doi:.
Ashley
,
W. S.
,
S.
Strader
,
T.
Rosencrants
, and
A. J.
Kremenec
,
2014
:
Spatiotemporal changes in tornado hazard exposure: The case of the expanding bull’s-eye effect in Chicago, Illinois
.
Wea. Climate Soc.
,
6
,
175
193
, doi:.
Borick
,
C. P.
, and
B. G.
Rabe
,
2014
:
Weather or not? Examining the impact of meteorological conditions on public opinion regarding global warming
.
Wea. Climate Soc.
,
6
,
413
424
, doi:.
Capra
,
C. M.
,
K.
Lanier
, and
S.
Meer
,
2008
: Attitudinal and behavioral measures of trust: A new comparison. Working paper, Dept. of Economics, Emory University, 48 pp. [Available online at http://www2.hawaii.edu/~culdiff/docs_10-12/HP%20docs_10-12/suren%20docs/Capra2007_TrustSurvey.pdf.]
Cassar
,
A.
,
A.
Healy
, and
C.
von Kessler
,
2017
:
Trust, risk, and time preferences after a natural disaster: Experimental evidence from Thailand
.
World Dev.
,
94
,
90
105
, doi:.
Coller
,
M.
, and
M. B.
Williams
,
1999
:
Eliciting individual discount rates
.
Exp. Econ.
,
2
,
107
127
, doi:.
Collins
,
M. L.
, and
N.
Kapucu
,
2008
:
Early warning systems and disaster preparedness and response in local government
.
Disaster Prev. Manage.
,
17
,
587
600
, doi:.
Drost
,
R.
,
2013
:
Memory and decision making: Determining action when the sirens sound
.
Wea. Climate Soc.
,
5
,
43
54
, doi:.
Drost
,
R.
,
M.
Casteel
,
J.
Libarkin
,
S.
Thomas
, and
M.
Meister
,
2016
:
Severe weather warning communication: Factors impacting audience attention and retention of information during tornado warnings
.
Wea. Climate Soc.
,
8
,
361
372
, doi:.
Feldstein
,
M. S.
,
1964
:
The social time preference discount rate in cost benefit analysis
.
Econ. J.
,
74
,
360
379
, doi:.
Fleming
,
D.
,
A.
Chong
, and
H. D.
Bejarano
,
2011
: Trust and trustworthiness in the aftermath of natural disasters: Experimental evidence from the 2010 Chilean earthquake. Institute for International Economic Policy Working Paper Series, George Washington University, 37 pp. [Available online at https://www2.gwu.edu/~iiep/assets/docs/papers/Chong_IIEPWP2011-15.pdf.]
Fothergill
,
A.
,
E. G.
Maestas
, and
J. D.
Darlington
,
1999
:
Race, ethnicity and disasters in the United States: A review of the literature
.
Disasters
,
23
,
156
173
, doi:.
Frederick
,
S.
,
G.
Loewenstein
, and
T.
O’Donoghue
,
2002
:
Time discounting and time preference: A critical review
.
J. Econ. Lit.
,
40
,
351
401
, doi:.
Gächter
,
S.
,
B.
Herrmann
, and
C.
Thöni
,
2004
:
Trust, voluntary cooperation, and socio-economic background: Survey and experimental evidence
.
J. Econ. Behav. Organ.
,
55
,
505
531
, doi:.
Glaeser
,
E. L.
,
D. I.
Laibson
,
J. A.
Scheinkman
, and
C. L.
Soutter
,
2000
:
Measuring trust
.
Quart. J. Econ.
,
115
,
811
846
, doi:.
Hamilton
,
L. C.
, and
B. D.
Keim
,
2009
:
Regional variation in perceptions about climate change
.
Int. J. Climatol.
,
29
,
2348
2352
, doi:.
Hardisty
,
D. J.
,
K. F.
Thompson
,
D. H.
Krantz
, and
E. U.
Weber
,
2013
:
How to measure time preferences: An experimental comparison of three methods
.
Judgment Decis. Making
,
8
(
3
),
236
249
.
Kahn
,
M.
,
2005
:
The death toll from natural disasters: The role of income, geography, and institutions
.
Rev. Econ. Stat.
,
87
,
271
284
, doi:.
Lee
,
G.-E.
,
S.
Loveridge
, and
J. A.
Winkler
,
2016
:
Does the public care about how climate change might affect agriculture?
Choices
,
31
(
1
),
1
5
.
Li
,
Y.
,
E. J.
Johnson
, and
L.
Zaval
,
2011
:
Local warming daily temperature change influences belief in global warming
.
Psychol. Sci.
,
22
,
454
459
, doi:.
Loewenstein
,
G.
,
1988
:
Frames of mind in intertemporal choice
.
Manage. Sci.
,
34
,
200
214
, doi:.
Loewenstein
,
G.
, and
E.
Angner
,
2003
: Predicting and indulging changing preferences. Time and Decision: Economic and Psychological Perspectives on Intertemporal Choice, G. Loewenstein, D. Read, and R. F. Baumeister, Eds., Russell Sage Foundation, 114–143.
Loveridge
,
S.
,
J.
Bokemeier
,
P.
Kakela
, and
E.
Satriawan
,
2010
:
Are all discount rates equal? A note on time preferences across public and private benefits in Michigan’s Upper Peninsula
.
J. Reg. Sci.
,
50
,
858
871
, doi:.
Marquart-Pyatt
,
S. T.
,
A. M.
McCright
,
T.
Dietz
, and
R. E.
Dunlap
,
2014
:
Politics eclipses climate extremes for climate change perceptions
.
Global Environ. Change
,
29
,
246
257
, doi:.
Marx
,
S. M.
,
E. U.
Weber
,
B. S.
Orlove
,
A.
Leiserowitz
,
D. H.
Krantz
,
C.
Roncoli
, and
J.
Phillips
,
2007
:
Communication and mental processes: Experimental and analytic processing of uncertain climate information
.
Global Environ. Change
,
17
,
47
58
, doi:.
Masten
,
A. S.
, and
J.
Obradović
,
2008
:
Disaster preparation and recovery: Lessons from research on resilience in human development
.
Ecol. Soc.
,
13
,
9
, doi:.
McCright
,
A. M.
,
R. E.
Dunlap
, and
C.
Xiao
,
2014
:
The impacts of temperature anomalies and political orientation on perceived winter warming
.
Nat. Climate Change
,
4
,
1077
1081
, doi:.
Meyer
,
R. J.
,
2006
: Why we under-prepare for hazards. On Risk and Disaster: Lessons from Hurricane Katrina, R. Daniels, D. Kettl, and H. Kunreuther, Eds., University of Pennsylvania Press, 153–174.
Mileti
,
D. S.
,
1999
: Disasters by Design: A Reassessment of Natural Hazards in the United States. Joseph Henry Press, 371 pp.
Mischel
,
W.
,
E. B.
Ebbesen
, and
A.
Raskoff Zeiss
,
1972
:
Cognitive and attentional mechanisms in delay of gratification
.
J. Pers. Soc. Psychol.
,
21
,
204
, doi:.
Nagele
,
D. E.
, and
J. E.
Trainor
,
2012
:
Geographic specificity, tornadoes, and protective action
.
Wea. Climate Soc.
,
4
,
145
155
, doi:.
National Weather Service
,
2016
: 76 years of weather fatalities. Accessed 25 April 2017. [Available online at http://www.nws.noaa.gov/om/hazstats/resources/weather_fatalities.pdf.]
National Weather Service
,
2017
: The tornado outbreak of May 20, 2013. Accessed on 25 April 2017. [Available online at https://www.weather.gov/oun/events-20130520.]
Ostrom
,
E.
,
1999
:
Coping with tragedies of the commons
.
Annu. Rev. Polit. Sci.
,
2
,
493
535
, doi:.
Paul
,
B. K.
, and
M.
Stimers
,
2014
:
Spatial analyses of the 2011 Joplin tornado mortality: Deaths by interpolated damage zones and location of victims
.
Wea. Climate Soc.
,
6
,
161
174
, doi:.
Pearce
,
L.
,
2003
:
Disaster management and community planning, and public participation: How to achieve sustainable hazard mitigation
.
Nat. Hazards
,
28
,
211
228
, doi:.
Potoski
,
M.
,
R.
Urbatsch
, and
C.
Yu
,
2015
:
Temperature biases in public opinion surveys
.
Wea. Climate Soc.
,
7
,
192
196
, doi:.
Sattler
,
D. N.
,
C. F.
Kaiser
, and
J. B.
Hittner
,
2000
:
Disaster preparedness: Relationships among prior experience, personal characteristics, and distress1
.
J. Appl. Soc. Psychol.
,
30
,
1396
1420
, doi:.
Senkbeil
,
J. C.
,
M. S.
Rockman
, and
J. B.
Mason
,
2012
:
Shelter seeking plans of Tuscaloosa residents for a future tornado event
.
Wea. Climate Soc.
,
4
,
159
171
, doi:.
Silver
,
A.
, and
J.
Andrey
,
2014
:
The influence of previous disaster experience and sociodemographics on protective behaviors during two successive tornado events
.
Wea. Climate Soc.
,
6
,
91
103
, doi:.
Simmons
,
K. M.
,
P.
Kovacs
, and
G. A.
Kopp
,
2015
:
Tornado damage mitigation: Benefit–cost analysis of enhanced building codes in Oklahoma
.
Wea. Climate Soc.
,
7
,
169
178
, doi:.
Smith
,
A.
,
1759
: The Theory of Moral Sentiments. A. Millar, 551 pp.
Stigler
,
G.
, and
G.
Becker
,
1977
:
De gustibus non est disputandum
.
Amer. Econ. Rev.
,
67
(
2
),
76
90
.
Strader
,
S. M.
,
W. S.
Ashley
,
T. J.
Pingel
, and
A. J.
Krmenec
,
2017
:
Observed and projected changes in United States tornado exposure
.
Wea. Climate Soc.
,
9
,
109
123
, doi:.
Thaler
,
R. H.
,
1981
:
Some empirical evidence on dynamic inconsistency
.
Econ. Lett.
,
8
,
201
207
, doi:.
Tippett
,
M. K.
,
C.
Lepore
, and
J. E.
Cohen
,
2016
:
More tornadoes in the most extreme U.S. tornado outbreaks
.
Science
,
354
,
1419
1423
, doi:.
Toya
,
H.
, and
M.
Skidmore
,
2014
:
Do natural disasters enhance societal trust?
Kyklos
,
67
,
255
279
, doi:.
Wagner
,
V. L.
,
M. S.
Radigan
,
P. J.
Roohan
,
J. P.
Anarella
, and
F. C.
Gesten
,
2005
:
Asthma in Medicaid managed care enrollees residing in New York City: Results from a post-World Trade Center disaster survey
.
J. Urban Health
,
82
,
76
89
, doi:.
Warner
,
J. T.
, and
S.
Pleeter
,
2001
:
The personal discount rate: Evidence from military downsizing programs
.
Amer. Econ. Rev.
,
91
,
33
53
, doi:.

Footnotes

© 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

1

The document reports the 10 year average number of US deaths per year at 110, while the 30 year average is 70.

2

In a Newsweek article, Recovery School District Superintendent Paul Vallas said “we used Katrina as an opportunity to build—not rebuild, but build—a new school system” (see http://news.yahoo.com/blogs/upshot/orleans-public-schools-stage-impressive.html).

3

We tested whether recruitment method altered the results by using a dummy to control for recruitment type. The estimated coefficient was not significant so we dropped the variable in the interest of parsimony.

4

These additional estimations are available upon request from the authors.

5

The difference in the percentage of mailings returned in Moore versus Yukon was insignificant.