Communicating Hurricane Evacuation to Tourists

Abstract

The article discusses how the role of gender, residence, and past experience with hurricanes affects tourists’ voluntary evacuation in the event of hurricanes. The study examines how the aforesaid variables influence the perceived credibility of the information source and how such perceptions are translated into voluntary evacuation decisions. The findings indicate that the aforesaid variables significantly influence the perception of the credibility of the information source and information search behaviors leading to a voluntary evacuation decision. Implications of the findings are also discussed.

Keywords

tourists crisis hurricanes evacuation communication Florida

Introduction

In major crises, individuals frequently make decisions based on information about risk probability in which warnings emanate from multiple sources, such as media outlets, government officials, or friends and relatives. Scholars have focused on communication as fundamental to crisis responses, with such information being crucial in understanding how people perceive threats and whether or not they engage in protective behavior (Comfort 2006; Rodriquez, Diaz, and Aguirre 2004; Whitehead et al. 2000). Depending on one’s past crisis experience (Burnside, Miller, and Rivera 2007; Burton, Kates, and White 1993), some information sources may be more influential than others, with a perceived lack of credibility of certain sources contributing to an individual’s reluctance to seek information (Sorensen and Sorensen 2007; West and Orr 2007; West 2001).

To date, most research in the area of information seeking during crises has focused on residents (Phillips and Morrow 2007). This is understandable, as most crisis impacts affect local residents. Recent developments, however, also indicate the need to examine other groups that are also affected by crises. One group that is frequently affected by crises is transient individuals, who are considered an at-risk group since they do not reside in nor have ties to the destination (Phillips and Morrow 2007; Pennington-Gray et al. 2013). One notable example of transient individuals is tourists. Unfortunately, while there are numerous studies that examined tourists’ information seeking and their decision making (e.g., Carneiro and Crompton 2010; Choi et al. 2012; Jun and Holland 2012), little attention has been given to tourists’ information seeking in the event of a crisis, especially those who are already in the destination when a crisis occurs (Drabek 2000).

Tourists are an at-risk group during a crisis because of insufficient knowledge to decipher communication messages and are typically in unfamiliar places and lack support systems accessible to them at home (Burby and Wagner 1996; Faulkner 2001; Matyas et al. 2011; World Tourism Organization 1998). Thus, despite the significance of crisis and emergency planning, not enough is known about how tourists see risks and what dictates their willingness to voluntarily evacuate during a major crisis. It is not clear what factors influence perceptions of vulnerability, how various information sources are evaluated, and how interpersonal factors such as gender, past experience with crises, and place of residency affect perceptions about crisis decision making.

In this study, we focused on hurricanes, which are a particular threat to Florida, because of its geographical location, in order to examine the factors that influence tourists’ preventive evacuation behaviors. Almost 50% of U.S. land-falling tropical cyclones in the past 50 years have made landfall in Florida, more than twice that of any other state. Every Florida county has experienced effects of hurricane-force winds (NOAA 2008) ranging from floods, damaging winds, and other storm impacts. Hurricane season, which typically begins on June 1 and ends on November 30, coincides with Florida’s high tourism season. Annually, Florida welcomes more than 83 million tourists and generates approximately $62 billion in tourism/recreation taxable sales (Visit Florida® 2011).

A Visit Florida® survey found that 20% of potential tourists were concerned with returning to the state during 2005’s hurricane season, indicating that if all these individuals stayed away, it would result in a $6.7-billion loss in expenditures (Pain 2006). Consequently, impacts of hurricane risks to tourists are often greater than those to nontourist residents in the event of a hurricane evacuation. Given the variability in information consumption, determining whom tourists listen to and what recommendations affect their perceptions of vulnerability when a hurricane is imminent is essential. To understand the dynamics of tourists’ behaviors in the event of a hurricane evacuation, this study utilized the risk belief model (Figure 1). It is argued that the risk belief model helped understand the process that an individual undertakes when considering preventive behaviors and associated factors that influence the process, in this case hurricane evacuation.

Figure 1.

Theoretical model.

Theoretical Framework

The risk belief model underlying this study was developed based on traditional risk communication models, by incorporating traditional risk communication components with emphasis on tourists’ interpersonal risk factors. The model includes four major aspects of the risk communication and prevention process: risk factors (intrapersonal, past experience, risk beliefs, and contextual factors), risk communication (sender, channel, message, receiver), risk prevention intention (appraisal, decision making), and risk prevention behavior (risk-reducing behaviors).

With all types of risk communication, there is a degree of uncertainty involved in conveying meaningful messages since the receiver may not receive the message and may not decipher messages as they were intended (Dash and Gladwin 2007). The failure to comprehend and decipher risk communication may hinder an individual from determining if the message is pertinent to their situation. Equally, the content, communication medium, and credibility of the person or agency communicating the information are vital factors in risk perception and consequent decision making by the public (Dow and Cutter 2000; Lindell and Perry 2004; Sorensen 2000). Risk communication should recognize that a variety of cognitive and affective processes influence comprehension, trust, perceived credibility, and comprehension (Lyon, Bernhardt, and Eroglu 2009; Peters, Covello, and McCallum 1997). Additionally, individual risk beliefs have consistently been found to influence decision-making phases and to be predictors of future behaviors (Sonmez and Graefe 1998).

Past studies have examined the relationships between risk perception variables such as past experience (Lepp and Gibson 2003), gender (Carr 2001; Kozak, Crotts, and Law 2007), race/ethnicity (Pennington-Gray, Kaplanidou, and Schroeder 2012), age (Floyd and Pennington-Gray 2004), and income (Floyd and Pennington-Gray 2004). In addition, risk communication starts from an original source that transmits the message directly to the ultimate receiver or through intermediaries such as traditional media and nontraditional media sources. Frequently, risk communication is made available to the general public through various formats. Authorities, friends and relatives, or public figures can communicate messages. Each source plays a distinctive role and affects decision making in a different way (Burnside, Miller, and Rivera 2007; Fodness and Murray 1999; Fuchs et al. 2013). In the case of health risk communication, media such as the Internet, brochures, and one-to-one messaging are the most common methods used. Past research has also found that different ethnic groups such as the African American population receive and interpret messages from different sources with differing levels of trust (Pennington-Gray, Kaplanidou, and Schroeder 2012).

Based on the aforesaid framework, this study focuses on the constructs related to “intrapersonal factors” within the risk belief model, which has rarely been examined in the context of tourist hurricane evacuation. More specifically, we extend the literature related to (1) gender, (2) past experience with hurricanes, and (3) place of residency. Our model is used to predict which of these variables affects (1) perceived credibility of information sources, (2) information search behavior regarding evacuation decision, and (3) final evacuation decision. Specifically, this study is framed by four interrelated questions: (1) To what extent do gender, past experience, and place of residence differ with regard to perceived credibility of information sources regarding hurricane evacuation? (2) What is the extent to which gender, past experience, and place of residence in information search behavior differ regarding hurricane evacuation? (3) If there is a difference, what is the extent of the difference in perceived credibility and information search? (4) What role does information search play in predicting tourists’ evacuation decisions?

Literature Review

Tourist Evacuation

While numerous studies on residents’ evacuation in the event of a hurricane have been conducted, empirical studies on tourists’ evacuation behavior is still understudied (Matyas et al. 2011; Villegas et al. 2012). Hurricane evacuations are disruptive, expensive, and often politically sensitive issues (Phillips and Morrow 2007). From officials’ perceptions, managing tourist evacuees is challenging and involves judgment calls. In Monroe County (the Keys), Florida, a tourist evacuation order is typically issued 48 hours before tropical-force winds are expected to approach the area, or 12 hours for mobile home residents. In 2008 alone, tourists in the Florida Keys were asked to evacuate twice, even though the storms were not considered to be Category 1 hurricanes yet, which resulted in false alarms. Tropical storm Fay in August 2008 was projected to hit the Keys as a tropical storm, with a slight chance of growing into a Category 1 hurricane. Tourists were asked to leave two days before its projected strike and the storm later shifted south, leaving the Keys without any serious damage.

The costs of a hurricane evacuation can exceed one million dollars per mile of coastline from direct costs and losses in commerce, tourism, and general productivity. Key West, Florida, lost approximately $1.5 million a day from commerce and tourism when an evacuation order was issued during the 2008 hurricane season (Miami Herald 2009). Additionally, the 2009 surveys of commercial accommodations in the Keys suggested that 30% of tourists evacuate impulsively before an evacuation order is issued, 40% of tourists evacuate in the first 12 hours after the order is issued, while the remaining 30% of tourists evacuate at the same time as the mobile home park residents (Baker et al. 2009). This suggests the need to examine the underlying factors that affect tourists’ evacuation behaviors.

Risk Factors and Evacuation

As indicated in the theoretical framework, interpersonal factors such as demographic background, personality traits, culture, and social interaction influence how individuals appraise risks. Factors such as age, gender, race, and family situations affect how people assess risks (Dreier 2006; Eisenman et al. 2007; Fothergill, Maestas, and Darlington 1999; Roehl and Fesenmeier 1992; Becken 2013). The basic assumption is that everyone is not equal in terms of risk and view risk and threats differently. Older people, females, and minority groups are considered to be more vulnerable to major crises including hurricanes. They often lack support and access to official information that might aid them in assessing threats, which undermines their ability to respond to crisis events appropriately.

Additionally, the proximity to crisis areas also influences risk appraisal. Notably, one’s residence influences information search behaviors, with those who are from the within the destination employing different strategies than those from outside the destination (Major 1998). For instance, international tourists are more likely to utilize social media during crises than domestic tourists (Pennington-Gray, Kaplanidou, and Schroeder 2012); thus, those who are not from the destination would assess risks and threats differently than those who are from the destination.

Johnson and Tversky (1983) argue that an individual’s experience with one risk could be transferred to their responses to other risks. This includes employing the same information strategies as the ones the individual employed in the past. Nonetheless, past studies in hurricane evacuation have inconclusive results regarding the effect of past hurricane experience on information search behaviors, with some studies finding that past experiences positively influence the modes of information seeking that individuals employ in the decision-making process (Johnson and Meischke 1993). People with experience are more likely to obtain information regarding how to get accurate hurricane information compared to those who have never experienced hurricanes (Gladwin and Peacock 1997). Conversely, Whitehead (2003) argued that those with past hurricane experience tend to use that experience to guide their evacuation decision and therefore exhibit lower information search behavior.

Risk Communication

Communicators cannot assume that all members of the general public in an area where a hurricane makes landfall have the requisite prior knowledge to comprehend and process hurricane forecasts, warnings, and watches or evacuation orders at the same speed or in the same manner. Phillips and Morrow (2007) argued that people go through seven steps between the time that a message is issued and action is taken: (1) warning receipt, (2) affirming its credibility, (3) confirmation, (4) determination of relevance, (5) deciding if action is required, (6) deciding if the action is feasible, and (7) deciding specific actions to be taken. This logical progression is interjected, however, if the warning is not completely understood. For the recipient to understand warning messages, they must be relevant to them.

Societal interaction that affects individual hazard responses has focused on the examination of the effects of warnings and other sources of information (Sorensen 2000). Historically, these studies have focused on reviewing various characteristics of warnings such as the type of message, language used, timeliness, and so forth (Wolshon et al. 2005). Recently, however, Dash and Gladwin (2007) argue that a warning by itself has no value since it is considered based on its perceived credibility, interpretation, and an individual’s aversion to risk. Likewise, Lindell, Lu, and Prater (2005) add that mass media would disseminate warnings and other information related to hurricanes continuously during the hurricane season, making warnings and hurricane information available to the general public.

Researchers have defined source credibility as the extent to which a message receiver perceives the message source to be truthful and based on expertise (Spencer et al. 1992). The effects of high source credibility include higher persuasion levels demonstrated by changes in attitude toward the message or the advocated behavior, less counterargument generation, and a higher likelihood to behave in a positive direction (Eagly and Chaiken 1993). In risk communication, source credibility is important for two reasons. Firstly, assessment of the credibility of a source greatly influences how much people agree with the advocated message (i.e., evacuation) (Burnside, Miller, and Rivera 2007). Secondly, different ethnic and cultural groups manifest different risk perceptions (Fothergill, Maestas, and Darlington 1999) based among other variables in their assessment of credibility of official and nonofficial sources, leading to diverse effects on risk perception and behavioral compliance (Baker et al. 2009).

Warnings come from a wide range of sources, such as weather services, media outlets, government officials, and friends or relatives, with variable credibility. Friends and family are often considered as trustworthy sources, while government agencies and media outlets have found that their credibility has declined in the past few years (West and Orr 2007). The public often views government as uncaring and mass media as exhibiting unfairness and bias in their coverage (West 2001). As a result, social interaction helps individuals to better digest available information, making social interaction more critical than only the warning itself. Within the aforesaid scope, individual beliefs about the credibility of risk information sources have been found to affect individual preventive behaviors (Dash and Gladwin 2007; Dow and Cutter 1998; Griffin, Dunwoody, and Neuwirth 1999). Driscoll and Salwen’s (1996) study on hurricane Andrew found that residents ranked television first in terms of expertise, followed by radio, newspapers, and peers. However, when the sources were ranked based on trustworthiness, television ranked first followed by radio, peers, and newspapers. This suggests that for residents, television is regarded as having the highest level of trustworthiness in the event of hurricanes.

In the context of tourists, little exploration has been done in this area. Prideaux, Coghlan, and Falco-Mammone (2007) in their study on Cyclone Larry found that domestic tourists have a higher perception of the credibility of television than international tourists. Past literature has also found that gender influences risk information management (Bateman and Edwards 2002; Enarson, Fothergill, and Peek 2006). West and Orr (2007) found that women from Rhode Island were more likely than men to report that media coverage, a government order, and a recommendation by a friend of relatives would prompt them to evacuate. In the context of social media usage, the effect of gender remains inconclusive, with some researchers (Kaye 2005) finding a significant difference between male and females while others (Palen 2008) found no significant difference in the use of social media in the event of a crisis.

It is critical to note that individual beliefs regarding credibility are not automatically translated into actual information seeking (Griffin, Dunwoody, and Zabala 1998). One possible explanation is that under threats, individuals do not have the luxury to seek a wide range of information sources to guide their decisions. This is commonly known as bounded rationality (Viscusi 1995) whereby psychological constraints bind human decision making, arguing that individuals would maximize the utility of alternatives that are available for them at a given time. Therefore, it is possible that although one information source is perceived to have high credibility, it may not be used due to the lack of availability during a crisis. Additionally, other researchers such as Dillard et al. (1996) contend that the discrepancy is attributed to political, social, and cultural factors.

Methodology

Survey Administration

Data were collected in September 2011, during the Atlantic hurricane season. The decision to conduct the study during the peak of hurricane season was based on the preposition that during the hurricane season, people are more likely to cognitively seek information about hurricanes. An assessable population for this study was tourists visiting the state of Florida, specifically Orlando and Fort Lauderdale Beach, which were chosen due to the high volume of tourists. The survey used an intercept approach because of the nonexistence of a list that contains tourists who were currently visiting Florida that could be used as a sampling frame. The intercepting approach was selected as it allowed mimicking probability-sampling approach if it was done appropriately (Czaja and Blair 2005). Czaja and Blair (2005) further suggest that in order to mimic probability-sampling approach, researchers need to sample every nth eligible individual.

At each site, random tourists were intercepted and asked to complete a questionnaire. A screening question was conducted to identify eligible tourists. To maximize randomness of the participants, only every third eligible tourist was invited to participate in the survey. One adult from each travel party was approached (alternating males and females). All surveys were self-administered and took approximately 15 minutes to complete.

Prior to the actual survey administration, training was conducted for student surveyors. A pilot study in the form of a focus group and qualitative interviews were administered to validate the findings and the design of the survey instrument as well as the time needed to complete the survey to minimize any systematic errors (Dillman, Smyth, and Christian 2009).

Operationalization

Six variables were used for this study: (1) Perceived credibility was measured by asking respondents to rate the level of credibility of 13 information sources regarding hurricane information (family and friends, TV stations, social clubs, local tourism office, radio stations, social network sites, locals, weather channel, local authority, newspaper, National Hurricane Center [NHC], hotel staff, and other tourists) with a 5-point Liker-type scale with 1 = not credible at all to 5 = very credible. (2) Information search was measured by two questions (a) asking respondents where they turn to for information in the event of hurricane evacuation while they are vacationing in the destination, the options were eight unique two-way communication sources (family and friends, locals, other tourists, local tourism office, local authority, social network sites, hotel staff, and social clubs) (b) asking where they seek information in the event of hurricane evacuation while they are vacationing in the destination, the options was five unique one-way communication sources (TV stations, radio stations, newspaper, NHC, and Weather Channel). Each question was measured using a 5-point Likert-type scale with 1 = never to 5 = always. (3) Gender was measured by asking respondents their gender; (4) Past experience was measured by asking respondents whether or not they have experienced hurricane impacts in the past; (5) Place of residence was measured by asking where the respondents reside (in-state, out-of-state, or international); (6) Evacuation decision making was measured by asking respondents about their likelihood to evacuate for each given scenario with 1 = very unlikely to 5 = very likely. There were eight hurricane scenarios based on a combination of projected path (through–offset destination), projected category in the destination (category 1 and category 4), and time to destination (36 hours to 48 hours) used for this study. Each respondent only responded to four random scenarios.

Data Analysis

Data analysis for this study involved two stages. First, descriptive statistics were performed on the variables, which allowed for a closer examination of the nature of the data patterns. Next, in the second stage, a multivariate analysis was performed to answer the aforesaid research questions. For research questions 1 and 2, a multivariate analysis of variance (MANOVA) was performed to examine the difference between the respective dependent variables and their independent variables. MANOVA allows for simultaneous examination of multiple independent variables to multiple dependent variables (Hair et al. 2010). The Tukey honestly significant difference (HSD) test was chosen as the post hoc test to compare each group to other groups individually (in-state, out-state, and international) to determine significant differences between groups. For research question 3, a t-test was used to determine the difference between perceived credibility and information search.

For research question 4, an ordered-probit model (McKelvey and Zavonia 1975) was employed to relate all variables to evacuation decisions. The ordered-probit model recognized the inherent ordering in the outcome variables of interest and allowed for calculation of the probability of each level of outcome as a function of explanatory factors. In the model, a positive parameter indicated that the corresponding factor was associated with a higher likelihood of evacuation, and a negative parameter indicated the opposite effect. The parameters of the model were estimated using the maximum likelihood estimator. Before estimating the model, the data set was converted from a person-based to a scenario-based data set to better reflect decisions based on hypothetical hurricane scenarios. The SPSS 18 statistical package was used to assist in the data analysis.

Results

Respondent Profiles

A total of 632 eligible tourists were approached at all sites. Five hundred and forty-four agreed to participate (response rate = 86%). Out of 544, 533 were deemed usable because of the completeness of the responses and therefore were used for this study. Among the 533 completed surveys, females encompassed 54%, Caucasians 69%, individuals with no experience with hurricane encompassed 52%, with 50% having bachelor degrees, 17% in-state tourists, 49% out-state tourists, and 34% international tourists with a median travel party of 3. The youngest respondent’s age was 20 years old and the oldest was 88 years old, with a media age of 44. More than half of the respondents (54%) indicated that they have been in the destination before. More than half of respondents did not travel with children (64%). Forty-four percent of respondents earned $50,000 to $99,000 annually. The frequency of variables used in the study was presented in Table 1.

Table 1.

Summary of Descriptive Statistics.

Variable	Frequency	Percentage	Mean	Standard Deviation
Gender
Male	247	46.3
Female	286	53.7

Residence
In-state	92	17.3
Out-state	260	48.8
International	181	34.0

Past experience
No	277	52.0
Yes	256	48.0

Source credibility
National Hurricane Center	533		4.57	.89
Weather Channel	533		4.50	.85
Local authority	533		4.37	.95
Television	533		4.20	.92
Radio stations	533		4.10	.92
Newspaper	533		3.71	1.02
Local tourism office	533		3.50	1.25
Hotel staff	533		3.35	2.11
Locals	533		3.13	1.18
Social network sites	533		2.81	1.16
Family and friends	533		2.76	1.14
Social groups	533		2.32	1.03
Other tourists	533		2.11	1.03

Information source
Weather Channel	533		4.51	1.01
National Hurricane Center	533		4.25	1.19
Local authority	533		4.17	1.12
Radio stations	533		3.78	1.25
Hotel staff	533		3.56	1.33
Local tourism office	533		3.15	1.44
Newspaper	533		3.10	1.30
Locals	533		2.93	1.17
Television	533		2.87	1.35
Family and friends	533		2.87	1.35
Social network sites	533		2.62	1.35
Social groups	533		2.05	1.13
Other tourists	533		2.04	1.10

Evacuation likelihood(533 × 4 scenarios)	2132		3.77	1.26

Perceived Credibility Differences

The results of a MANOVA analysis (Table 2) indicated significant effects of gender (Wilks’s lambda = .954, p = .028), place of residence (Wilks’ lambda = .915, p = .01), and past hurricane experience (Wilks’ lambda= .913, p = .001) on perceived credibility. No significant interaction effects were found. The subsequent test between subjects (Table 3) indicated that while both genders have a similar order of what were considered as credible sources, females perceived the following sources to have higher credibility than did males: family and friends (p = .034), TV (p = .031), local tourism office (p = .032), Weather Channel (p = .005), newspaper (p = .003), and hotel staff (p = .033). It is noteworthy that both genders viewed Weather Channel as the most credible source. The small standard errors scores indicated that the sample mean was close to the population mean; thus, the sample was considered representative (Kay, 2008).

Table 2.

Summary of MANOVA Analysis of Gender, Residence, and Past Experience on Perceived Credibility.

Effect	Wilks’s Lambda	DF	F Statistic	Significance (p)
Gender	.954	13	1.901	.028*
Residence	.915	26	1.781	.010*
Past experience	.913	13	3.724	.001*

Note: Only significant effects are listed. MANOVA = multivariate analysis of variance; DF = degree of freedom.

Significant at .05.

Table 3.

Test of Between-Subjects Effects of Gender on Perceived Credibility.

	Mean Score (SE)				Significance (p)
Dependent Variables	Male	Female	DF	F Statistic	Significance (p)
Family and friends	2.675 (0.084)	2.922 (0.080)	1	4.516	.034*
Television	4.091 (0.068)	4.293 (0.065)	1	4.658	.031*
Local tourism office	3.279 (0.091)	3.548 (0.086)	1	4.606	.032*
Weather Channel	4.366 (0.064)	4.612 (0.061)	1	7.823	.005*
Newspaper	3.585 (0.075)	3.897 (0.072)	1	8.970	.003*
Hotel staff	3.161 (0.082)	3.404 (0.078)	1	4.589	.033*

Note: Only significant effects are listed. SE = standard error; DF = degree of freedom.

Significant at .05.

The Tukey HSD test indicated that place of residence significantly affects use of local tourism offices (p = .03) and newspaper (p = .01). Closer examination revealed that there was a significant difference between tourists who were not from Florida (out-of-state tourists and international) and those who were from Florida, with both out-of-state tourists and international yielding a higher perception of credibility of the local tourism office. Likewise, a significant difference also existed between those who were from the United States (both in-state and out-state tourists) and international tourists, with international tourists indicating higher perceptions of credibility of newspapers (Table 4). Likewise, past experience with hurricanes affected family and friends (p = .027), local tourism office (p = .001), and hotel staff (p = .001), and those with past experience with hurricane perceived higher credibility on family and friends while those without experiences perceived higher credibility on local tourism office and hotel staff (Table 5).

Table 4.

Test of Between-Subjects Effects of Residence on Perceived Credibility.

	Mean Score (SE)					Significant (p)
Dependent Variables	In-State	Out-State	International	DF	F Statistic	Significant (p)
Local tourism office	3.143 (0.142)	3.504 (0.077)	3.594 (0.096)	2	3.578	.029*
Newspaper	3.775 (0.118)	3.576 (0.064)	3.872 (0.080)	2	4.389	.013*

Only significant effects are listed. SE = standard error; DF = degree of freedom.

Significant at .05.

Table 5.

Test of Between-Subjects Effects of Past Experience on Perceived Credibility.

	Mean Score (SE)		DF	F Statistic	Significance (p)
Dependent Variables	Without Experience	With Experience	DF	F Statistic	Significance (p)
Family and friends	2.670 (0.088)	2.928 (0.076)	1	4.899	.027*
Local tourism office	3.639 (0.095)	3.189 (0.082)	1	12.881	.001*
Hotel staff	3.546 (0.086)	3.019 (0.074)	1	21.639	.001*

Note: Only significant effects are listed. SE = standard error; DF = degree of freedom.

Significant at .05.

Information Search Differences

For research question 2, a MANOVA was conducted for information search. Table 6 indicated there were significant effects of gender (Wilks’s lambda = .944, p = .005), place of residence (Wilks’ lambda = .916, p = .012), and past experience (Wilks’s lambda = .893, p = .001). No interaction was found to be significant. Table 7 indicated that females utilized family (p = .003), locals (p = .016), local tourism office (p = .004), local authority (p = .01), hotel staff (p = .001), and NHC (p = .022) more than males. The small standard errors scores also indicated that the sample means was close to the population mean, thus the sample was also considered representative (Kay 2008).

Table 6.

Summary of MANOVA Analysis of Gender, Residence, and Past Experience on Information Search.

Effect	Wilks’s Lambda	DF	F Statistic	Significance (p)
Gender	.944	13	2.342	.005*
Residence	.916	26	1.750	.012*
Past experience	.893	13	4.684	.001*

Note: Only significant effects are listed. MANOVA = multivariate analysis of variance; DF = degree of freedom.

Significant at .05.

Table 7.

Test of Between-Subjects Effects of Gender on Information Search.

	Mean Score (SE)		DF	F Statistic	Significance (p)
Dependent Variables	Male	Female	DF	F Statistic	Significance (p)
Family and friends	2.645 (0.101)	3.060 (0.096)	1	8.914	.003*
Locals	2.744 (0.087)	3.035 (0.083)	1	5.849	.016*
Local tourism office	2.827 (0.103)	3.233 (0.098)	1	8.177	.004*
Local authority	3.968 (0.082)	4.260 (0.078)	1	6.640	.010*
Hotel staff	3.161 (0.095)	3.676 (0.090)	1	15.530	.001*
National Hurricane Center	4.081 (0.088)	4.360 (0.084)	1	5.312	.022*

Note: Only significant effects are listed. SE = standard error; DF = degree of freedom.

Significant at .05.

The Tukey HSD test indicated that international tourists utilized local tourism offices more than tourists from the United States (both in-state and out-state tourists). There was no significant difference between in-state and out-state tourists regarding the use of local tourism office. Likewise, a significant difference existed between those from the United States (both in-state and out-state tourists) and international tourists regarding the use of local authorities as an information source, with international tourists utilizing local authorities more than those from the United States. Regarding the use of social network as an information source, a significant difference existed between tourists from outside Florida (out-state and international tourists) and in-state tourists, with both out-state and international tourists using social networks as an information source of hurricane evacuation more than in-state tourists. A significant difference existed among three tourists groups with regard to the use of hotel staff as an information source of hurricane evacuation information, with international tourists demonstrating a greater willingness to approach hotel staff for information as compared to other groups. Finally, the test also revealed a significant difference between tourists from the United States and international tourists, with international tourists using newspapers more than both in-state and out-state tourists (Table 8). Additionally, as indicated in Table 9, those without experience sought information from other tourists (p = .007), local tourism offices (p = .001), social networking sites (p = .001), and hotel staff (p = .001) more than other groups while those with past hurricane experience used NHC more than those without experiences (p = .002).

Table 8.

Test of Between-Subjects Effects of Residence on Information Search.

	Mean Score (SE)			DF	F Statistic	Significance (p)
Dependent Variables	In-State	Out-State	International	DF	F Statistic	Significance (p)
Local tourism office	2.683 (0.161)	3.154 (0.087)	3.254 (0.109)	2	4.522	.011*
Local authority	3.875 (0.129)	4.178 (0.070)	4.289 (0.087)	2	3.577	.029*
Social networks	2.185 (0.154)	2.624 (0.084)	2.788 (0.105)	2	4.519	.011*
Hotel staff	3.052 (0.148)	3.551 (0.081)	3.651 (0.100)	2	5.904	.003*
Newspaper	2.793 (0.151)	2.993 (0.082)	3.283 (0.102)	2	4.265	.015*

Note: Only significant effects are listed. SE = standard error; DF = degree of freedom.

Significant at .05.

Table 9.

Test of Between-Subjects Effects of Past Experience on Information Search.

	Mean Score (SE)		DF	F Statistic	Significance (p)
Dependent Variables	Without Experience	With Experience	DF	F Statistic	Significance (p)
Other tourists	2.130 (0.084)	1.828 (0.073)	1	7.403	.007*
Local tourism office	3.387 (0.107)	2.673 (0.093)	1	25.291	.001*
Social networks	2.764 (0.103)	2.267 (0.089)	1	13.290	.001*
Hotel staff	3.695 (0.099)	3.141 (0.086)	1	17.915	.001*
National Hurricane Center	4.036 (0.092)	4.405 (0.079)	1	9.272	.002*

Note: Only significant effects are listed. SE = standard error; DF = degree of freedom.

Significant at .05.

Perceived Credibility and Information Search Differences

For research question 3, a paired t-test was conducted between perceived credibility of information sources and information search. The results in Table 10 indicated that a significant discrepancy existed between perceived credibility and their information search among 10 sources, with a significance level of .05. Only family, The Weather Channel, and other tourists were not different. The negative sign in the mean difference column indicated that the mean for a given information source with respect to information search was higher than the mean for the same information source with respect to perceived credibility. Likewise, the positive sign indicated that the mean for a given information source with respect to perceived credibility was higher than the mean for the same information source with respect to information search.

Table 10.

Summary of Paired t-Test for Perceived Credibility and Information Search.

Pair	Perceived Credibility	Information Search	Mean Difference (SE)	DF	t Statistic	Significance (p)
National Hurricane Center	4.57	4.25	0.319 (0.048)	532	6.663	.001*
Weather Channel	4.50	4.51	−0.009 (0.041)	532	−0.227	.821
Television	4.20	4.40	−0.203 (0.045)	532	−4.480	.001*
Radio	4.10	3.78	0.326 (0.053)	532	6.196	.001*
Local authority	4.37	4.17	0.197 (0.047)	532	4.185	.001*
Newspapers	3.71	3.10	0.610 (0.056)	532	10.916	.001*
Local tourism office	3.50	3.15	0.353 (0.059)	532	6.019	.001*
Hotel staff	3.35	3.56	−0.201 (0.052)	532	−3.879	.001*
Locals	3.13	2.92	0.203 (0.057)	532	3.527	.001*
Social networks	2.81	2.62	0.195 (0.052)	532	3.745	.001*
Family and friends	2.76	2.87	−0.111 (0.063)	532	−1.769	.077
Social group	2.32	2.05	0.272 (0.052)	532	5.243	.001*
Other tourists	2.11	2.04	0.073 (0.052)	532	1.408	.160

Significant at .05.

Evacuation Likelihood

For research question 4, the −2 log likelihood of the model at convergence was 4521.737 (χ² = 144.006, df = 17, p = .001) indicating a significant improvement from the baseline model. The model with all independent variables accounted for 23% of the variance in the evacuation likelihood. Table 11 outlines the result of the Ordered-Probit Model.

Table 11.

Results of the Ordered-Probit Model.

Variables	Parameter Estimate	Significance (p)
Information source
Family and friends	0.039	.239
Television	0.083	.165
Social group	0.018	.717
Local tourism office	0.116	.004*
Weather Channel	−0.018	.765
Local authority	0.133	.006*
Newspapers	−0.093	.020*
National Hurricane Center	0.010	.814
Hotel staff	−0.188	.001*
Radio	0.071	.111
Social networks	0.012	.783
Locals	−0.089	.024*
Other tourists	−0.008	.873
Female (ref = male)	0.462	.001*
Without past experience (ref= with)	0.146	.008*
Place of residence (ref = in-state)
International	0.991	.001*
Out-state	0.832	.001*
Thresholds
Evacuate = 1	−1.053	.001
Evacuate = 2	−0.066	.785
Evacuate = 3	1.033	.001
Evacuate = 4	2.061	.001

Note: –2 log Likelihood at convergence (n = 2132) = 4521.737 (χ² = 144.006, df = 17, significance = .001). Pseudo R² Nagelkerke = .23. Ref = reference group.

Significant at .05.

Tourists who utilized locals (β = –.089), hotel staff (β = –.188), and newspapers (β = –.093) were less likely to evacuate than their counterparts, while those who utilized a local tourism office (β = .116) and local authority (β = .133) would be more likely to evacuate. Females (β = .462) were more likely to evacuate than males. International tourists (β = .991) and out-of-state tourists (β = .832) were more likely to evacuate than in-state tourists. Likewise, those without past experience with hurricanes (β = .146) were more likely to evacuate than those with experiences.

Discussion

The study examined the differences as a result of gender, place of residence, and past experience with hurricanes in the perceived credibility of information sources and their information search regarding hurricane evacuation. It also examined how these variables affected their evacuation likelihood. This study found that gender, place of residence, and past experience had a significant influence on one’s perception of source credibility and information search, leading to evacuation decision making in the event of hurricanes.

The first research question found that while both genders demonstrated similar perceptions of the credibility of multiple sources, both genders also exhibited differences in terms of perceptions of credibility of the information sources regarding hurricanes evacuation, with females demonstrating a higher perceived credibility of most information sources that include family and friends, television, local tourism office, weather channel, newspaper, and hotel staff than did males. This is somewhat parallel to previous studies (Rodriquez, Diaz, and Aguirre 2004), which found a significant contrast in perceived credibility of information sources. As predicted, female tourists are more likely to view friends and relatives to be slightly more credible than are male tourists.

The mean score for both genders was less than 3, which indicates that family and friends were not generally viewed to be a credible source of hurricane evacuation information. These results were in contrast with past studies such as Dow and Cutter (1998) and Rodriquez, Diaz, and Aguirre (2004) who found that evacuation orders from friends carried more weight than local authorities. Nonetheless, the finding is parallel with West and Orr (2007) that found that only 29% of their sample viewed family and friends to be a credible source for hurricane evacuation information. The finding is understandable since credibility is parallel with expertise therefore, unlike residents with friends and relatives in the destination, tourists are unlikely to have family and relatives within the travel destination. International tourists from Japan visiting Florida for instance, would be less likely to view their family and friends in Japan to be credible sources of hurricane evacuation, as they are likely unaware of support systems in Florida.

Likewise, while both genders viewed television and the Weather Channel to be highly credible, females deemed these sources more credible than males. This finding is parallel to results indicated by West and Orr (2007). A similar pattern was also found in local tourism offices, newspapers and hotel staff options. The average mean of perceived credibility of these information sources were around 3, indicating an average credibility. In general, the contrast in gender in perceived credibility supported the notion that males and females interpret risks differently (Dow and Cutter 1998).

With regard to the effect of residence on perceived credibility, the results indicated that there were significant differences regarding local tourism offices and newspapers, with the average mean of perceived credibility of these two sources being approximately 3. The scores for international tourists and out-state tourists were significantly higher than in-state tourists. One possible explanation is that local tourism offices tend to cater to international tourists and out-of-state tourists. These two groups are more likely to utilize the local tourism office services (what to do, where to go, and where to eat) when they are vacationing at the destination. Those who are from neighboring areas were found to be less likely to consult local tourism offices in the destinations as they tend to exhibit higher familiarity with the destinations because they are from the same state. This might be interpreted as likelihood of turning to this source rather than the credibility of the source. More research is needed to flesh this finding out.

Interestingly, both in-state and out-state tourists viewed newspapers to be slightly less credible than international tourists. This is an intriguing finding. One constraint of this finding is that we did not ask participants which newspapers they were considering, for example, local newspapers may have different levels of credibility than national newspapers (i.e., USA Today). Another constraint is that we did not differentiate between print and online newspapers. As most online versions of newspaper are updated more frequently than the print version, further study needs to be conducted to fully understand the role of newspapers in the event of a hurricane.

The findings also found differences between those with past experience regarding hurricane impacts and those without experience, especially in their perception of credibility of family and friends, local tourism offices, and hotel staff. Those with past hurricane experience indicated slightly higher levels of credibility than those without experience. Various differences in credibility, however, occurred within types; for example, those with experience indicated higher perceptions of credibility of local tourism offices and hotel staff.

The findings on information search also indicated significance by gender, residence, and past experiences. Female tourists were found to demonstrate a higher propensity of utilizing family and friends, locals, local tourism offices, local authorities, hotel staff, and the National Hurricane Center than were males. This is parallel with a past study (Kaye 2005) that found that men and women interpret risk differently, with females tending to indicate a higher perception of risk. Therefore, they were more likely to find ways to alleviate the uncertainty and seek information from external sources.

Place of residence also affected where tourists seek information. Those who were not from the destination were found to have higher perceptions of risks regarding hurricanes, and therefore seek more information from multiple sources. This explains why among three groups of tourists, international tourists showed a higher likelihood to seek more information in the information source options followed by out-state tourists and in-state tourists. It seems that in-state tourists tend to rely on their own knowledge and therefore exhibited less effort to seek information regarding hurricane evacuation from external sources. In addition, the findings also indicated that international tourists use social networks more than in-state tourists. Therefore, it is important to inform international tourists as to which credible social networks they can follow to receive up-to-date information in a timely manner regarding hurricanes (e.g., Visit Florida’s Twitter page).

The significant discrepancy between perceived credibility and information search illustrated bounded rationality assertion, that despite their perception of credibility, tourists would utilize the available information sources in that time to obtain information regarding hurricane evacuation. Television still ranks first as a primary information source, although the score for perceived credibility is lower than the actual information search score. This suggests that traditional media system is still the most effective way to communicate hurricane evacuation information to tourists in the destination. Further investigation is needed to determine whether or not the rank of credibility changes by type of crises (i.e., health pandemic, terrorism, etc.). Moreover, newspapers had a lower perceived credibility than other information sources, suggesting that they may not be the most effective way to communicate hurricane evacuation to tourists. One of the reasons is that news cycles of newspapers might hinder updates and real-time development of the situation, making television more effective for timely information dissemination. Another example is radio: while radios have been frequently used to communicate hurricane advisories to residents (Whitehead 2003), it might not be as effective for tourists. One possible reason is that tourists are less likely to pay attention to the radio when they are vacationing, even if they rent a car with radio installed or they may not know which stations are local and, with the advent of satellite radio, fewer stations are actually local. Moreover, tourists would still utilize hotel staff to seek information about hurricane evacuation, even though our results suggest that the likelihood is lower than other sources. This reaffirms the need to continually train hotel staff on crisis communication techniques and strategies in order to be able to communicate with different types of tourists (Drabek 2000).

In major storms, some of the greatest loss of life occurred when individuals declined official recommendations to evacuate and thus lives were lost. The most interesting finding is that some of the information sources that were deemed to be credible and were used to gather information regarding hurricane evacuation were not significant factors in terms of the decision to evacuate. Television, The Weather Channel, and the National Hurricane Center, for instance, although they were deemed to be highly credible and frequently used to seek information did not significantly affect the likelihood to evacuate. Instead, evacuation decisions were influenced by information received from local tourism offices, local authorities, newspapers, hotel staff, and locals, with those who received information from local tourism offices and local authorities being more likely to evacuate than those who received information from newspapers, hotel staff, and locals.

This is best explained in the context that even in an era of public cynicism with government, there remains enough credibility in government that when an official evacuation order is issued for tourists, tourists take that seriously and is much more likely to evacuate. This illustrates that timely and well-published evacuation orders are important for tourists’ decision making on whether to stay or to leave. This also reinforces the need to have engagement by the tourism industry so they are communicating timely and important information to tourists.

It is also relevant to discuss social network sites in communicating hurricane evacuation. The use of social network sites has been highlighted in recent studies on risk and crisis communication. During Hurricane Irene (2011), social media played a major role in communicating the status of the hurricane, available shelters, and evacuation routes. Individuals took pictures of the hurricanes and reported their location on Twitter before journalists could get to the scene (Preston 2011). Interestingly however in this study, respondents indicated a lower perception of credibility and lower usage of social media. Nonetheless, recent studies indicate that new media may not have yet reached its full potential with regard to major crises. Nevertheless, social network sites may serve an essential role in posting crisis messages, particularly by “credible sources” such as the Red Cross. Further research needs to examine these sites with regard to credibility and influence on decision making.

The findings also illustrate the crucial role of hotel staff and local tourism offices (DMOs) in influencing evacuation decisions such that those who utilized hotel staff were more likely to remain in the destination and those who sought information from local tourism offices were also more likely to evacuate. Therefore, emergency agencies, DMOs, and the travel industry need to collaborate on messaging and be consistent with what is being communicated. In addition, hotels need to have plans in place to better assist their guests, including shelter designations, refund policies, and evacuation procedures.

As predicted, gender, place of residence, and past experience with hurricane impacts all significantly affect evacuation likelihood. These findings are consistent with other research in hurricane evacuation (Matyas et al. 2011; Villegas et al. 2012). Living outside the destination and being a female are furthermore linked to evacuation decisions. International tourists exhibit a higher likelihood to evacuate than in-state tourists and females. This indicates the vulnerability of these groups in the event of hurricanes.

Conclusion

To summarize, it was found that gender, place of residence, and past experience with hurricane impacts affect perceptions about hurricanes. Gender, place of residence, and past experience with hurricanes affect the credibility of the information source, information search behaviors, and evacuation decisions.

The findings of this study explained the intricacy of understanding human behavior during major crises. Individuals and entities involved in hurricane management and communication cannot adopt a “cookie-cutter” approach where everyone is treated equally in terms of the sources of and types of messages. There are significant differences in how much attention tourists pay to the U.S. Weather Channel, local authorities, mass media, and family and friends. Consequently, unless the emergency management agency and DMO build nuances into their hurricane communication to tourists, they will not be successful with all types of tourists.

As part of implementation, DMOs could conduct training or guidelines to aid hoteliers in a training program for staff, which would provide guidelines for strategies and communications for guests. The goal of training would be to provide information for tourists to make well-informed decisions. In addition, having a booklet about what to do in the event of a hurricane would be beneficial. Hotel and local DMOs may also need to maximize the use of social media. With recent developments in smart phones, wireless Internet technologies will allow them to be utilized en route. In the state of Florida, the Division of Emergency Management website also provides links to hotels in Florida, Alabama, and Georgia to facilitate booking hotel reservations in the event of an evacuation. In addition to such information, the website may also need to provide other information such as local shelter information, alternative evacuation routes, congestion, incident information, and services such as gas stations, rest area locations, etc.

As a further theoretical contribution, this study found a need to revisit the traditional risk communication model by including factors specifically relevant for tourists such as residency. With an increase in global crises, this study also contributes by developing a theoretical model that can be used to understand tourist behaviors during crises, which is currently lacking in the tourism literature. Further application of the theoretical model in this study to other potential crises such as earthquakes or tick-borne disease is warranted to test the robustness of the model.

Principal limitations exist that should be considered in interpreting these results. First, the data were collected one week following hurricane Irene (2011) that made landfall in the eastern seaboard of the United States, which triggered evacuations in several major cities including New York City. During this time, the public was bombarded with information about hurricanes through multiple channels including those in the state of Florida; thus, respondents’ responses might be affected by the information they received. Because of the complexity of data collection during a crisis, future efforts should seek to obtain audience responses at one point in time and as quickly as possible to limit the likelihood of hindsight bias and memory alteration in the responses. Second, because the study was conducted in Florida, further studies need to be replicated in other states that experience hurricanes to help understand the complexity of the situation.

Third, although some of the results can be applied to different types of crisis scenarios, caution should be used when making such decisions. The order of perceived credibility and information search may only apply to the context of hurricanes; other types of crises may yield different priorities because of the nature of the crisis. Further, evacuation during different crisis types may be different depending on the lead time to a crisis or the response time during a crisis. Nonetheless, these data are worthwhile in their outline of information source choices and issues of evacuation. Likewise, differences emerged in information seeking and media use that should prove valuable in message design and placement. Discovering how to best direct and construct messages based on these differences is a worthwhile area of study to provide practitioners with stronger recommendation about crisis responses and message designs. Moreover, future studies should also be conducted during a nonhurricane season to ensure the stability of the model as well as to include other variables that may also influence hurricane evacuation decision making to fully predict the actual preventive behaviors when a hurricane is eminent.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Data used in this study were collected for a project funded by the Holland American Line–Tourism Cares, Southeastern States Chapter of Travel and Tourism Research Association (SETTRA), and Eric Friedheim Tourism Institute at University of Florida.

Author Biographies

Ignatius Cahyanto, PhD is an Assistant Professor at Black Hills State University in South Dakota. His research interest is centered around crisis management for the tourism industry. He is particularly interested in understanding tourists’ behavior in the event of crises.

Lori Pennington-Gray, PhD is the current Director of the Tourism Crisis Management Institute at the University of Florida. Her research primarily focuses on crisis management strategies for the tourism industry.

References

Baker

Shaw

W. D.

Bell

Brody

Riddel

Woodward

E. T.

Neilson

(2009). “Explaining Subjective Risks of Hurricanes and the Role of Risks in Intended Moving and Locating Choice Models.” Natural Hazards Review, 10 (3): 102-12.

Bateman

J. M.

Edwards

(2002). “Gender and Evacuation: Closer Look at Why Women Are More Likely to Evacuate for Hurricanes.” Natural Hazard Review, 3 (3): 107-17.

Becken

(2013). “Measuring the Effect of Weather on Tourism: A Destination- and Activity-Based Analysis.” Journal of Travel Research, 52 (2): 156-67.

Burby

F. J.

Wagner

(1996). “Protecting Tourists from Death and Injury in Coastal Storms.” Disasters, 20 (1): 49-60.

Burnside

Miller

D. S.

Rivera

J. D.

(2007). “The Impact of Information and Risk Perception on the Hurricane Evacuation Decision Making of Greater New Orleans Residents.” Sociological Spectrum, 27 (6): 727-40.

Burton

Kates

White

(1993). The Environment as Hazard, 2nd edition. New York: Guildford.

Carneiro

M. J.

Crompton

J. L.

(2010). “The Influence of Involvement, Familiarity, and Constraints on the Search for Information about Destinations.” Journal of Travel Research, 49 (4): 451-70.

Carr

(2001). “An Exploratory Study of Gendered Differences in Young Tourists Perception of Danger within London.” Tourism Management, 22:565-70.

Choi

Lehto

X. Y.

Morrison

A. M.

Jang

S. C.

(2012). “Structure of Travel Processing and Information Use Patterns.” Journal of Travel Research, 51 (1): 26-40.

10.

Comfort

(2006). “Cities at Risk: Hurricane Katrina and the Drowning of New Orleans.” Urban Affairs Review, 41 (4): 501-16.

11.

Czaja

Blair

(2005). Designing Surveys: A Guide to Decisions and Procedures. Thousand Oaks, CA: Pine Forge Press.

12.

Dash

Gladwin

(2007). “Evacuation Decision Making and Behavioral Responses: Individual and Household.” Natural Hazards Review, 8 (3): 69-77.

13.

Dillard

J. P.

Plotnick

C. A.

Godbold

L. C.

Freimuth

V. S.

Edgar

(1996). “The Multiple Affective Consequences of AIDS PSAs: Fear Appeals Do More Than Scare People.” Communication Research, 2:44-72.

14.

Dillman

Smyth

Christian

L. M.

(2009). Internet, Mail and Mixed-Mode Surveys: The Tailored Design Method. Hoboken, NJ: Wiley.

15.

Dow

Cutter

L. S.

(1998). “‘Crying Wolf’: Repeat Responses to Hurricane Evacuation Orders.” Coastal Management, 26:237-51.

16.

Dow

Cutter

L. S.

(2000). “Public Orders and Personal Opinions: Household Strategies for Hurricane Risk Assessment.” Environmental Hazards, 2 (2): 143-55.

17.

Drabek

T. E.

(2000). “Disaster Evacuation Behavior: Tourist-Business Managers Rarely Act as Customers Expect.” Cornell Hotel and Restaurant Administration Quarterly, August:48-57.

18.

Dreier

(2006). “Katrina and Power in America.” Urban Affairs Review, 41 (4): 528-49.

19.

Driscoll

Salwen

M. B.

(1996). “Riding Out the Storm: Public Evaluations of News Coverage of Hurricane Andrew.” International Journal of Emergencies and Disasters, 14:293-303.

20.

Eagly

A. H.

Chaiken

(1993). The Psychology of Attitudes. San Diego, CA: Harcourt Brace.

21.

Eisenman

D. P.

Cordasco

K. M.

Asch

Golden

Glik

(2007). “Disaster Planning and Risk Communication with Vulnerable Communities: Lessons from Hurricane Katrina.” American Journal of Public Health, 97 (Suppl 1): 109-15.

22.

Enarson

Fothergill

Peek

(2006). “Gender and Disaster: Foundations and Directions.” In Handbook of Disaster Research, edited by Rodriquez

Quarentelli

E. L.

Dynes

R. R.

New York: Springer, pp. 130-46.

23.

Faulkner

(2001). “Towards a Framework for Tourism Disaster Management.” Tourism Management, 22 (2): 135-47.

24.

Floyd

Pennington-Gray

(2004). “Profiling Risk Perceptions of Tourists.” Annals of Tourism Research, 31 (4): 1051-54.

25.

Fodness

Murray

(1999). “A Model of Tourist Information Search Behavior.” Journal of Travel Research, 37:220-30.

26.

Fothergill

Maestas

Darlington

(1999). “Race, Ethnicity and Disasters in the United States.” Disasters, 23 (2): 156-73.

27.

Fuchs

Uriely

Reichel

Maoz

(2013). “Vacationing in a Terror-Stricken Destination: Tourists’ Risk Perceptions and Rationalizations.” Journal of Travel Research, 52 (2): 182-91.

28.

Gladwin

Peacock

W. G.

(1997). “Warning and Evacuation: A Night for Hard Houses.” In Hurricane Andrew: Ethnicity, Gender, and Sociology of Disaster, edited by Peacock

W. G.

Morrow

B. H.

Gladwin

London: Routledge, pp. 52-74.

29.

Griffin

R. J.

Dunwoody

Neuwirth

(1999). “Proposed Model of the Relationship of Risk Information Seeking and Processing to the Development of Preventive Behaviors.” Environmental Research Section A, 80:S230-S245.

30.

Griffin

R. J.

Dunwoody

Zabala

(1998). “Public Reliance on Risk Communication Channels in the Wake of a Cryptosporidium Outbreak.” Risk Analysis, 18:367-75.

31.

Hair

J. F.

Black

W. C.

Babin

B. J.

Anderson

R. E.

(2010). Multivariate Data Analysis, 7th edition. Upper Saddle River, NJ: Prentice Hall.

32.

Johnson

E. J.

Tversky

(1983). “Affect, Generalization, and the Perception Risk.” Journal of Personality and Social Psychology, 45:20-31.

33.

Johnson

J. D.

Meischke

(1993). “A Comprehensive Model of Cancer-Related Information Seeking Applied to Magazines.” Human Communication Research, 19:343-67.

34.

Jun

S. H.

Holland

(2012). “Information-Processing Strategies: A Focus on Pictorial Information Roles.” Journal of Travel Research, 51 (2): 205-18.

35.

Kay

(2008). “The Ongoing Discussion Regarding Standard Deviation and Standard Error.” Advances in Physiology Education, 32 (4): 334.

36.

Kaye

B. K.

(2005). “It’s a Blog, Blog, Blog, Blog World.” Atlantic Journal of Communication, 13 (2): 73-95.

37.

Kozak

Crotts

Law

(2007). “The Impact of the Perception of Risk on International Travelers.” International Journal of Tourism Research, 9 (4): 233-42.

38.

Lepp

Gibson

(2003). “Tourist Roles, Perceived Risk and International Tourism.” Annals of Tourism Research, 30 (3): 606-24.

39.

Lindell

M. K.

Prater

C. S.

(2005). “Household Decision Making and Evacuation in Response to Hurricane Lili.” Natural Hazards Review, 6 (4): 171-79.

40.

Lindell

M. K.

Perry

R. W.

(2004). Communicating Environmental Risk in Multiethnic Communities. Thousand Oaks, CA: Sage.

41.

Lyon

D. K.

Bernhardt

J. M.

Eroglu

(2009). “Social Marketing and Health Communication: From People to Places.” American Journal of Public Health, 99:2120-22.

42.

Major

A. M.

(1998). “The Utility of Situational Theory of Publics for Assessing Public Response to a Disaster Prediction.” Public Relations Review, 24 (4): 489-509.

43.

Matyas

Srinivasan

Cahyanto

Thapa

Pennington-Gray

Villegas

(2011). “Risk Perception and Evacuation Decisions of Florida Tourists under Hurricane Threats: A Stated Preference Analysis.” Natural Hazards, 58: DOI: 10.1007/s11069-011-9801-0.

44.

McKelvey

R. D.

Zavonia

(1975). “A Statistical Model for the Analysis of Ordinal Level Dependent Variables.” Journal of Mathematical Biology, 4:103-20.

45.

Miami Herald. (2009). “Keys May Ease Hurricane Evacuation Policy.” http://www.miamiherald.com (accessed June 3, 2011).

46.

NOAA (National Oceanic and Atmospheric Administration). (2008). “List of all recorded U.S.-landfalling Atlantic storms 1851-present.” http://www.aoml.noaa.gov/hrd/hurdat/ushurrlist18512008.txt (accessed November 11, 2008).

47.

Pain

(2006). Florida Tourism Flying High Despite’04 Hurricanes. Associated Press. http://www.usatoday.com/travel/news/2005-06-14-florida-tourism_x.htm (accessed February 15, 2012).

48.

Palen

(2008). “Online Social Media in Crisis Events.” Education Quarterly, 3:76-78.

49.

Pennington-Gray

Schroeder

Donohoe

Cahyanto

(2013). “Travelers’ Perceptions of Crisis Preparedness Certification in the United States.” Journal of Travel Research. DOI: 10.1177/0047287513496470.

50.

Pennington-Gray

Kaplanidou

Schroeder

(2012). “Drivers of Social Media Use among African Americans in the Event of a Crisis.” Natural Hazards. DOI: 10.1007/s11069-012-0101-0.

51.

Peters

Covello

McCallum

(1997). “The Determinants of Trust and Credibility in Environmental Risk Communication: An Empirical Study.” Risk Analysis, 17 (1): 43-54.

52.

Phillips

Morrow

(2007). “Social Science Research Needs: Focus on Vulnerable Populations, Forecasting, and Warnings.” Natural Hazards Review, 8 (3): 61-68.

53.

Preston

(2011, August 27). “Social Media State-by-State Guide for Hurricane Irene.” The New York Times. http://mediadecoder.blogs.nytimes.com/2011/08/27/social-media-state-by-state-guide-for-hurricane-irene/ (accessed May 28, 2012).

54.

Prideaux

Coghlan

Falco-Mammone

(2007). “Post Crisis Recovery: The Case of after Cyclone Larry.” Journal of Travel and Tourism Marketing, 23 (2/3/4): 163-74.

55.

Rodriquez

Diaz

Aguirre

(2004). “Communicating Risk and Warnings.” University of Delaware Disaster Research Center Preliminary Paper No. 337.

56.

Roehl

W. S.

Fesenmaier

D. R.

(1992). “Risk Perceptions and Pleasure Travel: An Exploratory Analysis.” Journal of Travel Research, 30 (4): 17-26.

57.

Sonmez

S. F.

Graefe

A. R.

(1998). “Determining Future Travel Behavior from Past Travel Experience and Perceptions of Risk and Safety.” Journal of Travel Research, 37:171-77.

58.

Sorensen

J. H.

(2000). “Hazard Warning Systems: Review of 20 Years of Progress.” Natural Hazards Review, 1 (2): 119-25.

59.

Sorensen

J. H.

Sorensen

B. V.

(2007). “Community Process: Warning and Evacuation.” In Handbook of Disaster Research, edited by Rodriquez

Quarentelli

E. L.

Dynes

R. R.

New York: Springer, pp. 182-99.

60.

Spencer

J. W.

Seydliz

Laska

Triche

(1992). “The Difference Influences of Newspaper and Television News Reports of a Natural Hazard on Response Behavior.” Communication Research, 19:299-325.

61.

Villegas

Matyas

Srinivasan

Cahyanto

Thapa

Pennington-Gray

(2012). “Cognitive and Affective Responses of Florida Tourists after Exposure to Hurricane Warning Messages.” Natural Hazards. DOI: 10.1007/S11069-012-0119-3.

62.

Viscusi

(1995). Fatal Tradeoffs: Public and Private Responsibilities for Risk. New York: Oxford University Press.

63.

Visit Florida@ (2011). “Market Research.” http://media.visitflorida.com/about/research/ (accessed April 27, 2012).

64.

West

(2001). The Rise and Fall of the Media Establishment. Boston: Bedford/St. Martin’s.

65.

West

D. M.

Orr

(2007). “Race, Gender, and Communications in Natural Disasters.” Policy Studies Journal, 35 (4): 569-86.

66.

Whitehead

J. C.

(2003). “One Million Dollars a Mile? The Opportunity Costs of Hurricane Evacuation.” Ocean and Coastal Management, 46:1069-89.

67.

Whitehead

J. C.

Edwards

Van Willigen

Maiolo

J. R.

Wilson

Smith

K. T.

(2000). “Heading for Higher Ground: Factors Affecting Real and Hypothetical Hurricane Evacuation Behavior.” Environmental Hazard, 2 (4): 133-42.

68.

Wolshon

Urbina

Wilmot

Levitan

(2005). “Review of Policies and Practices for Hurricane Evacuation.” Natural Hazards Review, 6 (3): 129-42.

69.

World Tourism Organization. (1998). Handbook on Natural Disaster Reduction in Tourist Areas. Madrid: World Tourism Organization.