Citation
Revisiting travel risk after three decades of study

Material Information

Title:
Revisiting travel risk after three decades of study The role of perceived risk, perceived efficacy, and risk reduction behaviors
Creator:
Schroeder, Ashley L
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (297 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Health and Human Performance
Tourism, Recreation, and Sport Management
Committee Chair:
PENNINGTON,LORI
Committee Co-Chair:
DONOHOE,HOLLY M
Committee Members:
KAPLANIDOU,KYRIAKI
THAPA,BRIJESH
KIOUSIS,SPIRO K
Graduation Date:
12/18/2015

Subjects

Subjects / Keywords:
Fear of crime ( jstor )
Hotels ( jstor )
Insurance risks ( jstor )
International travel ( jstor )
Mathematical dependent variables ( jstor )
Multiple regression ( jstor )
Psychology ( jstor )
Self ( jstor )
Tourism ( jstor )
Travel ( jstor )
Tourism, Recreation, and Sport Management -- Dissertations, Academic -- UF
crime -- crisis -- destination -- efficacy -- mexico -- risk -- tourism
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Health and Human Performance thesis, Ph.D.

Notes

Abstract:
Over the past three decades, the tourism literature has examined the role of risk perceptions in an effort to gain a better understanding of the factors that influence travel decision-making in these uncertain times. Taking the major criticisms of the travel risk research into consideration, this study adopted a theoretical lens to the study of travel risk. Through the development of the proposed conceptual model, this study sought to advance knowledge related to travel risk by integrating protection motivation theory and the risk-as-feelings hypothesis with the existing body of knowledge related to travel risk. Notably, a majority of the risk-related variables identified from theory have been understudied in the context of international travel (perceived severity, affective risk perceptions, self-efficacy, response efficacy, and engagement in a recommended risk reduction behavior). Accordingly, the purpose of this study was to reconceptualize and reoperationalize the risk-related constructs studied in the context of international travel, as well as to analyze the relationships between risk-related constructs. Specifically, this study sought to test the proposed conceptual model which considers the relationships between the risk-related constructs of perceived risk (perceived vulnerability, perceived severity, affective risk perceptions), perceived efficacy (self-efficacy, response efficacy), and engagement in a recommended risk reduction behavior. The case of Acapulco, Mexico was used to address the purpose and principal objectives of this study. Based on risk management, this study examined perceived risk in terms of crime-related risk while visiting the destination of Acapulco. Perceived efficacy and engagement in a recommended risk reduction behavior were examined in terms of behaviors that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco. An online survey was conducted in October 2015 to garner responses from American international travelers (n= 1,126). Overall, the findings suggested that self-efficacy and response efficacy were the strongest predictors of tourists intentions to engage in different recommended risk reduction behaviors to ensure their personal safety while visiting the destination. Between the two perceived efficacy variables, self-efficacy had the strongest influence on intentions to engage in a recommended risk reduction behavior. Implications, limitations and suggestions for future research were presented. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2015.
Local:
Adviser: PENNINGTON,LORI.
Local:
Co-adviser: DONOHOE,HOLLY M.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2016-12-31
Statement of Responsibility:
by Ashley L Schroeder.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Embargo Date:
12/31/2016
Classification:
LD1780 2015 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

REVISITING TRAVEL RISK AFTER THREE DECADES OF STUDY : THE ROLE OF PERCEIVED RISK, PERCEIVED EFFICACY, AND RISK REDUCTION BEHAVIORS By ASHLEY L. SCHROEDER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2015

PAGE 2

© 2015 Ashley L. Schroeder

PAGE 3

To destination management organizations around the globe

PAGE 4

ACKNOWLEDGMENTS Words cannot express my gratitude for the endless support of my mentor, Dr. Lori Pennington Gray. Throughout my graduate studies, you have challenged me to be the best that I can be, while providing constant support, encouragement, commitme nt, inspiration, and patience. Thank you for going above and beyond the role of my academic advisor; more importantly, thank you for being my mentor, colleague, and life long friend. I am also forever grateful to you for helping me find my true calling in life. For this, I thank you, as I recognize that we rarely meet someone in life who makes such a profound impact. As I prepare to embark on a new journey, I feel fortunate knowing that your mentorship has prepared me to have a successful career . I am also confident that our journey together will continue for many years to come. I would like to express my sincere gratitude to my committee members Dr. Holly Donohoe, Dr. Kyriaki Kaplanidou, Dr. Spiro Kiousis, and Dr. Brijesh Thapa. Thank you for reviewing my dissertation and providing me with valuable and thought provoking feedback. I would also like to thank faculty members and staff in the Department of Tourism, Recreation and Sport Management for their support and guidance. In addition, I would like to ack nowledge the financial support received from Tourism Cares and the Eric Friedheim Tourism Institute at the University of Florida. Without their financial support, this study would not have been possible. Furthermore, I would like to thank my colleagues Hany Kim, Becky Liu, Semih Yilmaz, and Shintaro Sato. Thank you for your support, encouragement, and camaraderie . Last, but certainly not least, I would like to thank my parents and sister for always encouraging me to follow my dreams. I truly have appreciated your unwavering su pport and reassurance throughout this journey.

PAGE 5

5 TABLE OF CONTENTS p age ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 12 LIST OF ABBR EVIATIONS ................................ ................................ ........................... 13 ABSTRACT ................................ ................................ ................................ ................... 14 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 16 Study Context ................................ ................................ ................................ ......... 16 Increase in Competition among Destinations ................................ ................... 16 Increase in Crises Aff ecting the Tourism Industry ................................ ............ 17 Sustained Tourism Crisis Affected Destination: The Case of Acapulco, Mexico ................................ ................................ ................................ ........... 18 Fundamentals of Destination Crisis Management ................................ .................. 23 Defining Tourism Crisis ................................ ................................ .................... 24 Types of Tourism Crisis ................................ ................................ .................... 27 Tourism Crisis Management Models ................................ ................................ 28 Travel Risk Research ................................ ................................ .............................. 29 Adopting a Theoretical Lens to the Study of Travel Risk ................................ ........ 33 Protection Mot ivation Theory ................................ ................................ ............ 33 Risk As Feelings Hypothesis ................................ ................................ ............ 35 Developing a Theory Based Conceptual Model for the Study of Travel Risk ... 35 Purpose ................................ ................................ ................................ .................. 38 Research Questions ................................ ................................ ............................... 39 Delimitations ................................ ................................ ................................ ........... 40 2 LITERATURE REVIEW ................................ ................................ .......................... 41 Destinat ion Choice Process ................................ ................................ .................... 41 Crime and Tourism ................................ ................................ ................................ . 44 Risk Related Constructs ................................ ................................ ......................... 45 Perceived Risk ................................ ................................ ................................ . 45 Perceived Effi cacy ................................ ................................ ............................ 54 Risk Reduction Behaviors ................................ ................................ ................ 55 Bivariate Relationships between Risk Related Constructs ................................ ..... 60 Perceived Risk and Perceived Efficacy ................................ ............................ 60 Perceived Risk and Risk Reduction Behaviors ................................ ................. 61 Perceived Efficacy and Risk Reduction Behaviors ................................ ........... 63

PAGE 6

6 Demographic Factors and Risk Reduction Behaviors ................................ ...... 65 Past Travel Experience and Risk Reduction Behaviors ................................ ... 66 Hypotheses ................................ ................................ ................................ ............. 66 3 METHODS ................................ ................................ ................................ .............. 69 Research Design ................................ ................................ ................................ .... 69 Data Collection ................................ ................................ ................................ ....... 69 Survey Instr ument ................................ ................................ ................................ ... 70 Pre Test ................................ ................................ ................................ ............ 70 Primary Study ................................ ................................ ................................ ... 72 Operationalization of Constructs ................................ ................................ ............. 75 Independent Variable: Demographic Factors ................................ ................... 76 Independent Variable: International Travel Specific Psychological Factors ..... 78 Independent Variable: Destination Specific Psychological Factors .................. 79 Independent Variable: Destination Specific Factors ................................ ......... 79 Independ ent Variable: Past Travel Experience ................................ ................ 80 Mediating Variable: Perceived Risk ................................ ................................ .. 81 Mediating Variable: Perceived Efficacy ................................ ............................ 83 Dependent Variable: Intentions to Engage in a Recommended Risk Reduction Behavior ................................ ................................ ....................... 86 Data Analysis ................................ ................................ ................................ .......... 87 Research Question 1: Do the Antecedents Predict Perceived Efficacy? .......... 88 Research Question 2: Does Perceived Risk Mediate the Relationship between the Antecedents and Perceived Efficacy? ................................ ...... 88 Research Question 3: Does Perceived Risk Predict Intentions to Engage in a Recommended Risk Reduction Behavior? ................................ ................. 89 Research Question 4: Does Perceived Efficacy Mediate the Relationship between Perceived Risk and Intentions to Engage in a Recommended Risk Reduction Behavior? ................................ ................................ ............. 90 Preparation for Data Analysis ................................ ................................ ................. 91 Missing Data Analysis ................................ ................................ ...................... 92 Outlier Analysis ................................ ................................ ................................ 94 Assessing Normality ................................ ................................ ......................... 96 Independent Variable: Internationa l Travel Specific Psychological Factors ..... 98 Mediating Variable: Perceived Risk ................................ ................................ 100 Descriptive Analysis ................................ ................................ .............................. 103 Indepe ndent Variable: Demographic Factors ................................ ................. 103 Independent Variable: International Travel Specific Psychological Factors ... 105 Independent Variable: Destination Specific Psychological Factors ................ 106 Independent Variable: Destination Specific Factors ................................ ....... 106 Independent Variable: Past Travel Experience ................................ .............. 107 Mediating Variable: Perceived Risk ................................ ................................ 108 Mediating Variable: Perceived Efficacy ................................ .......................... 109 Dependent Variable: Intentions to Engage in a Recommended Risk Reduction Behavior ................................ ................................ ..................... 113 Measurement Model ................................ ................................ ............................. 116

PAGE 7

7 Restatement of Research Questions and Hypotheses ................................ ... 116 4 RESULTS ................................ ................................ ................................ ............. 121 Register with the U.S. Department of State Prior to Travel ................................ ... 121 Research Question 1: Do the Antecedents Predict Perceived Efficacy? ........ 121 Research Question 2: Does Perceived Risk Mediate the Relationship between the Antecedents and Perceived Efficacy? ................................ .... 134 Research Question 3: Does Perceived Risk Predict Intentions to Engage in a Recommended Risk Reduction Behavior? ................................ ............... 155 Research Question 4: Does Perceived Efficacy Media te the Relationship between Perceived Risk and Intentions to Engage in a Recommended Risk Reduction Behavior? ................................ ................................ ........... 158 Purchase Travel Insurance Prior to Travel ................................ ............................ 164 Research Question 1: Do the Antecedents Predict Perceived Efficacy? ........ 164 Research Question 2: Does Perceived Risk Mediate the Relationship between the Antecedents and Perceived Efficacy? ................................ .... 174 Research Question 3: Does Perceived Risk Predict Intentions to Engage in a Recommended Risk Reduction Behavior? ................................ ............... 190 Research Question 4: Does Perceived Efficacy Mediate the Relationship between Perceived Risk and Intentions to Engage in a Recommended Risk Reduction Behavior? ................................ ................................ ........... 193 Avoid Traveling Outside of the Hotel Zone ................................ ........................... 198 Research Question 1: Do the Antecedents Predict Perceived Efficacy? ........ 198 Research Question 2: Doe s Perceived Risk Mediate the Relationship between the Antecedents and Perceived Efficacy? ................................ .... 208 Research Question 3: Does Perceived Risk Predict Intentions to Engage in a Recommended Risk Reduction Behavior? ................................ ............... 227 Research Question 4: Does Perceived Efficacy Media te the Relationship between Perceived Risk and Intentions to Engage in a Recommended Risk Reduction Behavior? ................................ ................................ ........... 230 Summary of Findings by Hypothesis ................................ ................................ .... 239 5 DISCUSSION AND CONCLUSIONS ................................ ................................ .... 242 Discussion of Findings ................................ ................................ .......................... 242 Conclusions and Implications ................................ ................................ ............... 244 Theoretical Implications ................................ ................................ .................. 244 Practical Implications ................................ ................................ ...................... 246 Limi tations and Suggestions for Future Research ................................ ................ 249 APPENDIX A INFORMED CONSENT ................................ ................................ ........................ 252 B SURVEY INSTRUMENT FOR THOSE WITH COUNTRY LEVEL DESTINATION PAST TRAVEL EXPERIENCE ................................ .................... 255

PAGE 8

8 C SURVEY INSTRUMENT FOR THOSE WITH COUNTRY LEVEL DESTINATION PAST TRAVEL EXPERIENCE ................................ .................... 268 D DEVELOPMENT OF DESTINATION CHOICE PROCESS MODEL ..................... 279 LIST OF REFERENCES ................................ ................................ ............................. 284 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 297

PAGE 9

9 LIST OF TABLES Table page 3 1 Operationalization of demographic factors. ................................ ........................ 77 3 2 Operationalization of international travel specific psychological factors. ............ 79 3 3 Operationalization of destination specific psychological factors. ........................ 79 3 4 Op erationalization of destination specific factors. ................................ ............... 80 3 5 Operationalization of past travel experience. ................................ ...................... 81 3 6 Operationalization of perceived risk. ................................ ................................ ... 83 3 7 Operationalization of perceived efficacy. ................................ ............................ 84 3 8 Operationalization of intentions to engage in a recommen ded risk reduction behavior. ................................ ................................ ................................ ............. 87 3 9 Underlying dimensions of international travel risk perceptions. .......................... 99 3 10 Underlying dimensions of perceived risk. ................................ ......................... 102 3 11 Descriptive analysis of demographic factors. ................................ ................... 104 3 12 Descriptive analysis of international travel risk perceptions. ............................. 105 3 13 Descriptive analysis of safety concerns. ................................ ........................... 106 3 14 Descriptive analysis of destination specific psychological factors. ................... 106 3 15 Descriptive analysis of destination specific factors. ................................ .......... 107 3 16 Descriptive analysis of past travel experience. ................................ ................. 108 3 17 Descriptive analysis of perceived risk. ................................ .............................. 108 3 18 Descriptive analysis of self efficacy. ................................ ................................ . 110 3 19 Descriptive analysis of response efficacy. ................................ ........................ 111 3 20 Descriptive analysis of intentions to engage in a recommended risk reduction behavior. ................................ ................................ ................................ ........... 114 4 1 Multiple regression of the relationship between the antecedents and self efficacy associated with regi stering with the U.S. Department of State prior to travel. ................................ ................................ ................................ ................ 125

PAGE 10

10 4 2 Multiple regression of the relationship between the antecedent s and response efficacy associated with registering with the U.S. Department of State prior to travel. ................................ ................................ .......................... 127 4 3 Multiple regressio n of the relationship between the antecedents and perceived vulnerability/affective risk perceptions. ................................ ............. 136 4 4 Multiple regression of the relationship between the antecedents and perceived severity. ................................ ................................ ............................ 137 4 5 Multiple regression of the relationship between perceived risk and self efficacy associated with registering with the U.S. Department of State prior to travel. ................................ ................................ ................................ ................ 140 4 6 Multiple regression of the relationship between perceived risk and response efficacy associated with registering with the U.S. Department of State pri or to travel. ................................ ................................ ................................ ................ 141 4 7 Multiple regression of the relationship between perceived risk and intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel. ................................ ........................... 157 4 8 Multiple regression of the relat ionship between perceived efficacy and intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel. ............................ 161 4 9 Multiple regression of the relationship between the antecedents and self efficacy associated with purchasing travel insurance prior to travel. ................ 166 4 10 Multiple regression of the relationship between the antecedents and response efficacy associated with purchasing travel insurance prior to travel. . 168 4 11 Multiple regression of the relationship between perceived risk and self efficacy associated with purchasing travel insurance prior to travel. ................ 177 4 12 Multiple regression of the relationship between perceived risk and response efficacy associated with purchasing travel in surance prior to travel. ................ 178 4 13 Multiple regression of the relationship between perceived risk and intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel. ................................ ................................ ................... 192 4 14 Multiple regression of the relationsh ip between perceived efficacy and intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel. ................................ ....................... 195 4 15 Multiple regression of the relationship between the antecedents and self efficacy associated with avoiding traveling outside of the hotel zone. .............. 200

PAGE 11

11 4 16 Multiple regression of the relationship between the antecedents and response efficacy associated with avoiding traveling outside of the hotel zone. ................................ ................................ ................................ ................. 202 4 17 Multiple regression of the relationship between perceived risk and self efficacy associated with avoiding traveling outside of the hotel zone. .............. 210 4 18 Multiple regression of the relationship between perceived risk and response efficacy associated with avoiding traveling outside of the hotel zone. .............. 211 4 19 Multiple regression of the relationship between the significant antecedents and perceived severity and self efficacy associated with avoiding traveling o utside of the hotel zone. ................................ ................................ ................. 213 4 20 Multiple regression of the relationship between the significant antecedents and perceived sever ity and response efficacy associated with avoiding traveling outside of the hotel zone. ................................ ................................ ... 214 4 21 Multiple regression of the relati onship between perceived risk and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone. ................................ ................................ ................. 22 9 4 22 Multiple regression of the relationship between perceived efficacy and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone. ................................ ................................ ... 232 4 23 Multiple regression of the relationship between perceived severity and perceived efficacy and intentions to engage in the recommended risk reduction beha vior of avoiding traveling outside of the hotel zone. .................. 234 4 24 Sobel test of the mediation of perceived efficacy in the relationsh ip between perceived severity and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone. .................. 235 4 25 Summary of findings by hypothesis. ................................ ................................ . 239

PAGE 12

12 LIST OF FIGURES Figure page 1 1 Map of Acapulco, Mexico (Google, 2015). ................................ .......................... 20 1 2 Conceptual model to understand the role of risk related constructs in the destination choice process. ................................ ................................ ................ 37 3 1 Measurement model. ................................ ................................ ........................ 117 D 1 Tourist Destination Choice Process, Adapted from LeBlanc (1989), Mansfeld (1992), and Sönmez & Graefe (1998a). ................................ ........................... 282

PAGE 13

13 LIST OF ABBREVIATIONS DMO Destination management organization EKB Engel, Kollat, & Blackwell HIT Human intelligence task KMO Kaiser Meyer Olkin m easure of s ampling a dequacy MTurk Amazon Mechanical Turk NTTO National Travel and Tourism Office PATA Pacific Asia Travel Association PCA Principal components analysis PMT Protection m otivation t heory SPSS Statistical Package for the Social Sciences TCMI Tourism Crisis Management Initiative UN / ISDR United Nations Office for Disaster Risk Reduction UNWTO United Nations World Tourism Organization WTTC World Travel & Tourism Council

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy REVISITING TRAVEL RIS K AFTER THREE DECADES OF STUDY: THE ROLE OF PERCEIVED RISK, PERCEIVED EFFICACY, AND RISK REDUCTION BEHAVIORS By Ashley L. Schroeder December 2015 Chai r: Lori Pennington Gray Major: Health and Human Performance Over the past three decades, the tourism literature has examined the role of risk perceptions in an effort to gain a better understanding of the factors that influence travel decision making in these uncertain times. Taking the major criticism s of the travel risk research into consideration, thi s study adopted a theoretical lens to the study of travel risk. T hrough the development of the proposed conceptual model, this study sought to advance knowledge related to travel risk by integrating protection motivation theory and the risk as feelings hyp othesis with the existing body of knowledge related to travel risk . Notably, a majority of the risk related variables identified from theory have been understudied in the context of international travel (perceived severity, affective risk perceptions, self efficacy, response efficacy, and engagement in a recommended risk reduction behavior ) . Accordingly, t he purpose of this study was to recon ceptualize and reoperationalize the risk related constructs studied in the context of international travel , as well as to analyze the relationships between r isk related constructs. S pecifically, this study sought to test the proposed conceptual model which considers the relationships between the risk related constructs of perceived risk (perceived vulnerability, perceiv ed

PAGE 15

15 severity, affective risk perceptions), perceived efficacy (self efficacy, response efficacy), and engagement in a recommended risk reduction behavior. The case of Acapulco, Mexico was used to address the purpose and principal objectives of this study. Based on risk management , this study examined perceived risk in terms of crime related risk while visiting the destination of Acapulco. P erceived efficacy and engagement in a recommended risk reduction behavior were examined in terms of behaviors that tour ists could engage in to ensure their personal safety while visiting the destination of Acapulco. An online survey was conducted in October 2015 to garner responses from American international travelers (n= 1,126). Overall, the findings suggested that self efficacy and response efficacy were the different recommended risk reduction behavior s to ensure their personal safety while visiting the destination. Between the two perceived efficacy variables, self efficacy had the strongest influence on intentions to engage in a recommended risk reduction behavior. Implications, l imitations and suggestions for future research were presented.

PAGE 16

16 CHAPTER 1 INTRODUCTION Study Context Increase in Competition among Destinations The tourism industry has grown substantially over the past six decades, from 25 million international tourist arrivals in 1950 to a record 1,087 million in 2013 ( United Nations World Tourism Organization [ UNWTO ] , 2014). During the same time period, international tourism receipts have also increased to $1 , 159 billion in 2013 (UNWTO, 2014). As tourist arrivals and tourism receipts have continued to grow at a rapid pace, destinations have recognized the socio economic contributions of tourism in terms of revenue , employment generation, business opportunities, and infrastructure development (UNWTO, 2014). Consequently, the amount of destinations entering into and investing in tourism has continued to increase (U NWTO, 2014). As a result of unprecedented expansion and diversification since 1950, tourism has established itself among the biggest and fastest growing economic sectors across the globe (UNWTO, 2014 ; World Travel & Tourism Council [ WTTC ] , 2015 ). Addition ally, tourism has become a significant economic sector for destinations at the local , state/ provincial, national, regional, and international levels (Goeldner & Ritchie, 2012) . Industry growth is not expected to slow down in the foreseeable future , as UNWTO (2014) estimates that international tourist arrivals will grow at a rate of 3.3% per year to a total of 1.8 billion internati onal tourist arrivals by the year 2030. Accordingly, competition among destination s is also not expected to slow down in the near future . In fact, traditional touris t destinations , such as Europe and North America, are faced with the challenge of expected declines in market sh are . One

PAGE 17

17 reason for this anticipated challenge is that UNWTO (2014) estimates that emerging destination s will experience increases in tourist arrivals at double the rate of traditional destinations (+4.4% vs. + 2.2%, respectively) from 2010 to 2030. Consequently , emerging destinations are expected to represent 57% of the market share by the year 2030 (UNWTO, 2014). Thus, the balance of market share is expected to shift because the tourism landscape is competitive and dynamic . As the destination landscape become s more competitive, it is more important than ev er to maintain existing tourist origin markets and t o grow new markets. A major barrier to maintaining and growing markets is any shock to the system that impacts either the image or the product of the destination. Increase in Crises Affecting the Tourism Industry Over the past decade, tourism has experienced a wide range of shocks to the system , such as the September 11 th terrorist attack s and The Great Recession of the late 2000s . C rises have significant effects on the tourism industry (UNWTO, 2011). These events are becoming more apparent as t here has been a documented increase in the quantity and severity of both natur e induced and human induced crises across the globe in modern times (Drabek, 2009). As a result, t ourism has been directly and indirectly affected by the increasing prevalence of crises. C ontemporary examples of crises affecting the tourism industry around the globe hav e included th e Ebola outbreak in West Africa, a major earthquake in Nepal, terrorist attacks in the popular destination of Paris , France , and the weakening of the euro. Th is increasing prevalence of crises has also highlighted the significant need to prioritize safety and security issues in the tourism industry ( Pacific Asia Travel Association [ PATA ] , 2011 ; Tourism Crisis Management Initiative [ TCMI ] , 2015 ). T his need is exacerbated by the fact that the

PAGE 18

18 sustainability of destinations can be threatened by real and perceived management failures (PATA, 2011 ; TCMI, 2015 ). In some cases, the tourism industry has been considered to be particularly vulnerable because of its rapid growth (Drabek, 1995; Faulkner, 2001; Murphy & Bayley, 1989). In addition , some destinations are located in especially risky areas (Faulkner, 2001 ; UNWTO, 2011 ) . For exampl e , c oastal destinations are inherently susceptible to natural disasters su ch as hurricanes and tsunamis. Furthermore , the tourism system is open by nature and, thus, is continuously affected by the external environment (Morrison, 2010). As a result, the tourism industry is often exceptionally vulnerable to negative influences fr om the external environment (Mansfeld, 1999). Given the significance of tourism to economies around the world, crises can have devastating impacts on the affected de stination, tourism system , economy , and tourism dependent community (TCMI, 2015 ; UNWTO, 2011 ). Sustained Tourism Crisis Affected Destination: The Case of Acapulco, Mexico Overall, Mexico is an established international destination with relatively stable international tourist arrivals (UNWTO, 2014). However, t he destination has been affected by a variety of tourism crises in recent times , including The Great Recession, the H1N1 outbreak (Linthicum, 2014 ; The Economist, 2009) , and a category 5 hurricane that hit the Pacific coast (The Weather Channel, 2015) . Mexico has also been faced with a su stained tourism crisis. The origin of this sustained tourism crisis is the Mexican Drug Wars, which resulted in approximately 60,000 deaths between 2006 and 2012 (Bender & Rosen, 2014). While drug related violence has generally not been an issue in s major international tourist destinations (U.S. Department of State, 201 5 ), the

PAGE 19

19 drug cartel violence Although tourist arrivals t o the country have been on the rise recently (UNWTO, 2014), the tourism industry in specific destinations within Mexico have taken a direct hit from the sustained tourism crisis. For example, l ocated on the Pacific coast of Mexico (Figure 1 1 ), Acapulco was prime international tourist destination s (Flannery, 2014). However, in the mid for cartel in Mexican s tate Guerrero (U.S. Department of State, 201 5 ). Further, according to the Citizen Council for Public Security and Criminal Justice, the municipality of Acapulco has the second highest homicide rate among urban areas in Latin America (Control Risks, 2014). The homicide rate in Acapulco is 142.88 murders per 100,000 residents (Control Risks, 2014). To provide a approximately 30% higher than the U.S. average (Rueda, 2013). Sources indicate that problems with crime h ave mainly occurred in Further, crimes against international tourists are considered to be rare (Rueda, 2013 ). However, the destination is faced with perceptions that the en tire city is affected by negatively impacted (Flannery, 2014; Rueda, 2013). In regards to the sustained tourism crisis in Acapulco, the ed , rime is an issue of perception.

PAGE 20

20 low Flannery , 2014). Figure 1 1 . Map of A capulco, Mexico (Google, 2015) . a 50% decline in tourism over a five year period ending in 2011 ( Datar & Douglas, 2012 ). A majority of the decline occurred in 2011, which coincided with a peak in drug related violence that totaled in approximately 1,000 deaths between January and August of 2011 (Datar & Douglas, 2012). As tourism declined, the effects also rippled through various sectors of the industry. T he average occupan cy rates fell from 56% in the first three quarters of 2006 to 44% in the first three quarters of 2011 (Rueda, 2013) . The number of cruise ships docking in Acapulco went from 180 in 2012 to 5 between the months of January and November in 2014 ( Associated Pr ess [ AP ] , 2014) . Furthermore, a local association that represents 200 local businesses on the main

PAGE 21

21 tourist strip (i.e., the Avenida Costera) reported that half of its members went out of business between 2011 and 2013 (Rueda, 2013). Thus, it is evident tha t the destination, which relies on tourism for approximately three fourths of its income (Rueda, 2013), has been seriously impacted by the sustained tourism crisis. Despite the sustained tourism crisis, the destination reported a record 15.7% increase in visitors during the summer 2014 season (Myers, 2014). This was the best summer in terms of tourism numbers in a decade (Myers, 2014). However, the increase was largely attributed to domestic tourism (Myers, 2014). Therefore, the statistics indicate d that t he market share for the former popular international tourist destination has shifted towards domestic tourists. In terms of international tourism, the United States has proven to be a significant tourist origin market for Acapulco (Myers, 2014). The U.S. i s also a key tourist origin market for Mexico as a whole. In 201 3 , Mexico welcomed 23,734,000 international tourists ( UNWTO , 201 4 ), with 20, 851 ,000 originating from the U.S. ( National Travel and Tourism Office [ NTTO ] , 201 4 ). Thus, the U.S. represent ed approximately 88% of international tourist arrivals to Mexico. Further, a pproximately 1 in 3 international travelers originating from the U.S. visited Mexico in 201 3 (NTTO, 201 4 ) , indicating that Mexico is one of the most popular international touris t des tinations for Americans. Given that the U.S. and Mexico share a border and a travel visa is not required for Americans traveling to Mexico for less than 180 days (U.S. Department of State, 201 5 ), access to Mexico is fairly easy for Americans. In fact, over 150,000 Americans cross the border daily (U.S. Department of State, 2015 ). Travel specific information related to the sustained tourism crisis is disseminated directly to U.S. residents by the U.S. Department of State in the form of a travel

PAGE 22

22 warning. On May 5, 2015, the U.S. Department of State (201 5 ) released an updated travel warning regarding risks to American tourists because of organized crime issues in Mexico. Information was given about conditions in the country in general, as well as more detailed information about conditions by state and the major tourist destinations within the states (U.S. Department of State, 2015). Notably, the section on the State of Guerrero remained the same as the travel warning issued on December 24, 2014 (U.S. Department of State, 2014). The section of the updated travel warning which focused on the state of Guerrero, in which Acapulco is located, follows: Guerrero: Acapulco, Ixtapa, Taxco and Zihuatanejo are major cities/travel destinations in Guerrero Defer non essential travel to all parts of the state, except for the cities of Acapulco, Ixtapa, and Zihuatanejo. Travel to Acapulco and Ixtapa/Zihuatanejo only by air or cruise ship, exercise caution, and remain in tourist areas. Travel in and out of Acapulco b y air and cruise ship is permitted for U.S. government personnel. U.S. government personnel are prohibited from traveling within Guerrero state Acapulco, as well as highway 200 betw een Acapulco and Ixtapa/Zihuatanejo. In Acapulco, defer non essential travel to areas further than two blocks inland of the Costera Miguel Aleman Boulevard, which parallels the popular beach areas. Lodging for U.S. government personnel is limited to the ho the Krystal Beach Acapulco hotel in the north and going south through Puerto Marquez, including the Playa Diamante area and ending at The Resort at Mundo Imperial hotel. In general, the popular tourist area of Diamante, just south of the city, has been less affected by violence. Any activity outside the hotel zone for U.S. government personnel is limited to the coastal area from La Quebrada to the beginning of the hotel zone and only during daylight hou rs. The state of Guerrero was the most violent state in Mexico in 2013, with 2,087 homicides and 207 reported cases of kidnapping, according to the Mexican Secretariado Ejecutivo Nacional de Seguridad Publica. Self defense groups operate independently of t he government in many areas of Guerrero. Armed members of these groups frequently maintain roadblocks and, although not considered hostile to foreigners or tourists, are suspicious of outsiders and should be consid ered volatile and unpredictable (U.S. Depa rtment of State, 201 5 ). According to the travel warning , the U.S. Department of State (2015) has not prohibit Americans from visiting Acapulco. Rather, U.S. citizens have been urged to remain

PAGE 23

23 cautious while visiting the destination and to avoid leaving Aca ( U.S. Department of State , 201 5 ). Acapulco, Mexico was used as a case to address the purpose, principal objectives, research questions , and associated hypotheses of this study . The destination was chosen based on tourism statistics which indicate that Acapulco is experiencing a sustained tourism crisis, as well as the state specific evaluations 5 ) travel warning for Mexico. Given the nature of the sustained tourism crisis, this s tudy focused on crime related risk. Accordingly, the survey questions used to measure the risk related constructs in this perceptions of crime related risk associated wit h visiting Acapulco . Further, given that the U.S. is a key tourist origin market for both the city level destination of Acapulco and the national level destination of Mexico , this study focused specifically on the perceptions and intentions of American int ernational tourists. Fundamentals of Destination Crisis Management The current environment of increased crises (Drabek, 2009) provides support for (1998) proposition that organizations should consider when crises may occur, what types o f crises may occur, and how the organization plans to manage different types of crises. Kash and Darling (1998) essentially advocated for a fundamental shift in mindset from complacency and reactiveness to proactive planning. Applied to the context of tour ism, this suggests that destinations can no longer be content and question if they will be affected by a crisis. Rather, destination management organizations [ DMOs ] must acknowledge that a crisis could occur at any destination at

PAGE 24

24 virtually any time (Faulkn er, 2001) , as well as that no destination is immune to crises (UNWTO, 2011). As crises are putting increasing pressure on destinations and tourism organizations to effectively manage their businesses and safeguard both their visitors and their images , the re i s a critical need for proactive destination crisis management (TCMI, 2015). Further, crisis management for destinations must be tailored specifically to the dynamic nature of the tourism system as a whole and the destination in particular (TCMI, 2015). Proactive destination crisis management is essential to the mitigation of the potential economic, environmental, and social impacts of tourism crises (TCMI, 2015). Given that these are also the three pillars of sustainable tourism development (Goeldner & Ritchie, 2012), destination crisis management should be considered to be a vital component of the sustainability of destinations and tourism dependent communities. Thus, functioning from a sustainable tourism development paradigm, proactive planning and po licy initiatives can moderate economic, environmental, and social impacts of tourism crises on the affected destination, the tourism system, and the tourism dependent community. Defining Tourism Crisis At the core of destination crisis management is unders tanding what constitutes a tourism crisis. an event is, to some extent, self inflicted through such problems as inept management structures and practices of a failure to adapt to chan definition of a crisis has been accepted by other tourism scholars, including Scott and Laws (2005) and Prideaux, Laws, and Faulkner (2003). Accordingly, tourism scholars

PAGE 25

25 have generally consider ed the key characteristic of a crisis to be that the organization is at least partially responsible for the situation. While tourism academics have taken an organizational attribution approach to defining a crisis, tourism practitioners have taken a differe nt approach to defining a tourism crisis . Specifically, PATA (2003) defined a tourism crisis as has the potential to affect long term confidence in an organization or a product, or which may interfere with its ability to continue operat Similarly, UNWTO (2005) defined a tourism crisis as confidence in a destination and interferes ( p. 11 ) . Accordingly , practitioners have focu sed on determining when a situation becomes a crisis in terms of consumer confidence and business operations within the tourism system. In comparing the two definitions proposed by PATA (2003) and UNWTO (2005) , the working definition of a tourism crisis f or a destination in this study wa s any situation normal tourism operations at the destination. The decision was made to defer to the tourism ation of a tourism crisis for a few reasons. First, this study focused on prov iding research based solutions for destination crisis management. Therefore, in order to provide solutions for practice, the conceptualization of a tourism crisis in this study n eeded to be consistent with the industry . Second, this definition recognizes that if tourists perceive a situation to be a crisis, it should be considered as a crisis and managed accordingly . Third, this definition recognizes that the tourism system can be influenced by unexpected situations in other related systems. For example,

PAGE 26

26 according to the academic definition of a crisis, a global financial recession would not have the characteristics of a tourism crisis. However, practitioners recognize that tourism does not operate in a vacuum. Therefore , what happens in the external and/or disrupts normal tourism operations at the destination. The academic tourism crisis management literature has lacked conceptual clarity in defining key terms such as crisis and disaster ( Carlsen & Liburd, 2008 ). Ritchie (2008) attributed this lack of conceptual clarity to overlaps between the two concepts. For example, Faulkner (2001) de collection of enterprises in the case of a tourist destination) is confronted with sudden according to Faulkner (2 001), the key difference between a crisis and a disaster is the attribution of the situation. Specifically, an organization is at least partially responsible for a crisis, while a disaster is unexpected and outside of the control of the organization (Faulk ner, 2001). Prideaux et al. catastrophic change that can normally only be responded to after the event, either by Accordingly, Prideaux et al. (2003) disaster. Taking both definitions into consideration, tourism academics have generally suggest ed that disaster management is conducted through reactive responses (Prideaux et al., 2003). community or population, causing deaths, injuries or damage to property, livelihoods or

PAGE 27

27 the environment, that exceeds the ability of the affected community to cope using its o United Nations Office for Disaster Risk Reduction [ UN/ISDR ] , 2004: 17). academics. Specifically, UN/ISDR (2004) focuse d on the magnitude of the effec ts of a situation on the society and the environment as being the defining characteristic s of a disaster. In addition, UN/ISDR (2015) has advocate d for proactive disaster risk reduction tion and disaster preparedness . This contrasts the academic literature which suggests that a disaster cannot be managed proactively (Faulkner, 2001; Prideaux et al., 2003). Types of Tourism Cris i s While a broad range of tourism crises have been identifie d (APEC International Centre for Sustainable Tourism, 2006; PATA, 2003 ; UNWTO, 2005), the types of tourism crises have traditionally been divided into two categories: nature induced and human induced (de Sausmarez, 2007; TCMI, 2015). Examples of nature ind uced tourism crises include hurricanes, earthquakes, tsunamis, floods, volcanoes, and other types of natural disasters. Alternatively, examples of human induced tourism crises include terrorism, crime, political instability, hostage situations, and physica l accidents. In addition to classifying a tourism crisis in terms of the root cause , a tourism crisis can also be classified from a temporal perspective. Parsons (1996) identified three crisis types based on a time constraint perspective. First, in the eve nt of immediate 1996 , p. 26 ). Accordingly, Parsons (1996) suggested that it is difficult to prepare for immediate crises. Second, emerging crises develop more slowly, but are not easily predictable (Parsons, 1996 ). According to Parsons (1996), the challenge in managing

PAGE 28

28 that indicate that a crisis, like a long dormant volcano, is ab However, if an issue is managed, there is the potential to prevent it from becoming a crisis (Ritchie, 2004). Third, sustained crises can extend over a time period from weeks to months to years (Parsons, 1996). Lew (2014) also adopte d a temporal perspective when he suggested that a crisis can be classified based on the magnitude of disturbance, which was defined as either slow/gradual or a sudden shock to the tourism system (Lew, 2014). Therefore, when developing crisis management strategies, destination crisis managers must consider the onset and the duration of the crisis (TCMI, 2015) , as well as the magnitude of the impact (Ritchie, 2004) . Tourism Crisis Management Models Several tourism crisis management models have been introduced over the past several de cades. These models have been based on four approaches: life cycle approach, strategic crisis management approach, action oriented crisis management approach, and integrated approach (Huang, Ts tourism disaster management framework serves as an example of a life cycle approach to tourism crisis management (Huang et al., 2007). Specifically, Faulkner (2001) suggested a six stage approach to tourism disaster management : pre event, prodromal, emergency, intermediate, long term (recovery), and resolution. framework for crisis and disaster management also adopted a life cycle approach to tourism crisis management and included the same six stage (2001) model. It is important to note that the life cycle approach to tourism crisis management has tended to focus on an organization as a unit of analysis.

PAGE 29

29 On the other hand, action oriented models have tended to focus on a destina tion oriented approach and is considered to be the most comprehensive tourism crisis management model (Pennington mode l potential crisis is detected and dealt with quickly PATA, 2003, p. 4). According to the 4 R model, reduction, readiness, response, and recov ery are the four phases of tourism crisis management (PATA, 2003). Therefore, based on Pennington Gray, Thapa, Kapla nidou, Ca hyanto, & McLaughlin, 2011, p. 312). Given that the reduction and readiness phases take place before a tourism crisis occurs, the two initial phases focus on proactive planning for tourism crises (Pennington Gray et al., 2011). Once a crisis has o ccurred, the response phase begins and the procedures and strategies planned in the two initial phases are implemented (Pennington Gray et al., 2011). Lastly, the recovery phase focuses on returning tourism operations at the destination to normal and resto Gray et al., 2011). Travel Risk Research Over the past three decades, the tourism literature has examined the role of risk perceptions in an effort to gain a better understanding of the factors that influence travel decision making in these uncertain times. However, major criticism s of the existing body of knowledge related to travel r isk stem from a lack of conceptual clarity and a lack of theoretical underpinnings .

PAGE 30

30 Overall , the conceptualization and measurement of travel risk perceptions ha ve not been consistent with the theoretical frameworks of travel risk studies. For example, Sön mez and Graefe ( 1998a , 1998 b ) indicated that protection motivation theory was one of the theories that guided their research. However, the conceptualization of perceived risk according to the theory was not applied in their study . Rather, Sönmez and Graefe ( 1998a Thus, even when travel risk studies have adopted a theoretical lens, the conceptualization of travel risk has lacked strong theoretical underpinnings. T he fields of health behavior (Maddux & Rogers, 1983; Rogers, 1975, 1983) and psychology ( Loewenstein , Weber, Hsee, & Welch , 2001 ) consider risk perceptions to be multidimensional. For example, protection motivation theory, one of the most established health behavior theories, suggests that risk perceptions consist of an evaluation of the perceived vulnerability to and perceived sev erity of a risk (Floyd, Prentice Dunn, & Rogers, 2000; Maddux & Rogers, 1983; Rogers, 1975, 1983). Several tourism studies have adopted the perceived vulnerability measure of risk perceptions (Floyd , Gibson, Pennington Gray, & Thapa , 2004; Kozak, Crotts, & Law, 2007; Law, 2006; Pennington Gray, Kaplanidou, & Schroeder, 2013; Pennington Gray , Schroeder , & Kaplanidou, 2011; Schroeder, Pennington Gray, Donohoe, & Kiousis , 2013a ; Schroeder, Pennington Gray, Kaplanidou, & Zhan, 2013b ; Schroeder & Pennington Gray , 2014), while few have measured perceived severity (Kozak et al., 2007; Law, 2006). Thus, while theory considers perceived risk to be multidimensional , travel risk studies have tended to consider risk perceptions to be unidimensional. Provided that

PAGE 31

31 protec tion motivation theory is often cited as the theoretical framework for travel risk studies, tourism scholars have generally had a problem with conceptualizing and measuring travel risk perceptions in accordance with the guiding theoretical framework of the ir studies. Often times, t he operationalization of travel risk has lacked any theoretical underpinnings. For example, simplistic measures such as a 1 5 scale of very safe very risky (Sönmez & Graefe, 1998a) have been used . Also, t he measurement of travel risk has been derived from either the travel risk literature or the destination image literature (Schroeder et al. , 2013b). Those who have turn ed to the destination image literature have suggest ed that travel risk is a component of destination image becaus e risk factors are included in destination image studies (Qi, Gibson, & Zhang, 2009). However, destination image has often been operationalized through a series of either Likert type scales or semantic differential scales (Echtner & Ritchie, 2003). Accordi ngly, the measures of perceived risk based on the destination image literature have tend ed to be simplistic in nature , as well as have lack ed theoretical under pinnings. In addition , while travel risk studies have considered risk perceptions to be a factor that influences decision making, the field of health behavior focuses on the cognitive processes from which risk is perceived ( Maddux & Rogers, 1983; Rogers, 1975, 1983 ). Psycholog y, on the other hand, considers both the cognitive and affective processes from which risk is perceived (Loewenstein et al., 2001; Slovic & Peters, 2006). Therefore, adapting the conceptualization of risk perceptions from the health behavior and psychology literature c an provide a deeper understanding of the processes that tourists go through when evaluating travel risk.

PAGE 32

32 Furthermore, the existing body of knowledge related to travel risk has primarily focused on risk perceptions . However, there are a variet y of other risk related constructs studied in other academic disciplines. Thus, in order to consider additional risk related constructs in the context of travel , there is a need for an interdisciplinary approach to the study of travel risk . Specifically, t he health behavior and psychology literature are more advanced and have a vast body of knowledge related to risk y decision making that should be integrated with the existing body of knowledge related to travel risk. For example, perceived efficacy is an im portant risk related construct in the fields of health behavior (Floyd et al., 2000; Maddux & Rogers, 1983; Rogers, 1975, 1983) and psychology (Bandura, 1977, 1982, 1986, 1992) that is understudied in the travel risk literature. Accordingly , there is a nee d to look outside of the tourism literature and to adapt risk related constructs from the fields of health behavior and psychology to the context of tourism in an effort to provide a deeper, more critical understanding of the role of risk in travel decisio n making . Therefore , future travel risk research must take an interdisciplinary approach by integrating knowledge from several different disciplines, including health behavior and psychology . In summary, the main criticism s of the existing body of knowledg e stem from a lack of conceptual clarity and a lack of theoretical underpinnings . Overall, the travel risk literature has been plagued with problems in going from theory to application. Accordingly, t here is a need to reconceptualize and reoperationalize t ravel risk perceptions. T he conceptualization and operationalization of travel risk perceptions requires a strong theoretical foundation and an interdisciplinary perspective. T here is also a need to consider additional risk related constructs which can be integrated with

PAGE 33

33 the existing travel risk literature to provide a deeper understanding of the role of risk in travel decision making. Thus , there is a need to adopt a theory based, interdis ciplinary approach to the reconceptualization and re operationalization of the risk related constructs studied in tourism. Particularly, the fields of health behavior and psychology can provide theory based conceptualizations and measures of risk related co nstructs which can be adapted to the context of international travel . Adopting a theory based, interdisciplinary approach to the conceptualization and operationalization of risk related constructs can provide a more holistic understanding of the role of ri sk in travel decision making. Adopting a Theoretical Lens to the Study of T ravel Risk Taking the major criticism s of the travel risk research into consideration, this study adopted a theoretical lens to the study of travel risk. T he theoretical lens of t h is study was guided by protection motivation theory and the risk as feelings hypothesis. Protection Motivation Theory Protection motivation theory [PMT] (Rogers, 1975, 1983) is considered to be one of the most prominent models in the field of health behav ior (Weinstein, 1993). Although originally developed as a theory of fear appeals (Rogers, 1975), PMT was later revised into a general attitudinal change model (Maddux & Rogers, 1983; Rogers, 1983). Particularly, by adding the self efficacy construct, PMT b ecame an attitudinal model which focuses on the cognitive processes which mediate behavioral change (Maddux & Rogers, 1983; Rogers, 1983). As a general attitudinal change model, PMT offers a framework for understanding the reason for attitudinal and behavi oral change in risky situations (Floyd et al. , 2000). Prentice Dunn and Rogers (1986) suggested that PMT is comprehensive enough to be applicable to any context involving risk.

PAGE 34

34 A core assumption of PMT is that individuals go through two cognitive processes when deciding whether or not to engage in a behavior to protect oneself from a risk (Rogers, 1983). First, individuals go through a threat appraisal process in which they evalu ate risk in terms of perceived severity and perceived vulnerability (Floyd et al., 2000). Perceived severity represents the perceived noxiousness of the event (Rogers, 1975). Perceived vulnerability represents the perceived likelihood that a threatened eve nt will occur (Rogers, 1975). Second, individuals go through a coping appraisal process in which they evaluate behaviors to cope with risk in terms of response efficacy and self efficacy (Floyd et al., 2000). Response efficacy represents the perception of the effectiveness of a recommended behavior in protecting oneself from a risk (Floyd et al., 2000). S elf efficacy represents the perception that an individual is able to successfully perform a recommended behavior in an effort to protect oneself from a r is k (Floyd et al., 2000) . The reason that PMT assumes that the threat appraisal process comes before the coping appraisal process is that an individual must perceive a risk before assessing whether they will engage in a behavior to reduce a risk or not (Flo yd et al., 2000). The outcome of the two cognitive mediational processes is that the threat appraisal and the coping appraisal processes come together to stimulate, maintain, and guide engagement in risk reduction behaviors (Floyd et al., 2000). It is impo rtant to note that PMT does not assume that decision makers are rational (Floyd et al., 2000). Rather, cognitive and motivational biases are believed to have an effect on all PMT constructs and the two cognitive evaluation processes (Floyd et al., 2000).

PAGE 35

35 Risk As Feelings Hypothesis The risk as feelings hypothesis, a concept originating from the field of psychology, suggests that both affective and cognitive risk perceptions directly influence decision making (Loewenstein et al. , 2001). Accordingly, emotion s such as anxiety, dread, fear, and worry directly influence reactions to risky situations ( Loewenstein et al., 2001 ). Further, cognitive evaluations are hypothesized to have affective consequences (Loewenstein et al., 2001). The emotions associated with t he affective consequences , in return , have an effect on cognitive evaluations ( Loewenstein et al., 2001 ). Accordingly, the risk as feelings hypothesis suggests that while cognitive risk perceptions and affective risk perceptions directly influence risky de cision making, there is also an interaction between these two types of risk perceptions. Therefore, w hile the risk as feelings hypothesis acknowledges that individuals perceive risk in two different ways (Slovic & Peters, 2006), it also acknowledges that c ognitive risk perceptions and affective risk perceptions may be associated with one another (Loewenstein et al., 2001). Develop ing a Theory Based Conceptual Model for the Study of Travel Risk Through the develop ment of the theory based proposed conceptual model, this study sought to advance knowledge related to travel risk by integrating the existing body of knowledge related to risk in the fields of health behavior and psychology with the travel risk literature. Specifically, thi s study integrated PMT and the risk as feelings hypothesis with the existing body of knowledge related to travel risk in an effort to reconceptualize and reoperationalize the risk related constructs studied in the context of international travel based on t heory .

PAGE 36

36 T he development of the proposed conceptual model occurred in three stages. The first stage of the development of the proposed conceptual model focused on the destination choice process , which was based on the consumer behavior and tourism decision m aking literature . A detailed description of the development of this stage of the proposed conceptual model is presented in Appendix D . The second stage of the development of the proposed conceptual model focused on the relationships between the risk relat ed constructs . This section of the proposed conceptual model was based on PMT and the risk as feelings hypothesis. Notably, of the risk related variables included in the proposed conceptual model, only perceived vulnerability has been studied extensively i n the travel risk literature. Therefore, the proposed conceptual model considered the role of the understudied risk related variables of perceived severity, affective risk perceptions, self efficacy, response efficacy, and engagement in a risk reduction be appraisal process was extended with the inclusion of affective risk perceptions , which were derived from the risk as feelings hypothesis . Further , in accordance with the core assumption of PMT, perceived risk and perceived efficacy were entered in a causal string in which individuals go through the threat appraisal process before the coping appraisal process (Floyd et al., 2000) . The third stage of the development of the proposed conceptual model focused on integrating the existing body of knowledge related to travel risk with the risk related constructs. Overall, the proposed conceptual model suggests that the risk related constructs serve as mediat ing variables in the destination choice process . The proposed conceptual m od el is presented in Figure 1 2.

PAGE 37

37 Figure 1 2 . C onceptual m odel to u nderstand the r ole of r isk r elated c onstructs in the d estination c hoice p rocess.

PAGE 38

38 Purpose T he purpose of this study was to recon cep tualize and reoperationalize the risk related constructs studied in the context of international travel , as well as to analyze the relationship s between the risk related constructs. More s pecifically, this study sought to test and refine the proposed conceptual model which considers the relationships between the risk related constructs of perceived risk (perceived vulnerability, perceived severity, affective risk perceptions) , perceived efficacy (self efficacy, response efficacy) , and engagement in a recommended risk reduction behavior. Given that a majority of the risk related variables (i.e., perceived severity, affective risk perceptions, self efficacy, response efficacy, and engagement in a recommended risk reduction behavior) have not been studied in the context of international travel, this study focused on engagement in a recommended risk reduction behavior as the dependent variable. Such efforts can provide a better understanding of the dynamic processes between the r isk related constructs, as w ell as provide conceptual clarity prior to the examination of the full conceptual model . In other words, the relationships between the risk related constructs were examined prior to an investigation into their role in the destination choice process . B ased on the purpose of this study , the 3 principal objectives of this study were: 1. To recon cep tualize the risk related constructs studied in the context of international travel ; 2. To reoperationalize the risk related constructs studied in the context of international travel ; 3. To test and refine the proposed conceptual model. As previously mentioned, t he case of Acapulco, Mexico was used to address the purpose and principal objectives of this study. sustained tourism crisis , this study examined perceived risk (perceived vulnerability,

PAGE 39

39 perceived severity, and affective risk perceptions) in terms of crime related risk while visiting the destination of Acapulco. In addition, perceived efficacy (self efficacy and response efficacy) and engagement in a recommended risk reduction behavior were examined in terms of behaviors that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco. Research Questions The research questions were developed in an effort to address the third principal objective of this study to test and refine the proposed conceptual model. The center of the proposed conceptual model focused on the mediating risk related variables of perceived vulnerability, percei ved severity, affective risk perceptions, self efficacy, response efficacy , and engagement in a recommended risk reduction behavior . Therefore, a mediation analysis approach was taken in the development of the research questions. Specifically, the research questions focused on the total and mediation models within the proposed conceptual model. Accordingly, the 4 research questions addressed by this study were: 1. Do the antecedents (demographic factors, international travel specific psychological factors, des tination specific psychological factors, destination specific factors, and past travel experience) predict perceived efficacy (self efficacy and response efficacy)? 2. Does perceived risk (perceived vulnerability, perceived severity, and affective risk percep tions) mediate the relationship between the antecedents (demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) and perceived efficacy (self efficacy and response efficacy)? 3. Does perceived risk (perceived vulnerability, perceived severity, and affective risk perceptions) predict intentions to engage in a recommended risk reduction behavior? 4. Does perceived efficacy (self efficacy and respo nse efficacy) mediate the relationship between perceived risk (perceived vulnerability, perceived severity, and

PAGE 40

40 affective risk perceptions) and intentions to engage in a recommended risk reduction behavior? Delimitations This study ha d a few delimit ations. First, this study contextualized travel risk in terms of risk perceived in the anticipation and planning stage and in relation to the on site stage of the travel experience. Accordingly, this study focused on risk perceived in relation to travel at the destination by prospective tourists to the destination . Second, this study used on ordered approach to PMT, based on the core assumption of the theory. Third, this study focused specifically on the destination of Acapulco, Mexico because travel risk p erceptions and behaviors have been found to be destination specific (Kozak et al. , 2007; Sönmez & Graefe, 1998a) . Acapulco was also selected because the destination is experiencing a sustained tourism crisis . Fourth , this study was also delimited to the travel risk type of crime , as risk management involves the identification, analysis, and evaluation of relevant risks to the specific destination (APEC International Centre for Sustainable Tourism, 2006). Crime was ch osen specifically because the destination of Acapulco is dealing with a sustained tourism crisis as a result of the Mexican Drug Wars . Further, crime was also chosen because of the nature of the F if th , the survey questions used to measure the risk related constructs in this study were asked in the destination of Acapulco and perceptions of crime related risk associated with visiting the destinat ion . Six th , the U.S. ( NTTO, 201 4), who have traveled internationally.

PAGE 41

41 CHAPTER 2 LITERATURE REVIEW This chapter presents a synthesis and critical evaluation of the existing body of knowledge related to the constructs examined in this study. This chapter is structured into five main sections: (1) destination choice process; (2) crime and tourism; ( 3 ) risk related constructs; (4) bivariate relationships between risk related constructs ; (5) hypotheses . The specific risk related constructs examined in this study we re perceived risk , perceived e fficacy , and engagement in recommended risk reduction behaviors . Destinati on Choice Process Travel decision making models have been proposed from two perspectives. From one perspective , travel decision making has been considered at the macro level (Clawson & Knetsch, 1966) . From the other perspective , travel decision making has also been considered at the micro level (Mansfeld, 1992; Moutinho, 1987; Sönmez & Graefe, 1998a; Um & Crompton, 1990). From the micro level perspective, researchers have tended to focus on the destination choice process (Mansfeld, 1992; Um & Crompton, 1990 ) because destination choice is central to the travel decision making process ( Fesenmaier & Jeng, 2000; Jeng & Fesenmaier, 2002 ). The existing body of knowledge related to the destination choice process is substantial and generally suggests that the process occurs in the shape of a funnel (Sirakaya & Woodside, 2005). Accordingly, individuals move through a process in which a number o f potential destinations are narrowed down until a specific destination is selected (Sirakaya & Woodside, 2005). Traditionally, travel decision making models have been based on the grand models from the consumer behavior literature (Sirakaya & Woodside, 2 005) . The most

PAGE 42

42 widely accepted model of consumer decision making was proposed by Engel, Blackwell, and Kollat (1978). The Engel, Kollat, and Blackwell [ EKB ] model proposes a five stage process of consumer decision making: (1) Need or problem recognition ; ( 2) Information search ; (3) Alternative evaluation ; (4) Purchase ; (5) Post purchase evaluation (Engel, Blackwell, & Kollat, 1978) . Recognizing the existence of choice sets , LeBlanc (1989) challenged the traditional five stage consumer decision making proce ss model. He argued that given the presence of awareness sets and subsets, the choice process does not occur solely in the purchase stage as suggested by the EKB model (LeBlanc, 1989) evoked sets exist then there is a choice process occurring before purchase choice (LeBlanc, 1989, p. 8). In other words, the selection of an evoked set is a distinct decision process that occurs before the final purchase choice. Considering a two stage choice process for choice sets, LeBlanc (1989) proposed a si x stage consumer decision making model: (1) Problem recognition ; (2) Search (internal and external) ; (3) Awareness set identification ; (4) Evaluation of the awareness set and choice processes leading to an evoked set ; (5) Evaluation and choice from an evok ed set ; (6) Post purchase evaluation. The six stage model was proposed for situations in which an individual was faced with an ext ensive decision making process ( LeBlanc, 1989 ) . Given that the destination choice process involves detailed planning long befo re a tourist actually embarks on a trip (Fesenmaier & Jeng, 2000; Jeng & Fesenm aier, 2002), the destination choice process is considered to be an extensive decision making process . Thus, it is not surprising that t he choice set framework has received considerable attention in travel decision making research

PAGE 43

43 (Sirakaya & Woodside, 2005). In fact, s everal destination choice process models have used the notion of choice sets as the ir foundation (Um & Crompton, 1990; Woodside & Lyonski, 1989). The use of choice sets as a framework for understanding the process that tourists go through when evaluating an extensive amount of destination alternatives has been generally accepted ( Ankomah, Crompton, & Baker, 1996). Specifically, the choice sets framewor k suggests that the destination choice process resembles a funnel, in which the number of alternatives decreases over time (Ankomah et al., 1996; Jang, Lee, Lee, & Hong, 2007). I t is important to note that a brand must be in the evoked set to be considered , however, being in the evoked set does not guarantee that the brand w ill be chosen (Le Blanc, 1989). Further , LeBlanc (1989) found that the decision strategies used during the evoked set selection and final purchase choice were independent of one another. In the context of destination risk management , this highlights the need to understand where risk plays an influential role in the destination choice process. In particular, DMOs need to know when risk is more influential in the decision making process . In summary , choice sets are considered to have substantial practical value in tourism (Sirakaya & Woodside, 2009). Choice set models do not explicitly consider the influence of other factors on the decision making process (Spiggle & Sewall, 1987). Ho wever, choice set models implicitly making process (Spiggle & Sewall, 1987, p. 102). C hoice set models also assume that psychological factors influe nce information search

PAGE 44

44 and evaluation processes and that there is an interplay between perceptions, images, and attitudes and the decision making process (Spiggle & Sewall, 1987). In the context of destination risk management, this could be exemplified in changing after engaging in information search. Given this interplay, the choice set decision making process is considered to be a process of transitions. For travel risk studies, this interplay highlights the fact that both the des tination choice process and perceived risk are not static. Rather, perceived risk and other factors can change as tourists go through the destination choice process. Often, the travel risk research has failed to acknowledge the dynamic and evolving nature of perceived risk throughout the t ravel decision making process. Crime and Tourism researched within th e context of tourism for several decades, t he literature has lacked a systematic approach to the study of crime and tourism (Pennington Gray, 2016) . Rather, a majority of studies have taken a case study or ad hoc approach (Pennington Gray, 2016) . Within this area of the tourism literature, Brunt, Mawby, and Hambly (2000) identified six research focuses: (1) tourist areas as areas of high crime; (2) tourists as victims; (3) tourists as offenders; (4) tourism generating higher levels of deviant or il legal activity; (5) terrorism and tourism; (6) policy responses to tourism and crime. Related to the second research focus, s cholars have suggested that tourists are particularly vulnerable and may be targets of crime for several reasons ( Mawby, Brunt, & H ambly, 1999; Tarlow, 2006) . For example, it is generally believed that touris ts frequently leave

PAGE 45

45 and cri minals can easily detect tourists in the destination based on their attire (e.g., fannypacks , cameras ) , as well as their inability to speak the native language of the destination (Tarlow, 2006). Crimes against tourists have the potential to tarnish percep tions of a safe destination (Pizam, 1996). In addition, crimes against tourists can garner intense media coverage (Mansfeld & Pizam, 2006). I t is important to recognize that media attention related to crime in the destination may be over exaggerated (Cryst al, 1993). In such cases, perceptions of crime may become heightened and, subsequently, tourism arrivals may decrease (Schiebler, Crotts, & H o llinger, 1995). Accordingly , DMOs are particularly concerned with crime because they seek to maintain an image as a safe destination (Mansfeld & Pizam, 2006; Pennington Gray, 2016) . In addition, DMOs are concerned with the subsequent impact that crime can have on tourist arrivals (Mansfeld & Pizam, 2006). Risk Related Constructs Perceived Risk Risk perceptions have tr aditionally been conceptualized in terms of cognitions (Rogers, 1975 , 1983 ). Given that cognitive processes are associated with deliberate numerical calculations ( Loewenstein et al. , 2001 ), cognitive risk perceptions are considered to be a function of the perceived vulnerability to ( Maddux & Rogers, 1983; Rogers, 1975 , 1983 ; Rohrmann, 1995; Severtson & Henriques, 2009) and the perceived severity of ( Maddux & Rogers, 1983; Rogers, 1975 , 1983 ; Severtson & Henriques, 2009) a risk. Perceived vulnerability represents the perceived likelihood that a threatened event will occur, while perceived severity represents the perceived level of

PAGE 46

46 harm to an individual that is associated with the event (Rogers , 1975). According to the health behavior literature, individuals assess the perceived vulnerability and perceived severity of a risk w hen faced with a threat ( Floyd et al. , 2000; Maddux & Rogers, 1983; Rogers, 1975 , 1983). The process in which cognitive r isk perceptions are formed is referred to as the threat appraisal process (Floyd et al. , 2000; Maddux & Rogers, 1983; Rogers, 1975 , 1983 ). Thus, the threat appraisal process produces risk perceptions ( Floyd et al. , 2000; Maddux & Rogers, 1983; Rogers, 1975 , 1983) , in which high levels of perceived severity and perceived vulnerability contribute to high overall risk perceptions (Slovic, 2000; Weinstein, 2003). The risk literature in the field of psychology has recently recognized that in addition to cognitive risk perceptions, affective risk perceptions also have a significant effect on decision making (Loewenstein et al., 2001 ; Slovic & Peters, 2006 ). Accordingly, Slovic and Peters (2006) suggested that individuals perceive risk in two different ways. On one hand, the risk as feelings hypothesis suggests that individuals 2006, p. 322). In terms of affect, it has been proposed that risky decisi on making is influenced at least to a degree by emotions (Loewenstein et al., 2001). Emotional responses to risk have been conceptualized in terms of the feelings of worry, fear, dread, and anxiety (Loewenstein et al., 2001). Thus, the risk as feelings hyp othesis highlights the importance of affective risk perceptions (Slovic & Peters, 2006). On the other hand, the risk as analysis perspective suggests that individuals perceive risk 6, p. 322). This perspective proposes that, in accordance with traditional models, individuals assess risk

PAGE 47

47 at a cognitive level (Loewenstein et al., 2001). Thus , th e risk as analysis perspective highlights the importance of cognitive risk perceptions. Not ably, the factors affecting the two assessment processes are believed to be different (Loewenstein et al., 2001). While the proximity of a risk is believed to influence affective risk perceptions, this factor is not believed to influence cognitive risk per ceptions (Loewenstein et al., 2001). The likelihood of a risk and outcome expectancies are believed to influence both affective and cognitive risk perceptions, although in different ways (Loewenstein et al., 2001). Thus, emotional responses can differ from cognitive assessments of a particular risk due to differences in their determining factors (Loewenstein et al., 2001). Despite differences in the factors affecting the two processes, it has been suggested that there is a relationship between cognitive and affective risk perceptions (Loewenstein et al., 2001). Accordingly, w hile cognitive risk assessments can contribute to affective consequences , feelings can also have an effect on cognitive assessments (Loewenstein et al., 2001). Therefore, the risk as feelings hypothesis does not suggest that there is only a direct relationship between cognitive and affective risk perceptions and risky decision making (Loewenstein et al., 2001). Rather, the risk as feelings hypothes is posits that the relationship between cognitive risk perceptions and associated behaviors may be at least partially mediated by affective risk perceptions (Chapman & Coups, 2006; Loewenstein et al., 2001). Ultimately, b y understanding that individuals ma y rely on both affective and cognitive risk perceptions, the psychology literature has advocated that individuals assess risk in terms of both what they think ( i.e., cognitions) and how they feel ( i.e., affect) about the risk (Slovic & Peters, 2006).

PAGE 48

48 As pr eviously mentioned, the conceptualization of travel risk perceptions has lacked conceptual clarity. In addition, the operationalization of travel risk has tended to be simplistic and has generally lacked a theoretical foundation, particularly when compar ed to the measures used in health behavior and psychology . For example, few travel risk studies have operationalized perceived vulnerability ( e.g., Floyd et al., 2004; Kozak et al., 2007; Law, 2006; Pennington Gray et al., 2011, 2013; Schroeder et al., 2013a , 2013 b; Schroeder & Pennington Gray, 2014 ) and perceived severity (Kozak et al., 2007; Law, 2006). Further, affective risk perceptions have been understudied in the context of travel risk. Villegas, Matyas, Srinivasan, Cahyanto, Thapa, and Pennington Gra y ( 2013 ) examined affective risk perceptions. However, affective risk perceptions were measured as an emotional risk outcome of viewing a hurricane forecast ( Villegas, Matyas, Srinivasan, Cahyanto, Thapa, & Pennington Gray, 2013 ). Thus, the affective measu re of perceived risk was associated with a message. Reisinger and Mavondo (2005) have studied travel anxiety , while tourist worries have also been studied (Brun, Wolff, & Larsen, 2011; Larsen, Brun, & Øgaard , 200 9 ; Wolff & Larsen, 2013) . However, both travel anxiety (Reisinger & Mavondo, 2005) and tourist worry (Brun et al., 2011; Larsen et al., 200 9 ; Wolff & Larsen, 2013) have been conceptualized as being distinct from perceived risk. Thus, the risk as feelings hypothesis has generally not been recogni zed in travel risk studies. In summary , the operationalization of travel risk perceptions has not comprehensively reflected the risk perceptions construct, as conceptualized by the body of knowledge related to risk in the fields of health behavior and psyc hology .

PAGE 49

49 T he direction of the effects of travel risk perceptions on travel decision making has been found to vary based on the specific type of risk (Pennington Gray et al., 2013; Schroeder, 2012; Schroeder et al., 2013a, 2013 b). For example, Schroeder et al. (2013b) found that perceptions of risk associated with natural disasters and unexpected weather positively affected intention s to travel to London to attend the 2014 Summer Olympic Games among American travelers, while perceptions of risk associated wi th increased crime negatively affected the intention to travel to London to attend the Olympic Games. These findings support the need to examine different types of travel risks separately (see Floyd et al., 2004). Travel risk perception s have also been fo und to have an effect on the evaluation of destination alternatives (Roehl & Fesenmaier, 1992; Sönmez & Graefe, 1998a, 1998 b). Particularly, certain destinations are associated with a higher degree of perceived risk than others (Sönmez & Graefe, 1998a, 199 8 b). Furthermore, Sönmez and Graefe (1998a) also found that travel risk perception s were a predictor of future travel behavior and suggested that perceptions of risk may be more influential than actual risk (e.g., Sönmez & Graefe, 1998a, 1998 b). As Fuchs a nd Reichel (2011) further fact, it does not exist in reality. In contrast, an unperceived risk will not affect consumer 7). Therefore, understanding tourists perception s may be more important that actual risk and is, therefore, essential to the sustainability of destinations and the economy.

PAGE 50

50 Contextualizing travel risk by the stage in the travel experience Althoug h travel risk perceptions have been defined in various ways, Tsaur, tourists during the process of purchasing and consuming traveling services and at the ited in Huang et al. , 2008, p. 206). This definition reveals that perceived risk coincides with different stages of the travel experience. Adapted from consists of the following five stages: (1) anticipation and planning; (2) travel to the destination; (3) on site (at the destination); (4) travel back home; (5) recollection of the trip definition is perceived during the pu rchasing process, this type of risk is perceived during the anticipation and planning stage of the travel experience. In contrast, given that the second type of travel risk identified is perceived during the consumption of travel at the destination (Tsaur, Tzeng, & Wang, 1997), this type of risk is perceived during the on site or in situ (at the destination) stage of the travel experience. While Tsaur et al. (1997) essentially advocated that travel risk can be perceived during two stages of the travel expe rience (i.e., anticipation/planning and on site), travel risk can, in fact, be perceived at any stage of the travel experience. For example, tourists can also perceive risk while in route to the destination and traveling home from the destination. Further, while tourists may be in one stage of the travel experience, they may perceive risk in relation to any stage of the travel experience. For example, in the anticipation and planning stage, tourists can perceive risk related to the on site experience. Speci fically, prospective tourists who are in the anticipation and planning

PAGE 51

51 stage are often asked about their perceptions of risk associated with either travel in general (Floyd et al, 2004; Sönmez & Graefe, 1998a, 1998 b) or at a specific destination (Schroeder , 2012; Schroeder et al., 2013a, 2013 b). Thus , there is a need to provide conceptual clarity to the travel risk literature by contextualizing the stage in which tourists are asked about risk and the stage in which risk is perceived in association with. The re is also a need to ensure that data collection takes these nuances into account. Contextualizing travel risk by the scale and scope of the destination Risk perception studies have tended to focus on one of two scales and scopes of a destination. Specif ically, studies have examined either risk perceptions associated with travel in general (Floyd et al., 2004; Floyd & Pennington Gray, 2004) or a specific destination (Schroeder, 2012; Schroeder et al., 2013a, 2013b). I n studying travel risk, there is a nee d to clearly contextualize the geographic scale and scope of a destination for several reasons. First, risk perceptions are considered to be situation specific (Dowling, 1986; Germunden, 1985; Jackson, Hourany, & Vidmar, 1972; Knowles, 1976; Knowles, Cutte r, Walsh, & Casey , 1973; MacCrimmon & Wehrung, 1986; Roehl & Fesenmaier, 1992). Second, risk perceptions have been found to vary by the specific destination , with some destinations perceived as being riskier than others (Kozak et al., 2007; Sönmez & Graefe , 1998a) . Third, tourists associate specific types of risk with specific destinations (Kozak et al., 2007) and levels of perceived risk have been found to vary by the type of risk itself (Pennington Gray et al., 2013; Schroeder, 2012; Schroeder et al., 201 3a, 2013 b). Lastly, destination risk perceptions are narrower in context and may not necessarily be directly influenced by the broader context of risk perceptions associated with international travel (Schroeder et al., 2013b). Therefore,

PAGE 52

52 there is a need to geographically contextualize perceived risk in terms of the scale and scope of the destination. Within destination risk perceptions, the geographic scale and scope of the destination has varied in size from the broad context of an international region (S önmez & Graefe, 1998b) or a country (Fuchs & Reichel, 2006a; Schroeder, 2012) to the narrow context of a city (Schroeder et al., 2013b; Schroeder & Pennington Gray, 2014). There has been a general tendency to focus on countries as the geographic scale and scope of destinations. However, for the same reasons mentioned above, there is a need to provide a narrower geographic scale and scope when examining destination risk perceptions. For example, perceived risk associated with travel in Florida is not likely to be the same as perceived risk associated with travel in Miami. For DMOs, understanding risk perceptions associated with the scale and scope destination choice is essential t o comprehensive destination risk management. One reason for this need is that tourists making and behaviors have been found to be influenced by the levels of perceived risk associated with travel to specific destinations (Schroeder et al. , 2013 b ; Schroeder & Pennington Gray, 2014; Sönmez & Graefe, 1998b) . Therefore, there is a need to provide conceptual clarity to the travel risk literature by geographically contextualizing travel risk in terms of the scale and scope of the destination. Types of t ravel risk In general, travel risk studies have often failed to recognize that the types of perceived risk associated with particular stages of the travel experience may be

PAGE 53

53 different. For example, Sönmez and Graefe (1998a) examined risk perceptions associ ated with the following ten types of risk: financial; functional/equipment; physical; psychological; social; satisfaction; time; health; political instability; terrorism . These ten types of risk were combined into a subscale of travel risk perceptions (Sönmez & Graefe, 1998a). However, in general, the types of risk studied may be grouped into two categories. Specifically, financial, psychological, satisfaction, social, and time risks are generally perceived in relation to the purchase process , which occurs prior to a trip. In contrast , equipment, physical, health, political instability, and terrorism risks are generally perceived in relation to the on site or at the destina tion stage of the travel experience. Interestingly, Roehl and Fesenmaier (1992) used factor analysis to determine the underlying structure of risk perceptions associated with equipment, financial, physical, psychological, satisfaction, social, and time ri sks in the context of travel in general and in the context of a destination (Roehl & Fesenmaier, 1992). Three distinct risk factors were found (Roehl & Fesenmaier, 1992). The first factor included perceived risk associated with physical and equipment risks in both the context of travel in general and a destination specifically (Roehl & Fesenmaier, 1992). Based on the definitions provided for these risk types, this factor is generally associated with the on site stage of the travel experience. The second fac tor included perceived risk associated with time, satisfaction, financial, and psychological risk in the context of travel in general. The third factor also included perceived risk associated with time, satisfaction, financial, and psychological risk, but in the context of risk that was present during last trip (Roehl & Fesenmaier, 1992). According to the definition s provided , these two

PAGE 54

54 factors are generally associated with the purchasing process and, thus, the anticipation and planning phase o (1992) findings provide empirical support for the notion that the types of perceived risks are generally different depending on the stage of the travel experience. In other words, the types of ri sk studied should depend on the contextualization of travel risk in terms of the stage in the travel experience. Perceived Efficacy When faced with a threat, individuals go through a c oping appraisal in which they assess their ability to engage in a behavi or to protect themselves from a risk (Floyd et al., 2000). In this manner , the coping appraisal process involves an evaluation of efficacy (Rogers, 1975). Efficacy has been conceptualized as being multi dimensional (Rogers, 1975). The first dimension is re sponse efficacy , which has been described as the perception of the effectiveness of a recommended behavior in protecting oneself from a risk (Floyd et al., 2000). Accordingly, response efficacy is an outcome expectancy for a recommended behavior (Maddux & Rogers, 1983). The second dimension is self efficacy , which has been defined as attain designated types of performances . Thus, self efficacy is have (Bandura, 1986). Bandura (1982) regarded perceived self efficacy as one of the most significant factors which influences behavioral change because it affects an self efficacy is considered to be a significant factor which has an effect on affective, cognitive, and motivational processes (Bandura, 1992). Im portantly, self efficacy is

PAGE 55

55 activity and situation specific (Bandura, 1977). Therefore, self efficacy should be treated as a mediating factor of behavior, as opposed to an antecedent factor ( Bandura, 1986; Bandura, Adams, Hardy, & Howells, 1980; Bandura, R eese, & Adams, 1982). In the context of risk , self efficacy has been described as the perception that an individual is able to successfully perform a recommended behavior in an effort to protect oneself from a risk (Floyd et al., 2000). In other words, self efficacy reflects the extent to which an individual feels competent to implement behavioral change (Rosenstock , Stretcher, & Becker , 1988) to mitigate risk. While the travel medicine literature has examined the efficacy of various med icines and vaccinations in the context of travel (e.g., Connor & Schwartz, 2005; Hilton , Kolakowski, Singer, & Smith , 2006), there is a paucity of research in the travel risk literature related to efficacy . A review of the travel risk literature revealed o n e study which examined efficacy ( Law , 2006). However, the conceptualization of efficacy as the 291) is not consistent with the conceptualization of response efficacy or self efficacy. In particular, the focus of efficacy was not on a recommended risk reduction behavior that a t ourist would engage in. Rather, the focus was on the effectiveness of th e media in providing information in the event of a crisis at the destination. Thus, while Law (2006) examined efficacy, the conceptualization and , consequently, operationalization were not consistent with the health behavior literature . Risk Reduction Beh aviors In the health behavior literature, a major research focus has been on understanding how individuals decide to behave when confronted with a number of

PAGE 56

56 different types of health related risk s (Floyd et al., 2000). Accordingly, r isk reduction behaviors have been a central focus of this research stream (Floyd et al., 2000; Maddux & Rogers, 1983; National Cancer Institute, 2005; Rogers, 1975, 1983). Motivation to engage in behaviors to protect oneself is presumed to be influenced by the expectation of an adverse health outcome, as well as the desire to either prevent the outcome or lessen its effects (Weinstein, 2003). The behavioral intention to perform a recommended risk reduction behavior, a construct commonly referred to as protection motivation (Boer & Seydel, 1996), is considered to be an intervening variable which stimulates, maintains, and guides engagement in a behavior to protect oneself from harm (Floyd et al., 2000; Maddux & Rogers, 1983; Rogers, 1975). Thus, intentions to engage in a recommende d risk reduction behavior tend to serve as a dependent variable in health behavior studies (Rogers & Prentice Dunn, 1997). Further, protection motivation is regarded as a relevant concept in the context of any risk, as long as there is an effective suggest ed risk reduction behavior that an individual is capable of performing (Floyd et al., 2000). The nature of the risk itself if considered to significantly affect the specific behaviors that an individual may engage in to protect oneself from harm ( Neuwirth, Dunwoody, & Griffin, 2000). The consumer behavior literature has also studied the concept of risk reduction behaviors. The existing body of knowledge in the field of consumer behavior suggests that an individual will either abandon a purchase or engage i n behaviors to reduce risk related to the purchase when their risk tolerance t hreshold has been met (Mitchell, Davies, Moutinho, & Vassos , 1999). Risk tolerance has been conceptualized as representing both the level of risk that a consumer is unable or unw illing to tolerate and

PAGE 57

57 Accordingly, the point at which consumers choose to engage in risk reduction behaviors (i.e., risk threshold) is directly influenced by their ri sk tolerance (Mitchell et al., 1999). In the consumer behavior literature, risk reduction is commonly explained as being a process in which consumers strive to decrease the uncertainty or negative outcomes of a decision (Mitchell et al., 1999). Thus, risk reduction behaviors have been examined in the context of purchase decisions (Mitchell et al., 1999). T he travel risk literature has tended to focus on avoidance of a destination ( Sönmez & Graefe, 1998a ) and extensive information search (Maser & Weiermair, 1998) as risk reduction behaviors. In particular, travel risk studies have suggest ed that a destination that is perceived as risky is likely to be substituted for a destination that is perceived to be safer ( Kozak et al., 2007; Sönmez & Graefe, 1998a) and that tourists are likely to modify their original travel plans if a crisis were to occur at their selected destination (Law, 2006). Further, scholars have generally suggested that tourists tend to search for destinations which are associated with lower lev els of perceived risk (Brin, 2006; McKercher, Hui, Hall, Timothy, & Duval, 2003; Uriely, Maoz, & Reichel, 2007). Accordingly, in general, the travel risk literature has traditionally considered tourists to be rational and risk averse decision makers (Fuchs , Uriely, Reichel, & Maoz, 2013). More recently, however, an emerging line of research has recognized that tourists may not necessarily abandon a purchase by way of avoiding a destination perceived to be risky or having a lower propensity to travel. Rathe r, recent travel risk research has both suggested ( Fuchs & Reichel, 2006a, 2006 b; 2011 ) and found (Cahyanto, 2012; Cahyanto et al., 2014; Matyas , Srinivasan, Cahyanto, Thapa,

PAGE 58

58 Pennington Gray, &Villegas , 2011; Villegas et al. , 2013) that travel risk percept ions are positively associated with risk reduction behaviors. Therefore, recent research has highlighted the need to understand the role of risk reduction behaviors in the context of the on site or in situ stage of the travel experience. A number of risk reduction strategies related to travel have been suggested . For example, Law (2006) examined whether the following risk reduction behaviors strengthen confidence to travel: safety of tourists by the local government, (iii) an increase of transparency of information related to risk incidents , and (iv) introduction of surveillance systems or These, however, do not reflect risk reduction strategies employed by tourists them selves. Rather, they represent strategies that the DMO could implement in an effort to persuade tourists to visit the destination despite high levels of perceived risk. Therefore, these risk reduction behaviors are not consistent with the health behavior o r consumer behavior literature because they are not behaviors that individual tourists could engage in to mitigate risk. In addition to DMO level risk reduction strategies, individual level risk reduction strategies have also been examined (see Fuchs & Re ichel, 2011; Lo, Cheung, & Law, 2011a ; Lo, Law, & Cheung, 2011 b for complete list). The types of risk reduction behaviors that the travel risk literature has suggested that tourists themselves may employ have primarily been associated with searching for in formation via a variety of sources, including friends and family (Fuchs & Reichel, 2011; Lo et al., 2011a, 2011b), travel agents (Fuchs & Reichel, 2011; Lo et al., 2011a, 2011b), the Internet (Fuchs & Reichel, 2011), and previous visitors to the destinatio n (Fuchs & Reichel, 2011).

PAGE 59

59 Additional strategies have included purchasing travel insurance (Lo et al., 2011a, 2011b) and choosing a popular destination (Fuchs & Reichel, 2011). The types of travel related risk reduction strategies have primarily been base d on those suggested by the consumer behavior literature. The approach of adopting risk reduction behaviors from the consumer behavior literature rather than adapting them to reflect the stages of the travel experience can be criticized for several reasons . First, the risk reduction behaviors studied in the field of consumer behavior have tended to focus on perceived risk associated with the purchasing process (Mitchell et al., 1999). Perceived travel risk, however, can also be associated with the four othe r stages of the travel experience. Accordingly, there is a need to examine risk reduction behaviors beyond the purchasing process or the anticipation and planning stage to the on site experience or during travel at the destination. In particular, tourists could engage in risk reduction behaviors during travel itself. This notion has been examined in a recently developed research agenda which has considered risk reduction behavior of hurricane evacuation (Cahyanto, 2012; Cahyanto et al., 2014; Cahyanto & Pennington Gray, 2014; Matyas et al., 2011; Villegas et al., 201 3 ). However, a review of the literature revealed that the role of risk reduction behaviors that prospective tourists may intend to engage in while at the destination i n mediating the relationship between travel risk perceptions and the destination choice process has not been examined. Second, the types of travel related risk reduction strategies have primarily been based on the seven risk types commonly examined in the consumer behavior literature equipment, financial, physical, psychological, satisfaction, social, and time risks

PAGE 60

60 (Brooker, 1983; Jacoby & Kaplan, 1972). Accordingly, a majority of these risk types are associated with the purchasing process (Roehl & Fesen maier, 1992). On the other hand, additional risk types including political instability, terrorism, and health ( Sönmez & Graefe , 1998a , 1998b) are commonly studied in relation to travel risk. These three risk types are often associated with the on site stag e of the travel experience. Thus, there is a need to distinguish between the risk reduction strategies that tourists may engage in during the purchasing process versus those that tourists may engage in during travel at the destination because the types of risk generally differ depending on the contex tualization of travel risk. Third, Fuchs and Reichel (2011) suggested that there may be a relationship between risk type and risk reduction behavior. Intuitively, this makes sense as the types of risk reduction behaviors that individuals may engage in are determined by the nature of the risk (Neuwirth et al., 2000). For example, a tourist may avoid drinking tap water at the destination to mitigate the risk of contracting a waterborne illness, but this would not be an appropriate behavior to mitigate the risk of being a victim of pickpocketing at the destination. In summary, there is a need to extend the notion of risk reduction behaviors beyond the anticipation and planning stage to other stages of the travel experience, such as the on site experience. In accordance, there is also a need to consider the broad range of risk types associated w ith travel at the destination and to ensure that risk reduction behaviors are tailored to the risk type. Bivariate Relationships between Risk Related Constructs Perceived Risk and Perceived Efficacy The relationship between the threat appraisal process, w hich includes an assessment of perceived vulnerability and perceived severity, and the coping appraisal

PAGE 61

61 process, which includes an assessment of response efficacy and self efficacy, has been found to vary (Neuwirth et al., 2000). Interaction effects have a lso been found between both perceived vulnerability and perceived severity and response efficacy when examining intentions to engage in a risk reduction behavior (Rogers & Mewborn, 1976). In particular, Rogers and Mewborn (1976) found that i ncremental incr ease s in perceived vulnerability and perceived severity w ere associated with increased intentions to engage in a risk reduction behavior only when response efficacy was high. However, interaction effects were not found in the low response efficacy conditio n (Rogers & Mewborn, 1976). Maddux and Rogers (1983) found interaction effects between self efficacy and both perceived vulnerability and response efficacy when examining intentions to engage in a risk reduction behavior . Given that there is a paucity of r esearch related to perceived efficacy in the travel risk literature, the relationship between risk perceptions and perceived efficacy has not been previously examined in this context. However, findings from the health behavior literature highlight the impo rtance of examining the relationships and potential interactions between the risk related constructs. Perceived Risk and Risk Reduction Behaviors C ognitive risk perceptions are considered to be important in predicting the likelihood that an individual will engage in recommended risk reduction behaviors (Rogers, 1975; Rosenstock, 1974). In evaluating whether an individual will engage in risk reduction behaviors, individuals go through a threat appraisal process in which they assess the perceived vulnerability to and perceived severity of a risk (Floyd et al. , 2000; Rogers, 1975). High levels of perceived severity and perceived vulnerability contribute to an increased likelihood to engage in a recommended risk reduction behavior to

PAGE 62

62 protect oneself from harm ( Floyd et al., 2000 ; Weinstein, 2003 ). Further, the risk as feelings hypothesis suggests that the relationship between affective and cognitive risk evaluations determines the associated behavior (Loewenstein et al., 2001). Specifically, affective risk perceptions may mediate the relationship between cognitive risk perceptions and engagement in risk reduction behaviors (Chapman & Coups, 2006; Loewenstein et al., 2001). In the context of service oriented products, risk perceptions have been found t o be positively associated with engagement in risk reduction behaviors (Mitchell, Moutinho, & Lewis, 2003). Thus, high risk perceptions have been associated with high intentions to engage in risk reduction strategies (Mitchell et al. , 2003) . High risk perc eptions have also been associated with engagement in a greater number of risk reduction behaviors ( Mitchell, 1993). In the context of tourism, the relationship between cognitive risk perceptions and risk reduction behaviors has been examined ( Cahyanto, 20 12; Cahyanto et al., 2014; Matyas et al., 2011; Villegas et al., 201 3). Research has found that tourists with high cognitive risk perceptions were significantly more likely to engage in the risk reduction behavior of evacuating in the event of a hurricane at the destination than those with low cognitive risk perceptions (Cahyanto et al., 2014; Matyas et al., 2011; Villegas et al., 2013). In other words, tourists at the destination have been found to be more likely to engage in behaviors to reduce perceived risk when cognitive risk perceptions were high ( Cahyanto, 2012; Cahyanto et al., 2014; Matyas et al., 2011; Villegas et al., 201 3). Review of the tourism literature revealed one study which has examined the relationship between affective risk perceptions a nd risk reduction behaviors. In a study

PAGE 63

63 on the risk reduction behavior of hurricane evacuation by tourists at the destination , Villegas et al. (2013) found that tourist s with high affective risk perceptions (measured as fear) were significantly more likely to evacuate than those with low affective risk perceptions , though cognitive risk perceptions had a stronger effect on the likelihood to evacuate than affective risk perceptions (Villegas et al., 2013) . The authors suggested that their finding provided su pport for the need to study both tourist affective risk perceptions (Villegas et al., 2013) . Thus, there is evidence that both cognitive and affective risk perceptions influence tourists reduction behaviors. Fu chs and Reichel (2006b) proposed that travelers may utilize a variety of strategies to decrease destination specific risk perceptions to a level in which individuals will choose to travel to the destination. Th erefore , risk reduction behaviors may mediate the relationship between destination risk perceptions and the destination choice process. However, Fuchs and Reichel (2006b) surveyed travelers visiting a risky destination Israel. Consequently, the potential mediating role of risk reduction behaviors in the r elationship between risk pe rceived in the anticipation and planning stage and in association with the on site/in situ stage and destination choice has not been empirically examined among prospective travelers. Perceived Efficacy and Risk Reduction Beh aviors Efficacy is considered to be an important construct in predicting engagement in risk reduction behaviors (Floyd et al. , 2000; National Cancer Institute, 2005; Rogers, 1975; Weinstein, 1993). The evaluation of perceived efficacy is considered to be c entral to c oping appraisal (Rogers, 1975) , a process in which an individual assesses their ability to engage in a behavior to protect oneself f rom a risk (Floyd et al., 2000) .

PAGE 64

64 Perceived e fficacy has been conceptual ized as being multi dimensional, in terms of response efficacy and self efficacy (Rogers, 1975). Research in the area of social learning theory has found that changes in self efficacy are positively associated with behavioral change (Bandura et al. , 1980). Self efficacy is considered to increase c oping effectiveness (Bandura, Cioffi, Taylor, & Brouillard, 1988) and to affect the amount and length of effort an individual will use when confronted with a threat (Bandura, 1977, 1982; Bandura & Schunk, 1981). Self efficacy also serves as a means of self regulation, as an individual will avoid behaviors associated with low self efficacy (Bandura, 1977, 1982, 1989). Understanding efficacy is essential in studies related to risk reduction behaviors because response efficacy and self efficacy directly affect intentions to engage in risk reduction behaviors ( Maddux & Rogers, 1983; Rogers, 1983). The health behavior literature suggests that w hen an individual believes that a recommended behavior will be effective in minimizing the risk and that they can effecti vely perform the recommended behavior, they are more likely to engage in the recommended risk reduction behavior (Floyd et al., 2000; Maddux & Rogers, 1983). In other words, when response efficacy and self efficacy are high, respectively, an individual is more likely to intend to engage in risk reduction behaviors (Floyd et al., 2000). T here is a paucity of research in the travel risk literature related to the role of efficacy i n travel decision making . T his gap is likely associated with the limited research related to the perceived efficacy construct . However , research in the health behavior and psychology literature suggests that efficacy is an important construct which warrants attention in the travel risk literature .

PAGE 65

65 Demographic Factor s and Risk Reduction Behaviors Demographic factors, including age and education level, have been found to influence engagement in the risk reduction behavior of information search in the consumer behavior literature (Mitchell, 1993). While the extent of in formation search decreased as individuals aged, the extent of information search increased among individuals with higher levels of education in specific product categories (Mitchell, 1993). In the context of travel, Lo et al. (2011a) found significant diff erences in the types of risk reduction behaviors that tourists engaged in based on annual household income, gender, and age, as well as household size. Age, gender, and household income have reduction behaviors in the event of a crisis during travel at the destination. Specifically, tourists who were older (Cahyanto et al., 2014), female (Cahyanto et al., 2014; Cahyanto & Pennington Gray, 2014), and had an annual household income of greater th an $125,000 (Cahyanto et al., 2014) were significantly more likely to evacuate in the event of a hurricane at the destination than their counterparts. destination affects their tra vel behaviors (Bao & McKercher, 2008; Lo & Lam, 2004). This was supported by differences found in the types of risk reduction behaviors engaged in based on whether a tourist intended to visit a short haul or a long haul destination (Lo et al., 2011a). Furt her, distance also has been found to affect intentions to engage in risk reduction behaviors while at the destination, as out of state and international tourists were found to be more likely to evacuate in the event of a hurricane at the destination than i n state tourists (Cahyanto et al., 2014; Cahyanto &

PAGE 66

66 Pennington Gray, 2014). Accordingly, place of residence has also been found to influence engagement in risk reduction behaviors. Past Travel Experience and Risk Reduction Behaviors Differences in the type s of risk reduction behaviors engaged in have been found based on the extent of past travel experience in general (Lo et al., 2011a), as well as the extent of past travel experience to the destination (Fuchs & Reichel, 2011). Fuchs and Reichel (2011) also found that, in general, first time visitors to the destination relied on extensive information search and a greater number of risk reduction behaviors than those who had visited the destination in the past. This may be attributed to the fact that tourists without experience tend to depend on external information search to reduce risk (Nysveen & Lexhagen, 2001). On the other hand, individuals increasingly rely on their intrinsic cues in an effort to mitigate risk as they gain knowledge (Brunel & Pichon, 2004 ). Thus, tourists with a greater extent of past experience increasingly rely on their knowledge or internal information search as a means to reduce risk. It should be noted, however, that travel is uncontrollable and unpredictable in nature. Consequently, an (Mitchell et al., 1999). Past travel experience with the destination has been found to affect intentions to engage in risk reduction behaviors (Cahyanto et al., 2014 ), as well. Specifically, tourists who had not previously visited the destination have been found to be more likely to evacuate in the event of a hurricane while traveling at the destination than those with past experience traveling to the destination (Cah yanto et al., 2014). Hypotheses Based on a review of the literature and the proposed conceptual model , the hypotheses addressed by this study were:

PAGE 67

67 1. Do the antecedents (demographic factors, international travel specific psychological factors, destination sp ecific psychological factors, destination specific factors, and past travel experience) predict perceived efficacy (self efficacy and response efficacy)? a) Demographic factors predict perceived efficacy. b) International travel specific psychological factors pr edict perceived efficacy. c) Destination specific psychological factors predict perceived efficacy. d) Destination specific factors predict perceived efficacy. e) Past travel experience predicts perceived efficacy. 2. Does perceived risk (perceived vulnerability, perc eived severity, and affective risk perceptions) mediate the relationship between the antecedents (demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past t ravel experience) and perceived efficacy (self efficacy and response efficacy)? a) Perceived risk mediates the relationship between demographic factors and perceived efficacy . b) Perceived risk mediates the relationship between international travel specific psyc hological factors and perceived efficacy . c) Perceived risk mediates the relationship between destination specific psychological factors and perceived efficacy . d) Perceived risk mediates the relationship between destination specific factors and perceived effica cy . e) Perceived risk mediates the relationship between past travel experience and perceived efficacy . 3. Does perceived risk (perceived vulnerability, perceived severity, and affective risk perceptions) predict intentions to engage in a recommended risk reducti on behavior? a) Perceived risk predicts intentions to engage in a recommended risk reduction behavior. 4. Does perceived efficacy (self efficacy and response efficacy) mediate the relationship between perceived risk (perceived vulnerability, perceived severity, and affective risk perceptions) and intentions to engage in a recommended risk reduction behavior?

PAGE 68

68 a) Perceived efficacy mediates the relationship between perceived risk and intentions to engage in a recommended risk reduction behavior.

PAGE 69

69 CHAPTER 3 METHODS This chapter explains the methods and procedures that were used to collect and analyze the data for this study . Specifically, this chapter is structured into eight main sections: ( 1 ) research design; ( 2 ) data collection; ( 3 ) survey instrument; ( 4 ) operationalization of constructs; ( 5 ) data analysis ; ( 6 ) preparation for data analysis; ( 7 ) descriptive analysis; ( 8 ) measurement model. Research Design A cross sectional design was utilized because this study sought to examine the extent to which existing group differences in the independent variables , as opposed to interventions, were associated with vari ation in the dependent variable (de Vaus, 2001, p. 50) in a natural setting . Further, cross section al designs can be employed to effectively test and modify an a priori model ( de Vaus, 2001, p. 18 0 ) . Therefore, a cross sectional design was deemed appropria te for addressing the purpose and principal objectives of this study. Data Collection Data for this study was collected via an online survey between September and October 2015 . The online survey was hosted by Qualtrics. Respondents were re cruited from an o nline panel Amazon Mechanical Turk [MTurk]. MTurk has approximately 500,000 individuals who voluntarily participate in surveys and other tasks in exchange for payment (Amazon Mechanical Turk, 2015). M T more diverse in terms of demographic factor s than student samples and other forms of online samples (Buhrmester, Kwang, & Gosling, 2011; Mason & Suri, 2012). Further,

PAGE 70

70 sampling methods (Behrend, Sharek, Meade, & Wiebe, 2011; Buhrmester et al., 2011; Mason & Suri, 2012; Paolacci , Chandler, & Ipeirotis, 2010). The population for this study was international tourists originating from the United States. The sample for this study was made up of U.S. residents who we re at least 18 years old and have traveled internationally (i.e., outside of the U.S.) within the past 5 years . MTurk allows requestors to define worker requirements prior to launching a study. The worker requirement settings can be used to ensure that respondents are qualified for participation in the study based on pre determined properties. T o ensure that the respondents originated from the U.S. , the worker requirements were adjusted when creating the project in MTurk so that the location of workers was lim ited to the United States . In addition, a screening question was included in the survey instrument to ensure that the respondents had traveled internationally within the past 5 years . Hair, Black, Babin, and Anderson (2010) sugg ested that the preferred minimum ratio of respondents to independent variables should be either 15:1 or 20:1 for multiple regression (p. 175). Considering the more conservative ratio of 20:1 and that there was a total of 1 4 antecedents and 6 risk related c onstructs in the conceptual model, the minimum sample size for th is study was 4 0 0 respondents. Survey Instrument Pre Test A pre test was conducted on the destination specific measures on September 29, 2015 . T he pre test primarily focused on the survey questions used to measure the perceived risk, perceived efficacy, and engagement in a recommended risk reduction behavior constructs . Further, while descriptive analys i s w as conducted for each of the risk related constructs, a primary aim of t he pre test was t o assess the validity and

PAGE 71

71 reliability of the 9 perceived risk items. Hair et al. (2010 ) suggested that the preferred minimum ratio of respondents to independent variables should be 10:1 for factor analysis (p. 102). Given that there were 9 perceived risk items , the minimum sample size for the pre test was 90. A total of 99 responses were collected for the pre test . In exchange for participating in the pre test , respondents were paid $0.50. On average, it took respondents 10 minutes and 49 seconds to complete the pre test survey instrument. Given that t he primary aim of the pre test was to assess the validity and reliability of the perceived risk items , the 9 perceived risk items underwent a three stage process of data analysis using 22. The first stage of data analysis evaluated the suitability of the data for prin cipal components analysis [PCA] by examining the correlation c oefficients, Kaiser Meyer value (Pallant, 2007, p. 181). PCA was used for two main reasons. First, protection motivation theory [PMT] does not include affective risk perc eptions as part of the threat appraisal process. Therefore, the 3 factor model of perceived risk has not been examined in PMT studies. Second , the risk as feeling hypothesis suggests that cognitive and affective risk perceptions can either have direct infl uences on behavior or these types of risk perceptions can be related to one another (Loewenstein et al., 2001). Thus, the risk as feelings hypothesis does not suggest that the three factor model of perceived risk will be supported in every setting. Consequ ently, the factor analysis of the perceived risk items should be exploratory in nature. Considering these two reasons, PCA was considered to be appropriate for this study.

PAGE 72

72 The second stage of data analysis focused on assessing construct validity by utiliz ing PCA to transform the items into a reduced set of linear combinations, while using all of the variance in the items (Pallant, 2007, p. 180). The third stage of data analysis evaluated the reliability or internal consistency of the components by examining Overall, t he results of the pre test indicated that the perceived risk scales were consistent with theoretical expectations and had good internal consistency . Consequently , the perceived risk scales were deemed valid and reliable and were included in the primary study . In addition, descriptive analysis of t he pre test survey questions also revealed that approximately seven in ten respondents (70.7%) had taken a trip to Mexico. This was co mpared with (2009) Portrait of American Travelers® report , which indicated that approximately 60% of American leisure travelers have taken a trip to Mexico. Thus, it was expected that approximately 60% of the sample would have past experience traveling to Mexico, while approximately 40% of the sample would not have past experience traveling to Mexico. Based on this comparison, the researcher decided to use proportional quota sampling for the primary study in an effort to ensure that the sample was representative of the population in terms of destination past travel experience at the country level. Thus, rather than having to weight the data based on country level destination past travel experience, the proportional quota selected was 60% of res pondents have taken a trip to Mexico and 40% of respondents have not taken a trip to Mexico. Primary Study In order to implement the proportional quota , 2 separate survey instruments were created in Qualtrics for the primary study. The first survey instru ment was developed for

PAGE 73

73 those with country level destination past travel experience. T he survey instrument for those with country level destination past travel experience included a total of 36 questions and was structured into the following 15 sections: (1 ) informed consent; (2) screening questions international past travel experience; (3) destination past travel experience; (4) Mexico specific psychological factors; (5) Acapulco specific psychological factors; (6) international travel specific psychologic al factors ; (7) threat appraisal ( perceived risk ) ; (8) objective knowledge; (9) coping appraisal ( perceived efficacy) ; (10) attention question; (11) engagement in risk reduction behavior s (intentions); (12) destination choice ; (13) demographic factors ; (14 ) destination specific factors ; (15) survey code. The informed consent is provided in Appendix A and the survey instrument for those with country level destination past travel experience is provided in Appendix B. The second survey instrument was developed for those without country level destination past travel experience. The survey instrument for those without country level destination past travel experience included a total of 30 questions and was structured into the following 14 sections: (1) informed consent; (2) screening questions international past travel experience; (3) Mexico specific psychological factors; (4) Acapulco specific psychological factors; (5) international travel specific psychological factors ; (6) threat appraisal (perceived risk); (7) objective knowledge; (8) coping appraisal (perceived efficacy); (9) attention question; (10) engagement in risk reduction behaviors (intentions); (11) destination choice; (12) demographic factors; (13) destinati on specific factors; (14) survey code. The survey instrument for those with out country level destination past travel experience is provided in Appendix C. The 6

PAGE 74

74 question difference between the two surveys stemmed from the difference in country level destin ation past travel experience. Specifically, those with country level destination past travel experience were asked to answer 6 additional survey questions about the nature of their past experience. The primary study was launched and data collection was com pleted on October 2, 2015. MTurk allows researchers to determine the human intelligence tasks [HIT] prior a single, self contained task that a Worker can work on, submit a n answer, and collect a T he selected number of assignments per HIT in MTurk was 480 for the survey instrument for those without country level destination past travel experience and 720 for the survey instrument for those with countr y level destination past travel experience. Therefore, the total sample size selected was 1,200. Given the 60/40 proportional quota for country level destination past travel experience , this estimated sample size ensured that the conceptual model could be tested among those without country level destination past travel experience in future studies . In exchange for participating in the primary study , respondents were paid $1.50. The average response time for both of the survey instruments was the same. On av erage, it took respondents 16 minutes and 3 seconds to complete the survey instrument s for the primary study. As previously mentioned, t he survey instrument s included screening question s to ensure that the respondents were representative of the target pop ulation . The first screening question specifically asked respondents if they have traveled internationally within the past 5 years. If a respondent indicated that they had not taken at least 1 trip

PAGE 75

75 outside of the U.S. with in the past 5 years, they were not considered to be representative of the target population. As a result, respondents who answered no to the screening question were not eligible for participation and were not permitted to complete the remainder of the survey. The second screening question focused on country level destination past travel experience. The screening question on both surveys asked respondents if they had ever taken a trip to Mexico. If a respondent indicated that they had never been to Mexico before, they were not consi dered to be representative of those with country level destination past travel experience. In addition, if a respondent indicated that they had been to Mexico before, they were not considered to be representative of those without country level destination past travel experience. In both cases, respondents who failed to adhere to the screening criteria were not eligible for participation and were not permitted to complete the remainder of the respective survey instruments . Operationalization of Const ructs G iven that this study focused specifically on American international tourists , place of residence and international past travel experience were controlled for based on the population of this study. Further , based on the purpose of this study and the sample , intentions to engage in a recommended risk reduction behavior served as the dependent variable . Accordingly, this study measured 1 2 independent variables , 5 mediating variables , and 1 dependent variable. The 1 2 independent variables were: (1) gender ; (2) age ; (3) household income ; (4) race/ethnicity ; (5) education level ; (6) life stage ; (7) international travel risk perceptions ; (8) safety concerns ; (9) destination image ; (1 0 ) presence of friends/relatives living in the destination ; (1 1 ) fluency in the nat ive language of the destination ; (12) destination past travel experience . The 5

PAGE 76

76 mediating variables were: (1) affective risk perceptions ; (2) perceived severity ; (3) perceived vulnerability ; (4) self efficacy ; (5) response efficacy . Independent Variable : Demographic Factors The demographic factors which were measured in this study we re gender, age, household income, race/ethnicity, education level, and life stage. Demographic factors were measured through a series of questions. The measures were derived fr om the questions and responses used by the United States Census Bureau . G ender was mea sured as a dichotomous variable ( 0: female; 1: male ) (Peterson, 2000) . The 5 categories which were used to measure age were : (1) 18 24; (2) 25 34; (3) 35 49; (4) 50 64; ( 5) 65 or older (Peterson, 2000). Household income was measured in terms of total annual household income in 2014. T he 8 categories which were used to measure household income were : (1) under $15,000; (2) $15,000 to $24,999; (3) $25,000 to $34,999; (4) $35, 000 to $49,999; (5) $50,000 to $74,999; (6) $75,000 to $99,999; (7) $100,000 to $124,999; (8) $125,000 and over (adapted from United States Census Bureau, 2012) . R ace/ethnicity was measured by asking respondents if they consider themselves to be: (1) White/Caucasian; (2) Black/African American; (3) Asian; (4) Native Hawaiian/ Pacific Islander; (5) Hispanic; (6) American Indian/Alaska Native; (7) Multi ethnic/mixed race; (8) Other ( United States Census B ureau, 2012). Education level was measured by the highest degree or level of school completed (Peterson, 2000). The 8 categories which were used to measure education level were: (1) Some high school credit, no degree; (2) High School Graduate high school diploma or the equivalent (for example: GED); (3) Some college credit, no

PAGE 77

77 (7) Professional degr ee (for example: MD, DDS, DVM, LLB, JD); (8) Doctoral Degree (for example: PhD, EdD) (Peterson, 2000) . Lastly, life stage was measured in terms of marital status and presence of children in the household. T he 5 categories which were used to measure marital status were : (1) Now married; (2) Separated; (3) Widowed; (4) Divorced; (5) Never married (Peterson, 2000). The 6 categories which were used to measure presence of children currently living in the household were : (1) 0; (2) 1; (3) 2; (4) 3; (5) 4; (6) 5 o r more (Peterson, 2000) . The survey questions which were used to measure demographic factor s are provided in Table 3 1. Table 3 1. Operationalization of demographic factors. Independent variable Survey question and items Source Gender Are you..? (0) f emale; (1) male Peterson, 2000 Age What is your age? (1) 18 24; (2) 25 34; (3) 35 49; (4) 50 64 ; (5) 65 or older Peterson, 2000 Household income Which statement best describes your total annual household income in 2014 (from all sources and before taxes)? (1) under $15,000; (2) $15,000 to $24,999; (3) $25,000 to $34,999; (4) $35,000 to $49,999; (5) $50,000 to $74,999; (6) $75,000 to $99,999; (7) $100,000 to $124,999; (8) $125,000 and over United States Census Bureau, 2012 Race/ethnicity Which of the following best describes your race or ethnic origin? (1) White/Caucasian; (2) Black/African American; (3) Asian; (4) Native Hawaiian/Pacific Islander; (5) Hispanic; (6) American Indian/Alaska Native; (7) Multi ethnic/mixed race; (8) Other United States Census Bureau, 2012 Education level What is the highest degree or level of school you have completed? If you are currently enrolled, mark the previous grade or highest degree received. (1) Some high school credit, no degree; (2) High School Graduate high school diploma or the equivalent (for example: GED); (3) Some college credit, no degr ee; (4) Associate degree (for example: MEng, MSW, MBA); (7) Professional degree (for example: MD, DDS, DVM, LLB, JD); (8) Doctoral Degree (for example: PhD, EdD) Peterson, 2000

PAGE 78

78 Table 3 1. Continued . Independent variable Survey question and items Source Life stage (1) Now married; (2) Separated; (3) Widowed; (4) Divorced; (5) Never married How many children (younger than the age of 18) are currently living in your household? (1) 0; (2) 1; (3) 2; (4) 3; (5) 4; (6) 5 or more Peterson, 2000 Independent Variable: I nternational Travel Specific Psychological Factors The international travel specific psychological factors which were measured in this study were international travel risk perceptions and safety concerns. The measures of both independent variables were ada pted from Floyd et al. (2004) , by editing the wording of the items to reflect the context of international travel. Both international travel risk perceptions and safety concerns were measured on a Likert type scale of 1 5 ( where 1= strongly disagree and 5= strongly agree). The 3 items which were used to measure international travel risk perceptions were: (1) I feel nervous about traveling internationally right now ; (2) Traveling internationally is risky right now ; (3) I would feel very uncomfortable traveling internationally right now (adapted from Floyd et al., 2004) . The 2 items which were used to measure safety concerns were: (1) Safety is the most important attribute an international destination can offer ; (2) Safety is a serious consideration whe n choosing an international tourist destination (adapted from Floyd et al., 2004) . The survey questions which were used to measure the international travel specific psychological factors of international travel risk perceptions and safety concerns are prov ided in Table 3 2.

PAGE 79

79 Table 3 2 . Operationalization of international travel specific psychological factors . Independent variable Survey question and items Source International travel risk perceptions Please indicate your level of agreement with the following statements, on a scale of 1 5 (where 1= strongly disagree and 5= strongly agree). (1) I feel nervous about traveling internationally right now; (2) Traveling internationally is risky right now; (3) I would feel very u ncomfortable traveling internationally right now Floyd et al., 2004 Safety concerns Please indicate your level of agreement with the following statements, on a scale of 1 5 (where 1= strongly disagree and 5= strongly agree). (1) Safety is the most important attribute an international destination can offer; (2) Safety is a serious consideration when choosing an international tourist destination Floyd et al., 2004 Independent Variable: Destination Specific Psychological F actors The destinatio n specific psychological factor which w as measured in this study was destination image. The measure of destination image was ad o pted from Papadimitriou, Apostolopoulou, and Kaplanidou (2015). Th e independent variable was measured on a Likert type scale of 1 5 (where 1= very negative and 5= very positive). The single item measure of destination image the overall image of the city as a tourist destination ( Papadimitriou, Apostolo poulou, & Kaplanidou , 2015 , p. 306) . The survey question which w as used to measure the destination specific psychological factor of destination image is provided in Table 3 3. Table 3 3 . Operationalization of destination specific psychological factors . Independent variable Survey question and items Source Destination image Please indicate your overall image of Acapulco as a tourist destination, on a scale of 1 to 5 (where 1= very negative and 5= very positive). Papadimitriou et al., 2015 Independent Variable: Destination Specific Factors The destination specific factors which were measured in this study we re presence of friends and/or relatives living in the destination and fluency in the native

PAGE 80

80 language of the destination. To measure presence of friends and/or relatives living in the destination , r espondents were asked to indicate whether or not they have friends and/or relatives currently living in Mexico ( 0: no; 1: yes). To measure fluency in the native language of the destination , r espondents were asked to indicate how well they sp oke Spanish. The 4 categories which were used to measure fluency in the language of the destination were adopted from the United States Census Bureau (2011): (1) not at all; (2) not well; (3) well; (4) very well. The survey question s which w ere used to measure the destination specific factors of presence of friends and/or relatives living in the destination and fluency in the native language of the destination are provided in Table 3 4 . Table 3 4 . Operation alization of destination specific factors . Independent variable Survey question and items Source Presence of friends and/or relatives living in destination Do you have friends and/or relatives currently living in Mexico? (0) no; (1) yes Fluency in the native language of the destination How well do you speak Spanish? (1) not at all; (2) not well; (3) well; (4) very well United States Census Bureau, 2011 Independent Variable: Past Travel Experience P ast travel experience was measured in the context of destination specific travel. Destination past travel experience was measured at both the country level and the city level . The measures of past travel experience were adopted from (2009) Po rtrait of American Travelers® report . At the country level, destination past travel experience was measured in terms of whether or not respondents have ever

PAGE 81

81 taken a trip to Mexico ( 0: no; 1: yes ) (MMGY Global, 2009) . At the city level, destination past tra vel experience was measured in terms of whether or not respondents have visited [Mexican destination] in their lifetime (0: no; 1: yes) (MMGY Global, 2009) . Given that this study focused specifically on the Mexican destination of Acapulco, the responses re lated to whether or not respondents have visited Acapulco in their lifetime served as the measure of city level destination past travel experience in this study. The survey question s which were used to measure destination past travel experience at the coun try level and the city level are provided in Table 3 5 . Table 3 5 . Operationalization of past travel experience . Independent variable Survey question and items Source Destination past travel experience Have you ever taken a trip to Mexico? (0) no; (1) yes Which of the following Mexican destinations have you visited in your lifetime? Select all that apply. Acapulco (0) no; (1) yes MMGY Global, 2009 Mediating Variable : Perceived Risk Based on the working definition of perceived risk utilized in this study, prospective tourists were asked about perceived risk associated with their personal safety while visiting the destination. Further, i n accordance with the conceptual model, both affe ctive (Loewenstein et al., 2001) and cognitive (Floyd et al., 2000; Rogers, 1975, 1983) risk perceptions were measured. A ffective risk perceptions w ere operationalized in terms of anxiety, fear, and worry (Loewenstein et al., 2001). T he decision was made n ot to use the Tourist Worry Scale (Larsen et al., 2009; Brun et al., 2011; Wolff & Larsen, 2013) because Larsen et al. (2009) did not conceptualize worry as being an affective dimension of risk perceptions.

PAGE 82

82 Rather , the operationalization of affective risk perceptions was based on the types of affective risk perceptions and the associated semantic differential scales that are commonly employed in the field of psychology. More specifically, Loewenstein et al. (2001) conceptualized affective risk perceptions in terms of anxiety, fear, and worry . Accordingly, this study operationalized affective risk perceptions in terms of anxiety, fear, and worry. To adapt the risk as feelings hypothesis to the context of international travel, a feelings when personal safety while visiting the destination . The 5 point semantic differential scales which were used to measure affective risk perceptions were : (1) re laxed anxious; (2) fearless fearful; (3) assured worried (Newby, 2014). As suggested by PMT , individuals go through a threat appraisal process in which they evaluate risk in terms of perceived severity and perceived vulnerability (Floyd et al., 2000). The refore, cognitive risk perceptions were measured by the 2 mediating variables of perceived severity and perceived vulnerability (Floyd et al., 2000; Rogers, 1975, 1983). The items used to measure perceived severity and perceived vulner ability were adapted from Witte, Cameron, Lapinski, and Nzyuko (1998). Both mediating variables were measured on a Likert type scale of 1 5 ( where 1= strongly disagree and 5= strongly agree). The 3 items which were used to measure perceived severity were: (1) If I were a victi m of crime while visiting Acapulco, I would experience serious negative consequences ; (2) It would have a serious negative impact on me if I were a victim of crime while visiting Acapulco ; (3) If I were a victim of crime while visiting Acapulco, it would b e harmful to my well being (adapted from Witte, Cameron, Lapinski, & Nzyuko, 1998) . The 3 items which were used to measure perceived vulnerability

PAGE 83

83 were: (1) It is likely that I will be a victim of crime while visiting Acapulco ; (2) I am at risk for being a victim of crime while visiting Acapulco ; (3) My chances of being a victim of crime while visiting Acapulco are high (adapted from Witte et al., 1998) . The order of the 6 items which were used to measure perceived severity and perceived vulnerability was r andomized in Qualtrics. The survey question s which w ere used to measure the mediating variables of affective risk perceptions , perceived severity , and perceived vulnerability are provided in Table 3 6 . Table 3 6 . Operationalization of perceived risk . Mediating variable Survey question and items Source Affective risk perceptions Using the rating scales below, please indicate how you feel when you think about your personal safety while visiting Acapulco for leisure purposes. (1) relaxed anxious; (2) fearless fearful; (3) assured worried Loewenstein et al., 2001; Newby, 2014 Perceived severity Please indicate your level of agreement with the following statements, on a scale of 1 5 (where 1= strongly disagree and 5= strongly agree). (1) If I were a victim of crime while visiting Acapulco, I would experience serious negative consequences; (2) It would have a serious negative impact on me if I were a victim of crime while visiting Acapulco; (3) If I were a victim of crime while visiting Acapulco, it would be harmful to my well being Witte et al., 1998 Perceived vulnerability Please indicate your level of agreement with the following statements, on a scale of 1 5 (where 1= strongly disagree and 5= strongly agree). (1) It is likely that I will be a victim of crime while visiting Acapulco; (2) I am at risk for being a victim of crime while visiting Acapulco; (3) My chanc es of being a victim of crime while visiting Acapulco are high Witte et al., 1998 Mediating Variable: Perceived Efficacy I n accordance with PMT , individuals go through a coping appraisal process in which they evaluate risk reduction behaviors in terms of response efficacy and self efficacy (Floyd et al., 2000). Thus, the perceived efficacy construct was measured by

PAGE 84

84 the 2 mediating variables of se lf efficacy and response efficacy ( Maddux & Rogers, 1983; Rogers, 19 83 ). The items which were used to measure both self efficacy and response efficacy were adapted from Witte et al. (1998) . Self efficacy was measured on a Likert type scale of 1 5 (where 1= very unconfident and 5= very confident) and in a recommended risk reduction behavior to ensure personal safety while visiting the destination (adapted from Witte et al., 1998) . Response efficacy was measured on a Likert type scale of 1 5 in the effectiveness of a recommended risk reduction behavior in ensuring safety while visiting the destination (adapte d from Witte et al., 1998) . The survey question s which were used to measure the mediating variables of self efficacy and response efficacy are provided in Table 3 7 . Table 3 7 . Operationalization of perceived efficacy . Mediating variable Survey question and items Source Self efficacy How confident are you in your ability to perform the following behaviors to ensure your personal safety while visiting Acapulco , on a scale of 1 5 (where 1= very un confident and 5= very confident)? (1) Remain in tourist areas; (2) Avoid traveling outside of the hotel zone; (3) Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras); (4) Maintain awareness of your surroundings; (5) Avoid situations in which you may be isolated or stand out as a potential victim; (6) Avoid walking around the destination by yourself; (7) Avoid using ATMs if they are not affiliated with a recognized international chain; (8) Only use recommended taxi companies; (9) Avoid carrying large amounts of cash; (10) Avoid drawing attention to yourself; (11) Avoid walking around the destination by yourself at night; (12) Search for more information online about how to stay safe prior to travel; (13) Register wi th the U.S. Department of State prior to travel; (14) Avoid using services like Uber; (15) Purchase travel insurance prior to travel Witte et al., 1998

PAGE 85

85 Table 3 7 . Continued . Mediating variable Survey question and items Source Response efficacy Please indicate how effective you believe the following behaviors would be in ensuring your personal safety while visiting Acapulco, on a scale of 1 5 (where 1= very ineffective and 5= very effective). (1) Remain in tourist areas; (2) Avoid traveling outside of the hotel zone; (3) Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras); (4) Maintain awareness of your surroundings; (5) Avoid situations in which you may be isolated or stand out as a potential victim; (6) Avoid walking around the destination by yourself; (7) Avoid using ATMs if they are not affiliated with a recognized international chain; (8) Only use recommended taxi companies; (9) Avoid carrying large a mounts of cash; (10) Avoid drawing attention to yourself; (11) Avoid walking around the destination by yourself at night; (12) Search for more information online about how to stay safe prior to travel; (13) Register with the U.S. Department of State prior to travel; (14) Avoid using services like Uber; (15) Purchase travel insurance prior to travel Witte et al., 1998 As perceived efficacy relates to a specific recommended risk reduction behavior, the items which were used to measure self efficacy and response efficacy consisted of different risk reduction behaviors that tourists may engage in to ensure their personal safety while visiting the destination. Several of the items were adopted from the U.S. Depar 15 items, or recommended risk reduction behaviors, measured were: (1) Remain in tourist areas; (2) Avoid traveling outside of the hotel zone; (3) Avoid displaying indicators of wealth (such as expensiv e looking jewelry, watches, or cameras); (4) Maintain awareness of your surroundings; (5) Avoid situations in which you may be isolated or stand out as a potential victim; (6) Avoid walking around the destination by yourself; (7) Avoid using ATMs if they a re not affiliated with a recognized international chain; (8) Only use recommended taxi companies; (9) Avoid carrying large amounts of cash; (10) Avoid drawing attention to

PAGE 86

86 yourself; (11) Avoid walking around the destination by yourself at night; (12) Searc h for more information online about how to stay safe prior to travel; (13) Register with the U.S. Department of State prior to travel; (14) Avoid using services like Uber; (15) Purchase travel insurance prior to travel. T he order of the 15 items which were used to measure the mediating variables of self efficacy and response efficacy was randomized in Qualtrics. Dependent Variable: Intentions to Engage in a Recommended Risk Reduction Behavior In this study, travel risk was contextualized as being perceived in the anticipation and planning stage and in relation to the on site experience. Accordingly, intentions to engage in a recommended r isk reduction behavior was contextualized in terms of behaviors that tourists could engage in during the on site experienc e. Specifically, respondents were asked to indicate their intentions to engage in recommended risk reduction behaviors (Floyd et al., 2000) to ensure their personal safety while visiting the destination. I ntentions to engage in a recommended risk reduction behavior was measured on a Likert type scale of 1 5 ( where 1= very unlikely and 5= very likely ). Consistent with the measurement of perceived efficacy , t he items which were used to measure intentions to engage in a recommended risk reduction behavior consisted of 15 different risk reduction behaviors that tourists may engage in to ensure their personal safety while visiting the destination. The order of the 15 items which were used to measure intentions to engage in a recommended risk reduction behavio r was randomized in Qualtrics. The survey question which was used to measure the dependent variable of intentions to engage in a recommended risk reduction behavior is provided in Table 3 8 .

PAGE 87

87 Table 3 8 . Operationalization of intentions to engage in a recom mended risk reduction behavior . D ependent variable Survey question and items Source I ntentions to engage in a recommended risk reduction behavior Please indicate the likelihood that you would engage in the following behaviors to ensure your personal safety while visiting Acapulco, on a scale of 1 5 (where 1= very unlikely and 5= very likely). (1) Remain in tourist areas; (2) Avoid traveling outside of the hotel zone; (3) Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras); (4) Maintain awareness of your surroundings; (5) Avoid situations in which you may be isolated or stand out as a potential victim; (6) Avoid walking around the destination by yourself; (7) Avoid using ATMs i f they are not affiliated with a recognized international chain; (8) Only use recommended taxi companies; (9) Avoid carrying large amounts of cash; (10) Avoid drawing attention to yourself; (11) Avoid walking around the destination by yourself at night; (1 2) Search for more information online about how to stay safe prior to travel; (13) Register with the U.S. Department of State prior to travel; (14) Avoid using services like Uber; (15) Purchase travel insurance prior to travel D ata Analysis Prior to ad dressing the r esearch questions and the associated hypotheses , the data was prepared for analysis. Specifically, missing data, outlier, and normality analysis was conducted (Hair et al., 2010; Pallant, 2007) using SPSS . In addition, factor analysis was con ducted on the multi item measures of international travel specific psychological factors and perceived risk in an effort to prepare for data analysis. Furthermore, descriptive analysis was performed for eac h of the independent, mediating, and dependent var iable s to assess the nature of the data patterns. Principal objective 3, to test the proposed conceptual model, was addressed through regression based path analysis . Since the proposed conceptual model included

PAGE 88

88 multiple mediator model was employed (Hayes, 2013, p. 125). In particular, t he serial multiple mediator model was used to X on Y while modeling a process in which X causes M 1 , which in turn causes M 2 , and so forth, concluding with Y as the final consequent (Hayes, 2013, p. 144). The specific data analysis techniques used to add ress the 4 research questions are provided below. Research Question 1: Do the Antecedents Predict Perceived Efficacy ? Research question 1 focused on examining if the antecedents (demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) predicted perceived efficacy ( self efficacy and response efficacy ). To test research question 1, a series of multiple regressions were conducted using SPSS. Specifically, each perceived efficacy variable was treated as a dependent variable in the multiple regression s . Thus, t he analysis of research question 1 provided the total effects of the antecedents (independent variables) on perceived efficacy (second set of mediating variables). Research Question 2: Does Perceived Risk Mediate the Relationship between the Antecedents a nd Perceived Efficacy ? Research question 2 focused on examining if perceived risk (perceived vulnerability, affective risk perceptions, and perceived severity) mediated the relationship between the antecedents (demographic factors, international travel spe cific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) and perceived efficacy (self efficacy and response efficacy). To test research question 2, a series of multiple regressions were conducted using SPSS.

PAGE 89

89 In the first series of multiple regressions, the relationships between the antecedents and perceived risk were examined. Thus, t he first series of multiple regressions provided the indirect paths between the anteced ents ( independent variables ) and perceived risk (first set of mediating variables). In the second series of multiple regressions, the relationships between perceived risk and perceived efficacy were examined. T hus, t he second series of multiple regressions provided the indirect paths between perceived risk ( first set of mediating variables) and perceived efficacy ( second set of mediating variables). If there were significant indirect paths in both the first and second series of multiple regressions , a thir d series of multiple regressions was conducted. In the third series of multiple regressions, the significant antecedents from the first series of multiple regressions and the significant perceived risk variables from the second series of multiple regressio ns were treated as independent variables. P erceived efficacy was treated as the dependent variable . T hus, t he third series of multiple regressions provided the direct effect of the significant antecedents (independent variables) and perceived risk (first s et of mediating variables) on perceived efficacy (second set of mediating variables). In addition to the regression based path analysis , if mediation was found, an online calculation of the Sobel test was conducted (Preacher, 2015) to determine the amount of mediation (Sobel, 1982) . Resea rch Question 3: Does Perceived Risk Predict Intentions to Engage in a Recommended Risk Reduction Behavior? Research question 3 focused on examining if perceived risk (perceived vulnerability, affect ive risk perceptions, and perceived severity) predicted intentions to engage in a recommended risk reduction behavior. To test research question 3, multiple

PAGE 90

90 regression was conducted using SPSS. The analysis of research question 3 provided the total effects of perceived risk (fir st set of mediating variables) on intentions to engage in a recommended risk reduction behavior (dependent variable). Research Question 4: Does Perceived Efficacy Mediate the Relationship between Perceived Risk and Intentions to Enga ge in a Recommended Risk Reduction Behavior? Research question 4 focused on examining if perceived efficacy (self efficacy and response efficacy) mediated the relationship between perceived risk (perceived vulnerability, affective risk perceptions, and per ceived severity) and intentions to engage in a recommended risk reduction behavior. To test research question 4 , a series of multiple regressions were conducted using SPSS. In the first series of multiple regressions, the relationships between perceived r isk and perceived efficacy were examined. T hus, t he first series of multiple regressions provided the indirect paths between perceived risk ( first set of mediating variables) and perceived efficacy ( second set of mediating variables). In the second series of multiple regressions, the relationships between perceived efficacy and intentions to engage in a recommended risk reduction behavior were examined. Thus, t he second series of multiple regressions provided the indirect paths between perceived efficacy ( s econd set of mediating variables) and intentions to engage in a recommended risk reduction behavior ( dependent variable ). If there were significant indirect paths in both the first and second series of multiple regressions , a third series of multiple regr essions was conducted. In the third series of multiple regressions, the significant perceived risk variables from the first series of multiple regressions and the significant perceived efficacy variables from the second series of multiple regressions were treated as independent variables. Intentions

PAGE 91

91 to engage in a recommended risk reduction behavior was treated as the dependent variable . Thus, the third series of multiple regressions provided the direct effect of the significant perceived risk variables ( fi rst set of mediating variables) and perceived efficacy ( second set of mediating variables) on intentions to engage in a recommended risk reduction behavior ( dependent variable ) . In addition to the regression based path analysis , if mediation was found, an online calculation of the Sobel test was conducted (Preacher, 2015) to determine the amount of mediation (Sobel, 1982) . Preparation for Data Analysis Based on the screening questions for international travel in the past 5 years and country level destinatio n past travel experience, a total of 521 respondents were eligible to complete the survey instrument for those without country level destination past travel experience . Based on the screening questions for international travel in the past 5 years and count ry level destination past travel experience, a total of 838 respondents were eligible to complete the survey instrument for those with country level destination past travel experience . Thus, a combined total of 1,3 59 respondents were qualified to participa te in the study based on the screening questions . In addition to ensuring that respondents represented the population in terms of international past travel experience and country level destination past travel experience , a question was also included withi n the survey instrument s to ensure sure that respondents were paying attention and reading the instructions provided . For the survey instrument for those without country level destination past travel experience, a total of 35 respondents in correctly answer ed the attention check question. For the survey instrument for those with country level destination past travel experience, a total of 117 respondents in correctly answered the attention check question. Accordingly, a

PAGE 92

92 combined total of 152 respondents who incorrectly answered the attention check question were deemed to be in eligible for participation and were not permitted to complete the remainder of the respective survey instruments. Consequently , th e responses of those individuals who answered the atten tion check question incorrectly were deleted from the respective data set s . The combined sample size was 1,207 following the preparation for data analysis based on the attention check question. Finally, given that 2 survey instruments were launched for 2 different samples (i.e., those without country level destination past travel experience and those with country level destination past travel experience), the MTurk IDs of respondents for both data sets were compared in Excel to ensure that the same individ ual did not complete both survey instruments. A total of 34 respondents entered their MTurk ID at the end of both survey instruments . T herefore, respondents who completed both survey instruments were deemed to be in eligible for participation and their resp onses were deleted from both data sets. As a result of this stage of the preparation for data analysis , a combined total of 1,173 respondents were qualified for participation in this study. Following the aforementioned stages of the preparation for data an alysis , the 2 data sets (survey instrument for those without country level destination past travel experience and survey instrument for those with country level destination past travel experience) were merged in SPSS. The merged data set was utilized for a ll subsequent analyses in SPSS . Missing Data Analysis Following the initial stages of the preparation for data analysis , missing data analysis was performed in an effort to prepare the data for analysis of the research questions and associated hypotheses . An initial descriptive analysis revealed that there

PAGE 93

93 w ere a total of 21 missing values. The extent of missing data was analyzed in 2 way s . First, the number and percentage of cases with missing data was analyzed for each variable that had missing data (Hair et al., 2010, p. 47). This m issing value analysis revealed 1 missing response for the assured worried semantic differential scale of affective risk perceptions ( 0.1% ) , 3 missing responses for gender ( 0.3% ) , 2 missing responses for age (0.2% ) , 2 missing responses for race/ethnicity ( 0.2% ) , 4 missing responses for education level ( 0.3% ) , 6 missing responses for marital status ( 0.5%), and 3 missing responses for presence of c hildren in the household ( 0.3% ). Second, the percentage of variables with missing data was analyzed for each case that had missing data (Hair et al., 2010, p. 47 ). This missing data analysis revealed that 18 cases had at least 1 missing response. More spec ifically, 16 cases had 1 missing response (1.39%), 1 case had 2 missing responses (2.78%), and 1 case had 3 missing responses (4.17%). T he results of the two step missing data analysis were compared to (2010) rule of thumb for the extent of missing data (p. 47) . Specifically, given that less than 10% of data was missing for an individual case and that the number of cases with no missing data was large enough for analysis , the cases with missing data were not deleted from the data set (Hair et al., 2010) . (Hair et al., 2010, p. 47). Consequently, the missing responses were coded as 99 system missing in SPSS . Si nce the extent of missing data was small (less than 10%), any method could be used to treat the missing data ( Hair et al., 2010, p. 47 ). Given that t he exclude cases listwise option is the least preferred method (Hair et al., 2010, p. 56) , the missing data

PAGE 94

94 was treated using the exclude cases pa irwise option in SPSS . Thus, the cases with missing data were only excluded when the variable with missing data was included in the analysis (Pallant, 2007, p. 57). Accordingly , the number of valid responses remained 1,1 73 after this stage of the preparation for data analysis . Outlier Analysis Following the missing data analysis , univariate outlier analysis was performed in an effort to prepare for data analysis . First, boxplots were reviewed for each Likert type scale measure to identify potential outliers (Pallant, 2007, p. 63) . Second, any case that was considered an outlier in the boxplot ( i.e., greater than 1.5 box lengths from the edge of the box; Pallant, 2007, p. 63) was examined to determine if the case fell outside of the range of 4 standard deviations (based on Hair et al., 2010, p. 67) . According to this rule of thumb, an outlier would be considered a case that is 4 or more standard deviations from the mean for studies with large sample sizes (based on Hair et al., 2010, p. 67). Based on this two step analysis of boxplots and standard deviations , a total of 44 responses were identified as being potential outliers. Further analysis was conducted to profile the potential outliers (Hair et al., 2010, p. 67). This analysis revealed that the 44 outlier responses were associated with 27 cases. Any case with 2 (n= 7), 3 (n= 3), or 5 (n=1) outlier responses w as examined further. For example , the data was assessed to determine if there were distinct patterns within each 1 for all items). Of the 11 cases with multipl e potential outlier responses, 7 cases were deemed a s outliers and were consequently deleted from the data set. The 2 0 other certain number of observations may occur normally in these outer ranges of the

PAGE 95

95 ir et al., 2010, p. 67). Therefore, only the cases that included truly distinctive responses were identified and treated as outliers (based on Hair et al., 2010, p. 67). As a result of this stage of the preparation for data analysis , the number of valid re sponses was 1,1 66 . After the identified outliers were deleted from the data set, the mean score for each variable was compared with the 5% trimmed mean to assess whether extreme values were greatly influencing the mean (Pallant, 2007, p. 59). For each of t he Likert type scale measures, the difference between the mean and the 5% trimmed mean was 0.11 or less . Accordingly, t his analysis indicated that the extreme values were not greatly influencing the mean (based on Pallant, 2007, p. 59). M ultivariate outlie rs were also detected through preliminary regression analysi s . Preliminary descriptive analysis revealed that the bottom 3 items for intentions to engage in a recommended risk reduction behavior were to register with the U.S. Department of State, purchase travel insurance prior to travel, and avoid traveling outside of the hotel zone . The decision was made to focus on the lowest ranking items (justification provided in the descriptive analysis of the dependent variable). Accordingly, preliminary regression analysis was employed for these three dependent variables. Mahalanobis D 2 was used to detect multivariate outliers (Pallant, 2007, p. 157) . A case is considered to be a multivariate outlier if the probability associated with its D 2 is 0.001 or less (Tabac hnick & Fidell, 20 13 , p. 74 ). Therefore, the probability of D 2 was computed in SPSS. Based on this analysis, 19 cases were identified as multivariate outliers for the dependent variable of intentions to register with the U.S. Department of State prior to t ravel, 9 cases were identified as multivariate outliers for the dependent

PAGE 96

96 variable of intentions to purchase travel insurance prior to travel, and 12 cases were identified as multivariate outliers for the dependent variable of intentions to avoid traveling outside of the hotel zone. Thus, in total, 40 cases were identified as being multivariate outliers because the probability of D 2 was less than 0.001 (based on Tabachnick & Fidell, 20 13 , p. 74 ) . Consequently, the 40 cases were deleted from the data set. Ac cordingly , the final number of valid responses for this study was 1,126. Assessing Normality Following the outlier analysis , the univariate normality of the data was assessed in an effort to prepare for data analysis . First, skewness and kurtosis values were examined to assess the distribution of the scores (Pallant, 2007, p. 56 ). T he skewness statistic was within the acceptable range of 2 to +2 (George & Mallery, 2010) for each of the Likert type scale measures . The k urtosis statistic was also within the acceptable range of 2 to +2 (George & Mallery, 2010) for a majority of the Likert type scale measures . T he kurtosis statistic was outside of the acceptable range of 2 to +2 (George & Mallery, 2010) for the following 6 measures: self efficacy item Avoid displaying indicators of wealth (kurtosis= 2. 35 ); self efficacy item Avoid carrying large amounts of cash (kurtosis= 3.2 2 ); self efficacy item Search for more information online about how to stay safe prior to travel (kurtosis= 2.9 0 ); intentions to engage in a recommended risk reduction behavior item Avoid displaying indicators of wealth (kurtosis= 2. 27 ); intentions to engage in a recommended risk reduction behavior item Maintain awareness of your surroundings (kurt osis= 2. 2 6 ); intentions to engage in a recommended risk reduction behavior item Avoid carrying la rge amounts of cash (kurtosis= 2.97 ). Thus, t he kurtosis statistics for these 6 items indicate d a peaked or leptokurtic distribution (based on Hair et al., 20 10, p. 71) . However, Tabachnick and

PAGE 97

97 Fidell (20 13 ) suggest ed that the risk of kurtosis resulting in an underestimation of the variance is reduced when the sample size is larger than 200 (p. 80). Further, tests used to assess kurtosis statistics are considered to be highly sensitive among large samples (Pallant, 2007, p. 56). The Kolmogorov Smirnov statistic was also examined in an effort to assess normality (Pallant, 2007, p. 62). The significanc e value of the Kolmogorov Smirnov test was p < 0.001 for each of the Likert type scale measures, which suggests that the assumption of normality was violated (Pallant, 2007, p. 62) because the distribution of each variable was significantly different from a normal distribution (Hair et al., 2010, p. 73) . The Kolmogorov Smirnov test, however, is considered to be highly sensitive among samples of greater than 1,000 cases (Hair et al., 2010, p. 74). Further, the normality of the distribution of the independent and dependent variables is not a required assumption for linear regression (Draper & Smith, 1998). Rather, the assumption of normality is related to the distribution of the error term for linear regression (Hair et al., 2010, p. 182). T he sample size shou ld also be considered when determining the severity of a non normal distribution (Hair et al., 2010, p. 71). Specifically, the negative influences of non normality may be reduced or negligible when t he sample size is large ( n= 200+; Hair et al., 2010, p. 7 as the sample sizes become large, the researcher can be less concerned about nonnormal variables , except as they might lead to other assumption violations that do have an impact in other ways (p. 7 2). For example, the non normal distribution of a variable may potentially lead to a violation of the heteroscedasticity assumption of linear

PAGE 98

98 regression (Hair et al., p. 77). Nevertheless , the assumptions of the individual tests were examined pr ior t o performing statistical analyses. I ndependent Variable: International Travel Specific Psychological Factors Following the normality analysis , factor analysis was performed on the multi item measures of international travel specific psychological factors and perceived risk in an effort to prepare for data analysis . The conceptual model for this study proposed a 2 factor model of international travel specific psychological factors. The 2 factors were international travel risk perceptions and safety concerns. To examine whether the 2 factor model of international travel specific psychological factors was supported, the 5 items underwent correlation analysis, principal component s analysis [PCA] , and reliability tests using SPSS. First, the correlation coefficients, Kaiser Meyer Olkin were examined to evaluate the suitability of the data for PCA (Pallant, 2007, p. 181). When all 5 items were entered together, a review o f the correlation coefficients indica ted low correlations between each of the 3 international travel risk perceptions items and the 2 safety concerns items (ranging from 0.1 8 0.3 5 ). Thus, given that a majority of the coefficients were less than 0.3, PCA of all 5 items was deemed to be in appropriate (based on Tabachnick & Fidell, 20 13 , p. 619 ). Accordingly, separate analyses were performed for the 3 international travel risk perception items and the 2 safety concern s items. International travel risk percept ions The 3 items which were used to measure international travel risk perceptions were subjected to PCA. The results of the PCA are presented in Table 3 9 . Prior to performing PCA, the suitability of the data for factor analysis was assessed. Inspection

PAGE 99

99 of the correlation matrix revealed high pair wise coefficients between each of the 3 items (ranging from 0.70 significance at the p < 0.001 l evel and the KMO value was 0.72 , supporting t he factorabil ity of the correlation matrix (based on Pallant, 2007, p. 190) . Next , PCA was conducted to transform the items into a reduced set of linear combinations, while using all of the variance in the items (Pallant, 2007, p. 180). PCA was applied with a Varimax rotation to verify the construct validity. PCA revealed the presence of 1 component with an initial eigenvalue exceeding 1.0 , explaining 84.8 6 % of the variance. A review of the component matrix indicated that the 3 items loaded strongly (0.8 8 0.9 5 ) on the 1 component. Table 3 9 . Underlying dimensions of international travel risk perceptions . Items Component 1 I feel nervous about traveling internationally right now 0.9 5 Traveling internationally is risky right now 0. 93 I would feel very uncomfortable traveling internationally right now 0. 8 8 Kaiser Meyer Olkin Measure of Sampling Adequacy 0.72 2 = 2494.05 , df=3, p < 0 .001 Eigenvalue 2.5 5 % variance explained 84.8 6 Mean 2.31 0.91 Note: Component 1= international travel risk perceptions T coefficient was examined to determine the internal consistency of the scale (Pallant, 0.91 for the international travel risk perceptions c omponent, which demonstrated that the scale had adequate reliability ( values for each dimension was greater than 0.7 ; Nunnally, 1978 ). Finally, the mean score of the component, as an alternative to the component score, was computed and

PAGE 100

100 treated as an individual independent variable in the analysis of the research questions and associated hypotheses. Safety concerns The 2 items which were used to measure safety concerns were subjected to PCA. Prior to performing PCA, the suitability of the data for factor analysis was assessed. Inspection of the correlation matrix revealed adequately high pair wise coefficients between the 2 items ( 0.6 2 sign ificance at the p < 0.001 level, however, the KMO valu e was 0. 50. Given that the KMO value was less than the suggested minimum of 0.6 , the data did not support the factorability of the correlation matrix (based on Pallant, 2007, p. 190) . Therefore, the 2 items which were used to measure safety concerns were e ntered independently in the analysis of the research questions and associated hypotheses . Mediating Variable: Perceived Risk The conceptual model for this study proposed a 3 factor model of perceived risk. The 3 factors were affective risk perceptions, pe rceived severity, and perceived vulnera bility. As previously mentioned, PCA was used for two main reasons. First, PMT does not include affective risk perceptions as part of the threat appraisal process. Therefore, the 3 factor model of perceived risk has n ot been examined in PMT studies. Second, the risk as feeling hypothesis suggests that cognitive and affective risk perceptions can either have direct influences on behavior or these types of risk perceptions can be related to one another (Loewenstein et al ., 2001). Thus, the risk as feelings hypothesis does not suggest that the three factor model of perceived risk will be supported in every setting. Considering these two reasons, PCA was considered to be appropriate for this study. Accordingly, t o examine whether the 3 factor model of

PAGE 101

101 perceived risk was supported, the 9 items used to measure perceived risk were subjected to PCA. The results of the PCA are presented in Table 3 1 0 . Prior to performing PCA, the suitability of the data for factor analysis was a ssessed. Inspection of the correlation matrix revealed relatively high pair wise coefficients between most of level and the KMO value was 0.8 3 , supporting the factor ability of the correlation matrix (based on Pallant, 2007, p. 190) . Next , PCA was applied with a Varimax rotation to verify the construct validity. PCA revealed the presence of 2 components with an initial eigenvalue exceeding 1.0, explaining 46.57 % and 2 1. 83 % of the variance, respectively. V arimax rotation, which converged in 3 iterations, provided the best defined factor structure. A review of the rotated component matrix indicated that 6 items loaded strongly (0.75 0.80) on component 1 , while 3 items lo aded strongly (0.88 0. 90 ) on component 2 . Specifically, the 3 items which were used to measure perceived vulnerability and the 3 items which were used to measure affective risk perceptions loaded on component 1 . Further, the 3 items which were used to meas ure perceived severity loaded on component 2 . Accordingly, component 1 was labeled perceived vulnerability/affective risk perceptions and component 2 was labeled perceived severity. The reliability of the 2 component s alpha 0.8 8 for both component 1 and component 2, which demonstrated that the scale s for perceived vulnerability/affective risk perceptions and perceived severity had adequate reliability

PAGE 102

102 each dimension was greater than 0.7; Nunnally, 1978). The mean scores of the components, as an alternative to the component score, were computed and treated as individual mediating variables in the analysis of the research questions and associated hypotheses. Following the computation of the mean score for component 1 (perceived vulnerability/affective risk perceptions) , the case with missing data on the assured worried sem antic differential scale of affective risk perceptions was coded as 99 system missing in SPSS. Table 3 1 0 . Underlying dimensions of perceived risk . Items Component 1 Component 2 Communalities My chances of being a victim of crime while visiting Acapulco are high 0.80 0.67 Relaxed anxious 0.80 0.64 A ssured worried 0.78 0.61 It is likely that I will be a victim of crime while visiting Acapulco 0.77 0.63 I am at risk for being a victim of crime while visiting Acapulco 0.76 0.63 F earless fearful 0.75 0.57 It would have a serious negative impact on me if I were a victim of crime while visiting Acapulco 0.90 0.82 If I were a victim of crime while visiting Acapulco, I would experience serious negative consequences 0.88 0.80 If I were a victim of crime while visiting Acapulco, it would be harmful to my well being 0.88 0.79 Kaiser Meyer Olkin Measure of Sampling Adequacy 0.8 3 2 = 5636.24 , df=3 6 , p < 0 .001 Eigenvalue 4.19 1.96 % variance explained 46.57 21. 83 Cumulative variance explained 46.57 68. 39 Mean 2.68 4.07 0.8 8 0.8 8 Note: Component 1= perceived vulnerability/affective risk perceptions; Component 2= perceived severity

PAGE 103

103 Descriptive Analysis Independent Variable: Demographic Factors Descriptive analysis of t he demographic factors is provided in Table 3 1 1 . There was a fairly even distribution of female (46. 9 %) and male (53. 1 %) respondents . The largest age group was 25 34 (47. 3 %), followed by 35 49 (27. 3 %) , 1 8 24 (15. 0 %) . Only approximately one in ten respondents were between the ages of 50 64 (9. 3 %) and 65 or older ( 1. 0%) . The largest household income groups were $50,000 to $74,999 (24. 7 %), $35,000 to $49,999 (16.2%), $75,000 to $99,999 ( 15.7 % ), and $25,000 to $34,999 (13. 1 %). Fewer than one in ten respondents were in each of the following household income groups: $15,000 to $24,999 (9. 6 %), $100,000 to $124,999 (9.5%), $125,000 or over (7.0%), and under $15,000 (4.3%). A pproximately three in fo ur respondents considered themselve s to be White/Caucasian ( 78.5 %) , followed by Asian (7. 5 %) , Hispanic ( 5.2 %), Black/African American (5. 2 %), multi ethnic/mixed race (2.8%), American Indian/Alaska Native (0. 4 %), Native Hawaiian/Pacific Islander (0. 4 %), and Other (0.1%) . Approximately two in approximately one in five respondents had some college credit but no degree (22.5%). An additional one in Associate degree (10.5%). A minority of respondents had a High school diploma (8.3%), professional degree (2.6%), Doctoral degree (1.3%), or some high school credit but no degree (0.6%). Life stage was measured by marital status and presence of children in the household. A majority of the sample had either never been married (47.9%) or was currently married (44.3%). A minority of respondents had either been divorced (5.8%), separated (1.4%), or widowed (0.5%). A pproximately three in five respondents did not

PAGE 104

104 have any children (younger than the age of 18) currently living in their household (64.2%). Of those with children currently living in the household, the largest groups were those with either 1 child (15.5%) o r 2 children (13.9%) currently living in their home. A minority of respondents had either 3 (4.8%), 4 (1.3%), or 5 or more (0.3%) children currently living in their home. Table 3 1 1 . Descriptive analysis of demographic factors . Independent variable Fre quency Valid % Gender (n= 1,123) Female 527 46.9 Male 596 53.1 Age (n= 1,124) 18 24 169 15.0 25 34 532 47.3 35 49 307 27.3 50 64 105 9.3 65 or older 11 1.0 Household income (n= 1,126) Under $15,000 48 4.3 $15,000 to $24,999 108 9.6 $25,000 to $34,999 147 13.1 $35,000 to $49,999 182 16.2 $50,000 to $74,999 278 24.7 $75,000 to $99,999 177 15.7 $100,000 to $124,999 107 9.5 $125,000 or over 79 7.0 Race/ethnicity (n= 1,124) White/Caucasian 882 78.5 Black/African American 59 5.2 Asian 84 7.5 Native Hawaiian/Pacific Islander 4 0.4 Hispanic 58 5.2 American Indian/Alaska Native 5 0.4 Multi ethnic/mixed race 31 2.8 Other 1 0.1 Education level (n= 1,122) Some high school credit, no degree 7 0.6 High School Graduate 93 8.3 Some college credit, no degree 252 22.5 Associate degree 118 10.5 484 43.1 124 11.1 Professional degree 29 2.6 Doctoral Degree 15 1.3

PAGE 105

105 Table 3 11 . Continued. Independent variable Frequency Valid % Marital status (n= 1,120) Now married 496 44.3 Separated 16 1.4 Widowed 6 0.5 Divorced 65 5.8 Never married 537 47.9 Presence of children in the household (n= 1,123) 0 721 64.2 1 174 15.5 2 156 13.9 3 54 4.8 4 15 1.3 5 or more 3 0.3 Independent Variable: International Travel Specific Psychological Factors Descriptive analysis of the international travel risk perceptions is provided in Table 3 1 2 . Based on PCA, this independent variable was measured as the mean score of the 3 internati onal travel risk perceptions items. The descriptive analysis revealed that respondents generally did not agree that international travel is risky, as well as that international travel makes them feel nervous and uncomfortable ( mean = 2.31). Table 3 1 2 . De scriptive analysis of international travel risk perceptions . Independent variable Mean St. deviation International travel risk perceptions (n= 1,1 26 ) 2.31 1.0 6 Descriptive analysis of safety concerns is provided in Table 3 1 3 . Overall, the respondents agreed that the safety of an international destination is a concern. mean = 4.2 3 ) . Approximately one in t wo respondents strongly agreed with this statement ( 44.8 %), while approximately two in five respondents agreed with this statement (37. 7 %). Fewer than one in twenty respondents either disagreed (3. 5 %) or strongly disagreed (0. 6 %) with this statement. Respo

PAGE 106

106 mean = 3.7 4 ). In particular, approximately one in three respondents agreed with this statement ( 35.5 %) and approximately three in ten respondents strongly agreed with this statement (27. 2 %). Approximately one in five respondents indicated a neutral response to this statement ( 24.2 %). Table 3 1 3 . Descriptive analysis of safety concerns . Independent variable 1 2 3 4 5 Mean St. deviation Safety concerns (n= 1,1 26 ) Safety is a serious consideration when choosing an international tourist destination 0.6 3.5 13.4 37.7 44.8 4.23 0.85 Safety is the most important attribute an international destination can offer 2.8 10.2 24.2 35.5 27.2 3.74 1.05 Independent Variable: Destination Specific Psychological Factors Descriptive analysis of destination image is provided in Table 3 1 4 . Respondents had a relatively positive image of Acapulco as a destination ( mean = 3.67). Specifically, approximately two in five respondents indicated a positive image of the destination ( 40.2 %) and approximately three in ten had a neutral image of Acapulco as a tourist destination (31. 7 %). Approximately one in five respondents had a very positive image of Acapulco as a tourist destination ( 18.8 %), while fewer than one in ten respondents had a negative (7. 7 %) or very negative (1. 5 %) image of Acapulco as a tourist destination. Table 3 1 4 . Descriptive analysis of destination specific psychological factors . Independent variable 1 2 3 4 5 Mean St. deviation Destination image (n= 1,1 26 ) 1.5 7.7 31.7 40.2 18.8 3.67 0.9 2 Independent Variable: Destination Specific Factors Descriptive analysis of presence of friends and /or relatives living in the destination and fluency in the native language of the destination is provided in Table 3 -

PAGE 107

107 1 5 . An overwhelming majority of respondents did not have friends and/or relatives living in Mexico ( 86.7 %) , while approximately one in eight r espondents had friends and/or relatives living in the destination (13.3%) . Approximately one in three respondents indicated that they were not fluent in Spanish ( 33.8 %) , the native language of the destination. Approximately one in two respondents indicate d that they speak Spanish, but not well ( 50.7 %). Fewer than one in five respondents consider themselves to be fluent ( 11.5 %) or very fluent ( 3.9 %) in Spanish, the native language of the destination. Table 3 1 5 . Descriptive analysis of destination specific factors . Independent variable Frequency Valid % Presence of friends/relatives living in the destination (n= 1,1 26 ) No 976 86.7 Yes 150 13.3 Fluency in the native language of the destination (n= 1,1 26 ) Not at all 381 33.8 Not well 571 50.7 Well 130 11.5 Very well 44 3.9 Independent Variable: Past Travel Experience Descriptive analysis of destination past travel experience is provided in Table 3 1 6 . As specified by the 60/40 proportional quota for country level destination past travel experience, approximately three in five respondents had taken a trip to Mexico ( 59.3 %) and approximately two in five respondents had never taken a trip to Mexico before (40. 7 %). D estination past travel experience was also measured at the city level. An overwhelming majority of respondents had never visited Acapulco before ( 87.7 %). Only approximately one in seven respondents had visited Acapulco before ( 12.3 %).

PAGE 108

108 Table 3 1 6 . Descriptive analysis of past travel experience . Independent variable Frequency Valid % Country level destination past travel experience (n= 1,1 26 ) No 458 40.7 Yes 668 59.3 City level destination past travel experience (n= 1,1 26 ) No 987 87.7 Yes 139 12.3 Mediating Variable: Perceived Risk Descriptive analysis of the perceived risk components is provided in Table 3 1 7 . Based on PCA, this mediating variable was measured for the 2 components of perceived vulnerability/affective risk perceptions and perceived severity. Specifically, component 1 was measured by the mean score of the 3 perceived vulnerability items and the 3 affective risk perceptions items. The descriptive analysis revealed that respondents generally did not agree that they were likely to be a victim of crime while visiting Acapulco and they did not indicate feelings of anxiety, fear, and worry when thinking about the ir personal safety while visiting the destination ( mean = 2.68). C omponent 2 was measured by the mean score of the 3 perceived severity items. Overall, the descriptive analysis revealed that respondents agreed that there would be serious negative consequenc es for themselves as individuals if they were a victim of crime while visiting Acapulco ( mean = 4.0 7 ). Thus, the descriptive analysis of perceived risk revealed that respondents agreed that the threat was serious, but did not believe it was likely to happen to them. Table 3 1 7 . Descriptive analysis of perceived risk . Mediating variable Mean St. deviation Perceived vulnerability/affective risk perceptions (n= 1,1 25 ) 2.68 0.7 8 Perceived severity (n= 1,1 26 ) 4.0 7 0.8 4

PAGE 109

109 Mediating Variable: Perceived Efficacy Descriptive analysis of self efficacy is provided in Table 3 1 8 . The mean scores ranged from 3.96 4.48 for 12 recommended risk reduction behavior s , indicating that respondents were confident in their ability to perform approximately two thirds of the recommended risk reduction behaviors to ensure their personal safety while visiting Acapulco. The top 3 recommended risk reduction behaviors in terms of self e fficacy were to avoid carrying large amounts of cash ( mean = 4.4 8 ), avoid displaying indicators of wealth ( mean = 4.4 4 ), and search for more information online about how to stay safe prior to travel ( mean = 4. 4 4 ). Following the top 3 recommended risk reduction behaviors, the behaviors that respondents indicated that they were confident in th eir ability to perform to ensure their personal safety while visiting Acapulco were to maintain awareness of their surroundings ( mean = 4.37), avoid walking around the destination by themselves at night ( mean = 4.3 3 ), avoid walking around the destination by themselves ( mean = 4.21), avoid using ATMs if they are not affiliated with a recognized international chain ( mean = 4.2 1 ), avoid situations in which they may be isolated or stand out as a potential victim ( mean = 4.18), avoid using services like Uber ( mean = 4 .15), avoid drawing attention to themselves ( mean = 4.12), only use recommended taxi companies ( mean = 4.1 1 ), and remain in tourist areas ( mean = 3.9 6 ) . The mean scores of 3 self efficacy items ranged from 3. 67 3. 78 , indicating that respondents were neutral in their ability to perform these recommended risk reduction behaviors to ensure their personal safety while visiting Acapulco. The bottom 3 recommended risk reduction behaviors in terms of self efficacy were to register with the U.S. Department of State p rior to travel (mean= 3.67), avoid traveling outside of the hotel zone (mean= 3.73), and purchase travel insurance prior to travel (mean= 3.78).

PAGE 110

110 None of the self efficacy items had a mean score below 3.0, indicating that respondents were neither unconfiden t nor very unconfident in their ability to perform any of the reduction behaviors to ensure their personal safety while visiting Acapulco. Table 3 1 8 . Descriptive analysis of self efficacy . Mediating variable 1 2 3 4 5 Mean St. deviation Self efficacy (n= 1,1 26 ) Avoid carrying large amounts of cash 1.0 2.0 8.4 25.2 63.4 4.48 0.81 Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras) 0.9 2.3 10.0 25.8 61.0 4.44 0.83 Search for more information online about how to stay safe prior to travel 1.4 2.4 9.4 24.8 62.0 4.44 0.87 Maintain awareness of your surroundings 0.7 2.0 10.6 33.2 53.5 4.37 0.81 Avoid walking around the destination by yourself at night 1.2 3.9 11.6 27.8 55.5 4.33 0.91 Avoid walking around the destination by yourself 1.9 5.2 14.2 28.1 50.7 4.21 0.99 Avoid using ATMs if they are not affiliated with a recognized international chain 1.4 5.6 13.9 28.3 50.7 4.21 0.98 Avoid situations in which you may be isolated or stand out as a potential victim 1.1 4.3 16.0 33.4 45.3 4.18 0.92 Avoid using services like Uber 2.4 7.3 16.0 21.1 53.2 4.15 1.08 Avoid drawing attention to yourself 0.8 5.2 17.5 34.2 42.3 4.12 0.93 Only use recommended taxi companies 1.1 5.8 16.7 33.9 42.5 4.11 0.95 Remain in tourist areas 2.8 7.6 18.9 32.2 38.4 3.96 1.07 Purchase travel insurance prior to travel 5.5 11.3 20.6 25.2 37.4 3.78 1.22 Avoid traveling outside of the hotel zone 4.3 12.3 22.1 28.8 32.6 3.73 1.16 Register with the U.S. Department of State prior to travel 5.8 14.5 21.2 23.6 34.9 3.67 1.25

PAGE 111

111 Descriptive analysis of response efficacy is provided in Table 3 19 . The mean scores ranged from 3.91 4.4 3 for 12 recommended risk reduction behavior s , indicating that respondents believed that approximately two thirds of the recommended risk reduction behaviors would be effective in ensuring their personal safety while visiting Acapulco. The top 3 recommended risk reduction behaviors in terms of respon se efficacy were to avoid walking around the destination by themselves at night ( mean= 4.4 3 ), maintain awareness of their surroundings ( mean= 4.41), and avoid situations in which they may be isolated or stand out as a potential victim ( mean= 4.35). Followi ng the top 3 recommended risk reduction behaviors, the behaviors that respondents indicated that they believed were effective in ensuring their personal safety while visiting the destination were to avoid displaying indicators of wealth ( mean= 4.32), avoid walking around the destination by themselves ( mean= 4.2 9 ), avoid carrying large amounts of cash ( mean= 4.22), avoid drawing attention to themselves ( mean= 4.16), avoid using ATMs if they are not affiliated with a recognized international chain ( mean= 4.0 3 ), only use recommended taxi companies ( mean= 3.9 8 ), remain in tourist areas ( mean= 3.9 6 ), avoid traveling outside of the hotel zone ( mean= 3.9 5 ), and search for more information online about how to stay safe prior to travel ( mean= 3.91). The mean scores of 3 response efficacy ranged from 3.13 3. 66, indicating that respondents were neutral in their beliefs that these recommended risk reduction behaviors would be effective in ensuring their personal safety while visiting Acapulco. The bottom 3 recommended risk reduction behaviors in terms of response efficacy were to register with the U.S. Department of State prior to travel ( mean= 3.13), purchase travel insurance prior to travel ( mean= 3.18), and avoid using services like Uber ( mean=

PAGE 112

112 3.66). None of the response efficacy items had a mean score below 3.0, indicating that respondents did not believe that any of the risk reduction behaviors were ineffective or very ineffective in ensuring their personal safety while visiting Acapulco. Table 3 19 . Descriptive analysis of response efficacy . Mediating variable 1 2 3 4 5 Mean St. deviation Response efficacy (n= 1,1 26 ) Avoid walking around the destination by yourself at night 0.3 2.2 9.9 29.5 58.2 4.43 0.78 Maintain awareness of your surroundings 0.4 2.3 10.0 30.6 56.7 4.41 0.79 Avoid situations in which you may be isolated or stand out as a potential victim 0.4 2.9 10.5 33.3 52.8 4.35 0.82 Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras) 0.4 3.9 11.3 32.5 52.0 4.32 0.85 Avoid walking around the destination by yourself 0.6 2.8 13.1 33.7 49.7 4.29 0.85 Avoid carrying large amounts of cash 1.1 4.8 13.8 31.9 48.5 4.22 0.93 Avoid drawing attention to yourself 0.5 4.8 15.9 35.6 43.2 4.16 0.90 Avoid using ATMs if they are not affiliated with a recognized international chain 1.3 5.9 19.8 34.8 38.2 4.03 0.97 Only use recommended taxi companies 1.2 5.0 22.1 38.7 33.0 3.98 0.92 Remain in tourist areas 1.4 8.0 19.2 36.1 35.3 3.96 1.00 Avoid traveling outside of the hotel zone 1.9 7.2 20.6 34.5 35.8 3.95 1.01 Search for more information online about how to stay safe prior to travel 1.5 7.9 22.9 33.1 34.5 3.91 1.01 Avoid using services like Uber 4.0 10.7 29.2 27.3 28.8 3.66 1.12 Purchase travel insurance prior to travel 11.6 20.9 25.3 22.3 19.9 3.18 1.29

PAGE 113

113 Table 3 19 . Continued . Mediating variable 1 2 3 4 5 Mean St. deviation Register with the U.S. Department of State prior to travel 11.7 20.2 29.7 20.2 18.3 3.13 1.26 Dependent Variable: Intentions to E ngage in a Recommended Risk Reduction Behavior Descriptive analysis of intentions to engage in a recommended risk reduction behavior is provided in Table 3 2 0 . The mean score s ranged from 4.0 4 4.56 for 11 recommended risk reductio n behavior s , indicating that respondents we re likely to engage in a majority of the behaviors to ensure their personal safety while visiting Acapulco. The top 3 recommended risk reduction behaviors that respondents indicated that they were most likely to e ngage in were to maintain awareness of their surroundings ( mean= 4.56 ), avoid carrying large amounts of cash ( mean= 4.54), and avoid displaying indicators of wealth ( mean= 4.53). Following the top 3 recommended risk reduction behaviors, the behaviors that respondents indicated that they were likely to engage in were to avoid situations in which they may be isolated or stand out as a potential victim ( mean= 4.3 9 ), avoid walking around the destination by themselves at night ( mean= 4.3 7 ), avoid drawing attention to themselves ( mean= 4.35), avoid using ATMs if they are not affiliated with a recognized international chain ( mean= 4.31), search for more information online about how to stay safe prior to travel ( mean= 4.2 7), avoid walking around the destination by themselves ( mean= 4.2 2 ), only use recommended taxi companies ( mean= 4.1 5 ), and avoid using services like Uber ( mean= 4.0 4 ). The mean scores of 4 recommended risk reduction behaviors ranged from 3.26 3.83, indica ting that respondents were neutral in their likelihood to engage in these

PAGE 114

114 behaviors. Respondents indicated that they were least likely to register with the U.S. Department of State prior to travel (mean= 3.26), purchase travel insurance prior to travel (me an= 3.31), avoid traveling outside of the hotel zone (mean= 3.61), and remain in tourist areas (mean= 3.83). None of the recommended risk reduction behaviors had a mean score below 3.0, indicating that respondents were neither unlikely nor very unlikely to engage any of the behaviors. Table 3 2 0 . Descriptive analysis of intentions to engage in a recommended risk reduction behavior . Mediating variable 1 2 3 4 5 Mean St. deviation Intentions to e ngage in a recommended risk reduction behavior (n= 1,1 26 ) Maintain awareness of your surroundings 0.2 0.9 7.5 25.3 66.1 4.56 0.69 Avoid carrying large amounts of cash 0.4 2.1 8.2 21.5 67.8 4.54 0.77 Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras) 0.2 2.4 8.0 23.5 65.9 4.53 0.76 Avoid situations in which you may be isolated or stand out as a potential victim 0.3 2.7 10.9 30.3 55.9 4.39 0.81 Avoid walking around the destination by yourself at night 1.1 4.1 11.5 23.0 60.3 4.37 0.92 Avoid drawing attention to yourself 0.6 2.0 13.2 30.6 53.6 4.35 0.83 Avoid using ATMs if they are not affiliated with a recognized international chain 1.4 3.2 12.4 28.9 54.1 4.31 0.91 Search for more information online about how to stay safe prior to travel 1.9 4.0 13.7 26.6 53.9 4.27 0.97 Avoid walking around the destination by yourself 2.1 6.0 13.2 25.6 53.1 4.22 1.02 Only use recommended taxi companies 1.2 4.4 17.1 32.5 44.8 4.15 0.94

PAGE 115

115 Table 3 20 . Continued . Mediating variable 1 2 3 4 5 Mean St. deviation Avoid using services like Uber 2.8 7.6 18.9 24.2 46.5 4.04 1.10 Remain in tourist areas 4.3 10.2 20.6 28.5 36.4 3.83 1.16 Avoid traveling outside of the hotel zone 6.7 14.2 21.9 26.1 31.1 3.61 1.24 Purchase travel insurance prior to travel 12.4 16.7 24.9 19.6 26.4 3.31 1.35 Register with the U.S. Department of State prior to travel 13.6 17.6 22.6 21.4 24.9 3.26 1.36 Based on PMT , the researcher made the decision not to reduce the intentions to engage in a recommended risk reduction behavior items using factor analys is. Specifically, PMT considers self efficacy, response efficacy, and intentions to engage in a recommended risk reduction behavior to be behavior specific. Accordingly, the measures must be contextualized in terms of the same recommende d risk reduction behavior when examining the hypothesized relationships between self efficacy, response efficacy, and intentions to engage in a recommended risk reduction behavior . Consequently, a lthough any of the 15 recommended risk reduction behavior i tem s could have been examined, it was beyond the scope of this dissertation to examine 15 dependent variable s . Thus, a s indicated in the multivariate outlier analysis section , the decision was made to focus on the bottom 3 items . The reas oning was that a m ajority of respondents were already very likely to engage in the top 9 recommended risk reduction behaviors. Therefore, understanding the predictors of intentions to engage in the bottom 3 recommended risk reduction behaviors could be used to develop awareness and education campaigns. Accordingly, the analysis of the research questions and associated hypotheses focused on self efficacy, response efficacy, and intentions to engage in the recommended risk reduction beha viors of registering with

PAGE 116

116 the U.S. Department of State prior to travel, purchasing travel insurance prior to travel, and avoiding traveling outside of the hotel zone. Measurement Model Based on the data collection and preparation for data analysis, the pro posed conceptual model was transformed into a measurement model . Specifically, the conceptual model included 3 perceived risk variables (perceived vulnerability, perceived severity, and affective risk perceptions). In an effort to reflect the results of PC A, the number of perceived risk variables was reduced to 2 in the measurement model (perceived vulnerability/affective risk perceptions and perceived severity). In addition, p lace of residence and international past travel experience were not included in t he measurement model because they were controlled for based on the target population of this study. Furthermore, i n accordance with the purpose of this study , the measurement model did not include the decision making process. Rather , t he measurement model reflected the dependent variable of intentions to engage in a recommended risk reduction behavior. The measurement model is presented in Figure 3 1. Restatement of Research Questions and Hypotheses In an effort to accurately reflect the measurement model, the research questions and associated hypotheses were restated below. Accordingly, the 4 research questions and 36 associated hypotheses addressed by this study were: 1. Do the antecedents (demographic factors, international travel specific psychological fact ors, destination specific psychological factors, destination specific factors, and past travel experience) predict perceived efficacy (self efficacy and response efficacy)? a) Demographic factors predict self efficacy. b) Demographic factors predict response eff icacy.

PAGE 117

117 Figure 3 1 . Measurement m odel. Note: * indicates that, in accordance with PMT , the mediating variables appear in a causal string in which individuals go through the threat appraisal process before the coping appraisal process (Floyd et al., 2000) c) International travel specific psychological factors predict self efficacy. d) International travel specific psychological factors predict response efficacy. e) Destination specific psychological factors predict self efficacy. f) Destination specific psychological factors predict response efficacy. g) Destination specific factors predict self efficacy.

PAGE 118

118 h) Destination specific factors predict response efficacy. i) Past travel experience predicts self efficacy. j) Past travel experience predicts response efficacy. 2. Does perceived risk (perceived vulnerability/affective risk perceptions and perceived severity) mediate the relationship between the antecedents (demographic factors, international travel specific psychological factors, destination specific psychological factors, destin ation specific factors, and past travel experience) and perceived efficacy (self efficacy and response efficacy)? a) Perceived vulnerability/affective risk perceptions mediates the relationship between demographic factors and self efficacy . b) Perceived vulnerab ility/affective risk perceptions mediates the relationship between demographic factors and response efficacy . c) Perceived vulnerability/affective risk perceptions mediates the relationship between international travel specific psychological factors and self efficacy . d) Perceived vulnerability/affective risk perceptions mediates the relationship between international travel specific psychological factors and response efficacy . e) Perceived vulnerability/affective risk perceptions mediates the relationship betw een destination specific psychological factors and self efficacy . f) Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific psychological factors and response efficacy . g) Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific factors and self efficacy . h) Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific factors and response efficacy . i) Perceived vulnerability/a ffective risk perceptions mediates the relationship between past travel experience and self efficacy . j) Perceived vulnerability/affective risk perceptions mediates the relationship between past travel experience and response efficacy . k) Perceived severity medi ates the relationship between demographic factors and self efficacy . l) Perceived severity mediates the relationship between demographic factors and response efficacy .

PAGE 119

119 m) Perceived severity mediates the relationship between international travel specific psycholo gical factors and self efficacy . n) Perceived severity mediates the relationship between international travel specific psychological factors and response efficacy . o) Perceived severity mediates the relationship between destination specific psychological factors and self efficacy . p) Perceived severity mediates the relationship between destination specific psychological factors and response efficacy . q) Perceived severity mediates the relationship between destination specific factors and self efficacy . r) Perceived severity mediates the relationship between destination specific factors and response efficacy . s) Perceived severity mediates the relationship between past travel experience and self efficacy . t) Perceived severity mediates the relationship between past travel experience and response efficacy . 3. Does perceived risk (perceived vulnerability/affective risk perceptions and perceived severity) predict intentions to engage in a recommended risk reduction behavior? a) Perceived vulnerability/affective risk perc eptions predicts intentions to engage in a recommended risk reduction behavior. b) Perceived severity predicts intentions to engage in a recommended risk reduction behavior. 4. Does perceived efficacy (self efficacy and response efficacy) mediate the relationshi p between perceived risk (perceived vulnerability/affective risk perceptions and perceived severity) and intentions to engage in a recommended risk reduction behavior? a) Self efficacy mediates the relationship between perceived vulnerability/affective risk perceptions and intentions to engage in a recommended risk reduction behavior. b) Self efficacy mediates the relationship between perceived severity and intentions to engage in a recommended risk reduction behavior. c) Response efficacy mediates the relationship between perceived vulnerability/ affective risk perceptions and intentions to engage in a recommended risk reduction behavior.

PAGE 120

120 d) Response efficacy mediates the relationship between perceived severity and intentions to engage in a recommended risk reduction behavior.

PAGE 121

121 CHAPTER 4 RESULTS This chapter presents the results of the 4 research questions and 36 associated hypotheses addressed by this study . Given that the measures of self efficacy, response efficacy, and intentions to engage in a recommended risk r eduction behavior were specific to an individual recommended risk reduction behavior, the results of the analysis of the research questions were divided into sections based on the specific recommended risk reduction behavior examined. Accordingly, t his cha pter is structured into four main sections: (1) register with the U.S. Department of State prior to travel; ( 2 ) purchase travel insurance prior to travel; ( 3 ) avoid traveling outside of the hotel zone ; (4) summary of findings by hypothesis . Register with the U.S. Department of State Prior to Travel Research Question 1: Do the Antecedents Predict Perceived Efficacy? Research question 1 focused on examining if the antecedents (demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) predicted perceived efficacy (self efficacy and response efficacy). To test research question 1, two multiple regressions were conducted using SPSS. Self efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the first multiple regression . R esponse efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the second multiple regression. The analysis of research question 1 provided the total effects of the antecedents (independen t variables) on perceived efficacy

PAGE 122

122 (second set of mediating variables) associated with the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel . To prepare the categorical variables for inclusion in the multi ple regression model s , groups with low frequencies were combined with other groups. First, the age groups of 50 64 and 65 or older were combined because of the low frequency of respondents in the age group of 65 or old. This age group was labeled 50 or old er. Second, the household income groups of under $15,000, $15,000 to $24,999, and $25,000 to $34,999 were combined. This household income group was labeled $34,999 or less. The household income groups of $100,000 to $124,999 and $125,000 or over were also combined. This household income group was labeled $100,000 or over. Third, 7 of the race/ethnicity groups were combined because approximately three in four respondents considered themselves to be White/Caucasian . Therefore, the following 7 race/ethnicity g roups were combined into 1 group: Black/African American, Asian, Native Hawaiian/Pacific Islander, Hispanic, American Indian/Alaska Native, multi ethnic/mixed race, and other. This race/ethnicity group was labeled other. Accordingly, race/ethnicity was a d ichotomous variable (0: other; 1: White/Caucasian). Fourth, the education level groups were combined to reflect whether respondents dichotomous variable (0: Associate degree or l Fifth, the marital status groups of separated, widowed, and divorced were combined with the currently married group. This marital status group was labeled have been married. Accordingly, marital status was a dichotomo us variable (0: never married; 1: have been married). Sixth, the presence of children in the household variable was

PAGE 123

123 recoded to reflect whether respondents currently had children living in their household or not. Accordingly, presence of children in the hou sehold was a dichotomous variable (0: no; 1: yes). Furthermore, in an effort t o prepare the categorical variables for inclusion in the multiple regression model s, dummy variables were created by utilizing the reference category approach (Hair et al., 2010, p. 87). Specifically, the reference group for age was 18 24, the reference group for household income was $34,999 or less, and the reference group for fluency in the native language of the destination was not at all. Accordingly , 3 age groups (25 34, 35 4 9, and 50 or older) , 4 household income groups ($35,000 $49,999, $50,000 $74,999, $75,000 $99,999, and $100,000 or over), and 3 fluency in the native language of the destination groups ( not well, well, very well ) were entered into the subsequent analyses . A preliminary multiple regression was run for the dependent variable of self efficacy associated with registering with the U.S. Department of State prior to travel in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Palla nt, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating th at there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was

PAGE 124

124 examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. In an effort to remedy this assumption violation, the dependent variable was transformed. As suggested by Tabachnick and Fidell (2013), the transformation method of reflect and log, NEWX= LG10(K X), was used because the distribution of the dependent variable was substantially negatively skewed (p. 89). Following the computation of the transformed dependent variable, the multiple regression was rerun. The Kolmogorov Smirnov statistic was significant at the p < 0.0 1 level, indicating that the assumption of the normality of error terms was still violated. However, practice, this assumption is rarely met, strictly speaking, primarily because of the Zeckhauser and Thompson (1970) suggested that normality of error terms is not cr itical in prediction . In addition, the F test results should be considered reliable even if the residuals are not normally distributed, as long as the sample size is large enough (Minitab, n.d.). Furthermore, regression analysis is considered to be robust, Therefore, the multiple regression was rerun with the non transforme d dependent variable . normality can influence the sampling variance in Accordingly, t o err on the side of caution, a more conservative signif icance level of p <

PAGE 125

125 0.01 was used when interpreting the results of the multiple regression which examined the relationship between the antecedents and self efficacy associated with registering with the U.S. Department of State prior to travel . This decisio n was made because increasing the significance level reduces the region of acceptance, which makes the hypothesis test more likely to reject the null hypothesis, thus increas ing the power of the test . The results of the m ultiple regress ion used to assess the ability of the antecedents (demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) to predict self efficacy asso ciated with registering with the U.S. Department of State prior to travel are presented in Table 4 1. T he total variance explained by the model as a whole was 4. 9 %, F (22, 1,09 4 )= 3. 612 , p < 0.001. T able 4 1 . Multiple regression of the relationship between the antecedents and self efficacy associated with registering with the U.S. Department of State prior to travel . Variable Sig. Gender .072 .016 Age: 25 34 .024 .607 Age: 35 49 .030 .517 Age: 50 or older .074 .071 Household income: $35,000 $49,999 .039 .263 Household income: $50,000 $74,999 .031 .402 Household income: $75,000 $99,999 .025 .491 Household income: $100,000 or over .051 .167 Race/ethnicity .033 .291 Education level .006 .841 Marital status .002 .949 Presence of children in the household .019 .585 International travel risk perceptions .050 .124 Safety is the most important attribute an international destination can offer .117 .003 ** Safety is a serious consideration when choosing an international tourist destination .119 .002 ** Destination image .052 .098 Presence of friends/relatives living in the destination .025 .438

PAGE 126

126 T able 4 1 . Continued . Variable Sig. Fluency in the native language of the destination: not well .025 .461 Fluency in the native language of the destination: well .008 .816 Fluency in the native language of the destination: very well .027 .404 Country level destination past travel experience .043 .189 City level destination past travel experience .014 .658 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.049 A preliminary multiple regression was run for the dependent variable of response efficacy associated with registering with the U.S. Department of State prior to travel in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. In an effort to remed y this assumption violation, the dependent variable was transformed. As suggested by Tabachnick and Fidell (2013), the transformation method of reflect and square root , NEWX= SQRT(K X), was used because the distribution of the dependent variable was modera tely negatively skewed (p. 89). Following the computation of the transformed

PAGE 127

127 dependent variable, the multiple regression was rerun. The Kolmogorov Smirnov statistic was significant at the p < 0.0 1 level, indicating that the assumption of the normality of e rror terms was still violated. However, based on the previous discussion of the normality of error terms assumption, the multiple regression was rerun with the non transformed dependent variable. T o err on the side of caution, a more conservative significa nce level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between the antecedents and response efficacy associated with registering with the U.S. Department of State prior to travel . The results of the multiple regression used to assess the ability of the antecedents (demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) to pr edict response efficacy associated with registering with the U.S. Department of State prior to travel are presented in Table 4 2 . The total variance explained by the model as a whole was 10.7%, F (22, 1,094)= 7.054, p < 0.001. T able 4 2 . Multiple regress ion of the relationship between the antecedents and response efficacy associated with registering with the U.S. Department of State prior to travel . Variable Sig. Gender .075 .010 Age: 25 34 .022 .623 Age: 35 49 .121 .007 ** Age: 50 or older .058 .146 Household income: $35,000 $49,999 .036 .279 Household income: $50,000 $74,999 .049 .172 Household income: $75,000 $99,999 .002 .946 Household income: $100,000 or over .045 .205 Race/ethnicity .086 .005 ** Education level .067 .032 Marital status .056 .130 Presence of children in the household .018 .600 International travel risk perceptions .012 .708 Safety is the most important attribute an international destination can offer .184 .000 ***

PAGE 128

128 T able 4 2 . Continued . Variable Sig. Safety is a serious consideration when choosing an international tourist destination .005 .892 Destination image .129 .000 *** Presence of friends/relatives living in the destination .028 .368 Fluency in the native language of the destination: not well .012 .710 Fluency in the native language of the destination: well .097 .003 ** Fluency in the native language of the destination: very well .046 .137 Country level destination past travel experience .029 .354 City level destination past travel experience .031 .311 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0. 107 Hypothesis 1a: Demographic factors predict self efficacy Hypothesis 1a posited that demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) would predict self efficacy . Multiple regression analysis indicated that there were n o statistically significant relationship s between the demographic factors and self efficacy associated with registering with the U.S. Department of State prior to travel at the p < 0.0 1 0.0 72 ), age 25 0.024), age 35 ), age 0.074 ), household income $35,000 $49,999 0.039 ) , household income $50,000 $74,999 0.031 ) , household income $75,000 $99,999 0.02 5 ) , household income $100,000 or over 51 0.033 ), education = 0.006 0.002 ), and presence of children in the household 0.019 ). Based on the results of the multiple regression, hypothesis 1a was rejected. Hypothesis 1b: Demographic factors predict response efficacy Hypothesis 1b posited that demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the

PAGE 129

129 household) would predict response efficacy. Multiple regression analysis indicated that there was a statistically signific ant relationship between the 35 49 age group and response efficacy associated with registering with the U.S. Department of State prior to 0.121). Accordingly, the results indicated that respondents between the ages of 35 a nd 49 had significantly lower levels of response efficacy associated with registering with the U.S. Department of State prior to travel than the reference group (age: 18 24) . Multiple regression analysis indicated that there was also a statistically significant relationship between race/ethnicity and response efficacy associated with registering with the U.S. Department of State prior to travel at the p < 0.0 86 ). Accordingly, the results indicated that White/Caucasian r espondents had significantly low er levels of response efficacy associated with registering with the U.S. Department of State prior to travel than respondents who considered themselves to be of a race/ethnicity other than White/Caucasian . In addition, mult iple regression analysis indicated that there were no statistically significant relationship s between the following demographic factors and response efficacy associated with registering with the U.S. Department of State prior to travel at the p < 0.01 leve l: 0.075), age 25 0.022 0.058 ), household income $35,000 $49,999 0.036 ) , household income $50,000 $74,999 0.049 ) , household income $75,000 $99,999 0.002 ) , household income $100,000 or over 0.045 0.067 0.056 ), 0.018 ). Based on the results of the multiple regression, hypothesis 1b was partially supported.

PAGE 130

130 Hypothesis 1c: International travel specific psychological factors predict self efficacy Hypothesis 1c posited that international travel specific psychological factors (international travel risk perceptions and safety concerns) would predict self efficacy. Multiple regression analysis indicated that there was no statistically significant relationship between international travel risk perceptions and self efficacy associated with registering with the U.S. Department of State prior to travel 0.050). In addition , multiple regression analysis indicated that there were statistically significant relationships between the two safety concerns items and self efficacy associated with registering with the U.S. Department of State prior to travel at the p < the results indicated that those with higher levels of efficacy as sociated with registering with the U.S. Department of State prior to travel than those with lower levels of agreement with the statements. Based on the results of the multiple regression, hypothesis 1c was partially supported. Hypothesis 1d: International travel specific psychological factors predict response efficacy Hypothesis 1d posited that international travel specific psychological factors (international travel risk perceptions and safety concerns) would predict response

PAGE 131

131 efficacy. Multiple regression analysis indicated that there was no statistically significant relationship between international travel risk perceptions and response efficacy associated with registering with the U.S. Department of State prior to travel at the p < In addition, multiple regression analysis indicated that there was a statistically associated with registering with the U.S. Department of State prior to travel at the p < 0.0 0 0.184). Accordingly, the results indicated that those with higher levels of agreement with destination can had significantly higher levels of response efficacy associated with registering with the U.S. Department of State prior to travel than those with lower levels of agreement with the statement. Furthermore, multiple regression analysi s indicated that there was efficacy associated with registering with the U.S. Dep artment of State prior to travel at 0.005). Based on the results of the multiple regression, hypothesis 1d was partially supported. Hypothesis 1e: Destination specific psychological factors predict self efficacy Hypothesis 1e posit ed that destination specific psychological factors (destination image) would predict self efficacy . Multiple regression analysis indicated that there was no statistically significant relationship between destination image and self efficacy associated with registering with the U.S. Department of State prior to travel at the p <

PAGE 132

132 0.052 ). Based on the results of the multiple regression, hypothesis 1e was r ejected. Hypothesis 1f: Destination specific psychological factors predict response efficacy Hypothesis 1f posited that destination specific psychological factors (destination image) would predict response efficacy. Multiple regression analysis indicated that there was a statistically significant relationship between destination image and response efficacy associated with registering with the U.S. Department of State prior to travel at those with a more positive destination image had significantly higher levels of respo nse efficacy associated with registering with the U.S. Department of State prior to travel than t hose with a more negative destination image. Based on the results of the multiple regression, hypothesis 1f was not rejected. Hypothesis 1g: Destination specif ic factors predict self efficacy Hypothesis 1g posited that destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of the destination) would predict self efficacy . Multiple regression analysis indicated that there were no statistically significant relationships between the destination specific factors and self efficacy associated with registering with the U.S. Department of State prior to travel at the p < 0.0 1 lev el: presence of friends and/or relatives living in the destination 0.025 ), fluency in the native language of ), fluency 0.00 8 ), and fluency in the native language o 27 ). Based on the results of the multiple regression, hypothesis 1g was rejected.

PAGE 133

133 Hypothesis 1h: Destination specific factors predict response efficacy Hypothesis 1h posited that destination specific factors (presence o f friends/relatives living in the destination and fluency in the native language of the destination) would predict response efficacy . Multiple regression analysis indicated that there was a statistically significant relationship between the fluency in the native language of the destination: well group and response efficacy associated with registering with the U.S. Department of State prior to travel 0.097). Accordingly, the results indicated that those who spoke Spanish well had si gnificantly higher levels of response efficacy associated with registering with the U.S. Department of State prior to travel than the reference group (fluency in the native language of the destination: not at all) . In addition, m ultiple regression analysis indicated that there were no statistically significant relationships between the following destination specific factors and response efficacy associated with registering with the U.S. Department of State prior to travel at the p < 0.0 1 level: presence of friends and/or 0.028 0.012 ), and 0.046 ). Based on the results of the multiple regres sion, hypothesis 1 h was partially supported . Hypothesis 1i: Past travel experience predicts self efficacy Hypothesis 1i posited that past travel experience (country level destination past travel experience and city level destination past travel experience ) would predict self efficacy . Multiple regression analysis indicated that there were no statistically significant relationships between the two types of past travel experience and self efficacy

PAGE 134

134 associated with registering with the U.S. Department of State prior to travel at the p < 0.0 1 level: country level destination past travel experience 0.0 43 ) and city level destination past travel experience 0.014 ). Based on the results of the multiple regression, hypothesis 1 i was rejected. Hypothesis 1j: P ast travel experience predicts response efficacy Hypothesis 1j posited that past travel experience (country level destination past travel experience and city level destination past travel experience ) would predict response efficacy . Multiple regression an alysis indicated that there were no statistically significant relationships between the two types of past travel experience and response efficacy associated with registering with the U.S. Department of State prior to travel at the p < 0.0 1 level: country l evel destination past travel experience 0.0 29 ) and city level destination past travel experience 0.031 ). Based on the results of the multiple regression, hypothesis 1 j was rejected. Research Question 2: Does Perceived Risk Mediate the Relationship between the Antecedents and Perceived Efficacy? Research question 2 focused on examining if perceived risk (perceived vulnerability, affective risk perceptions, and perceived severity) mediated the relationship between the antecedents (demographic factors , international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) and perceived efficacy (self efficacy and response efficacy). To test research question 2, a series of multiple regressions were conducted using SPSS.

PAGE 135

135 Indirect paths between the antecedents and perceived risk In the first series of multiple regressions, the relationships between the antecedents ( demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience ) and perceived risk ( perceived vulnerability / affective risk perceptions and perceived severity ) were examined. P erceived vulnerabili ty / affective risk perceptions was the dependent variable in the first multiple regression . Perceived severity was the dependent variable in the second multiple regression. The first series of multiple regressions provided the indirect paths between the ant ecedents ( independent variables ) and perceived risk (first set of mediating variables). A preliminary multiple regression was run for the dependent variable of perceived vulnerability / affective risk perceptions in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a co mbination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (V IF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmo gorov Smirnov statistic was not significant at the p < 0.0 5 level, indicating that the assumption of the normality of error terms was not violated.

PAGE 136

136 The results of the multiple regression used to assess the ability of the antecedents (demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) to predict perceived vulnerability/affective risk perceptions are presented in Table 4 3 . The total variance explained by the model as a whole was 39 .9%, F (22, 1,094)= 34.745 , p < 0.001. T able 4 3 . Multiple regression of the relationship between the antecedents and perceived vulnerability/affective risk perceptions . Variable Stan Sig. Gender .014 .559 Age: 25 34 .034 .360 Age: 35 49 .042 .263 Age: 50 or older .063 .053 Household income: $35,000 $49,999 .040 .150 Household income: $50,000 $74,999 .049 .095 Household income: $75,000 $99,999 .023 .416 Household income: $100,000 or over .020 .489 Race/ethnicity .001 .965 Education level .013 .604 Marital status .021 .494 Presence of children in the household .030 .285 International travel risk perceptions .341 .000 *** Safety is the most important attribute an international destination can offer .003 .921 Safety is a serious consideration when choosing an international tourist destination .113 .000 *** Destination image .460 .000 *** Presence of friends/relatives living in the destination .019 .462 Fluency in the native language of the destination: not well .033 .211 Fluency in the native language of the destination: well .007 .800 Fluency in the native language of the destination: very well .003 .893 Country level destination past travel experience .066 .011 * City level destination past travel experience .018 .481 Note: * indicates p < 0.05; ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0. 399 A preliminary multiple regression was run for the dependent variable of perceived severity in order to test the assumptions of multiple regression. The sample size

PAGE 137

137 assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of s ingularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of th e tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no mult ivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. Based on the previous discu ssion of the normality of error terms assumption, to err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between the antecedents an d perceived severity . The results of the multiple regression used to assess the ability of the antecedents (demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) to predict perceived severity are presented in Table 4 4 . The total variance explained by the model as a whole was 13.7 %, F (22, 1,094)= 9.071 , p < 0.001. T able 4 4 . Multiple regression of the relationship between the anteceden ts and perceived severity . Variable Sig. Gender .084 .003 ** Age: 25 34 .011 .807 Age: 35 49 .016 .725 Age: 50 or older .016 .680 Household income: $35,000 $49,999 .053 .111

PAGE 138

138 T able 4 4 . Continued . Variable Sig. Household income: $50,000 $74,999 .043 .225 Household income: $75,000 $99,999 .013 .718 Household income: $100,000 or over .011 .744 Race/ethnicity .003 .917 Education level .047 .125 Marital status .014 .689 Presence of children in the household .003 .928 International travel risk perceptions .029 .347 Safety is the most important attribute an international destination can offer .060 .109 Safety is a serious consideration when choosing an international tourist destination .280 .000 *** Destination image .098 .001 ** Presence of friends/relatives living in the destination .049 .113 Fluency in the native language of the destination: not well .034 .287 Fluency in the native language of the destination: well .017 .609 Fluency in the native language of the destination: very well .042 .175 Country level destination past travel experience .036 .246 City level destination past travel experience .003 .912 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0. 137 Indirect paths between perceived risk and perceived efficacy In the second series of multiple regressions, the relationships between perceived risk ( perceived vulnerability / affective risk perceptions and perceived severity) and perceived efficacy (self efficacy and response efficacy) were examined . Self efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the fir st multiple regression . R esponse efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the second multiple regression. The second series of multiple regressions provided the indirect paths between perceived risk (first set of mediating variables) and perceived efficacy (second set of mediating variables).

PAGE 139

139 A preliminary multiple regression was run for the dependent variable of self efficacy a ssociated with registering with the U.S. Department of State prior to travel in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 20 10, p. 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis d istance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was viola ted. Based on the previous discussion of the normality of error terms assumption, to err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between perceived risk and self efficacy associated with registering with the U.S. Department of State prior to travel. The results of the multiple regression used to assess the ability of perceived risk (perceived vulnerability / affective risk perceptions and perceived severity) to predict self efficacy associated with registering with the U.S. Department of State prior to travel are presented in Table 4 5 . The total variance explained by the model as a whole was 0.4 %, F (2, 1,122 )= 3.198 , p < 0.0 5 .

PAGE 140

140 T able 4 5 . Multiple regression of the relationship between perceived risk and self efficacy associated with registering with the U.S. Department of State prior to travel . Variable Sig. Perceived vulnerability/affective risk perceptions .010 .742 Perceived severity .071 .022 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.004 A preliminary multiple regression was run for the dependent variable of response efficacy associated with registering with the U.S. Department of State prior to travel in order to test the assumptions of multiple regression. The sample size assumption wa s not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables ( Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicati ng that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. Based on the previous discussion of the normality of error terms assumption, to err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between perceived risk and response efficacy associated with registering with the U.S. Department of State prior to travel .

PAGE 141

141 The results of the multiple regression used to assess the ability of perceived risk (perceived vulnerability / affective risk perceptions and perceived severity) to predict response efficacy associated with registering with the U.S. Department of State prior to travel are presented in Table 4 6 . The model was not significant at the p < 0.05 level and the total variance explained by the model as a whole was 0.0 %, F (2, 1,122 )= 0.815 , p > 0.0 5 . T able 4 6 . Multiple regression of the relationship between perceived risk and response efficacy associated with registering with the U.S. Department of State prior to travel . Variable Sig. Perceived vulnerability/affective risk perceptions .034 .277 Perceived severity .009 .762 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.000 Given that none of the indirect paths between the perceived risk variables (first set of mediating variables) and perceived efficacy (second set of mediating variables) were signific ant in the second series of multiple regressions, a third series of multiple regressions was not conducted. In addition, since mediation was not found, an online calculation of the Sobel test was not conducted. Hypothesis 2a: Perceived vulnerability/affective risk perceptions mediates the relationship between demographic factors and self efficacy Hypothesis 2a posited that perceived vulnerability/affective risk perceptions mediated the relationship between demographic factors (gender, a ge, household income, race/ethnicity, education level, marital status, and presence of children in the household) and self efficacy . The first series of m ultiple regression analysis indicated that there were no statistically significant relationships betwe en the demographic factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco at the p < 0.05 level: 0. 014 ), age 25

PAGE 142

142 0.034 ), age 35 0.0 42 ), age 0.0 63 ), household income $35,000 $49,999 0.040 ) , household income $50,000 $74,999 0.0 49 ) , household income $75,000 $99,999 0.02 3 ) , household income $100,000 or over 20 ), 0.0 01 = 0.0 13 21 ), and 30 ). The second series of multiple regressions indicated that there was no statistically significant relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and self efficacy associated with registering with the U.S. Department of State prior to travel at the p < 0.01 level 0.010 ) . Based on the results of the two series of multipl e regression s , hypothesis 2 a was rejected. Hypothesis 2b: Perceived vulnerability/affective risk perceptions mediates the relationship between demographic factors and response efficacy Hypothesis 2b posited that perceived vulnerability/affective risk perc eptions mediated the relationship between demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) and response efficacy . The relationships between demographic factors a nd perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco w ere discussed in the section on hypothesis 2a. The second series of multiple regressions indicated that there w as no statistically sign ificant relationship s between the perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and response efficacy associated with registering with the U.S. Department of State prior to travel at the p <

PAGE 143

143 0.01 level 0.034 ) . Based on the results of the two series of multiple regressions, hypothesis 2b was rejected. Hypothesis 2c: Perceived vulnerability/affective risk perceptions mediates the relationship between international travel specific psych ological factors and self efficacy Hypothesis 2c posited that perceived vulnerability/affective risk perceptions mediated the relationship between international travel specific psychological factors (international travel risk perceptions and safety concer ns) and self efficacy . The first series of multiple regression analysis indicated that there w as a statistically significant relationship between international travel risk perceptions and perceived vulnerability/ affective risk perceptions associated with crime while visiting the destination of Acapulco the results indicated that those with higher levels of international travel risk perceptions had significantly highe r levels of perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco than those with lower levels of international travel risk perceptions. The first series of multiple regression analysis indicat ed that there was no perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco regression analysis also indicated that there was a statistically significant relationship on when choosing an perceived vulnerability/affective risk perceptions

PAGE 144

144 associated with crime while visiting the destination of Acapulco at the p < 0.001 level the results indicated that those with higher levels of agreement perceived vulnerability/affective risk perceptions associated with crime while visi ting the destination of Acapulco than those with lower levels of agreement with the statement. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and self efficacy a ssociated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2 a. Based on the results of the two series of multiple regressions, hypothesis 2c was rejected. Hypothesis 2d: Perceived vulnerability/a ffective risk perceptions mediates the relationship between international travel specific psychological factors and response efficacy Hypothesis 2d posited that perceived vulnerability/affective risk perceptions mediated the relationship between internati onal travel specific psychological factors (international travel risk perceptions and safety concerns) and response efficacy . The relationships between international travel specific psychological factors and perceived vulnerability/affective risk perceptio ns associated with crime while visiting the destination of Acapulco w ere discussed in the section on hypothesis 2c. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulc o and response efficacy associated with registering with the U.S. Department of State prior to travel was

PAGE 145

145 discussed in the section on hypothesis 2b . Based on the results of the two series of multiple regressions , hypothesis 2d was rejected. Hypothesis 2e: Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific psychological factors and self efficacy Hypothesis 2e posited that perceived vulnerability/affective risk perceptions mediated the relationship between destination specific psychological factors (destination image) and self efficacy . The first series of multiple regression analysis indicated that there was a statistically significant relationship between destination image and perceived vulnerabili ty/affective risk perceptions associated with crime while visiting the destination of Acapulco 0.460). Accordingly, the results indicated that those with a more positive destination image had significantly lower levels of percei ved vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco than those with a more negative destination image. The relationship between perceived vulnerability/affective risk perceptions associated with cri me while visiting the destination of Acapulco and self efficacy associated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2 a. Based on the results of the two series of multiple regressions , hyp othesis 2e was rejected. Hypothesis 2f: Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific psychological factors and response efficacy Hypothesis 2f posited that perceived vulnerability/affective risk perceptions mediated the relationship between destination specific psychological factors (destination image) and response efficacy. The relationships between destination -

PAGE 146

146 specific psychological factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco w ere discussed in the section on hypothesis 2e. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and response efficacy associated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2b . Based on the results of the two series of multiple regressions, hypothesis 2f was rejected. Hypothesis 2g: Pe rceived vulnerability/affective risk perceptions mediates the relationship between destination specific factors and self efficacy Hypothesis 2g posited that perceived vulnerability/affective risk perceptions mediated the relationship between destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of the destination) and self efficacy. The first series of multiple regression analysis indicated that there were no statistically significant relationship s between the destination specific factors and perceived vulnerability/affective risk perceptions associated with crime whil e visiting the destination of Acapulco at the p < 0.05 level: presence of friends and/or 0.019), fluency in the native language of the 0.033), fluency in the native language of the destina tion: well The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and self effic acy associated with registering with the U.S. Department of State prior to travel was

PAGE 147

147 discussed in the section on hypothesis 2 a. Based on the results of the two series of multiple regressions, hypothesis 2g was rejected. Hypothesis 2h: Perceived vulnerabil ity/affective risk perceptions mediates the relationship between destination specific factors and response efficacy Hypothesis 2 h posited that perceived vulnerability/affective risk perceptions mediated the relationship between destination specific factor s (presence of friends/relatives living in the destination and fluency in the native language of the destination) and response efficacy . The relationships between destination specific factors and perceived vulnerability/affective risk perceptions associate d with crime while visiting the destination of Acapulco w ere discussed in the section on hypothesis 2 g . The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and respon se efficacy associated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2b . Based on the results of the two series of multiple regressions, hypothesis 2h was rejected. Hypothesis 2i: Perceived vu lnerability/affective risk perceptions mediates the relationship between past travel experience and self efficacy Hypothesis 2 i posited that perceived vulnerability/affective risk perceptions mediated the relationship between past travel experience (count ry level destination past travel experience and city level destination past travel experience ) and self efficacy. The first series of multiple regression analysis indicated that there was a statist ically significant relationship between country level desti nation past travel experience and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco at the p < 0.0 5 level 0.0 66 ) . Accordingly, the results

PAGE 148

148 indicated that those with higher levels of country level destination past travel experience had significantly lower levels of perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco than those wit h lower levels of country level destination past travel experience. The first series of multiple regression analysis indicated that there was not a statistically significant relationship between city level destination past travel experience and perceived v ulnerability/ affective risk perceptions associated with crime while visiting the destination of Acapulco at the p < 0.05 level 0.0 18 ). The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting th e destination of Acapulco and self efficacy associated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2 a. Based on the results of the two series of multiple regressions, hypothesis 2i was rejec ted. Hypothesis 2j: Perceived vulnerability/affective risk perceptions mediates the relationship between past travel experience and response efficacy Hypothesis 2j posited that perceived vulnerability/affective risk perceptions mediated the relationship b etween past travel experience (country level destination past travel experience and city level destination past travel experience ) and response efficacy . The relationships between past travel experience and perceived vulnerability/affective risk perception s associated with crime while visiting the destination of Acapulco w ere discussed in the section on hypothesis 2i. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and response efficacy associated with registering with the U.S. Department of State prior to travel was

PAGE 149

149 discussed in the section on hypothesis 2b . Based on the results of the two series of multiple regressions, hypothesis 2j was rejected. Hypothesis 2k: Perceived severity mediates the relationship between demographic factors and self efficacy Hypothesis 2k posited that perceived severity mediated the relationship between demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) and self efficacy . The first series of multiple regression analysis indicated that there was a statistically significant relationship between gender and perceived severity associated with crime wh ile visiting 0.084). Accordingly, the results indicated that females had significantly higher levels of perceived severity associated with crime while visiting the destination of Acapulco than males. The first series of multiple regression analysis indicated that there were no statistically significant relationships between the following demographic factors and perceived severity associated with crime while visiting the destination of Acapulco at the p < 0.0 1 level: age 25 0.011 ), age 35 ), household income $35,000 0.0 53 ), household income $50,000 0.04 3 ), household income $75,000 0.0 1 3), household income $100,000 or ov er 11 0.00 3 47 0.0 14 0 3). The second series of multiple regressions indicated that there was no statistically significant relatio nship between perceived severity associated with crime while visiting the destination of Acapulco and self efficacy associated with registering with the U.S.

PAGE 150

150 Based on the results of the two series of multiple regressions, hypothesis 2k was rejected. Hypothesis 2l: Perceived severity mediates the relationship between demographic factors and response efficacy Hypothesis 2l posited that perceived severity mediated the relationship between d emographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) and response efficacy . The relationships between demographic factors and perceived severity associated with crime while visiting the destination of Acapulco w ere discussed in the section on hypothesis 2k. The second series of multiple regressions indicated that there was no statistically significant relationship s between perceived severity associated with crime whil e visiting the destination of Acapulco and response efficacy associated with registering with the U.S. Department of State prior to travel at the p < 0.01 level 0.009 ) . Based on the results of the two series of multiple regressions, hypothesis 2l was rejected. Hypothesis 2m: Perceived severity mediates the relationship between international travel specific psychological factors and self efficacy Hypothesis 2m posited that perceived severity mediated the relationship between international travel specific psychological factors (international travel risk perceptions and safety concerns) and self efficacy . The first series of multiple regression analysis indicated that there were no statistically significant relationships between the following international travel specific psychological factors and perceived severity associated with crime while visiting the destination of Acapulco at the p < 0.01 level: international travel

PAGE 151

151 most important The first series of multiple regression analysis also indicated that there was a con associated with crime while visiting the destination of Acapulco at the p < 0.001 level reement crime while visiting the destination of Acapulco than those with lower l evels of agreement with the statement. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and self efficacy associated with registering with the U.S. Department of State prior to travel was discuss ed in the section on hypothesis 2k . Based on the results of the two series of multiple regressions, hypothesis 2m was rejected. Hypothesis 2n: Perceived severity mediates the relationship between international travel specific psychological factors and resp onse efficacy Hypothesis 2n posited that perceived severity mediated the relationship between international travel specific psychological factors (international travel risk perceptions and safety concerns) and response efficacy . The relationships between international travel specific psychological factors and perceived severity associated with crime while visiting the destination of Acapulco w ere discussed in the section on hypothesis 2m.

PAGE 152

152 The relationship between perceived severity associated with crime w hile visiting the destination of Acapulco and response efficacy associated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2l . Based on the results of the two series of multiple regressions, hyp othesis 2n was rejected. Hypothesis 2o: Perceived severity mediates the relationship between destination specific psychological factors and self efficacy Hypothesis 2o posited that perceived severity mediated the relationship between destination specific psychological factors (destination image) and self efficacy . The first series of multiple regression analysis indicated that there was a statistically significant relationship between destination image and perceived severity associated with crime while vis iting the destination of Acapulco 0.098). Accordingly, the results indicated that those with a more positive destination image had significantly lower levels of perceived severity associated with crime while visiting the destination of Acapulco than those with a more negative destination image. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and self efficacy associated w ith registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2k . Based on the results of the two series of multiple regressions, hypothesis 2o was rejected. Hypothesis 2p: Perceived severity mediates the relationship between destination specific psychological factors and response efficacy Hypothesis 2 p posited that perceived severity mediated the relationship between destination specific psychological factors (destination image) and response efficacy.

PAGE 153

153 The relationships between destination specific psychological factors and perceived severity associated with crime while visiting the destination of Acapulco w ere discussed in the section on hypothesis 2o. The relationship between perceived severity associate d with crime while visiting the destination of Acapulco and response efficacy associated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2l . Based on the results of the two series of multiple re gressions, hypothesis 2p was rejected. Hypothesis 2q: Perceived severity mediates the relationship between destination specific factors and self efficacy Hypothesis 2q posited that perceived severity mediated the relationship between destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of the destination) and self efficacy. The first series of multiple regression analysis indicated that there were no statistically significant relationships between the destination specific factors and perceived severity associated with crime while visiting the destination of Acapulco at the p < 0.01 level: presence of 0.049), fluency in the native langu 0.034), fluency in the native language of the 0.042). The relationship between perceived severity associated with crim e while visiting the destination of Acapulco and self efficacy associated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2k .

PAGE 154

154 Based on the results of the two series of multiple regressions, hypo thesis 2q was rejected. Hypothesis 2r: Perceived severity mediates the relationship between destination specific factors and response efficacy Hypothesis 2r posited that perceived severity mediated the relationship between destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of the destination) and response efficacy . The relation ships between destination specific factors and perceived severity associated with crime while visiting the destination of Acapulco w ere discussed in the section on hypothesis 2q. The relationship between perceived severity associated with crime while visi ting the destination of Acapulco and response efficacy associated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2l . Based on the results of the two series of multiple regressions, hypothesis 2 r was rejected. Hypothesis 2s: Perceived severity mediates the relationship between past travel experience and self efficacy Hypothesis 2 s posited that perceived severity mediated the relationship between past travel experience (country level destination past travel experience and city level destination past travel experience ) and self efficacy. The first series of multiple regression analysis indicated that there were no statist ically significant relationships between the two types of past travel experien ce and perceived severity associated with crime while visiting the destination of Acapulco at the p < 0.01 level: country level

PAGE 155

155 destination past travel experience 0.0 36 ) and city level destination past travel experience 0.003 ). The relationship be tween perceived severity associated with crime while visiting the destination of Acapulco and self efficacy associated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2k . Based on the results of the two series of multiple regressions, hypothesis 2s was rejected. Hypothesis 2t: Perceived severity mediates the relationship between past travel experience and response efficacy Hypothesis 2t posited that perceived severity mediated the relationship between past travel experience (country level destination past travel experience and city level destination past travel experience ) and response efficacy . The relationships between past travel experience and perceived severity associated with crime while v isiting the destination of Acapulco w ere discussed in the section on hypothesis 2s. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and response efficacy associated with registering with the U.S . Department of State prior to travel was discussed in the section on hypothesis 2l . Based on the results of the two series of multiple regressions, hypothesis 2t was rejected. Research Question 3: Does Perceived Risk Predict Intentions to Engage in a Recommended Risk Reduction Behavior? Research question 3 focused on examining if perceived risk (perceived vulnerability / affective risk perceptions and perceived severity) predicted intentions to engage in a recommended risk reduction behavior. To test res earch question 3, one multiple regression was conducted using SPSS. Intentions to engage in a

PAGE 156

156 recommended risk reduction behavior to ensure personal safety while visiting the destination of Acapulco was the dependent variable in the multiple regressi on . The analysis of research question 3 provided the total effects of perceived risk (first set of mediating variables) on intentions to engage in the recommended risk reduction behavior (dependent variable) of registering with the U.S. Department of State prior to travel . A preliminary multiple regression was run for the dependent variable of intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel in order to test the assumptions of m ultiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were n ot a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. In an effort to remedy this assumption violation, the dependent variable was transformed. As suggested by Tabachnick and Fidell (2013), the transformation method of reflect and square root , NEWX= SQRT(K X), was used because the distribution of the dependent variable was

PAGE 157

157 moderately negatively skewed (p. 89). Following the computation of the transformed dependen t variable, the multiple regression was rerun. The Kolmogorov Smirnov statistic was significant at the p < 0.0 1 level, indicating that the assumption of the normality of error terms was still violated. However, based on the previous discussion of the norma lity of error terms assumption, the multiple regression was rerun with the non transformed dependent variable. T o err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regressi on which examined the relationship between perceived risk and intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel . The results of the multiple regression used to assess the abili ty of perceived risk (perceived vulnerability / affective risk perceptions and perceived severity) to predict intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel are presented in Ta ble 4 7 . The total variance explained by the model as a whole was 1.0 %, F (2, 1,122 )= 6.915 , p < 0.01. T able 4 7 . Multiple regression of the relationship between perceived risk and intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel . Variable Sig. Perceived vulnerability/affective risk perceptions .109 .000*** Perceived severity .003 .912 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.010 Hypothesis 3a: Perceived vulnerability/affective risk perceptions predicts intentions to engage in a recommended risk reduction behavior Hypothesis 3a posited that perceived vulnerability/affective risk perceptions would predict intentions to engage in a recommended risk reduction behavior . Multiple regression analysis indicated that there was a statistically significant relationship

PAGE 158

158 between perceived vulnerability/affective risk perceptions associated with crime wh ile visiting the destination of Acapulco and intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel at the p < 0.0 01 level 0.109 ). Accordingly, the results indicated that those with higher levels of perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were significantly more likely to register with the U.S. Department of State prior to travel than those with low er levels of perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco . Based on the results of the multiple regression, hypothesis 3a was not rejected. Hypothesis 3b: Perceived severity predicts in tentions to engage in a recommended risk reduction behavior Hypothesis 3b posited that perceived severity would predict intentions to engage in a recommended risk reduction behavior . Multiple regression analysis indicated that there was no statistically s ignificant relationship between perceived severity associated with crime while visiting the destination of Acapulco and intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel at the p < 0.0 1 level 0.003 ). Based on the results of the multiple regression, hypothesis 3b was rejected. Research Question 4: Does Perceived Efficacy Mediate the Relationship between Perceived Risk and Intentions to Engage in a Recommended Risk Reduction Be havior? Research question 4 focused on examining if perceived efficacy (self efficacy and response efficacy) mediated the relationship between perceived risk ( perceived vulnerability / affective risk perceptions and perceived severity) and intentions to engage

PAGE 159

159 in a recommended risk reduction behavior. To test research question 4, a series of multiple regressions were conducted using SPSS. Indirect path s between perceived risk and perceived efficacy In the first series of multiple regr essions, the relationships between perceived risk ( perceived vulnerability / affective risk perceptions and perceived severity) and perceived efficacy (self efficacy and response efficacy) were examined. Self efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the first multiple regression . The results of the first multiple regression can be found in Table 4 5. R esponse efficacy associated wi th a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the second multiple regression. The results of the second multiple regression can be found in Table 4 6. Th e first series of multiple regressions provided the indirect paths between perceived risk (first set of mediating variables) and perceived efficacy (second set of mediating variables). Indirect path s between perceived efficacy and intentions to engage in a recommended risk reduction behavior In the second series of multiple regressions, the relationships between perceived efficacy (self efficacy and response efficacy) and intentions to engage in a recommended risk reduction behavior were examined. Intenti ons to engage in a recommended risk reduction behavior to ensure personal safety while visiting the destination of Acapulco was the dependent variable in the multiple regression . T he second series of multiple regressions provided the indirect paths b etween perceived

PAGE 160

160 efficacy (second set of mediating variables) and intentions to engage in a recommended risk reduction behavior (dependent variable). A preliminary multiple regression was run for the dependent variable of intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent v ariable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007 , p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. Based on the previous discussion of the normality of error terms assumption, to err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression wh ich examined the relationship between perceived efficacy and intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel. The results of the multiple regression used to assess the abilit y of perceived efficacy (self efficacy and response efficacy) to predict intentions to engage in the

PAGE 161

161 recommended risk reduction behavior of registering with the U.S. Department of State prior to travel are presented in Table 4 8 . The total variance explain ed by the model as a whole was 47.4%, F ( 2, 1,123)= 508.164, p < 0.001. T able 4 8 . Multiple regression of the relationship between perceived efficacy and intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel . Variable Sig. Self efficacy .433 .000*** Response efficacy .400 .000*** Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0. 474 Given that none of the indirect paths between the perceived risk variables (first set of mediating variables) and perceived efficacy (second set of mediating variables) were significant in the first series of multiple regressions, a third series of multipl e regressions was not conducted. In addition, since mediation was not found, an online calculation of the Sobel test was not conducted. Hypothesis 4a: Self efficacy mediates the relationship between perceived vulnerability/affective risk perceptions and intentions to engage in a recommended risk reduction behavior Hypothesis 4a posited that self efficacy would mediate the relationship between perceived vulnerability/affective risk perceptions and intentions to engage in a recommended risk reduction behav ior. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and self efficacy associated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2a . The second series of multiple regressions indicated that there was a statistically significant relationship between self efficacy associated with registering with the U.S. Department of State prior to travel and intentions to engage in the recommended risk

PAGE 162

162 reduction behavior of registering with the U.S. Department of State prior to travel at the p < 0.0 0 1 level 0. 433 ) . Accordingly, the results indicated that those with higher levels of self efficacy associated with registering with the U.S. Department of State prior to travel were significantly more likely to register with the U.S. Department of State prior to travel than those with lower levels of self efficacy. Based on the results of the two series of multip le regressions, hypothesis 4a was rejected. Hypothesis 4b: Self efficacy mediates the relationship between perceived severity and intentions to engage in a recommended risk reduction behavior Hypothesis 4 b posited that self efficacy would mediate the rela tionship between perceived severity and intentions to engage in a recommended risk reduction behavior. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and self efficacy associated with registerin g with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2k . The relationship between self efficacy associated with registering with the U.S. Department of State prior to travel and intentions to engage in the recomme nded risk reduction behavior of registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 4a. Based on the results of the two series of multiple regressions, hypothesis 4b was rejected. Hypothesis 4c: Response efficacy mediates the relationship b etween perceived vulnerability/ affective risk perceptions and intentions to engage in a recommended risk reduction behavior Hypothesis 4 c posited that response efficacy would mediate the relationship between perceived vulnerability/affective risk perceptions and intentions to engage in a recommended risk reduction behavior. The relationship between perceived vulnerability/ affective risk perceptions associated with crime while visiting the destination of

PAGE 163

163 Acapulco and re sponse efficacy associated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2b . The second series of multiple regressions indicated that there was a statistically significant relationship betwee n response efficacy associated with registering with the U.S. Department of State prior to travel and intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel at the p < 0.001 level ( = 0.400 ) . Accordingly, the results indicated that those with higher levels of response efficacy associated with registering with the U.S. Department of State prior to travel were significantly more likely to register with the U.S. Department of State prior to travel than those with lower levels of response efficacy. Based on the results of the two series of multiple regressions, hypothesis 4c was rejected. Hypothesis 4d: Response efficacy mediates the relationship between perceived severity and intentions t o engage in a recommended risk reduction behavior Hypothesis 4 d posited that response efficacy would mediate the relationship between perceived severity and intentions to engage in a recommended risk reduction behavior. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and response efficacy associated with registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 2l . The relationship between response efficacy associated with registering with the U.S. Department of State prior to travel and intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel was discussed in the section on hypo thesis 4c. Based on the results of the two series of multiple regressions, hypothesis 4d was rejected.

PAGE 164

164 Purchase Travel Insurance Prior to Travel Research Question 1: Do the Antecedents Predict Perceived Efficacy? Research question 1 focused on examining i f the antecedents (demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) predicted perceived efficacy (self efficacy and response effi cacy). To test research question 1, two multiple regressions were conducted using SPSS. Self efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent vari able in the first multiple regression . R esponse efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the second multiple regression. The a nalysis of research question 1 provided the total effects of the antecedents (independent variables) on perceived efficacy (second set of mediating variables) associated with the recommended risk reduction behavior of purchasing travel insurance prior to t ravel . A preliminary multiple regression was run for the dependent variable of self efficacy associated with purchasing travel insurance prior to travel in order to test the assumptions of multiple regression. The sample size assumption was not violated be cause there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem

PAGE 165

165 with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. In an effort to remedy this assumption violation, the dependent variable was transformed. As suggested by Tabachnick and Fidell (2013), the transformation method of r eflect and log, NEWX= LG10(K X), was used because the distribution of the dependent variable was substantially negatively skewed (p. 89). Following the computation of the transformed dependent variable, the multiple regression was rerun. The Kolmogorov Smi rnov statistic was significant at the p < 0.0 1 level, indicating that the assumption of the normality of error terms was still violated. However, based on the previous discussion of the normality of error terms assumption, the multiple regression was rerun with the non transformed dependent variable. T o err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationsh ip between the antecedents and self efficacy associated with purchasing travel insurance prior to travel . The results of the multiple regression used to assess the ability of the antecedents (demographic factors, international travel specific psychologica l factors, destination specific psychological factors, destination specific factors, and past travel experience) to predict self efficacy associated with purchasing travel insurance prior to travel are presented in Table 4 9 . The total variance explained b y the model as a whole was 7.0 %, F (22, 1,094)= 4.833 , p < 0.001.

PAGE 166

166 T able 4 9 . Multiple regression of the relationship between the antecedents and self efficacy associated with purchasing travel insurance prior to travel . Variable Standardized Coefficient Sig. Gender .110 .000 *** Age: 25 34 .066 .150 Age: 35 49 .077 .095 Age: 50 or older .073 .071 Household income: $35,000 $49,999 .017 .615 Household income: $50,000 $74,999 .014 .708 Household income: $75,000 $99,999 .027 .446 Household income: $100,000 or over .053 .144 Race/ethnicity .028 .359 Education level .063 .049 Marital status .053 .156 Presence of children in the household .007 .849 International travel risk perceptions .093 .004 ** Safety is the most important attribute an international destination can offer .110 .005 ** Safety is a serious consideration when choosing an international tourist destination .120 .001 ** Destination image .064 .038 Presence of friends/relatives living in the destination .037 .251 Fluency in the native language of the destination: not well .000 .999 Fluency in the native language of the destination: well .007 .824 Fluency in the native language of the destination: very well .017 .593 Country level destination past travel experience .038 .242 City level destination past travel experience .030 .341 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.0 70 A preliminary multiple regression was run for the dependent variable of response efficacy associated with purchasing travel insurance prior to travel in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity w as also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance

PAGE 167

167 inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate out liers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. In an effort to remedy this assumption violation, the dependent variable was transformed. As suggested by Tabachnick and Fidell (2013), the transformation method of reflect and square root , NEWX= SQRT(K X), was used because the distribution of the dependent variable was moderately negatively skewed (p. 89). Following the computation of the transformed dependent variable, the multiple regression was rerun. The Kolmogorov Smirnov statistic wa s significant at the p < 0.0 1 level, indicating that the assumption of the normality of error terms was still violated. However, based on the previous discussion of the normality of error terms assumption, the multiple regression was rerun with the non tra nsformed dependent variable. T o err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between the antecedents and response efficacy a ssociated with purchasing travel insurance prior to travel . The results of the multiple regression used to assess the ability of the antecedents (demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) to predict response efficacy associated with purchasing travel insurance prior to travel are presented in Table 4 10 . The total variance explained by the model as a whole was 8.0 %, F (22, 1 ,094)= 5.417 , p < 0.001.

PAGE 168

168 T able 4 10 . Multiple regression of the relationship between the antecedents and response efficacy associated with purchasing travel insurance prior to travel . Variable Sig. Gender .077 .010 Age: 25 34 .108 .017 Age: 35 49 .176 .000 *** Age: 50 or older .114 .005 ** Household income: $35,000 $49,999 .009 .801 Household income: $50,000 $74,999 .007 .852 Household income: $75,000 $99,999 .036 .308 Household income: $100,000 or over .064 .079 Race/ethnicity .091 .003 ** Education level .075 .019 Marital status .007 .859 Presence of children in the household .013 .706 International travel risk perceptions .021 .508 Safety is the most important attribute an international destination can offer .098 .012 Safety is a serious consideration when choosing an international tourist destination .019 .619 Destination image .168 .000 *** Presence of friends/relatives living in the destination .003 .930 Fluency in the native language of the desti nation: not well .006 .847 Fluency in the native language of the destination: well .058 .084 Fluency in the native language of the destination: very well .040 .210 Country level destination past travel experience .054 .092 City level destination past travel experience .010 .734 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0. 080 Hypothesis 1a: Demographic factors predict self efficacy Hypothesis 1a posited that demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) would predict self efficacy . Multiple regression analysis indicated that there was a statistically significant relationship between gender and self efficacy associated with purchasing travel insurance prior to travel at the p < 0.0 0 0.1 10 ). Accordingly, the results indicated that females had significantly higher levels of self efficacy associated with purchasing travel insurance prior to travel th an males.

PAGE 169

169 In addition, m ultiple regression analysis indicated that there were no statistically significant relationships between the following demographic factors and self efficacy associated with purchasing travel insurance prior to travel at the p < 0.0 1 level: age 25 0.066 ), age 35 .077 3 ), household income $35,000 $49,999 17 ) , household income $50,000 $74,999 14 ) , household income $75,000 $99,999 0.027 ) , household income $100,000 or over 05 3 28 0.0 63 0.0 53 07 ). Based on the results of the multiple regression, hypothesis 1a was partially supported . Hypothesis 1b: Demographi c factors predict response efficacy Hypothesis 1b posited that demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) would predict response efficacy. Multiple regres sion analysis indicated that there was a statistically significant relationship between the 35 49 age group and response efficacy associated with purchasing travel insurance prior to travel at the p < 0.0 0 0.1 76 ). Accordingly, the results indicated that respondents between the ages of 35 and 49 had significantly lower levels of response efficacy associated with purchasing travel insurance prior to travel than the reference group (age: 18 24) . Multiple regression analysis also indicated that there was a statistically significant relationship between the 50 or older age group and response efficacy associated with purchasing travel insurance prior to travel 0.1 14 ). Accordingly, the results indicated that respondents a ges 50 or older had significantly lower levels of response efficacy associated with purchasing travel insurance prior to travel than the

PAGE 170

170 reference group (age: 18 24) . Multiple regression analysis indicated that there was also a statistically significant re lationship between race/ethnicity and response efficacy associated with purchasing travel insurance prior to travel 0.0 91 ). Accordingly, the results indicated that White/Caucasian respondents had significan tly lower levels of response efficacy associated with purchasing travel insurance prior to travel than respondents who considered themselves to be of a race/ethnicity other than White/Caucasian In addition, multiple regression analysis indicated that there were no statistically significant relationships between the following demographic factors and response efficacy associated with purchasing travel insurance prior to travel at the p < 0.01 level: 0.07 7 ), age 25 0. 108 ), household income $35,000 $49,999 0.0 09 ) , household income $50,000 $74,999 07 ) , household income $75,000 $99,999 36 ) , household income $100,000 or over 0.0 64 ), education level 0.0 75 0.007 ), and presence of children in the hou 0.01 3 ). Based on the results of the multiple regression, hypothesis 1b was partially supported. Hypothesis 1c: International travel specific psychological factors predict self efficacy Hypothesis 1c posited that international travel specific psychological factors (international travel risk perceptions and safety concerns) would predict self efficacy. Multiple regression analysis indicated that there was a statistically significant relati onship between international travel risk perceptions and self efficacy associated with purchasing travel insurance prior to travel 0.0 93 ). Accordingly, the results indicated that those with lower levels of international travel ri sk

PAGE 171

171 perceptions had significantly higher levels of self efficacy associated with purchasing travel insurance prior to travel than those with higher levels of international travel risk perceptions. In addition, multiple regression analysis indicated that the re were statistically significant relationships between the two safety concerns items and self efficacy associated with purchasing travel insurance prior to travel is the most important attribute an international destination 0 ) and 0.1 20 ). Accordingly, the results indicated that those with higher levels of agreement with ribute an international destination can efficacy associated with purchasing travel insurance prior to travel than thos e with lower levels of agreement with the statements. Based on the results of the multiple regression, hypothesis 1c was not rejected . Hypothesis 1d: International travel specific psychological factors predict response efficacy Hypothesis 1d posited that international travel specific psychological factors (international travel risk perceptions and safety concerns) would predict response efficacy. Multiple regression analysis indicated that there was no statistically significant relationship between interna tional travel risk perceptions and response efficacy associated with purchasing travel insurance prior to travel 0.0 21 ).

PAGE 172

172 In addition, multiple regression analysis indicated that there were no statistically s ignificant relationships between the two safety concerns items and response efficacy associated with purchasing travel insurance prior to travel 09 8 ) and 0.019 ). Based on the results of the multiple regression, hypothesis 1d was rejected . Hypothesis 1e: Destination specific psychological factors predict self ef ficacy Hypothesis 1e posited that destination specific psychological factors (destination image) would predict self efficacy. Multiple regression analysis indicated that there was no statistically significant relationship between destination image and sel f efficacy associated with purchasing travel insurance prior to travel 0.0 64 ). Based on the results of the multiple regression, hypothesis 1e was rejected. Hypothesis 1f: Destination specific psychological factors predict response efficacy Hypothesis 1f posited that destination specific psychological factors (destination image) would predict response efficacy. Multiple regression analysis indicated that there was a statistically significant relationship between destination image a nd response efficacy associated with purchasing travel insurance prior to travel at the p < 0.001 level 68 ). Accordingly, the results indicated that those with a more positive destination image had significantly higher levels of response efficacy associated with purchasing travel insurance prior to travel than t hose with a more negative destination image. Based on the results of the multiple regression, hypothesis 1f was not rejected.

PAGE 173

173 Hypothesis 1g: Destination specific factors predict self effi cacy Hypothesis 1g posited that destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of the destination) would predict self efficacy . Multiple regression analysis indicated that there wer e no statistically significant relationships between the destination specific factors and self efficacy associated with purchasing travel insurance prior to travel at the p < 0.0 1 level: presence of friends and/or 0. 037 ), fluency in the native language of 00 ), fluency in the 0.00 7 ), and fluency in the native language 1 7 ). Based on the results of the mult iple regression, hypothesis 1g was rejected. Hypothesis 1h: Destination specific factors predict response efficacy Hypothesis 1h posited that destination specific factors (presence of friends/relatives living in the destination and fluency in the native l anguage of the destination) would predict response efficacy . Multiple regression analysis indicated that there were no statistically significant relationships between the destination specific factors and response efficacy associated with purchasing travel insurance prior to travel at the p < 0.0 1 level: presence of friends and/or 0.003 ), fluency in the native language of ), fluency in the native language of the destination: wel 0.058 ), and fluency in the native 0.040 ). Based on the results of the multiple regression, hypothesis 1 h was rejected .

PAGE 174

174 Hypothesis 1i: Past travel experience predicts self efficacy Hypothesis 1i posited th at past travel experience (country level destination past travel experience and city level destination past travel experience ) would predict self efficacy . Multiple regression analysis indicated that there were no statistically significant relationships between the two types of past travel experience and self efficacy associated with purchasing travel insurance prior to travel at the p < 0.0 1 level: country level destination past travel experience 0.0 38 ) and city level destination past travel experience 0.0 30 ). Based on the results of the multiple regression, hypothesis 1 i was rejected. Hypothesis 1j: Past travel experience predicts response efficac y Hypothesis 1j posited that past travel experience (country level destination past travel experience and city level destination past travel experience ) would predict response efficacy . Multiple regression analysis indicated that there were no statistical ly significant relationships between the two types of past travel experience and response efficacy associated with purchasing travel insurance prior to travel at the p < 0.0 1 level: country level destination past travel experience 0.0 54 ) and city leve l destination past travel experience 0.01 0 ). Based on the results of the multiple regression, hypothesis 1 j was rejected. Research Question 2: Does Perceived Risk Mediate the Relationship between the Antecedents and Perceived Efficacy? Research question 2 focused on examining if perceived risk (perceived vulnerability, affective risk perceptions, and perceived severity) mediated the relationship between the antecedents (demographic factors, international travel specific psychological fac tors, destination specific psychological factors, destination specific

PAGE 175

175 factors, and past travel experience) and perceived efficacy (self efficacy and response efficacy). To test research question 2, a series of multiple regressions were conducted using SPS S. Indirect paths between the antecedents and perceived risk In the first series of multiple regressions, the relationships between the antecedents ( demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience ) and perceived risk ( perceived vulnerability / affective risk perceptions and perceived severity ) were examined. P erceived vulnerability / affective risk perceptions was the depend ent variable in the first multiple regression . The results of the first multiple regression can be found in Table 4 3. Perceived severity was the dependent variable in the second multiple regression. The results of the second multiple regression can be fou nd in Table 4 4. The first series of multiple regressions provided the indirect paths between the antecedents ( independent variables ) and perceived risk (first set of mediating variables). Indirect paths between perceived risk and perceived efficacy In t he second series of multiple regressions, the relationships between perceived risk ( perceived vulnerability / affective risk perceptions and perceived severity) and perceived efficacy (self efficacy and response efficacy) were examined . Self efficacy associa ted with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the first multiple regression . R esponse efficacy associated with a behavior that tourists could engag e in to ensure their personal safety while visiting the destination of Acapulco was the

PAGE 176

17 6 dependent variable in the second multiple regression. The second series of multiple regressions provided the indirect paths between perceived risk (first set of mediati ng variables) and perceived efficacy (second set of mediating variables). A preliminary multiple regression was run for the dependent variable of self efficacy associated with purchasing travel insurance prior to travel in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independe nt variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. Based on the previous discussion of the normality of error terms assumption, to e rr on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between perceived risk and self efficacy associated with purchasing travel insur ance prior to travel . The results of the multiple regression used to assess the ability of perceived risk (perceived vulnerability / affective risk perceptions and perceived severity) to predict self efficacy associated with purchasing travel insurance prio r to travel are presented in

PAGE 177

177 Table 4 11 . The model was not significant at the p < 0.05 level and the total variance explained by the model as a whole was 0.3 %, F ( 2 , 1,122 )= 2.707 , p > 0.0 5 . T able 4 11 . Multiple regression of the relationship between perc eived risk and self efficacy associated with purchasing travel insurance prior to travel . Variable Sig. Perceived vulnerability/affective risk perceptions .044 .164 Perceived severity .069 .028 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.003 A preliminary multiple regression was run for the dependent variable of response efficacy associated with purchasing travel insurance prior to travel in order to test the assumptions of multiple regression . The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a combination o f other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statist ics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirn ov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. Based on the previous discussion of the normality of error terms assumption, to err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between perceived risk and response efficacy associated with purchasing travel insurance prior to travel .

PAGE 178

178 The results of the multiple regression used to assess the ability of perceived risk (perceived vulnerability / affective risk perceptions and perceived severity) to predict response efficacy associated with purchasing travel insurance prior to travel are presented in Table 4 12 . The model was not significant at the p < 0.05 level and the total variance explained by the model as a whole was 0.2 %, F (2, 1,122 )= 2.095 , p > 0.0 5 . T able 4 12 . Multiple regression of the relationship between perceived risk and response efficacy ass ociated with purchasing travel insurance prior to travel . Variable Sig. Perceived vulnerability/affective risk perceptions .052 .094 Perceived severity .019 .544 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.002 Given that none of the indirect paths between the perceived risk variables (first set of mediating variables) and perceived efficacy (second set of mediating variables) were significant in the second series of multiple regressions, a third series of multiple regressions was not conducted. In addition, since mediation was not found, an online calculation of the Sobel test was not conducted. Hypothesis 2a: Perceived vulnerability/affective risk percepti ons mediates the relationship between demographic factors and self efficacy Hypothesis 2a posited that perceived vulnerability/affective risk perceptions mediated the relationship between demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) and self efficacy . The relationships between demographic factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Aca pulco were discussed in the section on hypothesis 2a for the risk reduction behavior of registering with the U.S. Department of State prior to travel.

PAGE 179

179 The second series of multiple regressions indicated that there was no statistically significant relations hip between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and self efficacy associated with purchasing travel insurance prior to travel at the p < 0.01 level 0. 044 ) . Based on the results of the two series of multiple regressions, hypothesis 2a was rejected. Hypothesis 2b: Perceived vulnerability/affective risk perceptions mediates the relationship between demographic factors and response efficacy Hypothesis 2b p osited that perceived vulnerability/affective risk perceptions mediated the relationship between demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) and response ef ficacy . The relationships between demographic factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2a for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The second series of multiple regressions indicated that there was no statistically significant relationship s between the perceived vulnerability/affective risk perceptions associated with crim e while visiting the destination of Acapulco and response efficacy associated with purchasing travel insurance prior to travel at the p < 0.01 level 0.052 ) . Based on the results of the two series of multiple regressions, hypothesis 2b was rejected.

PAGE 180

180 Hypothesis 2c: Perceived vulnerability/affective risk perceptions mediates the relationship between international travel specific psychological factors and self efficacy Hypothesis 2c posited that perceived vulnerability/affective risk perceptions mediated the relationship between international travel specific psychological factors (international travel risk perceptions and safety concerns) and self efficacy . The relationships between international travel spe cific psychological factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco w ere discussed in the section on hypothesis 2 c for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and self efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2 a. Based on the results of the two series of multiple regressions, hypothesis 2c was rejected. Hypothesis 2d: Perceived vulnerability/affective risk perceptions mediates the relationship between international tr avel specific psychological factors and response efficacy Hypothesis 2d posited that perceived vulnerability/affective risk perceptions mediated the relationship between international travel specific psychological factors (international travel risk percep tions and safety concerns) and response efficacy . The relationships between international travel specific psychological factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were dis cussed in the section on hypothesis 2 c for the risk reduction behavior of registering with the U.S. Department of State prior to travel .

PAGE 181

181 The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the d estination of Acapulco and response efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2b . Based on the results of the two series of multiple regressions , hypothesis 2d was rejected. Hypothesis 2 e: Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific psychological factors and self efficacy Hypothesis 2e posited that perceived vulnerability/affective risk perceptions mediated the relationship be tween destination specific psychological factors (destination image) and self efficacy . The relationship s between destination specific psychological factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the des tination of Acapulco were discussed in the section on hypothesis 2e for the risk reduction behavior of registering with the U.S. Department of State prior to travel . The relationship between perceived vulnerability/affective risk perceptions associated wit h crime while visiting the destination of Acapulco and self efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2 a. Based on the results of the two series of multiple regressions, hypothesis 2e wa s rejected. Hypothesis 2f: Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific psychological factors and response efficacy Hypothesis 2f posited that perceived vulnerability/affective risk perceptions mediated the relationship between destination specific psychological factors

PAGE 182

182 (destination image) and response efficacy. The relationships between destination specific psychologic al factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2e for the risk reduction behavior of registering with the U.S. Department of Sta te prior to travel . The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and response efficacy associated with purchasing travel insurance prior to travel was discusse d in the section on hypothesis 2b . Based on the results of the two series of multiple regressions, hypothesis 2f was rejected. Hypothesis 2g: Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific factors and self efficacy Hypothesis 2g posited that perceived vulnerability/affective risk perceptions mediated the relationship between destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of t he destination) and self efficacy. The relationships between destination specific factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2g for the risk reduction behavior of registering with the U.S. Department of State prior to travel . The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and self efficac y associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2 a. Based on the results of the two series of multiple regressions, hypothesis 2g was rejected.

PAGE 183

183 Hypothesis 2h: Perceived vulnerability/affective risk p erceptions mediates the relationship between destination specific factors and response efficacy Hypothesis 2h posited that perceived vulnerability/affective risk perceptions mediated the relationship between destination specific factors (presence of frien ds/relatives living in the destination and fluency in the native language of the destination) and response efficacy . The relationships between destination specific factors and perceived vulnerability/affective risk perceptions associated with crime while v isiting the destination of Acapulco were discussed in the section on hypothesis 2g for the risk reduction behavior of registering with the U.S. Department of State prior to travel . The relationship between perceived vulnerability/affective risk perception s associated with crime while visiting the destination of Acapulco and response efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2b . Based on the results of the two series of multiple regressio ns, hypothesis 2h was rejected. Hypothesis 2i: Perceived vulnerability/affective risk perceptions mediates the relationship between past travel experience and self efficacy Hypothesis 2 i posited that perceived vulnerability/affective risk perceptions mediated the relationship between past travel experience (country level destination past travel experience and city level destination past travel experience ) and self efficacy. The relationships between past travel experience and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2i for the risk reduction behavior of registeri ng with the U.S. Department of State prior to travel.

PAGE 184

184 The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and self efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2 a. Based on the results of the two series of multiple regressions, hypothesis 2i was rejected. Hypothesis 2j: Perceived vulnerability/affective risk perceptions mediates the relationship between past travel experience and response efficacy Hypothesis 2j posited that perceived vulnerability/affective risk perceptions mediated the relationship between past travel experience (country level destination past travel experience and city level destinatio n past travel experience ) and response efficacy . The relationships between past travel experience and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were discussed in the section on hypot hesis 2i for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and resp onse efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2b . Based on the results of the two series of multiple regressions, hypothesis 2j was rejected. Hypothesis 2k: Perceived severity mediates the relationship between demographic factors and self efficacy Hypothesis 2k posited that perceived severity mediated the relationship between demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presenc e of children in the household) and self efficacy . The

PAGE 185

185 relationships between demographic factors and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2k for the risk reduction b ehavior of registering with the U.S. Department of State prior to travel. The second series of multiple regressions indicated that there was no statistically significant relationship between perceived severity associated with crime while visiting the dest ination of Acapulco and self efficacy associated with purchasing travel insurance prior to travel 069 ). Based on the results of the two series of multiple regressions, hypothesis 2k was rejected. Hypothesis 2l: Perceived severity mediates the relationship between demographic factors and response efficacy Hypothesis 2l posited that perceived severity mediated the relationship between demographic factors (gender, age, household income, race/ethnicity, education level, marit al status, and presence of children in the household) and response efficacy . The relationships between demographic factors and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2 k for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The second series of multiple regressions indicated that there was no statistically significant relationship s between perceived severity associated with c rime while visiting the destination of Acapulco and response efficacy associated with purchasing travel insurance prior to travel at the p < 0.01 level 0.019 ) . Based on the results of the two series of multiple regressions, hypothesis 2l was rejected.

PAGE 186

186 Hypothesis 2m: Perceived severity mediates the relationship between international travel specific psychological factors and self efficacy Hypothesis 2m posited that perceived severity mediated the relationship between international travel specific psycho logical factors (international travel risk perceptions and safety concerns) and self efficacy . The relationships between international travel specific psychological factors and perceived severity associated with crime while visiting the destination of Acap ulco were discussed in the section on hypothesis 2m for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived severity associated with crime while visiting the destination of Acapul co and self efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2k . Based on the results of the two series of multiple regressions, hypothesis 2m was rejected. Hypothesis 2n: Perceived severity me diates the relationship between international travel specific psychological factors and response efficacy Hypothesis 2n posited that perceived severity mediated the relationship between international travel specific psychological factors (international travel risk perceptions and safety concerns) and response efficacy . The relationships between international t ravel specific psychological factors and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2m for the risk reduction behavior of registering with the U.S. Department of State pri or to travel. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and response efficacy associated with purchasing travel

PAGE 187

187 insurance prior to travel was discussed in the section on hypothesis 2l . Bas ed on the results of the two series of multiple regressions, hypothesis 2n was rejected. Hypothesis 2o: Perceived severity mediates the relationship between destination specific psychological factors and self efficacy Hypothesis 2o posited that perceived severity mediated the relationship between destination specific psychological factors (destination image) and self efficacy . The relationships between destination specific psychological factors and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2o for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived severity a ssociated with crime while visiting the destination of Acapulco and self efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2k . Based on the results of the two series of multiple regressions, hyp othesis 2o was rejected. Hypothesis 2p: Perceived severity mediates the relationship between destination specific psychological factors and response efficacy Hypothesis 2 p posited that perceived severity mediated the relationship between destination speci fic psychological factors (destination image) and response efficacy. The relationships between destination specific psychological factors and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2o for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and response efficacy associate d with purchasing travel

PAGE 188

188 insurance prior to travel was discussed in the section on hypothesis 2l . Based on the results of the two series of multiple regressions, hypothesis 2p was rejected. Hypothesis 2q: Perceived severity mediates the relationship betwee n destination specific factors and self efficacy Hypothesis 2q posited that perceived severity mediated the relationship between destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of th e destination) and self efficacy. The relationships between destination specific factors and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2q for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and self efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2k . Based on the results of the two series of multiple regressions, hypothesis 2q was rejected. Hypothesis 2r: Perceived severity mediates the relationship between destination specific factors and response efficacy Hypothesis 2r posited that perceived severity mediated the relationship between destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of the destination) and response efficacy . The relation ships between destination specific factors and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2q for the risk reduction behavior of registering with the U.S. Department of State prior to travel.

PAGE 189

189 The relationship between perceived severity associated with crime while visiting the destination of Acapulco and response efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2l . Based on the results of the two series of multiple regressions, hypothesis 2r was rejected. Hypothesis 2s: Perceived severity m ediates the relationship between past travel experience and self efficacy Hypothesis 2 s posited that perceived severity mediated the relationship between past travel experience (country level destination past travel experience and city level destination p ast travel experience ) and self efficacy. The relationships between past travel experience and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2s for the risk reduction behavio r of registering with the U.S. Department of State prior to travel. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and self efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2k . Based on the results of the two series of multiple regressions, hypothesis 2s was rejected. Hypothesis 2t: Perceived severity mediates the relationship between past travel experience and response efficacy Hypothesis 2t posited that perceived severity mediated the relationship between past travel experience (country level destination past travel experience and city level destination past travel experience ) and response efficacy . The relationships between pas t travel experience and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2s for the risk reduction behavior of registering with the U.S. Department of State prior to travel.

PAGE 190

190 Th e relationship between perceived severity associated with crime while visiting the destination of Acapulco and response efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2l . Based on the results of the two series of multiple regressions, hypothesis 2t was rejected. Research Question 3: Does Perceived Risk Predict Intentions to Engage in a Recommended Risk Reduction Behavior? Research question 3 focused on examining if perceived risk (perceived vu lnerability / affective risk perceptions and perceived severity) predicted intentions to engage in a recommended risk reduction behavior. To test research question 3, one multiple regression was conducted using SPSS. Intentions to engage in a recommended risk reduction behavior to ensure personal safety while visiting the destination of Acapulco was the dependent variable in the multiple regression . The analysis of research question 3 provided the total effects o f perceived risk (first set of mediating variables) on intentions to engage in the recommended risk reduction behavior (dependent variable) of purchasing travel insurance prior to travel . A preliminary multiple regression was run for the dependent variable of intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity s tatistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating

PAGE 191

191 that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. In an effort to remedy this assumption violation, the dependent variable was transformed. As suggested by Tabachnick and Fidell (2013), the transformation method of reflect and square root , NEWX= SQRT(K X), was u sed because the distribution of the dependent variable was moderately negatively skewed (p. 89). Following the computation of the transformed dependent variable, the multiple regression was rerun. The Kolmogorov Smirnov statistic was significant at the p < 0.0 1 level, indicating that the assumption of the normality of error terms was still violated. However, based on the previous discussion of the normality of error terms assumption, the multiple regression was rerun with the non transformed dependent varia ble. T o err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between perceived risk and intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel . The results of the multiple regression used to assess the ability of perceived risk (perceived vulnerability / affective risk perceptions and perceived severity) to predict intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel are presented in Table 4 13 . The model was not significant at

PAGE 192

192 the p < 0.05 level and the total variance explained by the model as a whole was 0.2 %, F ( 2, 1,122 )= 1.883, p > 0.0 5 . T able 4 13 . Multiple regression of the relationship between perceived risk and intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel . Variable Standardized Coef Sig. Perceived vulnerability/affective risk perceptions .045 .153 Perceived severity .025 .418 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.002 Hypothesis 3a: Perceived vulnerability/affective risk perceptions predicts intentions to engage in a recommended risk reduction behavior Hypothesis 3a posited that perceived vulnerability/affective risk perceptions would predict intentions to engage in a recommended risk reduction behavior . Multiple regression analysis indicated that there was no statistically significant relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and intentions to engage in the recommended risk reduction behavi or of purchasing travel insurance prior to travel at the p < 0.0 1 level 0. 045 ). Based on the results of the multiple regression, hypothesis 3a was rejected. Hypothesis 3b: Perceived severity predicts intentions to engage in a recommended risk reduction behavior Hypothesis 3b posited that perceived severity would predict intentions to engage in a recommended risk reduction behavior . Multiple regression analysis indicated that there was no statistically significant relationship between perceived severity associated with crime while visiting the destination of Acapulco and intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel at the p < 0.0 1 level 0.025 ). Based on the results of the multiple regr ession, hypothesis 3b was rejected.

PAGE 193

193 Research Question 4: Does Perceived Efficacy Mediate the Relationship between Perceived Risk and Intentions to Engage in a Recommended Risk Reduction Behavior? Research question 4 focused on examining if perceived effica cy (self efficacy and response efficacy) mediated the relationship between perceived risk ( perceived vulnerability / affective risk perceptions and perceived severity) and intentions to engage in a recommended risk reduction behavior. To test research questi on 4, a series of multiple regressions were conducted using SPSS. Indirect paths between perceived risk and perceived efficacy In the first series of multiple regressions, the relationships between perceived risk ( perceived vulnerability / affective risk p erceptions and perceived severity) and perceived efficacy (self efficacy and response efficacy) were examined. Self efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the first multiple regression . The results of the first multiple regression can be found in Table 4 11. R esponse efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the second multiple regression. The results of the second multiple regression can be found in Table 4 12. The first series of multiple regressions provided the indirect paths between perceived risk (first set of mediating variables) and perceived efficacy (second set of mediating variables). Indirect paths between perceived efficacy and intentions to engage in a recommended risk reduction behavior In the second series of multiple regressions, the relationships between perceived efficacy (self efficacy and response efficacy) and intentions to engage in a

PAGE 194

194 recommended risk reduction behavior were examined. Intentions to engage in a recommended risk re duction behavior to ensure personal safety while visiting the destination of Acapulco was the dependent variable in the multiple regression . T he second series of multiple regressions provided the indirect paths between perceived efficacy (second set of mediating variables) and intentions to engage in a recommended risk reduction behavior (dependent variable). A preliminary multiple regression was run for the dependent variable of intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of si ngularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multi variate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. Based on the previous discus sion of the normality of error terms assumption, to err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the

PAGE 195

195 relationship between perceived efficacy and intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel . The results of the multiple regression used to assess the ability of perceived efficacy ( self efficacy and response efficacy ) to predict intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel are presented in Table 4 14 . The total variance explained by the model as a whole was 48.5 %, F (2, 1,123 )= 531.053 , p < 0.0 01 . T able 4 14 . Mult iple regression of the relationship between perceived efficacy and intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel . Variable Sig. Self efficacy .470 .000*** Response efficacy .365 .000*** Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0. 485 Given that none of the indirect paths between the perceived risk variables (first set of mediating variables) and perceived efficacy (second set of mediating variables) were significant in the first series of multiple regressions, a third series of multiple regressions was not conducted. In addition, since mediation was not foun d, an online calculation of th e Sobel test was not conducted. Hypothesis 4a: Self efficacy mediates the relationship between perceived vulnerability/affective risk perceptions and intentions to engage in a recommended risk reduction behavior Hypothesis 4a posited that self efficacy would mediate the relationship between perceived vulnerability/affective risk perceptions and intentions to engage in a recommended risk reduction behavior. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the

PAGE 196

196 destination of Acapulco and self efficacy associated with purchasing travel insurance prior to travel were discussed in the section on hypothesis 2a . The second series of multiple regressions indi cated that there was a statistically significant relationship between self efficacy associated with purchasing travel insurance prior to travel and intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to trav el at the p < 0.001 level 0.470 ) . Accordingly, the results indicated that those with higher levels of self efficacy associated with purchasing travel insurance prior to travel were significantly more likely to purchase travel insurance prior to travel than those with lower levels of self efficacy. B ased on the results of the two series of multiple regressions, hypothesis 4a was rejected. Hypothesis 4b: Self efficacy mediates the relationship between perceived severity and intentions to engage in a recom mended risk reduction behavior Hypothesis 4 b posited that self efficacy would mediate the relationship between perceived severity and intentions to engage in a recommended risk reduction behavior. The relationship between perceived severity associated wit h crime while visiting the destination of Acapulco and self efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2k . The relationship between self efficacy associated with purchasing travel insura nce prior to travel and intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel was discussed in the section on hypothesis 4a. Based on the results of the two series of multiple regressions, hypothesis 4b was rejected.

PAGE 197

197 Hypothesis 4c: Response efficacy mediates the relationship b etween perceived vulnerability/ affective risk perceptions and intentions to engage in a recommended risk reduction behavior Hypothesis 4 c posited that response efficacy would mediate the relationship between perceived vulnerability/affective risk perceptions and intentions to engage in a recommended risk reduction behavior. The relationship between perceived vulnerability/affective risk perceptions associated with crime while v isiting the destination of Acapulco and response efficacy associated with re purchasing travel insurance prior to travel was discussed in the section on hypothesis 2b . The second series of multiple regressions indicated that there was a statistically sign ificant relationship between response efficacy associated with purchasing travel insurance prior to travel and intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel at the p < 0.001 level ( ) . Accordingly, the results indicated that those with higher levels of response efficacy associated with purchasing travel insurance prior to travel were significantly more likely to purchase travel insurance prior to travel than those with lower levels of r esponse efficacy. Based on the results of the two series of multiple regressions, hypothesis 4c was rejected. Hypothesis 4d: Response efficacy mediates the relationship between perceived severity and intentions to engage in a recommended risk reduction beh avior Hypothesis 4 d posited that response efficacy would mediate the relationship between perceived severity and intentions to engage in a recommended risk reduction behavior. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and response efficacy associated with purchasing travel insurance prior to travel was discussed in the section on hypothesis 2l .

PAGE 198

198 The relationship between response efficacy associated with p urchasing travel insurance prior to travel and intentions to engage in the recommended risk reduction behavior of purchasing travel insurance prior to travel was discussed in the section on hypothesis 4c. Based on the results of the two series of multiple regressions, hypothesis 4d was rejected. Avoid Traveling Outside of the Hotel Zone Research Question 1: Do the Antecedents Predict Perceived Efficacy? Research question 1 focused on examining if the antecedents (demographic factors, international travel sp ecific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) predicted perceived efficacy (self efficacy and response efficacy). To test research question 1, two multiple regressions we re conducted using SPSS. Self efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the first multiple regression . R esponse efficacy associ ated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the second multiple regression. The analysis of research question 1 provided the total effects of th e antecedents (independent variables) on perceived efficacy (second set of mediating variables) associated with the recommended risk reduction behavior of avoiding traveling outside of the hotel zone . A preliminary multiple regression was run for the depen dent variable of self efficacy associated with avoiding traveling outside of the hotel zone in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable ( Hair et al., 2010, p.

PAGE 199

199 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all with in the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156) . Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. In an effort to remedy this assumption violation, the dependent variable was transformed. As suggested by Tabachnick and Fidell (2013), the transformation method of r eflect and log, NEWX= LG10(K X), was used because the distribution of the dependent variable was substantially negatively skewed (p. 89). Following the computation of the transformed dependent variable, the multiple regression was rerun. The Kolmogorov Smi rnov statistic was significant at the p < 0.0 1 level, indicating that the assumption of the normality of error terms was still violated. However, based on the previous discussion of the normality of error terms assumption, the multiple regression was rerun with the non transformed dependent variable. T o err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between the antecedents and se lf efficacy associated with avoiding traveling outside of the hotel zone . The results of the multiple regression used to assess the ability of the antecedents (demographic factors, international travel specific psychological factors,

PAGE 200

200 destination specific psychological factors, destination specific factors, and past travel experience) to predict self efficacy associated with avoiding traveling outside of the hotel zone are presented in Table 4 15 . The total variance explained by the model as a whole was 9.3 %, F (22, 1,094)= 6.175 , p < 0.001. T able 4 15 . Multiple regression of the relationship between the antecedents and self efficacy associated with avoiding traveling outside of the hotel zone . Variable Sig. Gender .057 .051 Age: 25 34 .061 .175 Age: 35 49 .047 .305 Age: 50 or older .041 .313 Household income: $35,000 $49,999 .009 .795 Household income: $50,000 $74,999 .055 .126 Household income: $75,000 $99,999 .075 .035 Household income: $100,000 or over .069 .056 Race/ethnicity .027 .367 Education level .085 .008 ** Marital status .065 .081 Presence of children in the household .006 .854 International travel risk perceptions .008 .810 Safety is the most important attribute an international destination can offer .175 .000 *** Safety is a serious consideration when choosing an international tourist destination .114 .002 ** Destination image .029 .347 Presence of friends/relatives living in the destination .083 .009 ** Fluency in the native language of the destination: not well .009 .790 Fluency in the native language of the destination: well .013 .694 Fluency in the native language of the destination: very well .004 .891 Country level destination past travel experience .006 .845 City level destination past travel experience .027 .387 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.0 93 A preliminary multiple regression was run for the dependent variable of response efficacy associated with avoiding traveling outside of the hotel zone in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p.

PAGE 201

201 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. In an effort to remedy this assumption violation, the dependent va riable was transformed. As suggested by Tabachnick and Fidell (2013), the transformation method of reflect and log, NEWX= LG10(K X), was used because the distribution of the dependent variable was substantially negatively skewed (p. 89). Following the comp utation of the transformed dependent variable, the multiple regression was rerun. The Kolmogorov Smirnov statistic was significant at the p < 0.0 1 level, indicating that the assumption of the normality of error terms was still violated. However, based on t he previous discussion of the normality of error terms assumption, the multiple regression was rerun with the non transformed dependent variable. T o err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting t he results of the multiple regression which examined the relationship between the antecedents and response efficacy associated with avoiding traveling outside of the hotel zone . The results of the multiple regression used to assess the ability of the ante cedents (demographic factors, international travel specific psychological factors,

PAGE 202

202 destination specific psychological factors, destination specific factors, and past travel experience) to predict response efficacy associated with avoiding traveling outside of the hotel zone are presented in Table 4 16 . The total variance explained by the model as a whole was 7.5 %, F (22, 1,094)= 5. 119 , p < 0.001. T able 4 16 . Multiple regression of the relationship between the antecedents and response efficacy associated w ith avoiding traveling outside of the hotel zone . Variable Sig. Gender .020 .498 Age: 25 34 .049 .282 Age: 35 49 .071 .121 Age: 50 or older .039 .336 Household income: $35,000 $49,999 .019 .570 Household income: $50,000 $74,999 .080 .028 Household income: $75,000 $99,999 .048 .185 Household income: $100,000 or over .068 .061 Race/ethnicity .003 .930 Education level .063 .050 Marital status .039 .296 Presence of children in the household .023 .516 International travel risk perceptions .022 .485 Safety is the most important attribute an international destination can offer .104 .008 ** Safety is a serious consideration when choosing an international tourist destination .176 .000 *** Destination image .133 .000 *** Presence of friends/relatives living in the destination .050 .116 Fluency in the native language of the destination: not well .018 .579 Fluency in the native language of the destination: well .005 .874 Fluency in the native languag e of the destination: very well .030 .352 Country level destination past travel experience .011 .742 City level destination past travel experience .041 .186 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0. 0 75 Hypothesis 1a: Demographic factors predict self efficacy Hypothesis 1a posited that demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) would predict self efficacy . Multiple regression analysis indicated that there

PAGE 203

203 was a statistically significant relationship between education level and self efficacy associated with avoiding traveling outside of the hotel zone 0. 085 ). According ly, the results indicated that those with an Associate degree or lower had significantly higher levels of self efficacy associated with avoiding traveling outside of the hotel zone than . In addition, multiple regre ssion analysis indicated that there were no statistically significant relationships between the following demographic factors and self efficacy associated with avoiding traveling outside of the hotel zone at the p < 0.01 level: gender 0.057), age 25 3 0.06 1 ), age 35 0.047 0.041 ), household income $35,000 $49,999 0.009 ) , household income $50,000 $74,999 0.055 ) , household income $75,000 $99,999 0.075 ) , household income $100,000 or over 0.069 ), race/eth 7 65 ), and 0.00 6 ). Based on the results of the multiple regression, hypothesis 1a was partially supported. Hypothesis 1b: Demographic factors predict response efficacy Hypot hesis 1b posited that demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) would predict response efficacy. Multiple regression analysis indicated that there were no statistically significant relationships between the demographic factors and response efficacy associated with avoiding traveling outside of the hotel zone at the p < 0.049), age 35 0.071), age 50 or 0.039), household income $35,000 $50,000

PAGE 204

204 household incom 0.003), education 0.063 0.039 ), and presence of children in the household 0.023 ). Based on the results of the multiple regression, hypothesis 1b was rejected. Hypothes is 1c: International travel specific psychological factors predict self efficacy Hypothesis 1c posited that international travel specific psychological factors (international travel risk perceptions and safety concerns) would predict self efficacy. Multip le regression analysis indicated that there was no statistically significant relationship between international travel risk perceptions and self efficacy associated with avoiding traveling outside of the hotel zone 0.0 08 ). In ad dition, multiple regression analysis indicated that there were statistically significant relationships between the two safety concerns items and self efficacy associated with avoiding traveling outside of the hotel zone at the p < 0.0 0 1 level and p < 0.01 level s , respectively 75 14 ). Accordingly, the results indicated that those with higher levels of vels of self efficacy associated with avoiding traveling outside of the hotel zone than those with lower levels of agreement with the statements. Based on the results of the multiple regression, hypothesis 1c was partially supported .

PAGE 205

205 Hypothesis 1d: Intern ational travel specific psychological factors predict response efficacy Hypothesis 1d posited that international travel specific psychological factors (international travel risk perceptions and safety concerns) would predict response efficacy. Multiple re gression analysis indicated that there was no statistically significant relationship between international travel risk perceptions and response efficacy associated with avoiding traveling outside of the hotel zone 0.02 2 ). In addition, multiple regression analysis indicated that there were statistically significant relationships between the two safety concerns items and response efficacy associated with avoiding traveling outside of the hotel zone at the p < 0 .01 level and p < the results indicated that those with higher levels of significantly higher levels of response efficacy associated with avoiding traveling outside of the hotel zone than those with lower levels of agreement with the statements. Based on the results of the multiple regression, hypothesis 1d was partially suppor ted. Hypothesis 1e: Destination specific psychological factors predict self efficacy Hypothesis 1e posited that destination specific psychological factors (destination image) would predict self efficacy. Multiple regression analysis indicated that there was no statistically significant relationship between destination image and self efficacy

PAGE 206

206 associated with avoiding traveling outside of the hotel zone 0.0 29 ). Based on the results of the multiple regression, hypothesis 1e was rejected. Hypothesis 1f: Destination specific psychological factors predict response efficacy Hypothesis 1f posited that destination specific psychological factors (destination image) would predict response efficacy. Multiple regression analysis indicated that there was a statistically significant relationship between destination image and response efficacy associated with avoiding traveling outside of the hotel zone at the p < 0.001 33 ). Accordingly, the results indicated that those with a mo re positive destination image had significantly higher levels of response efficacy associated with avoiding traveling outside of the hotel zone than t hose with a more negative destination image. Based on the results of the multiple regression, hypothesis 1 f was not rejected. Hypothesis 1g: Destination specific factors predict self efficacy Hypothesis 1g posited that destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of the destination) w ould predict self efficacy . Multiple regression analysis indicated that there was a statistically significant relationship between presence of friends and/or relatives living in the destination and self efficacy associated with avoiding traveling outside o f the hotel zone at the p < 0.0 1 level 0.0 83 ) . Accordingly, the results indicated that those without friends and/or relatives living in the destination had significantly higher levels of self efficacy associated with avoiding traveling outside of the hotel zone than those with friends and/or relative s living in the destination. In addition, m ultiple regression analysis indicated that there were no statistically significant relationships between the following destination specific factors and self -

PAGE 207

207 efficacy associated with avoiding traveling outside of the hotel zone at the p < 0.0 1 level: fluency in the native language of 0.009 ), fluency in 0.013 ), and fluency in the nat ive 0.004 ). Based on the results of the multiple regression, hypothesis 1g was partially supported . Hypothesis 1h: Destination specific factors predict response efficacy Hypothesis 1h posited that destination spe cific factors (presence of friends/relatives living in the destination and fluency in the native language of the destination) would predict response efficacy . Multiple regression analysis indicated that there were no statistically significant relationships between the destination specific factors and response efficacy associated with avoiding traveling outside of the hotel zone at the p < 0.0 1 level: presence of friends and/or relatives living in the destination 0.0 50 ), fluency in the native language o f 0.018 ), 0.0 05 ), and fluency in the 0.030 ). Based on the results of the multiple regression, hypothesis 1 h was rej ected . Hypothesis 1i: Past travel experience predicts self efficacy Hypothesis 1i posited that past travel experience (country level destination past travel experience and city level destination past travel experience ) would predict self efficacy . Multipl e regression analysis indicated that there were no statistically significant relationships between the two types of past travel experience and self efficacy associated with avoiding traveling outside of the hotel zone at the p < 0.0 1 level: country level d estination past travel experience 0.0 06 ) and city level destination past travel

PAGE 208

208 experience 0.0 27 ). Based on the results of the multiple regression, hypothesis 1 i was rejected. Hypothesis 1j: Past travel experience predicts response efficacy Hypothesis 1j posited that past travel experience (country level destination past travel experience and city level destination past travel experience ) would predict response efficacy . Multiple regression analysis indicated that there were no statistically significant relationships between the two types of past travel experience and response efficacy associated with avoiding traveling outside of the hotel zone at the p < 0.0 1 level: country level destination past travel experience 0.0 11 ) and city level destination past travel experience 0.0 41 ). Based on the results of the multiple regression, hypothesis 1 j was rejected. Research Question 2: Does Perceived Risk Mediate the Relationship between the Antecedents and Perceived Efficacy? Research question 2 focused on examining if perceived risk (perceived vulnerability, affective risk perceptions, and perceived severity) mediated the relationship between the antecedents (demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel experience) and perceived efficacy (self efficacy and response efficacy). To test research question 2, a series of multiple regressions were conducted using SPSS. Indirect path s between the antecedents and perceived risk In the first series of multiple regressions, the relationships between the antecedents ( demographic factors, international travel specific psychological factors, destination specific psychological factors, destination specific factors, and past travel

PAGE 209

209 experience ) and perceived risk ( perceived vulnerability / affective risk perceptions and perce ived severity ) were examined. P erceived vulnerability / affective risk perceptions was the dependent variable in the first multiple regression . The results of the first multiple regression can be found in Table 4 3. Perceived severity was the dependent varia ble in the second multiple regression. The results of the second multiple regression can be found in Table 4 4. The first series of multiple regressions provided the indirect paths between the antecedents ( independent variables ) and perceived risk (first s et of mediating variables). Indirect paths between perceived risk and perceived efficacy In the second series of multiple regressions, the relationships between perceived risk ( perceived vulnerability / affective risk perceptions and perceived severity) an d perceived efficacy (self efficacy and response efficacy) were examined . Self efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the fi rst multiple regression . R esponse efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the second multiple regression. The second series of multiple regressions provided the indirect paths between perceived risk (first set of mediating variables) and perceived efficacy (second set of mediating variables). A preliminary multiple regression was run for the dependent variable of self efficacy associated with avoiding traveling outside of the hotel zone in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p.

PAGE 210

210 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits . Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. Ba sed on the previous discussion of the normality of error terms assumption, to err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship b etween perceived risk and self efficacy associated with avoiding traveling outside of the hotel zone . The results of the multiple regression used to assess the ability of perceived risk (perceived vulnerability / affective risk perceptions and perceived sev erity) to predict self efficacy associated with avoiding traveling outside of the hotel zone are presented in Table 4 17 . The total variance explained by the model as a whole was 1.6 %, F ( 2 , 1,122 )= 10.342 , p < 0.0 01 . T able 4 17 . Multiple regression of t he relationship between perceived risk and self efficacy associated with avoiding traveling outside of the hotel zone . Variable Sig. Perceived vulnerability/affective risk perceptions .036 .243 Perceived severity .141 .000*** Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.016 A preliminary multiple regression was run for the dependent variable of response efficacy associated with avoiding traveling outside of the hotel zone in order to test the

PAGE 211

211 assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the varia nce inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2 013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. Based on the previous discussion of the normality of error terms assumption, to err on the si de of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between perceived risk and response efficacy associated with avoiding traveling outside of the hotel zone . The results of the multiple regression used to assess the ability of perceived risk (perceived vulnerability / affective risk perceptions and perceived severity) to predict response efficacy associated with avoiding traveling outside of the hotel zone are presented in Table 4 18 . The total variance explained by the model as a whole was 1.8 %, F (2, 1,122 )= 11.169 , p < 0.0 01 . T able 4 1 8 . Multiple regression of the relationship between perceived risk and response efficacy associated with avoid ing traveling outside of the hotel zone . Variable Sig. Perceived vulnerability/affective risk perceptions .067 .031 Perceived severity .145 .000*** Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.018

PAGE 212

212 Direct effects of antecedents and perceived risk on perceived efficacy Significant indirect paths were found between a few of the antecedents (part of the independent variables) and perceived severity ( part of the first set of mediating variables) . Signi ficant indirect paths were also found between perceived severity (part of the first set of mediating variables ) and the perceived efficacy variables (second set of mediating variables ). Therefore, a third series of multiple regressions was conducted. In th e third series of multiple regressions, the relationships between the significant antecedents and perceived severity and self efficacy associated with avoiding traveling outside of the hotel zone were examined. Self efficacy associated with avoiding travel ing outside of the hotel zone was the dependent variable in the first multiple regression . Response efficacy associated with avoiding traveling outside of the hotel zone was the dependent variable in the second multiple regression . T he third series of multiple regressions provided the direct effect of the significant antecedents (part of the independent variables) and perceived severity ( part of the first set of mediating variables) on p erceived efficacy (second set of mediating variables). A prelimina ry multiple regression was run for the dependent variable of self efficacy associated with avoiding traveling outside of the hotel zone in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In additio n, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance

PAGE 213

213 inflation factor (VIF) statistics were less than 10, indicating that there was not a problem wi th multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating th at the assumption of the normality of error terms was violated. Based on the previous discussion of the normality of error terms assumption, to err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the re sults of the multiple regression which examined the relationship between gender, the second safety concerns item, destination image, and perceived severity and self efficacy associated with avoiding traveling outside of the hotel zone. The results of the multiple regression used to assess the ability of gender, the second safety concerns item, destination image, and perceived severity to predict self efficacy associated with avoiding traveling outside of the hotel zone are presented in Table 4 19 . The total variance explained by the model as a whole was 6.3 %, F ( 4 , 1,1 18 )= 19.945 , p < 0.001. Table 4 19. Multiple regression of the relationship between the significant antecedents and perceived severity and self efficacy associated with avoiding tr aveling outside of the hotel zone . Variable Sig. Gender .063 .031 Safety is a serious consideration when choosing an international tourist destination .216 .000*** Destination image .046 .113 Perceived severity .054 .084 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.063 A preliminary multiple regression was run for the dependent variable of response efficacy associated with avoiding traveling outside of the hotel zone in order to test the assumptions of multiple regression. The sample size assumption was not violated

PAGE 214

214 because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independe nt variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. Based on the previous discussion of the normality of error terms assumption, to e rr on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between gender, the second safety concerns item, destination image, and perceive d severity and response efficacy associated with avoiding traveling outside of the hotel zone. The results of the multiple regression used to assess the ability of gender, the second safety concerns item, destination image, and perceived severity to predi ct response efficacy associated with avoiding traveling outside of the hotel zone are presented in Table 4 20 . The total variance explained by the model as a whole was 7.2 %, F (4, 1,118)= 22.690 , p < 0.001. T able 4 20 . Multiple regression of the relationship between the significant antecedents and perceived severity and response efficacy associated with avoiding traveling outside of the hotel zone . Variable Sig. Gender .024 .403 Safety is a serious consideration when choosing an international tourist destination .215 .000***

PAGE 215

215 T able 4 20 . Continued . Variable Sig. Destination image .132 .000*** Perceived severity .069 .025 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.072 S ince mediation was not found, an online calculation of the Sobel test was not conducted. Hypothesis 2a: Perceived vulnerability/affective risk perceptions mediates the relationship between demographic factors and self efficacy Hypothesis 2a posited that perceived vulnerability/affective risk perceptions mediated the relationship between demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) and self efficacy. The relationships between demographic factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2a for the r isk reduction behavior of registering with the U.S. Department of State prior to travel. The second series of multiple regressions indicated that there was no statistically significant relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and self efficacy associated with avoiding traveling outside of the hotel zone at the p < 0.01 level 0.0 36 ) . Based on the results of the two series of multiple regress ions, hypothesis 2a was rejected. Hypothesis 2b: Perceived vulnerability/affective risk perceptions mediates the relationship between demographic factors and response efficacy Hypothesis 2b posited that perceived vulnerability/affective risk perceptions m ediated the relationship between demographic factors (gender, age, household

PAGE 216

216 income, race/ethnicity, education level, marital status, and presence of children in the household) and response efficacy . The relationships between demographic factors and percei ved vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2a for the risk reduction behavior of registering with the U.S. Department of State prior to travel. T he second series of multiple regressions indicated that there was no statistically significant relationship between the perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and response efficac y associated with avoiding traveling outside of the hotel zone at the p < 0.01 level 0.06 7 ) . Based on the results of the two series of multiple regressions, hypothesis 2b was rejected. Hypothesis 2c: Perceived vulnerability/affec tive risk perceptions mediates the relationship between international travel specific psychological factors and self efficacy Hypothesis 2c posited that perceived vulnerability/affective risk perceptions mediated the relationship between international tra vel specific psychological factors (international travel risk perceptions and safety concerns) and self efficacy . The relationships between international travel specific psychological factors and perceived vulnerability/affective risk perceptions associate d with crime while visiting the destination of Acapulco w ere discussed in the section on hypothesis 2 c for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and self efficacy associated with avoiding traveling outside of the hotel zone was discussed in the

PAGE 217

217 section on hypo thesis 2 a. Based on the results of the two series of multiple regressions, hypothesis 2c was rejected. Hypothesis 2d: Perceived vulnerability/affective risk perceptions mediates the relationship between international travel specific psychological factors and response efficacy Hypothesis 2d posited that perceived vulnerability/affective risk perceptions mediated the relationship between international travel specific psychological factors (international travel risk perceptions and safety concerns) and respo nse efficacy . The relationships between international travel specific psychological factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2 c for the risk reduction behavior of registering with the U.S. Department of State prior to travel . The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and response efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2b . Based on the results of the two series of multiple regressions , hypothesis 2d was rejected. Hypothesis 2e: Perceived vulnerability/affectiv e risk perceptions mediates the relationship between destination specific psychological factors and self efficacy Hypothesis 2e posited that perceived vulnerability/affective risk perceptions mediated the relationship between destination specific psycholo gical factors (destination image) and self efficacy . The relationships between destination specific psychological factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were discussed in the section on

PAGE 218

218 hypothesis 2e for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destinat ion of Acapulco and self efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2 a. Based on the results of the two series of multiple regressions, hypothesis 2e was rejected. Hypothesis 2f: Perceiv ed vulnerability/affective risk perceptions mediates the relationship between destination specific psychological factors and response efficacy Hypothesis 2f posited that perceived vulnerability/affective risk perceptions mediated the relationship between destination specific psychological factors (destination image) and response efficacy. The relationships between destination specific psychological factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the desti nation of Acapulco were discussed in the section on hypothesis 2e for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and response efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2b . Based on the results of the two series of multiple regressions, hypothesis 2 f was rejected.

PAGE 219

219 Hypothesis 2g: Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific factors and self efficacy Hypothesis 2g posited that perceived vulnerability/affective risk perceptions mediated the relationship between destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of the destination) and self efficacy. The relationships between destination specific factors and perceived vulne rability/affective risk perceptions associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2g for the risk reduction behavior of registering with the U.S. Department of State prior to travel . The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and self efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypo thesis 2 a. Based on the results of the two series of multiple regressions, hypothesis 2g was rejected. Hypothesis 2h: Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific factors and response efficacy H ypothesis 2h posited that perceived vulnerability/affective risk perceptions mediated the relationship between destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of the destination) and response efficacy . The relationships between destination specific factors and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2g for the risk re duction behavior of registering with the U.S. Department of State prior to travel.

PAGE 220

220 The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and response efficacy associate d with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2b . Based on the results of the two series of multiple regressions, hypothesis 2h was rejected. Hypothesis 2i: Perceived vulnerability/affective risk perceptions mediates the relationship between past travel experience and self efficacy Hypothesis 2 i posited that perceived vulnerability/affective risk perceptions mediated the relationship between past travel experience (country level destination past travel experience and city level destination past travel experience ) and self efficacy. The relationships between past travel experience and perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were d iscussed in the section on hypothesis 2i for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and self efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2 a. Based on the results of the two series of multiple regressions, hypothesis 2i was rejected. Hypothesis 2j: Perceived vulnerability/affective risk perceptions mediates the relationship between past travel experience and response efficacy Hypothesis 2j posited that perceived vulnerability/affective risk perceptions mediated the relationship between past travel experience (country level destination past travel experience and city level destination past travel experience ) and response

PAGE 221

221 efficacy . The relationships between past travel experience and perceived vulnerability/affective risk perceptions associated with c rime while visiting the destination of Acapulco were discussed in the section on hypothesis 2i for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and response efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2b . Based on the results of the two series of multip le regressions, hypothesis 2j was rejected. Hypothesis 2k: Perceived severity mediates the relationship between demographic factors and self efficacy Hypothesis 2k posited that perceived severity mediated the relationship between demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) and self efficacy . The relationships between demographic factors and perceived severity associated with crime while visiting the des tination of Acapulco were discussed in the section on hypothesis 2k for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The second series of multiple regressions indicated that there was a statistically signif icant relationship between perceived severity associated with crime while visiting the destination of Acapulco and self efficacy associated with avoiding traveling outside of the hotel zone at the p < 0.0 0 114 ). Accordingly, the results indica ted that those with higher levels of perceived severity associated with crime while visiting the

PAGE 222

222 destination of Acapulco had significantly higher levels of self efficacy associated with avoiding traveling outside of the hotel zone. The third series of mult iple regressions indicated that, in the combined model, there was no statistically significant relationship between perceived severity associated with crime while visiting the destination of Acapulco and self efficacy associated with avoiding traveling out Based on the results of the three series of multiple regressions, hypothesis 2k was rejected. Hypothesis 2l: Perceived severity mediates the relationship between demographic factors and response effi cacy Hypothesis 2l posited that perceived severity mediated the relationship between demographic factors (gender, age, household income, race/ethnicity, education level, marital status, and presence of children in the household) and response efficacy . The relationships between demographic factors and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2k for the risk reduction behavior of registering with the U.S. Department of Sta te prior to travel. The second series of multiple regressions indicated that there was a statistically significant relationship between perceived severity associated with crime while visiting the destination of Acapulco and response efficacy associated w ith avoiding traveling outside of the hotel zone at the p < 0.0 0 145 ). Accordingly, the results indicated that those with higher levels of perceived severity associated with crime while visiting the destination of Acapulco had significantly hi gher levels of response efficacy associated with avoiding traveling outside of the hotel zone.

PAGE 223

223 The third series of multiple regressions indicated that, in the combined model, there was no statistically significant relationship between perceived severity as sociated with crime while visiting the destination of Acapulco and response efficacy associated with avoiding traveling outside of the hotel zone ). Based on the results of the three series of multiple regressions , hypothesis 2l was rejected. Hypothesis 2m: Perceived severity mediates the relationship between international travel specific psychological factors and self efficacy Hypothesis 2m posited that perceived severity mediated the relationship between international trave l specific psychological factors (international travel risk perceptions and safety concerns) and self efficacy . The relationships between international travel specific psychological factors and perceived severity associated with crime while visiting the de stination of Acapulco were discussed in the section on hypothesis 2m for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived severity associated with crime while visiting the dest ination of Acapulco and self efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2k . The results of the combined model were discussed in the section on hypothesis 2k. Based on the results of the three series of multiple regressions, hypothesis 2m was rejected. Hypothesis 2n: Perceived severity mediates the relationship between international travel specific psychological factors and response efficacy Hypothesis 2n posited that perceived severity m ediated the relationship between international travel specific psychological factors (international travel risk perceptions and safety concerns) and response efficacy . The relationships between international travel specific psychological factors and percei ved severity associated with crime while

PAGE 224

224 visiting the destination of Acapulco were discussed in the section on hypothesis 2m for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceiv ed severity associated with crime while visiting the destination of Acapulco and response efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2l . The results of the combined model were discussed in the section on hypothesis 2 l . Based on the results of the three series of multiple regressions, hypothesis 2n was rejected. Hypothesis 2o: Perceived severity mediates the relationship between destination specific psychological factors and self efficacy Hypothesis 2o posited that perceived severity mediated the relationship between destination specific psychological factors (destination image) and self efficacy . The relationships between destination specific psychological factors and perceived severity a ssociated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2o for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived severity ass ociated with crime while visiting the destination of Acapulco and self efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2k. The results of the combined model were discussed in the section on hypothesis 2k. Based on the results of the three series of multiple regressions, hypothesis 2o was rejected.

PAGE 225

225 Hypothesis 2p: Perceived severity mediates the relationship between destinatio n specific psychological factors and response efficacy Hypothesis 2 p posited that perceived severity mediated the relationship between destination specific psychological factors (destination image) and response efficacy. The relationships between destinat ion specific psychological factors and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2o for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and response efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2l . The results of the combined model were discussed in the section on hypothesis 2l. Based on the results of the three series of multiple regressions, hypothesis 2p was rejected. Hypothesis 2q: Perceived severity mediates the relationship between destination spec ific factors and self efficacy Hypothesis 2q posited that perceived severity mediated the relationship between destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of the destination) and self efficacy. The relationships between destination specific factors and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2q for the risk reduction behavior of registering wit h the U.S. Department of State prior to travel. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and self efficacy associated with avoiding traveling outside

PAGE 226

226 of the hotel zone was discussed in the section on hypothesis 2k . Based on the results of the two series of multiple regressions, hypothesis 2q was rejected. Hypothesis 2r: Perceived severity mediates the relationship between destination specific factors and response efficacy Hypothesis 2r pos ited that perceived severity mediated the relationship between destination specific factors (presence of friends/relatives living in the destination and fluency in the native language of the destination) and response efficacy . The relationships between destination specific factors and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2q for the risk reduction behavior of registering with the U.S. Dep artment of State prior to travel. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and response efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2l . Based on the results of the two series of multiple regressions, hypothesis 2r was rejected. Hypothesis 2s: Perceived severity mediates the relationship between past travel experience and self efficacy Hypothesis 2 s posited that perceived severity mediated the relationship between past travel experience (country level destination past travel experience and city level destination past travel experience ) and self efficacy. The relationships between past travel experience and perceived severi ty associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2s for the risk reduction behavior of registering with the U.S. Department of State prior to travel.

PAGE 227

227 The relationship between perceived severit y associated with crime while visiting the destination of Acapulco and self efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2k . Based on the results of the two series of multiple regressions, hypothesis 2s was rejected. Hypothesis 2t: Perceived severity mediates the relationship between past travel experience and response efficacy Hypothesis 2t posited that perceived severity mediated the relationship between past travel experience (country l evel destination past travel experience and city level destination past travel experience ) and response efficacy . The relationships between past travel experience and perceived severity associated with crime while visiting the destination of Acapulco were discussed in the section on hypothesis 2s for the risk reduction behavior of registering with the U.S. Department of State prior to travel. The relationship between perceived severity associated with crime while visiting the destination of Acapulco and response efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2l . Based on the results of the two series of multiple regressions, hypothesis 2t was rejected. Research Question 3: Does Perceived Ri sk Predict Intentions to Engage in a Recommended Risk Reduction Behavior? Research question 3 focused on examining if perceived risk (perceived vulnerability / affective risk perceptions and perceived severity) predicted intentions to engage in a recommended risk reduction behavior. To test research question 3, one multiple regression was conducted using SPSS. Intentions to engage in a recommended risk reduction behavior to ensure personal safety while visiting the destination of Acapulco was the dependent variable in the multiple regression . The analysis of research question 3 provided the total effects of perceived risk (first set of

PAGE 228

228 mediating variables) o n intentions to engage in the recommended risk reduction behavior (dependent variable) of avoiding traveling outside of the hotel zone . A preliminary multiple regression was run for the dependent variable of intentions to engage in the recommended risk red uction behavior of avoiding traveling outside of the hotel zone in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Spec ifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was e xamined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. In an e ffort to remedy this assumption violation, the dependent variable was transformed. As suggested by Tabachnick and Fidell (2013), the transformation method of reflect and log, NEWX= LG10(K X), was used because the distribution of the dependent variable was substantially negatively skewed (p. 89). Following the computation of the transformed dependent variable, the multiple regression was rerun. The Kolmogorov Smirnov statistic was significant at the p < 0.0 1 level, indicating that the assumption of the norma lity of error terms was still violated. However, based on the previous discussion of

PAGE 229

229 the normality of error terms assumption, the multiple regression was rerun with the non transformed dependent variable. T o err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between perceived risk and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone . The results of the multiple regression used to assess the ability of perceived risk (perceived vulnerability / affective risk perceptions and perceived severity) to predict intentions to engage in the recommended risk reduction behavior of avoi ding traveling outside of the hotel zone are presented in Table 4 2 1 . The total variance explained by the model as a whole was 1.7 %, F (2, 1,122 )= 10.533 , p < 0.001. T able 4 2 1 . Multiple regression of the relationship between perceived risk and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone . Variable Sig. Perceived vulnerability/affective risk perceptions .091 .003** Perceived severity .077 .014 Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.017 Hypothesis 3a: Perceived vulnerability/affective risk perceptions predicts intentions to engage in a recommended risk reduction behavior Hypothesis 3a posited that perceived vulnerability/affective risk perceptions would predict intentions to engage in a recommended risk reduction behavior . Multiple regression analysis indicated that there was a statistically significant relationship between perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco and intentions to engage in the recommended risk reduction behavior of avoiding traveling ou tside of the hotel zone at the p < 0.0 1 level 0.091 ). Accordingly, the results indicated that those with higher levels of perceived

PAGE 230

230 vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco were significa ntly more likely to avoid traveling outside of the hotel zone than those with lower levels of perceived vulnerability/affective risk perceptions associated with crime while visiting the destination of Acapulco . Based on the results of the multiple regressi on, hypothesis 3a was not rejected. Hypothesis 3b: Perceived severity predicts intentions to engage in a recommended risk reduction behavior Hypothesis 3b posited that perceived severity would predict intentions to engage in a recommended risk reduction b ehavior . Multiple regression analysis indicated that there was no statistically significant relationship between perceived severity associated with crime while visiting the destination of Acapulco and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone at the p < 0.0 1 level 0.077 ). Based on the results of the multiple regression, hypothesis 3b was rejected. Research Question 4: Does Perceived Efficacy Mediate the Relationship between Perceived Risk and Intentions to Engage in a Recommended Risk Reduction Behavior? Research question 4 focused on examining if perceived efficacy (self efficacy and response efficacy) mediated the relationship between perceived risk ( perceived vulnerability / affective risk perceptions and perceived severity) and intentions to engage in a recommended risk reduction behavior. To test research question 4, a series of multiple regressions were conducted using SPSS. Indirect paths between perceived risk and perceived effic acy In the first series of multiple regressions, the relationships between perceived risk ( perceived vulnerability / affective risk perceptions and perceived severity) and perceived

PAGE 231

231 efficacy (self efficacy and response efficacy) were examined. Self efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the first multiple regression . The results of the first multiple regression can be found in T able 4 17. R esponse efficacy associated with a behavior that tourists could engage in to ensure their personal safety while visiting the destination of Acapulco was the dependent variable in the second multiple regression. The results of the second multipl e regression can be found in Table 4 18. The first series of multiple regressions provided the indirect paths between perceived risk (first set of mediating variables) and perceived efficacy (second set of mediating variables). Indirect paths between perc eived efficacy and intentions to engage in a recommended risk reduction behavior In the second series of multiple regressions, the relationships between perceived efficacy (self efficacy and response efficacy) and intentions to engage in a recommended ris k reduction behavior were examined. Intentions to engage in a recommended risk reduction behavior to ensure personal safety while visiting the destination of Acapulco was the dependent variable in the multiple regression . T he second series of multipl e regressions provided the indirect paths between perceived efficacy (second set of mediating variables) and intentions to engage in a recommended risk reduction behavior (dependent variable). A preliminary multiple regression was run for the dependent va riable of intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone in order to test the assumptions of multiple regression. The sample size assumption was not violated because there were at least 20 cases per independent

PAGE 232

232 variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a co mbination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than 0.10 and all of the variance inflation factor (V IF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kolmo gorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. Based on the previous discussion of the normality of error terms assumption, to err on the side of caution, a more c onservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between perceived efficacy and intentions to engage in the recommended risk reduction behavior of avoiding travelin g outside of the hotel zone . The results of the multiple regression used to assess the ability of perceived efficacy ( self efficacy and response efficacy ) to predict intentions to engage in the recommended risk reduction behavior of avoiding traveling out side of the hotel zone are presented in Table 4 2 2 . The total variance explained by the model as a whole was 41.3 %, F (2, 1,123 )= 396.502 , p < 0.0 01 . T able 4 2 2 . Multiple regression of the relationship between perceived efficacy and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone . Variable Sig. Self efficacy .474 .000*** Response efficacy .270 .000*** Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.4 13

PAGE 233

233 Direct effects of perceived risk and perceived efficacy on intentions to engage in a recommended risk reduction behavior S ignificant indirect paths were found between perceived severity (part of the first set of mediating variables) and the perceived efficacy variables ( second set of mediating variables) . Significant indirect paths were also found between the perceived efficacy variables (second set of mediating variables ) and intentions to engage in a recommended risk reduction behavior (dependent variable). Therefore , a third series of multiple regressions was conducted. In the third series of multiple regressions, the relationships between perceived severity , self efficacy , and response efficacy and intentions to engage in a recommended risk reduction behavior were examined. Intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone was the dependent variable in the multiple regression . T he third series of multiple regressions provided the direct effect of perceived severity ( part of the first set of mediating variables) and perceived efficacy (second set of mediating variables) on intentions to engage in a recommended risk reduction be havior (dependent variable). A preliminary multiple regression was run for the dependent variable of intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone in order to test the assumptions of multip le regression. The sample size assumption was not violated because there were at least 20 cases per independent variable (Hair et al., 2010, p. 175). The assumption of singularity was also not violated because the independent variables included were not a combination of other independent variables (Pallant, 2007, p. 149). In addition, the collinearity statistics were all within the accepted limits. Specifically, all of the tolerance statistics were greater than

PAGE 234

234 0.10 and all of the variance inflation factor (VIF) statistics were less than 10, indicating that there was not a problem with multicollinearity (Pallant, 2007, p. 156). Mahalanobis distance was examined and no multivariate outliers were identified (based on Tabachnick & Fidell, 2013, p. 128). The Kol mogorov Smirnov statistic was significant at the p < 0.01 level, indicating that the assumption of the normality of error terms was violated. Based on the previous discussion of the normality of error terms assumption, to err on the side of caution, a more conservative significance level of p < 0.01 was used when interpreting the results of the multiple regression which examined the relationship between perceived severity, self efficacy, and response efficacy and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone . The results of the multiple regression used to assess the ability of perceived severity and perceived efficacy ( self efficacy and response efficacy ) to predict intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone are presented in Table 4 2 3 . The total variance explained by the model as a whole was 41.2 %, F ( 3 , 1,12 2 )= 264.194 , p < 0.0 01 . T able 4 2 3 . Mul tiple regression of the relationship between perceived severity and perceived efficacy and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone . Variable Sig. Percei ved severity .009 . 683 Self efficacy .473 .000*** Response efficacy .269 .000*** Note: ** indicates p < 0.01; *** indicates p < 0.001; Adjusted R 2 = 0.4 12

PAGE 235

235 Sobel test Given that mediation was found, i n addition to the regression based path analysis , an online calculation of the Sobel test was conducted (Preacher, 2015) . The results of the Sobel test are presented in Table 4 2 4 . T able 4 2 4 . Sobel test of the mediation of perceived efficacy in the relationship between perceived severity and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone . Hypothesis Test statistic Std. error Sig. 4b (mediator= self efficacy; IV= perceived severity) 4.408 0.022 .000*** 4d (mediator= response efficacy; IV= perceived severity) 4.285 .014 .000*** Note: ** indicates p < 0.01; *** indicates p < 0.001 Hypothesis 4a: Self efficacy mediates the relationship between perceived vulnerability/affective risk perceptions and intentions to engage in a recommended risk reduction behavior Hypothesis 4a posited that self efficacy would mediate the relationship between perceived vulnerability/affective risk perceptions and intentions to engage in a recommended risk reduction behavior. The relationship between perceived vulnerability/affective risk perceptions and self efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2a . The second series of multiple regressions indicated that there was a statistically significant relationship between self efficacy associated with avoiding traveling outside of the hotel zone and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone at the p < 0.001 level 0.474 ) . Accordingly, the results indicated that those with higher levels of self efficacy associated with avoiding traveling outside of the hotel zone were significantly more likely to avoid traveling outside of the hotel zone than those with lower levels of self efficacy. Based on the results of the two series of multiple regressions, hypothesis 4a was rejected.

PAGE 236

236 Hypothesis 4b: Self efficacy mediates the relationship between perceived s everity and intentions to engage in a recommended risk reduction behavior Hypothesis 4b posited that self efficacy would mediate the relationship between perceived severity and intentions to engage in a recommended risk reduction behavior. The first serie s of multiple regressions provided the indirect paths between perceived risk (first set of mediating variables) and perceived efficacy (second set of mediating variables). The relationship between perceived severity associated with crime while visiting the destination of Acapulco and self efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2k . The relationship between self efficacy associated with avoiding traveling outside of the hotel zone and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 4a. T he third series of multiple regressions provided the direct effect of perceived severity (first set of mediating variables) and perceived efficacy (second set of mediating variables) on intentions to engage in a recommended risk reduction behavior (dependent variable). In the combined model, m ultiple regression analysis indicated that there w as no statistically significant relationship between perceived severity associated with crime while visiting the destination of Acapulco and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone at the p < 0.01 level 0.009 ) . In the combined model, m ultiple regression analysis indicated that there was a statistically significant relationship between self efficacy associated with avoiding traveling outside of the hotel zone and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone

PAGE 237

237 at the p < 0.001 level 473 ). Furthermore, the Sobel test for hypothesis 4b was significant at the p < 0.001 level. Based on the results of the three series of multiple regressions and the calculation of the Sobel test, hypothesis 4b was not rejected. Hypothesis 4c: Response efficacy mediates the relationship between perceived vu lnerability/ affective risk perceptions and intentions to engage in a recommended ris k reduction behavior Hypothesis 4 c posited that response efficacy would mediate the relationship between perceived vulnerability/affective risk perceptions and intentions to engage in a recommended risk reduction behavior. The r elationship between perceiv ed vulnerability/ affective risk perceptions associated with crime while visiting the destination of Acapulco and response efficacy associated with avoiding traveling outside of the hotel zone was discussed in the section on hypothesis 2b . The second series of multiple regressions indicated that there was a statistically significant relationship between response efficacy associated with avoiding traveling outside of the hotel zone and intentions to engage in the recommended risk reduction be havior of avoiding traveling outside of the hotel zone at the p < 0.001 level ( 270 ) . Accordingly, the results indicated that those with higher levels of response efficacy associated with avoiding traveling outside of the hotel zone were significantly more likely to avoid traveling outside of the hotel zone than those with lower levels of response efficac y. Based on the results of the two series of multiple regressions, hypothesis 4c was rejected. Hypothesis 4d: Response efficacy mediates the relationsh ip between perceived severity and intentions to engage in a recommended risk reduction behavior Hypothesis 4d posited that response efficacy would mediate the relationship between perceived severity and intentions to engage in a recommended risk reduction

PAGE 238

238 behavior. The first series of multiple regressions provided the indirect paths between perceived risk (first set of mediating variables) and perceived efficacy (second set of mediating variables). The relationship between perceived severity associated wit h crime while visiting the destination of Acapulco and response efficacy associated with avoiding traveling outside of the hotel zone were discussed in the section on hypothesis 2l . The relationship between response efficacy associated with registering with the U.S. Department of State prior to travel and intentions to engage in the recommended risk reduction behavior of registering with the U.S. Department of State prior to travel was discussed in the section on hypothesis 4 c. T he third series of multiple regressions provided the direct effect of perceived severity (first set of mediating variables) and perceived efficacy (second set of mediating variables) on intentions to engage in a recommended risk reduction behavior (de pendent variable). As mentioned in the discussion of the findings of hypothesis 4b, the third series of multiple regressions indicated that was no statistically significant relationship between perceived severity associated with crime while visiting the de stination of Acapulco and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone at the p < 0.01 level. In the combined model, multiple regression analysis also indicated that there was a statistical ly significant relationship between response efficacy associated with avoiding traveling outside of the hotel zone and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the hotel zone at the p < 0.001 level = 0.269). Furthermore, the Sobel test for hypothesis 4 d was significant at the p <

PAGE 239

239 0.001 level. Based on the results of the three series of multiple regressions and the calculation of the Sobel test, hypothesis 4 d was not rejected. Summary of Findings by Hypothesis A summary of the findings by hypothesis is provided in Table 4 25. T able 4 25. Summary of findings by hypothesis. Hypothesis Register Insurance Hotel 1a Demographic factors predict self efficacy. 1b Demographic factors predict response efficacy. 1c International travel specific psychological factors predict self efficacy. X 1d International travel specific psychological factors predict response efficacy. 1e Destination specific psychological factors predict self efficacy. 1f Destination specific psychological factors predict response efficacy. X X X 1g Destination specific factors predict self efficacy. 1h Destination specific factors predict response efficacy. 1i Past travel experience predicts self efficacy. 1j Past travel experience predicts response efficacy. 2a Perceived vulnerability/affective risk perceptions mediates the relationship between demographic factors and self efficacy. 2b Perceived vulnerability/affective risk perceptions mediates the relationship between demographic factors and response efficacy. 2c Perceived vulnerability/affective risk perceptions mediates the relationship between international travel specific psychological factors and self efficacy. 2d Perceived vulnerability/affective risk perceptions mediates the relationship between international travel specific psychological factors and response efficacy. 2e Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific psychological factors and self efficacy.

PAGE 240

240 Table 4 25. Continued . Hypothesis Register Insurance Hotel 2f Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific psychological factors and response efficacy. 2g Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific factors and self efficacy. 2h Perceived vulnerability/affective risk perceptions mediates the relationship between destination specific factors and response efficacy. 2i Perceived vulnerability/affective risk perceptions mediates the relationship between past travel experience and self efficacy. 2j Perceived vulnerability/affective risk perceptions mediates the relationship between past travel experience and response efficacy. 2k Perceived severity mediates the relationship between demographic factors and self efficacy. 2l Perceived severity mediates the relationship between demog raphic factors and response efficacy. 2m Perceived severity mediates the relationship between international travel specific psychological factors and self efficacy. 2n Perceived severity mediates the relationship between international travel specific psychological factors and response efficacy. 2o Perceived severity mediates the relationship between destination specific psychological factors and self efficacy. 2p Perceived severity mediates the relationship between destination specific psychological factors and response efficacy. 2q Perceived severity mediates the relationship between destination specific factors and self efficacy. 2r Perceived severity mediates the relationship between destination specific factors and response efficacy. 2s Perceived severity mediates the relationship between past travel experience and self efficacy.

PAGE 241

241 Table 4 25. Continued . Hypothesis Register Insurance Hotel 2t Perceived severity mediates the relationship between past travel experience and response efficacy. 3a Perceived vulnerability/affective risk perceptions predicts intentions to engage in a recommended risk reduction behavior. X X 3b Perceived severity predicts intentions to engage in a recommended risk reduction behavior. 4a Self efficacy mediates the relationship between perceived vulnerability/affective risk perceptions and intentions to engage in a recommended risk reduction behavior. 4b Self efficacy mediates the relationship between perceived severity and intentions to engage in a recommended risk reduction behavior. X 4c Response efficacy mediates the relationship between perceived vulnerability/ affective risk perceptions and intentions to engage in a recommended risk reduction behavior. 4d Response efficacy mediates the relationship between perceived severity and intentions to engage in a recommended risk reduction behavior. X Note: X indicates failed to reject hypothesis; indicates partial support for hypothesis; Register indicates the risk reduction behavior of registering with the U.S. Department of State prior to travel; Insurance indicates the risk reduction behavior of purchasing travel insurance prior to travel; Hotel indicates the risk reduction behavior of avoiding traveling outside of the hotel zone

PAGE 242

242 CHAPTER 5 DISCUSSION AND CONCLUSIONS This chapter presents the discussion of findings of this study, as well as theoretical and practical implications. Specifically, this chapter is structured into three main sections: (1) discussion of findings ; ( 2 ) conclusions and implications ; ( 3 ) limitations and suggestions for future research . Discussion of Findings There were a few distinct patterns in the findings of this study . First , perceived efficacy (self efficacy and response efficacy) was only found to mediate the relationship between perceived severity and intentions to engage in the recommended risk reduction behavior of avoiding traveling outside of the h otel zone. Significant mediation was not found for the recommended risk reduction behaviors of registering with the U.S. Department of State prior to travel and purchasing travel insurance prior to travel. The difference in findings in this study may be ba sed on the stage of the tourism experience in which tourists would engage in the recommended risk reduction behavior. Specifically, the two non significant findings involved risk reduction behaviors that tourists would engage in during the anticipation and planning stage of the tourism experience. In contrast, the significant finding involved a risk reduction behavior that tourists would engage in during the on site (at the destination) experience. Second, self efficacy and response efficacy associated wit h a recommended risk reduction behavior were positively associated with intentions to engage in a recommended risk reduction behavior destination of Acapulco . Notably, self efficacy was a stronger predict or of intentions to engage in a recommended risk reduction behavior. Accordingly, when respondents

PAGE 243

243 were more confident in their ability to perform a recommended risk reduction behavior, they were more likely to engage in a recommended risk reduction behavi or to ensure their personal safety while visiting the destination. In addition, when respondents believed that a recommended risk reduction behavior was more effective, they were more likely to engage in the recommended risk reduction behaviors to ensure t heir personal safety while visiting the destination. These findings were consistent with the hypothesized relationships of PMT (Floyd et al., 2000), as well as past PMT studies (Floyd et al., 2000; Norman , Boer, & Seydel , 2005). Specifically, PMT hypothes izes that one needs to believe that they can perform a recommended risk reduction behavior if they are to engage in the behavior (Floyd et al., 2000). PMT also hypothesizes that one needs to believe that a risk reduction behavior will be effective in reduc ing the risk if they are to engage in the behavior (Floyd et al., 2000). Furthermore, self efficacy and response efficacy were found to be the strongest predictors of intentions to engage in a recommended risk reduction behavior. This supported the existin g body of knowledge, as the variables associated with the coping appraisal process have been found to be the strongest predictors of intentions to engage in a risk reduction behavior (Floyd et al., 2000; Norman et al., 2005). Third , perceived vulnerability/affective risk perceptions was a significant predictor of intentions to engage in two of the three recommended risk reduction behaviors. The positive associations between perceived vulnerability/affective risk perceptions associated with crim e and intentions to engage in the recommended risk reduction behaviors of registering with the U.S. Department of State prior to travel and avoiding

PAGE 244

244 traveling outside of the hotel zone were consistent with PMT (Floyd et al., 2000; Norman et al., 2005 ). The findings also supported the existing body of knowledge associated with travel risk. Previous studies have found that tourists with high levels of cognitive risk perceptions ( Cahyanto et al., 2014; Matyas et al., 2011; Villegas et al., 2013 ) and high level s of affective risk perceptions ( Villegas et al., 2013 ) were more likely to engage in the risk reduction behavior of evacuating in the event of a hurricane at the destination. Thus, the findings suggest ed that perceived vulnerability/affective risk percept ions associated with crime serve as a source of motivation to engage in a recommended risk reduction behavior destination . Conclusions and Implications Theoretical Implications By adopting a theoretical lens, the proposed conceptual model made several original contributions to the academic body of knowledge related to travel risk. First, this study contributed to the academic literature by recognizing that the conceptualization o f travel risk has been inadequate. This study also acknowledged and stated that travel risk can be perceived at any stage and in relation to any stage of the travel experience. Second , this study advance d knowledge and understanding in the travel risk lite rature by filling gaps related to the conceptualization and operationalization of risk related constructs. In particular, the novel risk related construct of perceived efficacy was examined. Notably, this novel risk related construct was found to be the st rongest predictor of intentions to engage in a recommended risk reduction behavior. In addition, travel risk perceptions were reconceptualized and reoperationalized in terms of cognitive and affective risk perceptions . Notably, perceived severity and affec tive risk

PAGE 245

245 perceptions have been understudied in tourism. Third, this study advance d knowledge by examining the bivariate relationships between risk related constructs in the context of travel in an effort to provide conceptual clarity , as well as controlli ng for mediating variables in a hierarchical model . Fourth, by extending PMT to the setting of international travel and crime related risk , the robustness of the theory was examined beyond health behavior . Particularly, perceived vulnerability, perceived severity, self efficacy, and response efficacy were found to be positively associated with intentions to engage in r ecommended risk reduction behavior s to ensure their personal safety while visiting Acapulco . Accordingly, the hypothesized relationships of PMT were supported within the setting of international travel. In addition, the coping appraisal process was found to be the strongest predictor of intentions to engage in a recommended risk reduction behavior. While this supported previous PMT studies, it also highlighted that the tourism literature has fallen behind risk related research in other fields. Specifically, perceived efficacy has been a key component of PMT for decades. For example, the coping appraisal process has been found to be the stronges t predictor in the prevention of smoking (Floyd et al., 2000). However, this study was the first to examine self efficacy and response efficacy in association with a recommended risk reduction beha vior in the tourism literature. Fifth , the threat appraisal process of PMT was extended by examining the role of affective risk perceptions. Thus , this study sought to boost the predictive ability of the risk perceptions construct in relation to intentions to enga ge in risk reduction behaviors while visiting the destination by recognizing the risk as feelings hypothesis . Further, perceived vulnerability /affective risk perceptions emerged as a component of perceived

PAGE 246

246 risk in this study . This supported the risk as fee lings hypothesis, which suggests that affective risk perceptions and cognitive risk perceptions may be related (Loewenstein et al., 2001) . In sum mary , this study integrated the existing body of knowledge related to risk in the fields of health behavior an d psychology with the travel risk literature. The findings of this study supported PMT and the risk as feelings hypothesis. As a result , this study supported the applicability of these theories within the context of international travel . Finally, by adopti ng a theoretical lens to the study of travel risk, this study provide d novel insights into the conceptualization, operationalization, and application of risk related constructs in the context of international travel. Practical Implications The findings fr om this study have the potential to make a significant impact on risk management . Specifically, the major finding of this study was that self efficacy and response efficacy were the strongest predictors of intentions to engage in a r ecommended risk reduction behavior t o ensure personal safety while visiting the destination of Acapulco . In other words, perceived efficacy associated with a recommended risk reduction behavior was found to drive behavior al intentions which, in turn, are considered to be the strongest predictors of engagement in a behavior . This is positive news for DMOs because this suggests that they can play a role in managing risk through education campaigns targeted at confidence in their ability to engage in a risk reduction behavior, as well as their beliefs in the effectiveness of a recommended risk reduction behavior. Accordingly, the role of the DMO should be to empower tourists and make them a ware of the fact that they can do something to protect themselves from risk at the destination .

PAGE 247

247 Given that two of the recommended risk reduction behaviors were associated with the anticipation and planning stage of the tourism experience and one of the re commended risk reduction behaviors was associated with the on site (at the destination) stage of the tourism experience, a tiered approach to perceived efficacy education is recommended. For example, during the anticipation and planning stage, DMOs could e ducate tourists through their website, guide books , tour operators , travel agents, etc . In an effort to educate tourists booking through third party websites, the DMO could also invest in pop up ads that would provide tourists with educational information and a link back to the DMO website for additional information. Further, during the on site (at the destination) experience, DMOs could educate tourists through signage in the airport and throughout the destination, as well as brochures in hotel room s . On a global level, few destinations have taken th is recommended educational approach. A rare exception is Japan. In Japanese hotel rooms and on the Okinawa Convention & Visitors (2015) website , brochures are provided which present information about p otential risks at the destination, as well as ways to mitigate the risks. For example, tourists are provided with information about different types of risk ranging from jellyfish stings to tsunamis and earthquakes. Unfortunately , a majority of DMOs have ta ken the approach of not acknowledging risks at the destination . Overall, destinations have indicated that they do not want to provide tourists with information about mitigating risks because they believe this will cause tourists to have heightened risk per ceptions and, consequently, they will choose not to travel to the destination. However, this study found that perceived vulnerability/affective risk perceptions and

PAGE 248

248 perceived severity associated with crime related risk were positively associated with inten tions to engage in risk reduction behaviors visiting the destination of Acapulco . This means that prospective tourists were found to be more likely to engage in behaviors to protect themselves when they ha d higher levels of risk perceptions. Therefore, this study suggests that perceived risk associated with crime can be seen as a source of motivation to protect oneself. Taken further, the finding s from this study suggest that prospective tourists may be able to overcome their risk perceptions by way of engaging in a recommended risk reduction behavior. For practice, the implication is that destinations that are perceived to be risky may not necessarily be avoided. Thus, rather than ignoring that risks exist, the DMO should educate tourists about what they can do to mitigate the risks at the destination. G th is study specifically he global economy by way of safeguarding the tourism industry in these turbulent times. The destination may make more informed decisions by having a greater understanding of the perceptions, attitudes, and behaviors of prospective tourists from the top ori gin market. Thus, this study will provide Mexico, in general, and its major tourist destination of Acapulco, specifically, with an understanding of the programs and services that tourists from the key tourist origin market of the U.S. would engage in which may move them towards being likely to travel to the destination despite perceived risk. Accordingly, the f indings can be used to guide the development of targeted, comprehensible messaging to efficacy and response efficacy in an e ffort to ensure their personal safety while visiting the destination . Messages can

PAGE 249

249 also be developed in an effort to increase intentions to travel to Acapulco despite perceived risks by way of highlightin g the risk reduction behaviors that tourists could e ffectively engage in to ensure their personal safety . Thus , the findings can be used to develop communication strategies for Acapulco , with consideration of actual risks, in an effort to draw more U.S. tourists to the destination . Such initiatives can be u tilized to moderate environmental, social, and economic impacts of the sustained tourism crisis to the destination as a surrogate to the host community . Limitations and Suggestions for Future Research This study served as a first step in reconceptualizing and analyzing the relationship between risk related constructs in tourism . Accordingly, there are several areas that future studies could explore in an effort to move this research agenda forward. First, future studies should test and refine the full proposed conceptual model. Specifically, future studies should consider the role of the risk related constructs in the destination choice process. Based on the proposed conceptual model, s uch research s hould examine the role of risk related constructs in different stages of the destination choice process , as well as the stage at which the risk related constructs might be the most influential. This research would address gaps in the travel risk literature and challenge Sönmez and Graefe ( 1998a ) model of international tourism decision making process , as travel risk studies have traditionally measured the destination choice process in terms of intentions to travel. Second, PMT studies have traditionally e xamined only the main effects of perceived vulnerability, perceived severity, self efficacy, and response efficacy ( Norman et al., 2005 ). However, interaction effects have been found between the perceived risk and perceived efficacy variables (Maddux & Rog ers, 1983; Rogers & Mewborn, 1976).

PAGE 250

250 Accordingly, given that examining the interaction effects w as beyond the scope of this study, future studies should exami ne potential combined effects of perceived vulnerability, perceived severity, self efficacy, and re sponse efficacy on engagement in a risk reduction behavior . Such research could examine if interaction terms are significant predictors of engagement in a risk reduction behavior, as well as how much the introduction of interaction terms into the model con tributes to the variance in engagement in a risk reduction behavior . Third, this study focused specifically on crime related risk because this wa s the most prevalent type of risk for the destination of Acapulco, Mexico. To understand if the findings of th is study can be replicated in and generalized to other settings, f uture studies should test the proposed conceptual model among different types of risk and in the context of different destinations. Particularly, in accordance with risk management, the type of risk examined should be based on the needs of the destination. In addition, the risk reduction behaviors should be tailored to the type of risk. Furthermore, different types of risk should also be examined from a temporal perspective. While thi s study focused on a sustained tourism crisis, future studies should also test the proposed conceptual model in the contexts of immediate and emerging crises. Fourth , experience. However, p ast experience with a product is one of three dimensions of consumer knowledge (Brucks, 1985). Accordingly, future research should also consider objective , or actual , knowledge about the risk. I n addition to objective knowledge of the risk in general, obje ctive knowledge may also be explored within the domain of actual knowledge of the risk at the destination. For example, respondents could have been

PAGE 251

251 asked a series of questions to determine if they were knowledgeable about crime related risk in Acapulco. Su ch research should also explore the placement of objective knowledge within the proposed conceptual model. For example, objective knowledge of the risk may be directly associated with perceived risk and perceived efficacy or objective knowledge of the risk may mediate the relationship between perceived risk and perceived efficacy. The examination of the role of objective knowledge of the risk c ould contribute to the refinement of the proposed conceptual model.

PAGE 252

252 APPENDIX A INFORMED CONSENT

PAGE 253

253

PAGE 254

254

PAGE 255

255 APPENDIX B SURVEY INSTRUMENT FOR THOSE WITH COUNTRY LEVEL DESTINATION PAST TRAVEL EXPERIENCE SECTION #1. INFORMED CONSENT Protocol Title The role of risk related constructs in the destination choice process: Reconceptualizations and analyses Please read this consent document carefully before you decide to participate in this study. Purpose of the research study of Acapulco, Mexico as a tourist destination. What you w ill be asked to do in this study You will be asked a series of questions that are related to your past travel history, perceptions of Acapulco, and travel intentions. Time required It should take you approximately 15 20 minutes to complete this survey. R isks and benefits There is a minimal risk that security of any online data may be breached, but the survey host (Qualtrics) uses strong encryption and other data security methods to protect your information. Despite this possibility, the risks to your phys ical, emotional, social, professional, or financial well being are considered to be minimal. There are also no foreseen benefits to you for taking this survey. Compensation You will be paid $1.5 0 compensation for participating in this study. Confidential ity Your identity will be kept confidential to the extent provided by the law. Your MTurk Worker ID will be used only for the purpose of awarding compensation and will not be shared with anyone outside of the research team. It will not be linked with your survey responses and will be removed from the data set once compensation has been provided. Voluntary participation Your participation in this study is completely voluntary. There is no penalty for not participating. Right to withdraw from the study You have the right to withdraw from this study at any time without consequence.

PAGE 256

256 Who to contact if you have questions about the study If you have any questions, you are encouraged to ask them. Questions can be addressed to Ashley Schroeder, Doctoral Candidate, at (352) 294 1658 or Dr. Lori Pennington Gray , Professor, at (352) 294 1657. Both can be reached by mail at University of Florida, Tourism Crisis Management Initiative, P.O. Box 118208, Gainesville, FL 32611 8208. Who to contact about your rights as a r esearch participant in this study Questions regarding your rights as a participant in this study should be directed to the University of Florida, UF IRB 02 Office, P.O. Box 112250, Gainesville, FL 32611 2250 or (352) 392 0433 . Agreement 1. I have read the pr option below signifies that I voluntarily agree to participate in this study. o AGREE SECTION #2. SCREENING QUESTIONS INTERNATIONAL PAST TRAVEL EXPERIENCE 2. Have you taken a trip outside of the U.S. within the past 5 years? o Yes o No that you do not qualify to participate in this study. Thank you for understanding. [SKIP TO END OF SURVEY] 3. Approximately how many trips outside of the U.S. have you taken in the past 5 years? o 1 o 2 o 3 o 4 o 5 o 6 o 7 o 8 o 9 o 10 o 11 o 12 o 13 o 14 o 15 or more 4. Have you ever taken a trip to Mexico? o Yes o No that you do not qualify to participate in this study. Thank you for understanding. [SKIP TO END OF SURVEY]

PAGE 257

257 SECTION #3. DESTINATION PAST TRAVEL EXPERIENCE 5. Approx imately how many trips to Mexico have you taken in your lifetime? o 1 o 2 o 3 o 4 o 5 o 6 o 7 o 8 o 9 o 10 o 11 o 12 o 13 o 14 o 15 or more 6. Which of the following Mexican destinations have you visited in your lifetime? Select all that apply. o Acapulco o Cancun o Cozumel o Guadalajara o Huatulco o Ixtapa Zihuatanejo o Los Cabos o Manzanillo o Mazatlan o Merida o Monterrey o Mexico City o Morelia o Oaxaca o Playa del Carmen o Puebla o Puerto Vallarta o Riviera Nayarit o Riviera Maya o San Miguel de Allende o Tampico o Taxco o Tijuana o Veracruz o Zacatecas o Other ______ Thinking about your MOST RECENT TRIP TO MEXICO, answer the following set of questions. 7. What year did you last visit Mexico? Write the year in four digit form (for example: 2014). ______ 8. What was the purpose of your most recent trip to Mexico? o Leisure o Business o Visiting friends and /or relatives 9. Was your most recent trip to Mexico part of a cruise itinerary? o Yes o No 10. On your most recent trip to Mexico, did you travel as part of an organized tour? o Yes, single destination tour o Yes, multi destination tour o No

PAGE 258

258 SECTION #4. MEXICO SPECIFIC PSYCHOLOGICAL FACTORS When answering this set of questions, please think about MEXICO AS A TOURIST DESTINATION. 11. Please indicate your overall image of Mexico as a tourist destination, on a scale of 1 to 5 (where 1= very negative and 5= very positive). Destination image 1 2 3 4 5 12. Are you interested in visiting Mexico for leisure purposes within the next year? o Yes o No o No t sure SECTION #5. ACAPULCO SPECIFIC PSYCHOLOGICAL FACTORS Now please think about ACAPULCO, MEXICO AS A TOURIST DESTINATION when you answer the following set of questions. 13. Please indicate your overall ima ge of Acapulco as a tourist destination, on a scale of 1 to 5 (where 1= very negative and 5= very positive). Destination image 1 2 3 4 5 14. Are you interested in visiting Acapulco for leisure purposes within the next year? o Yes o No o Not sure SECTION #6. INTERNATIONAL TRAVEL SPECIFIC PSYCHOLOGICAL FACTORS When you answer the following questions, please think about INTERNATIONAL TRAVEL. In other words, think about traveling outside of the United States. 15. Please indicate your level o f agreement with the following statements, on a scale of 1 5 (where 1= strongly disagree and 5= strongly agree). International travel risk perceptions 1 2 3 4 5 I feel nervous about traveling internationally right now Traveling internationally is risky right now I would feel very uncomfortable traveling internationally right now [Randomize the order of all choices]

PAGE 259

259 16. Please indicate your level of agreement with the following statement s , on a scale of 1 5 (where 1= strongly disagree and 5= strongly agree). Safety concerns 1 2 3 4 5 Safety is the most important attribute an international destination can offer Safety is a serious consideration when choosing an international tourist destination [Randomize the order of all choices] SECTION #7. THREAT APPRAISAL ( PERCEIVED RISK ) Please think about YOUR PERSONAL SAFETY WHILE VISITING ACAPULCO, MEXICO when you answer the following set of questions. 17. U s ing the rating scales below , please indicate how you feel when you think ab out your personal safety while visiting A capulco for leisure purposes. 1 2 3 4 5 Relaxed Anxious Excited Bored Fearless Fearful Assured Worried 18. Please indicate your level of agreement with the following statements, on a scale of 1 5 (where 1= strongly disagree and 5= strongly agree). Cognitive risk perceptions 1 2 3 4 5 If I were a victim of crime while visiting Acapulco, I would experience serious negative consequences It would have a serious negative impact on me if I were a victim of crime while visiting Acapulco If I were a victim of crime while visiting Acapulco, it would be harmful to my well being It is likely that I will be a victim of crime while visiting Acapulco I am at risk for being a victim of crime while visiting Acapulco My chances of being a victim of crime while visiting Acapulco are high [Randomize the order of all choices]

PAGE 260

260 SECTION # 8 . OBJECTIVE KNOWLEDGE (RISK REDUCTION BEHAVIOR KNOWLEDGE) 19. Have you heard of the following behaviors that tourists may engage in to ensure their personal safety while traveling ? Risk reduction behavior knowledge Have Heard Before Never Heard Before Remain in tourist areas Avoid travel ing outside of the hotel zone Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras) Maintain awareness of your surroundings Avoid situations in which you may be isolated or stand out as a potential victim Avoid walking around the destination by yourself Avoid using ATMs if they are not affiliated with a recognized international chain Only use recommended taxi companies Avoid carrying large amounts of cash Avoid drawing attention to yourself Avoid walking around the destination by yourself at night Search for more information online about how to stay safe prior to travel Register with the U.S. Department of State prior to travel Avoid using services like Uber Purchase travel insurance prior to travel [Randomize the order of all choices]

PAGE 261

261 SECTION # 9 . COPING APPRAISAL (PERCEIVED EFFICACY) This set of questions is related to BEHAVIORS THAT TOURISTS MAY ENGAGE IN TO ENSURE THEIR PERSONAL SAFETY WHILE VISITING ACAPULCO, MEXICO. 20. How confident are you in your ability to perform the following behaviors to ensure your personal safety while visiting Acapulco , on a scale of 1 5 (where 1= very unconfident and 5= very confident)? Self efficacy 1 2 3 4 5 Remain in tourist areas Avoid travel ing outside of the hotel zone Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras) Maintain awareness of your surroundings Avoid situations in which you may be isolated or stand out as a potential victim Avoid walking around the destination by yourself Avoid using ATMs if they are not affiliated with a recognized international chain Only use recommended taxi companies Avoid carrying large amounts of cash Avoid drawing attention to yourself Avoid walking around the destination by yourself at night Search for more information online about how to stay safe prior to travel Register with the U.S. Department of State prior to travel Avoid using services like Uber Purchase travel insurance prior to travel [Randomize the order of all choices]

PAGE 262

262 21. Please indicate how effective you believe the following behaviors would be in ensuring your personal safety while visiting A capulco, on a scale of 1 5 (where 1= very ineffective and 5= very effective). Response efficacy 1 2 3 4 5 Remain in tourist areas Avoid travel ing outside of the hotel zone Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras) Maintain awareness of your surroundings Avoid situations in which you may be isolated or stand out as a potential victim Avoid walking around the destination by yourself Avoid using ATMs if they are not affiliated with a recognized international chain Only use recommended taxi companies Avoid carrying large amounts of cash Avoid drawing attention to yourself Avoid walking around the destination by yourself at night Search for more information online about how to stay safe prior to travel Register with the U.S. Department of State prior to travel Avoid using services like Uber Purchase travel insurance prior to travel [Randomize the order of all choices]

PAGE 263

263 SECTION #10. ATTENTION QUESTION Recent research on decision making shows that choices are affected by context. Differences in how people feel, their previous knowledge and experiences, and their environment can affect their choices. To help us understand how people make decisions, we are interested in information about you. Specifically, we are interested in whether you have actually taken the time to read directions in this survey. If not, some results may not tell us very much about decision m aking in the real world. To show that you have read these directions, please ignore the question below about how you are currently feeling and , instead , check only the none of th e above option as your answer. 22. Which of the following words best describes ho w you are currently feeling ? S elect all that apply . o Interested o Distressed o Excited o Upset o Strong o Guilty o Scared o Hostile o Enthusiastic o Proud o Irritable o Alert o Ashamed o Inspired o Nervous o Determined o Attentive o Jittery o Active o Afraid o None of the above indicates that you are not paying attention to this survey. We are sorry but we cannot accept your responses. Thank you for understanding. [SKIP TO END OF SURVEY ]

PAGE 264

264 SECTION # 11. ENGAGEMENT IN RISK REDUCTION BEHAVIORS (INTENTIONS) When answering the following question, think about YOUR PERSONAL SAFETY WHILE VISITING ACAPULCO, MEXICO. 23. Please indicate the likelihood that you would engage in the following behaviors to ensure your personal safety while visiting Acapulco, on a scale of 1 5 (where 1= very unlikely and 5= very likely). Risk reduction behavior s (intention) 1 2 3 4 5 Remain in tourist areas Avoid travel ing outside of the hotel zone Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras) Maintain awareness of your surroundings Avoid situations in which you may be isolated or stand out as a potential victim Avoid walking around the destination by yourself Avoid using ATMs if they are not affiliated with a recognized international chain Only use recommended taxi companies Avoid carrying large amounts of cash Avoid drawing attention to yourself Avoid walking around the destination by yourself at night Search for more information online about how to stay safe prior to travel Register with the U.S. Department of State prior to travel Avoid using services like Uber Purchase travel insurance prior to travel [Randomize the order of all choices] SECTION #1 2 . DESTINATION CHOICE 24. What is the likelihood that you will VISIT ACAPULCO for leisure purposes within the next year, on a scale of 1 to 5 (where 1= very unlikely and 5= very likely)? Likelihood to visit 1 2 3 4 5 25. Please indicate the likelihood that you will AVOID VISITING ACAPULCO for leisure purposes within the next year, on a scale of 1 to 5 (where 1= very unlikely and 5= very likely). Likelihood to avoid visiting 1 2 3 4 5

PAGE 265

265 SECTION #1 3 . DEMOGRAPHIC FACTORS The following set of questions is about your demographics. Your responses to these questions will be kept in the strictest confidence and used for statistical purposes only. 26. o Male o Female 27. What is your age? o 18 to 24 o 25 to 34 o 35 to 49 o 50 to 64 o 65 or older 28. Which statement best describes your total annual household income in 2014 (from all sources and before taxes)? o Under $15,000 o $15,000 to $24,999 o $25,000 to $34,999 o $35,000 to $49,999 o $50,000 to $74,999 o $75,000 to $99,999 o $100,000 to $124,999 o $125,000 and over 29. Which of the following best describes your race or ethnic origin? o White/Caucasian o Black/African American o Asian o Native Hawaiian/Pacific Islander o Hispanic o American Indian/Alaska Native o Multi ethic/mixed race o Other 30. What is the highest degree or level of school you have completed? If you are currently enrolled, mark the previous grade or highest degree received. o Some high school credit, no degree o High School Graduate high school diploma or the equivalent (for example: GED) o Some college credit, no degree o Associate degree (for exampl e: AA, AS) o BA, AB, BS) o MA, MS, MEng, MSW, MBA) o Professional degree (for example: MD, DDS, DVM, LLB, JD) o Doctoral Degree (for example: PhD, EdD) 31. o Now married o Separated o Widowed o Divorced o Never married

PAGE 266

266 32. How many children (younger than the age of 18) are currently living in your household? o 0 o 1 o 2 o 3 o 4 o 5 or more 33. What is the zip code of your permanent residence? ______ SECTION #1 4 . DESTINATION SPECIFIC FACTORS proceed to the last page and obtain your personal survey code. 34. Do you have friends and / or relatives currently living in Mexico ? o Yes o No 35. How well do you speak Spanish? o Not at all o Not well o Well o Very well

PAGE 267

267 SECTION #1 5 . SURVEY CODE 36. Please enter your MTurk Worker ID here. ______ Your personal survey code can be found below. Thank you very much for your valuable input.

PAGE 268

268 APPENDIX C SURVEY INSTRUMENT FOR THOSE WITH COUNTRY LEVEL DESTINATION PAST TRAVEL EXPERIENCE SECTION #1. INFORMED CONSENT Protocol Title The role of risk related constructs in the destination choice process: Reconcept ualizations and analyses Please read this consent document carefully before you decide to participate in this study. Purpose of the research study of Acapulco, Mexico as a tourist destination. What you will be asked to do in this study You will be asked a series of questions that are related to your past travel history, perceptions of Acapulco, and travel intentions. Time required It should take you approximately 15 20 minutes to complete this survey. Risks and benefits There is a minimal risk that security of any online data may be breached, but the survey host (Qualtrics) uses strong encryption and other data security methods to protect your information. Despite th is possibility, the risks to your physical, emotional, social, professional, or financial well being are considered to be minimal. There are also no foreseen benefits to you for taking this survey. Compensation You will be paid $1.5 0 compensation for part icipating in this study. Confidentiality Your identity will be kept confidential to the extent provided by the law. Your MTurk Worker ID will be used only for the purpose of awarding compensation and will not be shared with anyone outside of the research team. It will not be linked with your survey responses and will be removed from the data set once compensation has been provided. Voluntary participation Your participation in this study is completely voluntary. There is no penalty for not participating. Right to withdraw from the study You have the right to withdraw from this study at any time without consequence.

PAGE 269

269 Who to contact if you have questions about the study If you have any questions, you are encouraged to ask them. Questions can be addressed to Ashley Schroeder, Doctoral Candidate, at (352) 294 1658 or Dr. Lori Pennington Gray , Professor, at (352) 294 1657. Both can be reached by mail at University of Florida , Tourism Crisis Management Initiative, P.O. Box 118208, Gainesville, FL 32611 8208. Who to contact about your rights as a research participant in this study Questions regarding your rights as a participant in this study should be directed to the Univers ity of Florida, UF IRB 02 Office, P.O. Box 112250, Gainesville, FL 32611 2250 or (352) 392 0433 . Agreement 1. option below signifies that I voluntarily agree to participate in this study. a. AGREE SECTION #2. SCREENING QUESTIONS INTERNATIONAL PAST TRAVEL EXPERIENCE 2. Have you taken a trip outside of the U.S. within the past 5 years? o Yes o No ates that you do not qualify to participate in this study. Thank you for understanding. [SKIP TO END OF SURVEY] 3. Approximately how many trips outside of the U.S. have you taken in the past 5 years? o 1 o 2 o 3 o 4 o 5 o 6 o 7 o 8 o 9 o 10 o 11 o 12 o 13 o 14 o 15 or more 4. Have you ever taken a trip to Mexico? o Yes o No YES that you do not qualify to participate in this study. Thank you for understanding. [SKIP TO END OF SURVEY]

PAGE 270

270 SECTION # 3 . MEXICO SPECIFIC PSYCHOLOGICAL FACTORS When answering this set of questions, please think about MEXICO AS A TOURIST DESTINATION. 5. Please indicate your overall image of Mexico as a tourist destination, on a scale of 1 to 5 (where 1= very negative and 5= very positive). Destination image 1 2 3 4 5 6. Are you interested in visiting Mexico for leisure purposes within the next year? o Yes o No o No t sure SECTION # 4 . ACAPULCO SPECIFIC PSYCHOLOGICAL FACTORS Now please think about ACAPULCO, MEXICO AS A TOURIST DESTINATION when you answer the following set of questions. 7. Please indicate your overall image of Acapulco as a tourist destination, on a scale of 1 to 5 (where 1= very negative and 5= very positive). Destination image 1 2 3 4 5 8. Are you interested in visiting Acapulco for leisure purposes within the next year? o Yes o No o Not sure SECTION # 5 . INTERNATIONAL TRAVEL SPECIFIC PSYCHOLOGICAL FACTORS When you answer the following questions, please think about INTERNATIONAL TRAVEL. In other words, think about traveling outside of the United States. 9. Please indicate your level of agreement with the following statements, on a scale of 1 5 (where 1= strongly disagree and 5= s trongly agree). International travel risk perceptions 1 2 3 4 5 I feel nervous about traveling internationally right now Traveling internationally is risky right now I would feel very uncomfortable traveling internationally right now [Randomize the order of all choices]

PAGE 271

271 10. Please indicate your level of agreement with the following statement s , on a scale of 1 5 (where 1= strongly disagree and 5= strongly agree). Safety concerns 1 2 3 4 5 Safety is the most important attribute an international destination can offer Safety is a serious consideration when choosing an international tourist destination [Randomize the order of all choices] SECTION # 6 . THREAT APPRAISAL ( PERCEIVED RISK ) Please think about YOUR PERSONAL SAFETY WHILE VISITING ACAPULCO, MEXICO when you answer the following set of questions. 11. U s ing the rating scales below , please indicate how you feel when you think about your personal safety while visiting A capulco for leisure purposes. 1 2 3 4 5 Relaxed Anxious Excited Bored Fearless Fearful Assured Worried 12. Please indicate your level of agreement with the following statements, on a scale of 1 5 (where 1= strongly disagree and 5= strongly agree). Cognitive risk perceptions 1 2 3 4 5 If I were a victim of crime while visiting Acapulco, I would experience serious negative consequences It would have a serious negative impact on me if I were a victim of crime while visiting Acapulco If I were a victim of crime while visiting Acapulco, it would be harmful to my well being It is likely that I will be a victim of crime while visiting Acapulco I am at risk for being a victim of crime while visiting Acapulco My chances of being a victim of crime while visiting Acapulco are high [Randomize the order of all choices]

PAGE 272

272 SECTION # 7 . OBJECTIVE KNOWLEDGE ( RISK REDUCTION BEHAVIOR KNOWLEDGE ) 13. Have you heard of the following behaviors that tourists may engage in to ensure their personal safety while traveling ? Risk reduction behavior knowledge Have Heard Before Never Heard Before Remain in tourist areas Avoid travel ing outside of the hotel zone Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras) Maintain awareness of your surroundings Avoid situations in which you may be isolated or stand out as a potential victim Avoid walking around the destination by yourself Avoid using ATMs if they are not affiliated with a recognized international chain Only use recommended taxi companies Avoid carrying large amounts of cash Avoid drawing attention to yourself Avoid walking around the destination by yourself at night Search for more information online about how to stay safe prior to travel Register with the U.S. Department of State prior to travel Avoid using services like Uber Purchase travel insurance prior to travel [Randomize the order of all choices]

PAGE 273

273 SECTION # 8 . COPING APPRAISAL ( PERCEIVED EFFICACY) This set of questions is related to BEHAVIORS THAT TOURISTS MAY ENGAGE IN TO ENSURE THEIR PERSONAL SAFETY WHILE VISITING ACAPULCO , MEXICO . 14. How confident are you in your ability to perform the following behaviors to ensure your persona l safety while visiting Acapulco , on a scale of 1 5 (where 1= very unconfident and 5= very confident)? Self efficacy 1 2 3 4 5 Remain in tourist areas Avoid travel ing outside of the hotel zone Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras) Maintain awareness of your surroundings Avoid situations in which you may be isolated or stand out as a potential victim Avoid walking around the destination by yourself Avoid using ATMs if they are not affiliated with a recognized international chain Only use recommended taxi companies Avoid carrying large amounts of cash Avoid drawing attention to yourself Avoid walking around the destination by yourself at night Search for more information online about how to stay safe prior to travel Register with the U.S. Department of State prior to travel Avoid using services like Uber Purchase travel insurance prior to travel [Randomize the order of all choices]

PAGE 274

274 15. Please indicate how effective you believe the following behaviors would be in ensuring your personal safety while visiting A capulco, on a scale of 1 5 (where 1= very ineffective and 5= very effective). Response efficacy 1 2 3 4 5 Remain in tourist areas Avoid travel ing outside of the hotel zone Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras) Maintain awareness of your surroundings Avoid situations in which you may be isolated or stand out as a potential victim Avoid walking around the destination by yourself Avoid using ATMs if they are not affiliated with a recognized international chain Only use recommended taxi companies Avoid carrying large amounts of cash Avoid drawing attention to yourself Avoid walking around the destination by yourself at night Search for more information online about how to stay safe prior to travel Register with the U.S. Department of State prior to travel Avoid using services like Uber Purchase travel insurance prior to travel [Randomize the order of all choices]

PAGE 275

275 SECTION # 9 . ATTENTION QUESTION Recent research on decision making shows that choices are affected by context. Differences in how people feel, their previous knowledge and experiences, and their environment can affect their choices. To help us understand how people make decisions, we are interested in information about you. Spe cifically, we are interested in whether you have actually taken the time to read directions in this survey. If not, some results may not tell us very much about decision making in the real world. To show that you have read these directions, please ignore t he question below about how you are currently feeling and , instead , check only the none of th e above option as your answer. 16. Which of the following words best describes how you are currently feeling ? S elect all that apply . o Interested o Distressed o Ex cited o Upset o Strong o Guilty o Scared o Hostile o Enthusiastic o Proud o Irritable o Alert o Ashamed o Inspired o Nervous o Determined o Attentive o Jittery o Active o Afraid o None of the above indicates that you are not paying attention to this survey. We are sorry but we cannot accept your responses. Thank you for understanding. [SKIP TO END OF SURVEY]

PAGE 276

276 SECTION # 1 0 . ENGAGEMENT IN RISK REDUCTION BEHAVIOR S (INTENTIONS) When answering the following question, think about YOUR PERSONAL SAFETY WHILE VISITING ACAPULCO , MEXICO . 17. Please indicate the likelihood that you would engage in the following behaviors to ensure your personal safety while visiting Acapulco, on a scale of 1 5 (where 1= very unlikely and 5= very likely). Risk reduction behavior s (intention) 1 2 3 4 5 Remain in tourist areas Avoid travel ing outside of the hotel zone Avoid displaying indicators of wealth (such as expensive looking jewelry, watches, or cameras) Maintain awareness of your surroundings Avoid situations in which you may be isolated or stand out as a potential victim Avoid walking around the destination by yourself Avoid using ATMs if they are not affiliated with a recognized international chain Only use recommended taxi companies Avoid carrying large amounts of cash Avoid drawing attention to yourself Avoid walking around the destination by yourself at night Search for more information online about how to stay safe prior to travel Register with the U.S. Department of State prior to travel Avoid using services like Uber Purchase travel insurance prior to travel [Randomize the order of all choices] SECTION #1 1 . DESTINATION CHOICE 18. What is the likelihood that you will VISIT ACAPULCO for leisure purposes within the next year, on a scale of 1 to 5 (where 1= very unlikely and 5= very likely)? Likelihood to visit 1 2 3 4 5 19. Please indicate the likelihood that you will AVOID VISITING ACAPULCO for leisure purposes within the next year, on a scale of 1 to 5 (where 1= very unlikely and 5= very likely). Likelihood to avoid visiting 1 2 3 4 5

PAGE 277

277 SECTION #1 2 . DEMOGRAPHIC FACTORS The following set of questions is about your demographics. Your responses to these questions will be kept in the strictest confidence and used for statistical purposes only. 20. o Male o Female 21. What is your age? o 18 to 24 o 25 to 34 o 35 to 49 o 50 to 64 o 65 or older 22. Which statement best describes your total annual household income in 2014 (from all sources and before taxes)? o Under $15,000 o $15,000 to $24,999 o $25,000 to $34,999 o $35,000 to $49,999 o $50,000 to $74,999 o $75,000 to $99,999 o $100,000 to $124,999 o $125,000 and over 23. Which of the following best describes your race or ethnic origin? o White/Caucasian o Black/African American o Asian o Native Hawaiian/Pacific Islander o Hispanic o American Indian/Alaska Native o Multi ethic/mixed race o Other 24. What is the highest degree or level of school you have completed? If you are currently enrolled, mark the previous grade or highest degree received. o Some high school credit, no degree o High School Graduate high school diploma or the equivalent (for example: GED) o Some college credit, no degree o Associate degree (for example: AA, AS) o BA, AB, BS) o or example: MA, MS, MEng, MSW, MBA) o Professional degree (for example: MD, DDS, DVM, LLB, JD) o Doctoral Degree (for example: PhD, EdD) 25. o Now married o Separated o Widowed o Divorced o Never married

PAGE 278

278 26. How many children (younger than the age of 18) are currently living in your household? o 0 o 1 o 2 o 3 o 4 o 5 or more 27. What is the zip code of your permanent residence? ______ SECTION #1 3 . DESTINATION SPECIFIC FACTORS more questions left to answer before you can proceed to the last page and obtain your personal survey code. 28. Do you have friends and / or relatives currently living in Mexico? o Yes o No 29. How well do you speak Spanish? o Not at all o Not well o Well o Very well SECTION #1 4 . SURVEY CODE 30. Please enter your MTurk Worker ID here. ______ Your personal survey code can be found below. Thank you very much for your valuable input.

PAGE 279

279 APPENDIX D DEVELOPMENT OF DESTINATION CHOICE PROCESS MODEL The destination choice process, a central decision in the travel decision making process ( Fesenmaier & Jeng, 2000; Jeng & Fesenmaier, 2002 ), was an important component of t he proposed conceptual model . The foundation of this section of the m making process for choice sets, which recognizes that individuals go through two evaluation stages when making a purchase choice. Accordingly, the six stages that consumers go through during purchase decisions ar e: (1) problem recognition; (2) search (internal and external); (3) awareness set identification; (4) evaluation of the awareness set and choice processes leading to an evoked set; (5) evaluation and choice from an evoked set; (6) post purchase evaluation (LeBlanc, 1989). making model was supplemented with concepts from choice sets. In particular, Mansfeld (1992) suggested that the information search process occurs in two stages. Thus, a second information search stage was integrated (evaluation of the awareness set and choice processes leading to an evoked s et) stages (1992) model also considered whether or not prospective tourists felt that they had adequate information to eliminate destination alternatives. Accordingly, this was integrated between stages four (evaluation of the awarenes s set and choice processes leading to an evoked set) and five (evaluation and choice from an evoked set) of making model for choice sets. This addition specifically suggests that if a prospective tourist does not believe that they have an

PAGE 280

280 adequate amount of information to make elimination s , they may revert back to the second information search stage to gather more information . In addition, travel at the destination was added in between stages five (evaluation and choice from an evo ked set) and six (post destination choice and travel are two distinct stages of the destination choice process (Mansfeld, 1992). A feedback loop was also added from the post travel evaluati on of destination choice to the beginning of the proposed destination choice process, as suggested by Mansfeld (1992). A time lag was added between the final desti nation choice stage and travel at the destination, as suggested by Sönmez and Graefe (1998a). The addition of this path recognizes that there is a ti me lag between the anticipation/ planning stage and the on site / in situ stage of the travel experience. During this time lag, there is the potential that something may happen at the destination (Sönmez & Graefe, 1998a) , such as a tourism crisis (e.g., natural disaster, crimes against tourists) . If something were to occur, tourists might choose to proceed with their travel plans, cancel their travel plans, or return to a previous stage in the decision making process in search of an alternative destination (Sönmez & Graefe, 1998a). Lastly, double ended arrows were included between a major ity of the stages of the proposed conceptual model of the tourist destination choice process . These arrows were added in an effort to recognize that the decision making process is dynamic . In other words, individuals do not have to proceed from the top of the model to the bottom in a linear fashion. Rather, the proposed conceptual model of the tourist destination

PAGE 281

281 choice process suggests that individuals can move back up to previous stages. Thus , the proposed model suggests that the process is fluid. The pr oposed conceptual model of the tourist destination choice process is presented in Figure D 1. The proposed tourist destination choice process model was color coded to indicate whether the component of the model was derived from Le stage d ecision model of international tourism decision making process (green). In summary, the proposed tourist destination choice process model suggests that prospective tourists go through nine stages when choosing a tourist destination. First, the tourist destination choice process begins when prospective tourists are motivated to travel. Once prospective tourists are motivated to travel, they search for information to a result of the information search process, prospective tourists then form a n early consideration set of destination alternatives. Next, prospective tourists go through a second information search in order to ensure that they have adequate information to evaluate the early consideration set of destination alternatives. Following the second information search, prospective tourists evaluate the destination alternatives and narrow the early consideration set down to form a late consideration set. Once the late consideration set of destination alternatives is formed, prospective touri sts may engage in an additional information search if they do not believe that they have adequate information to make a final evaluation and the final destination choice.

PAGE 282

282 Figure D 1. Tourist Destination Choice Process, Adapted from Le Blanc (1989), Mans feld (1992), and Sönmez & Graefe (1998a) .

PAGE 283

283 If prospective tourists believe that they have adequate information, they advance to the final evaluation and make a final destination choice from the late consideration set of destination alternatives. Travel at the destination follows the final destination choice, however, there is a time lag between these two stages. Finally, once tourists return home from their trip to the destination, they make a post travel evaluation of their destination choice. Notably, thr oughout the nine stages of the destination choice process, there is the potential that something may happen at the destination which may individuals can move back up to previous stages of the destination choice process.

PAGE 284

284 LIST OF REFERENCES Amazon Mechanical Turk (2015). Service summary . Retrieved March 9, 2015 from https://requester.mturk.com/tour Ankomah, P. K., Crompto n, J. L., & Baker, D. (1996). Influence of cognitive distance in vacation choice. Annals of Tourism Research , 23 (1), 138 150. APEC International Centre for Sustainable Tourism (AICST) (2006, December). Tourism risk management An authoritative guide to managing crises in tourism . Singapore: Asia Pacific Economic Cooperation (APEC). tourism. Daily Mail . Retrieved March 12, 2015, from http://www.dailymail.co.uk/wires/ap/article 2831985/Mexico Violent protests hit Acapulco tourism.html Bandura, A. (1977). Self efficacy: Toward a unifying theory of behavioral c hange. Psychological Review , 84 , 191 215. Bandura, A. (1982). Self efficacy mechanism in human agency. American Psychologist , 37 , 122 147. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prenti ce Hill. Bandura, A. (1989). Regulation of cognitive processes through perceived self efficacy. Developmental Psychology , 25 (5), 729 735. Bandura, A. (1992). Exercise of personal agency through the self efficacy mechanism. In R. Schwarzer (Ed.), Self efficacy: Thought Control of Action (pp. 3 38). Washington D.C.: Hemisphere. Bandura, A., & Schunk, D. H. (1981). Cultivating competence, self efficacy, and intrinsic interest through proximal self motivation. Journal of Personality and Social Psychol ogy , 41 , 586 598. Bandura, A., Adams, N. E., Hardy, A. B., & Howells, G. N. (1980). Tests of the generality of self efficacy theory. Cognitive Therapy and Research , 4 , 39 66. Bandura, A., Cioffi, D., Taylor, C. B., & Brouillard, M. E. (1988). Perceived sel f efficacy in coping with cognitive stressors and opioid activation. Journal of Personality and Social Psychology , 55 , 479 488. Bandura, A., Reese, L., & Adams, N. E. (1982). Microanalysis of action and fear arousal as a function of differential levels of perceived self efficacy. Journal of Personality and Social Psychology , 43 , 5 21.

PAGE 285

285 Bao, Y. F., & McKercher, B. (2008). The effect of distance on tourism in Hong Kong: A comparison of short haul and long haul visitors. Asia Pacific Journal of Tourism Resear ch , 13 (2), 101 111. Behrend, T. S., Sharek, D. J., Meade, A. W., & Wiebe, E. N. (2011). The viability of crowdsurfing for survey research. Behavior Research Methods , 43 (3), 800 813. Bender, J., & Rosen, A. (2014, October 24). Mexico's Drug War is entering a dark phase. Business Insider . Retrieved March 12, 2015, from http://www.businessinsider.com/mexicos drug war is entering a dangerous phase 2014 10 Black, H. (1979). (5 th ed.). St. Paul, MN: West. Boer, H., & Seydel, E. R. (1996). Protection motivation theory. In M. Connor and P. Norman (Eds.), Predicting Health Behavior: Research and Practice with Social Cognition Models (pp. 9 5 120). Buckingham: Open University Press. Brin, E. (2006). Politically oriented tourism in Jerusalem. Tourist Studies , 6 (3), 215 243. Brooker, G. (1983). An assessment of an expanded measure of perceived risk. In T. C. Kinnear (Ed.), Advances in Consumer Research (pp. 439 441). Provo, UT: Association for Consumer Research. Brucks, M. (1985). The effects of product class knowledge on information search behavior. Journal of Consumer Research , 12 , 1 16. Brun, W., Wolff, K., & Larsen, S. (2011). Tourist worri es after terrorist attacks: Report from a field experiment. Scandinavian Journal of Hospitality and Tourism , 11 (3), 387 394. Brunel, O., & Pichon, P. (2004). Food related risk reduction strategies: Purchasing and consumption processes. Journal of Consumer Behaviour , 3 (4), 360 374. Brunt, P., Mawby, R., & Hambly, Z. (2000). Tourist victimization and the fear of crime on holiday. Tourism Management , 21 (4), 417 424. M echanical T urk a new source of inexpensive, yet high quality, data? Perspectives on Psychological Science , 6 (1), 3 5. Cahyanto, I. (2012). Effects of individual characteristics, travel related variables, information search and risk appraisal (Unpublished doctoral dissertation). University of Florida, Gainesville, FL.

PAGE 286

286 Cahyanto, I., & Pennington Gray, L. (2014). Communicating hurricane evacuation to tourists: Gender, past experience with hurricanes, and place of residence. Jo urnal of Travel Research , doi: 10.1177/0047287513517418 Cahyanto, I., Pennington Gray, L., Thapa, B., Srinivasan, S., Villegas, J., Matyas, C., & hurricane evacuation decision making. Journal of Destination Marketing & Management , 2 (4), 253 265. Carlsen, J. C., & Liburd, J. J. (2008). Developing a research agenda for touri sm crisis management, market recovery and communications. Journal of Travel & Tourism Research , 23 (2), 265 276. Chapman, G. B., & Coups, E. J. (2006). Emotional and preventive health behavior: Worry, regret, and influenza vaccination. Health Psychology , 25 , 82 90. Clawson, M., & Knetsch, J. L. (1966). Economics of outdoor recreation . Baltimore, MD: John Hopkins Press. Connor, B. A., & Schwartz, E. (2005). Typhoid and paratyphoid fever in travelers. The Lancet Infectious Diseases , 5 (1), 623 628. Crystal, S. (1993). Welcome to downtown U.S.A. Meetings and Conventions , 28 (3), 42 59. Datar, R., & Douglas, J. (2012, February 20). Mexico's Acapulco resort takes a dive over drugs war. BBC News . Retrieved March 12, 2015, from http://news.bbc.co.uk/2/hi/programmes/fast_track/9696962.stm de Sausmarez, N. (2007). Crisis management, tourism and sustainability: The role of indicators. Journal of Sustainable Tourism , 15 (6), 700 714. de Vaus, D. (2001). Research design in social research . London, UK: Sage. Dowling, G. R. (1986). Perceived risk: the concept and its measurement. Psychology & Marketing , 3 , 193 210. Drabek, T. (2009). The human side of disaster . Boca Raton, FL: CRC Press. Drabek, T. E. (1995). Disaster responses within the tourism industry. International Journal of Mass Emergencies and Disasters , 13 (1), 7 23. Draper, N. R., & Smith, H. (1998). Applied r egression a nalysis (3rd ed.), New York, NY: Wiley. Echtner, C. M., & Rit chie, J. R. B. (2003). The meaning and measurement of destination image. The Journal of Tourism Studies , 14 (1), 37 48.

PAGE 287

287 Engel, J. F., Blackwell, R. D., & Kollat, D. T. (1978). Consumer behavior . Hinsdale, IL: Dryden, 1978. Faulkner, B. (2001). Towards a fra mework for tourism disaster management. Tourism Management , 22 , 135 147. Fesenmaier, D. R., & Jeng, J. (2000). Assessing structure in the pleasure trip planning process. Tourism Analysis , 5 , 13 27. Flannery, N. P. (2014, December 30). Can Acapulco bounce back? Forbes . Retrieved March 12, 2015, from http://www.forbes.com/sites/nathanielparishflanne ry/2014/12/30/can acapulco bounce back/ Floyd, D. L., Prentice Dunn, S., & Rogers, R. W. (2000). A meta analysis of research on protection motivation theory. Journal of Applied Social Psychology , 30 (2), 407 429. Floyd, M. F., & Pennington Gray, L. (2004). Profiling risk perceptions of tourists. Annals of Tourism Research , 31 (4), 1051 1054. Floyd, M. F., Gibson, H., Pennington Gray, L., & Thapa, B. (2004). The effect of risk perceptions on intentions to travel in the aftermath of September 11, 2001. Journal of Travel & Tourism Marketing, 15 (2 3), 19 38. Fuchs, G., & Reichel, A. (2006a). Tourist destination risk perception: The case of Israel. Journal of Hospitality and Leisure Marketing , 14 (2), 81 106. Fuchs, G., & Reichel, A. (2 006b). Correlates of destination risk perception and risk reduction strategies. In M. Kozak, & L. Andreu (Eds.), Progress in tourism marketing (pp. 161 170). Amsterdam: Elsevier. Fuchs, G., & Reichel, A. (2011). An exploratory inquiry into destination risk perceptions and risk reduction strategies of first time vs. repeat visitors to a highly volatile destination. Tourism Management , 32 , 266 276. Fuchs, G., Uriely, N., Reichel, A., & Maoz, D. (2013). Vacation in a terror stricken perceptions and rationalizations. Journal of Travel Research , 52 (2), 182 191. George, D., & Mallery, M. (2010). SPSS for Windows s tep by s tep: A s imple g uide and r eference, 17.0 update . Boston , MA : Pearson. Germunden, H. G. (1985). Perceived risk and info rmation search. International Journal of Research in Marketing , 2 , 79 100.

PAGE 288

288 Goeldner, C., & Ritchie, B. (2012). Tourism: Principles, practices, philosophies (12th ed.). New York , NY : John Wiley & Sons, Inc. Google (n.d.). [Google Maps of Acapulco, Mexico]. Retreived April 1, 2015, from https://www.google.com/maps/place/Acapulco,+Guerrero,+Mexi co/@16.8357713 , 100.0023555,11z/data=!3m1!4b1!4m2!3m1!1s0x85ca5785aece50c9:0x9801d8f7 8a90a4e3 . Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate data analysis (7 th ed.). Upper Saddle River, NJ: Prentice Hall. Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression based approach . New York: NY: The Guilford Press. Hilton, E., Kolakowski, P., Singer, C., & Smith, M. (2006). Efficacy of Lactobacilus GG as a diarrh eal preventive in travelers. Journal of Travel Medicine , 4 (1), 41 43. Huang, Y., Tseng, Y., & Petrick, J. F. (2008). Crisis management planning to restore tourism after disasters: A case study of Taiwan. Journal of Travel & Tourism Marketing , 23 (2 4), 203 221. Jackson, D. N., Hourany, L., & Vidmar, N. J. (1972). A four dimensional interpretation of risk taking. Journal of Personality , 40 , 483 501. Jacoby, J., & Kaplan, L. B. (1972). The components of perceived risk. In M. Venkatesan (Ed.), Proceedings, 3 rd Annual Conference (pp. 382 393). Chicago, IL: Association for Consumer Research. Jang, H., Lee, S., Lee, S., & Hong, S. (2007). Expanding the individual choice sets Tourism Management , 28 , 129 9 1314. Jeng, J., & Fesenmaier, D. (2002). Conceptualizing the travel decision making hierarchy: A review of recent developments. Tourism Analysis , 7 , 15 32. Kash, T.J., & Darling, J. (1998). Crisis management: Prevention, diagnosis and intervention. Leade rship & Organization Development Journal, 19 (4), 179 186. Knowles, E. S. (1976). Searching for motivations in risk taking and gambling. In W. R. Eadington (Ed.), Gambling and society (pp. 295 322). Springfield, IL: Charles C. Thomas. Knowles, E. S., Cutter , H. S. G., Walsh, D. H., & Casey, N. A. (1973). Risk taking as a personality trait. Social Behavior and Personality , 1 , 123 136.

PAGE 289

289 Kozak, M., Crotts, J.C., & Law, R. (2007). The impact of the perception of risk on international travellers. International Jou rnal of Tourism Research, 9 (4), 233 242. Larsen, S., Brun, W, & Øgaard , T. (200 9 ). What tourists worry about construction of a scale measuring tourist worries. Tourism Management , 30 , 260 265. Law, R. (2006). The perceived impact of risks on travel decisions. International Journal of Tourism Research , 8 (4), 289 300. LeBlanc, R. P. (1989). An exploratory investigation into the relationship between evoked set decisions and final choice process. In J. M. Hawes and J. Thanopoulos (Eds.), Developments in Marketing Science, Volume 12 (pp. 7 11). Akron, OH: Academy of Marketing Science. Lew, A. A. (2014). Scale, change and resilience in community tourism planning. Tourism Geographies: An International J ournal of Tourism, Space, Place and Environment , 16 (1), 14 22. Linthicum, K. (2014, October 13). Los Cabos, Mexico, starts to recover from Hurricane Odile. LA Times . Retrieved March 12, 2015, from http://www.latimes.com/world/mexico americas/la fg mexico cabo storm 20141014 story.html reduction strategies in leisure travel. Journal of Travel & Tourism Marketing , 28 (3), 240 260. Lo, A. S., Law, R., & Cheung, C. (2011b). Segmenting leisure travelers by risk reduction strategies. Journal of Travel & Tourism Marketing , 28 (8), 828 839. Lo, A., & Lam, T. (2004). Long haul and sh ort haul outbound all inclusive package tours. Asia Pacific Journal of Tourism Research , 9 (2), 161 176. Loewenstein, G. F., Weber, E. U., Hsee, C. K., & Welch, E. S. (2001). Risk as feelings. Psychological Bulletin , 127 , 267 286. MacCrimmon, K. P., & Wehrung, D. A. (1986). Taking risks . New York: The Free Press. Maddux, J. E., & Rogers, R. W. (1983). Protection motivation and self efficacy: A revised theory of fear appeals and attitude change. Journal of Experimental Social Psychology , 19 , 469 479. Ma nsfeld, Y. (1992). From motivation to actual travel. Annals of Tourism Research , 19 , 399 419.

PAGE 290

290 Mansfeld, Y. (1999). Cycles of war, terror and peace: Determinants and management of crisis and recovery of the Israeli tourism industry. Journal of Travel Research , 38 (1), 30 36. Maser, B., & Weiermair, K. (1998). Travel decision making: F rom the vantage point of perceived risk and information preferences. Journal of Travel & Tourism Marketing , 7 (4), 107 119. Mason, W., & Suri, S. (2012). Conducting behavior turk. Behavior Research Methods , 44 (1), 1 23. Matyas, C. J., Srinivasan, S., Cahyanto, I., Thapa, B., Pennington Gray, L., & Villegas, J. (2011). Risk perception and evacuation decisions of Florida tourists under hurrican e threats: A stated preference analysis. Natural Hazards , 59 , 871 890. McKercher, B., Hui, E., Hall, C., Timothy, D., & Duval, D. (2003). Terrorism, economic uncertainty and outbound travel from Hong Kong. Safety & Security in Tourism , 15 (2 4), 99 115. Mi nitab (n.d.). How important are normal residuals in regression analysis? [Blog post]. Retrieved from http://blog.minitab.com/blo g/adventures in statistics/how important are normal residuals in regression analysis Mitchell, V. W. (199 3 ). Factors affecting consumer risk reduction: A review of current evidence . Management Research News , 16 ( 9 ), 6 21 . Mitchell, V. W., Davies, F., Mouti nho, L., & Vassos, V. (1999). Using neural networks to understand service risk in the holiday product. Journal of Business Research , 46 (2), 167 181. Mitchell, V. W., Moutinho, L., & Lewis, B. R. (2003). Risk reduction in purchasing organizational professio nal services. The Service Industries Journal , 23 (5), 1 19. Morrison, A. M. (2010). Hospitality & Travel Marketing (4 th ed.). Clifton Park, NY: Delmar Cengage Learning. Moutinho, L. (1987). Consumer behaviour in tourism. European Journal of Marketing , 21(10), 5 44. Murphy, P. E., & Bayley, R. (1989). Tourism and disaster planning. Geographical Review , 79 (1), 36 46. Myers, G. N. (2014, September 23). A rising Sun Triangle. Travel Weekly . Retrieved March 12, 2015, from http://www.travelweekly.com/Mexico Travel/Insights/A rising Sun Triangle

PAGE 291

291 National Cancer Institute (2005). Theory at a glance: A guide for health promotion practice . Bethesda, MD: NIH Publication. Neuwi rth, K., Dunwoody, S., & Griffin, R.J. (2000). Protection motivation and risk communication. Risk Analysis , 20 (5), 721 734. Norman, P., Boer, H., & Seydel, E. (2005). Protection motivation theory. In M. Conner & P. Norman (Eds.), Predicting health behaviour: Research and practice with social cognition models (pp. 81 126). Buckingham, UK: Open University Press. National Travel and Tourism Office [ NTTO ] (2014). 2013 United States resident travel abroad . Washington, D.C.: NTTO. Nunnally, J. (1978). Psychometric methods . New York, NY: McGraw Hill. Nysveen, H., & Lexhagen, M. (2001). Swedish and Norwegian tourism websites: The importance of reservation services and value added services. Scandinavian Journal of Hospitality and Tourism , 1 (1), 38 44. Ok inawa Convention & Visitors Bureau (2015). For a safe and enjoyable stay in Okinawa: Safety guide for visitors to Okinawa. Okinawa Convention & Visitors Bureau . Retrieved November 1, 2015, from http://en.okinawastory.jp/assets/uploads/2015/03/SafeEnjoyableStay2015_en.pdf Pallant, J. (2007). SPSS survival manual (3rd ed.). New York , NY : McGraw Hill. Paolacci, G., Chandler, J., & Ipeirotis, P. G. (2010). Running exp eriments on Amazon Mech a nical Turk. Judgment and Decision Making , 5 (5), 411. Papadimitriou, D., Apostolopoulou, A., & Kaplanidou, K. (2015). Destination personality, affective image, and behavioral intentions in domestic urban tourism. Journal of Travel Research , 54 (3), 302 315. Parsons, W. (1996). Crisis management. Career Development International , 1 (5), 26 28. Pacific Asia Travel Association [ PATA ] (2003). . Bangkok, Thailand: PATA. Pacific Asia Travel Association [PATA] (2 012). Bounce back: Tourism, risk, crisis and recovery management guide . Bangkok, Thailand: PATA. Pennington Gray, L., Cahyanto, I., Thapa, B., McLaughlin, E., Willming, C., & Blair, S. (2010). Destination management organizations and tourism crisis managem ent plans in Florida. Tourism Review International , 13 , 247 261.

PAGE 292

292 Pennington Gray, L., Kaplanidou, K., & Schroeder, A. (2013). Drivers of social media use among African Americans in the event of a crisis. Natural Hazards , 66 (1), 77 95. Pennington Gray, L., Schroeder, A., & Kaplanidou, K. (2011). Examining the influence of past travel experience, general web searching behaviors, and risk perceptions on future travel intentions. International Journal of Safety and Security in Tourism/Hospitality , 1 (1), 64 89. Pennington Gray, L., Thapa, B., Kaplanidou, K., Ca hyanto, I., & McLaughlin, E. (2011). Crisis planning and preparedness in the United States tourism industry. Cornell Hospitality Quarterly , 52 (3), 312 320. Peterson, R. A . (200 0 ). Constructing effective questionnaires . London, UK: Sage . Preacher, K. J. (2015). Calculation for the Sobel test . Retrieved November 15, 2015 from http://quantpsy.org/sobel/sobel.htm Prentice Dunn, S., & Rogers, R. (1986). Prot ection motivation theory and preventive health: Beyond the health belief model. Health Education Research , 1 , 153 161. Prideaux, B., Laws, E., & Faulkner, B. (2003). Events in Indonesia: Exploring the limits to formal tourism trends forecasting methods in complex crisis situations. Tourism Management , 24 (4), 475 487. Qi, C.X., Gibson, H.J., & Zhang, J.J. (2009). Perceptions of risk and travel intentions: The case of China and the Beijing Olympic Games. Journal of Sport & Tourism, 14 (1), 43 67. Reisinger, Y., & Mavondo, F. (2005). Travel anxiety and intentions to travel internationally: implications of travel risk perceptions. Journal of Travel Research , 43 , 212 225. Ritchie, B. (2004). Chaos, crises and disasters: a strategic approach to crisis management in the tourism industry. Tourism Management , 25 , 669 683. Ritchie, B. (2008). Tourism disaster planning and management: From response and recovery to reduction and readiness. Current Issues in Tourism , 11 (4), 315 348. Roehl, W. S. & Fesenmaier , D. R. (1992). Risk perceptions and pleasure travel: An exploratory analysis. Journal of Travel Research , 30 , 17 26. Rogers, R. W. (1975). A protection motivation theory of fear appeals and attitude change. The Journal of Psychology , 91 (1), 93 114.

PAGE 293

293 Roger s, R. W. (1983). Cognitive and physiological processes in fear appeals and attitude change: A revised theory of protection motivation. In J. Cacioppo & R. Petty (Eds.), Social psychophysiology . New York, NY: Guilford Press. Rogers, R. W., & Mewborn, C. R. (1976) . Fear a ppeals and a ttitude c hange: Effects of a t hreat's n oxiousness, p robability of o ccurrence and the e fficacy of c oping responses. Journal of Personality and Social Psychology , 34 , 54 61 . Rogers, R. W., & Prentice Dunn, S. (1997). Protection moti vation theory. In D. Gochman (Ed.), Handbook of health behavior research: Vol 1. Determinants of health behavior: Personal and social (pp. 113 132). New York, NY: Plenum. Rohrmann, B. (1995). Effective risk communication for fire preparedness: A conceptual framework. The Australian Journal of Emergency Management , 10 , 43 48. Rosenstock, I. M. (1974). Historical origins of the health belief model. Health Education Monographs , 2 , 328 335. Rosenstock, I., Stretcher, V., & Becker, M. (1988). Social learning theory and the health belief model. Health Education Quarterly , 15 , 175 183. ABC News . Retrieved March 12, 2015 , from http://abcnews.go.com/ABC_Univision/ABC_Univision/acapulco worlds nicest dangerous resort city/story?id=18652375 Santiago, E. (2015). What should I do if my data is not normal? [Blog post]. Retrieved from http://blog.minitab.com/blog/ understanding statistics and its application/what should i do if my data is not normal v2 A. Pizam & Y. Mansfeld (Eds.), Tourism, Crime, and International Sec urity Issues . Chichester, UK: Wiley. Schroeder, A. (2012). media use Schroeder, A., & Pennington Gray, L. (2014). Perce ptions of crime at the Olympic Games: What role does media, travel advisories, and social media play? Journal of Vacation Marketing , 20 (3), 225 237. Schroeder, A., Pennington Gray, L., Donohoe, H., & Kiousis, S. (2013a). Using social media in times of cris is. Journal of Travel & Tourism Marketing, Special issue on Social Media, 30 (1 2), 126 143.

PAGE 294

294 Schroeder, A., Pennington Gray, L., Kaplanidou, K., & Zhan, F. (2013b). Destination risk perceptions among U.S. residents for London as the host city of the 2012 Summer Olympic Games. Tourism Management , 38 (October), 107 119. ng of system effects. Journal of Travel & Tourism Marketing , 19 (2/3), 149 58. Severtson, D. J., & Henriques, J. B. (2009). The effect of graphics on environmental health risk beliefs, emotions, behavioral intentions, and recall. Risk Analysis , 29 (11), 1549 1565 Sirakaya, E., & Woodside, A. G. (2005). Building and testing theories of decision making by travelers. Tourism Management , 26 , 815 832. Slovic, P. (2000). The Perception of Risk . London, UK: Earthscan. Slovic, P., & Peters, E. (2006). Risk perception and affect. Current Directions in Psychological Science , 15 (6), 322 325. Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equation models. In S. Leinhart (Ed.), Sociological methodology (pp. 290 312). San Francisco, C A: Jossey Bass. Sönmez, S. F., & Graefe, A. R . (1998a). Influence of terrorism risk on foreign tourism decisions. Annals of Tourism Research, 25 (1), 112 144. Sönmez, S. F., & Graefe, A. R. (1998b). Determining future travel behavior from past travel expe rience and perceptions of risk and safety. Journal of Travel Research, 37 (2), 171 177. Spiggle, S., & Sewall, M. A. (1987). A choice sets model of retail selection. Journal of Marketing , 51 (2), 97 111. StatTrek (2015). Power of a hypothesis test [Blog po st]. Retrieved from http://stattrek.com/hypothesis test/power of test.aspx?Tutorial=AP Tabachnick, B. G., & Fidell, L. S. (2013). Using multivariate statistics (6 th ed.). Boston, MA: Pearson. Tourism Crisis Management Initiative [ TCMI ] (2015). Welcome . Retrieved from http://tcmi.hhp.ufl.edu/ The Economist (2009, November 21). Mexico's economy: A different kind of recession. The Economist . Retrieved March 12, 2015, from http://www.economist.com/node/14917758

PAGE 295

295 The Weather Channel (2015, November 3). Hurricane Patricia recap: Strongest landfalling Pacific hurricane on record . The Weather Channel . Retrieved November 5, 2015, from http://www.weather.com/storms/hurricane/news/hurricane patricia mexico coast Tsaur, S. H., Tzeng, G. H., & Wang, K. C. (1997). Evaluating tourist risks from fuzzy perspectives. Annals of Tourism Research , 24 (4), 796 812. Um, S., & Crompton, J. L. (1990). Attitude determinants in tourism destination choice. Annals of Tourism Research , 17, 432 448. United Natio ns World Tourism Organization [ UNWTO ] (2014). UNWTO tourism highlights: 2014 edition . Madrid, Spain, UNWTO. Uriely, N., Maoz, D., & Reichel, A. (2007). Rationalizing terror related risks: The case of Israeli tourists in Sinai. International Journal of Tour ism Research , 9 (10), 1 8. U.S. Department of State (2014, December 24). Mexico Travel Warning. U.S. Department of State . Retrieved March 1, 2015, from http://travel.state.gov/content/passports/english/alertswarnings/mexico travel warning.html U.S. Department of State (2015, May 5). Mexico Travel War ning. U.S. Department of State . Retrieved May 11, 2015, from http://travel.state.gov/content/passports/english/alertswarnings/mexico travel warning. html Villegas, J., Matyas, C., Srinivasan, S., Cahyanto, I., Thapa, B., & Pennington Gray, L. (201 3 ). Cognitive and affective responses of Florida tourists after exposure to hurricane warning messages. Natural Hazards , 66 (1), 97 116. Weinstein, N. D. (19 93). Testing four competing theories of health protective behavior. Health Psychology , 12 (4), 324 333. Witte, K., Cameron, K. A., Lapinski, M. K., & Nzyuko, S. (199 8 ). A theoretically based evaluation of HIV/AIDS prevention campaigns along the trans Africa highway in Kenya. Journal of Health Communication , 3 , 345 363. Wolff, K., & Larsen, S. (2013). Tourist worries Here and now vs. there and then: The effect of item wording in the Tourist Worry Scale. Tourism Management , 35 , 284 287. Woodside, A. G., & L ysonski, S. (1989). A general model of traveler destination choice. Journal of Travel Research , 27 (4), 8 14. World Travel & Tourism Council [ WTTC ] (2015). Home . Retrieved from http://www.wttc.org/

PAGE 296

296 Zeckhauser, R. & Thompson, M. (1970). Linear regression with non normal error terms. The Review of Economics and Statistics , 52 (3), 280 286.

PAGE 297

297 BIOGRAPHICAL SKETCH Ashley L. Schroeder earned a Bachelor of Arts in Business Administration from the University of Florida in Ma y 2007. Upon graduation, she relocated to New York City and had a successful career in the fashion industry. In 2010, she returned to the University of Florida . S he earned a Master of Science in recreation, parks and tourism from the Department of Tourism, Recreation, and Sport Management in May 2012. She earned a Ph.D. in health and human performance from the College of Health and Human Performance at the University of Florida in December 2015. Ashley is affiliated with the Tourism Crisis Management Initi ative at the University of Florida. Throughout her academic tenure, she has been devoted to an active, focused research agenda in the area of destination management. Her overall research agenda has focused on proactive tourism crisis management from the su pply and demand perspectives. On the supply side, her focus has been on comprehensive tourism crisis management, risk management, destination resilience, and collaborative networks. On the demand side, her focus has been on understanding the perceptions, a ttitudes, intentions, and behaviors of tourists in times of real and perceived crisis. In the area of tourism crisis management , Ashley has a strong publication record. In addition to publishing in top peer reviewed journals, she has also written several technical reports and has given research presentations in the United States, Belgium, and Canada. Further, Ashley has an established record of externally funded research projects. She received several research grants to complete her dissertation and has al so been a Co PI on two research grants for the local visitor and convention bureau.