Cognitive and affective responses of Florida tourists after exposure to hurricane warning messages

MISSING IMAGE

Material Information

Title:
Cognitive and affective responses of Florida tourists after exposure to hurricane warning messages
Series Title:
Natural Hazards
Physical Description:
Journal Article
Creator:
Matyas, Corene
Publication Date:

Notes

Abstract:
Tourists are particularly vulnerable to natural disasters such as hurricanes since they might be less informed and prepared than residents of disaster-prone areas. Thus, understanding how the traits of a tropical cyclone as well as specific characteristics of tourists influence affective and cognitive responses to a hurricane warning message is a critical component in disaster planning. Using scenarios that presented tropical cyclones with different relevant characteristics (such as category at landfall), tourists’ knowledge, experience with hurricanes, trip traits, and the location of the survey (coastal or inland), this study contributes to the literature on sociological issues related to natural disasters. The findings suggest that risk perceptions and fear are influenced differently by the traits of the hurricanes and tourists’ knowledge and experience. Risk is strongly influenced by the projected category of the hurricane at landfall, while fear is not as sensitive to this extremely relevant trait of cyclones. The results also suggest that the influence of risk and fear on evacuation likelihood is strong and positive. This study shows the value of studying cognitive and affective responses to uncertain events.
Acquisition:
Collected for University of Florida's Institutional Repository by the UFIR Self-Submittal tool. Submitted by Corene Matyas.
Publication Status:
Published

Record Information

Source Institution:
University of Florida Institutional Repository
Holding Location:
University of Florida
Rights Management:
All rights reserved by the submitter.
System ID:
IR00003677:00001


This item is only available as the following downloads:


Full Text

PAGE 1

ORIGINALPAPERCognitiveandaffectiveresponsesofFloridatourists afterexposuretohurricanewarningmessagesJorgeVillegas€CoreneMatyas€SivaramakrishnanSrinivasan€IgnatiusCahyanto€BrijeshThapa€LoriPennington-GrayReceived:12August2011/Accepted:13February2012/Publishedonline:9March2012 SpringerScience+BusinessMediaB.V.2012Abstract Touristsareparticularlyvulnerabletonaturaldisasterssuchashurricanessince theymightbelessinformedandpreparedthanresidentsofdisaster-proneareas.Thus, understandinghowthetraitsofatropicalcycloneaswellasspeciccharacteristicsof touristsinuenceaffectiveandcognitiveresponsestoahurricanewarningmessageisa criticalcomponentindisasterplanning.Usingscenariosthatpresentedtropicalcyclones withdifferentrelevantcharacteristics(suchascategoryatlandfall),tourists'knowledge, experiencewithhurricanes,triptraits,andthelocationofthesurvey(coastalorinland), thisstudycontributestotheliteratureonsociologicalissuesrelatedtonaturaldisasters. Thendingssuggestthatriskperceptionsandfearareinuenceddifferentlybythetraitsof thehurricanesandtourists'knowledgeandexperience.Riskisstronglyinuencedbythe projectedcategoryofthehurricaneatlandfall,whilefearisnotassensitivetothis extremelyrelevanttraitofcyclones.Theresultsalsosuggestthattheinuenceofriskand fearonevacuationlikelihoodisstrongandpositive.Thisstudyshowsthevalueofstudying cognitiveandaffectiveresponsestouncertainevents. Keywords Hurricanes Tourists Fear Risk Evacuation Florida J.Villegas( & ) DepartmentofBusinessAdministration,UniversityofIllinoisatSpringeld,OneUniversityPlaza, MSUHB4054,Springeld,IL62703,USA e-mail:jvill2@uis.edu C.Matyas DepartmentofGeography,UniversityofFlorida,Gainesville,FL,USA S.Srinivasan DepartmentofCivilandCoastalEngineering,UniversityofFlorida,Gainesville,FL,USA I.Cahyanto B.Thapa L.Pennington-Gray TourismCrisisManagementInstitute,EricFriedheimTourismInstitute,UniversityofFlorida, Gainesville,FL,USA I.Cahyanto B.Thapa L.Pennington-Gray DepartmentofTourism,Recreation,andSportManagement,UniversityofFlorida, Gainesville,FL,USA123NatHazards(2013)66:97116 DOI10.1007/s11069-012-0119-3

PAGE 2

1Introduction Whenfacedwithanaturaldisaster,touristsasalarge,dynamicpopulationareparticularly atrisksincetheymightnothavetheknowledge,materialresources,orsocialnetworksthat residentsutilizeintheirdecision-makingprocess(Burnsideetal. 2007 ;PerryandGreen 1982 ;MiletiandSorensen 1988 ).Manynaturaldisasterssuchasearthquakesoccurwith littleadvancenotice;however,improvementsinhurricaneforecasts(Rappaportetal. 2009 )allowthoseinpotentiallyaffectedareastoprepareforthestorm'sarrivaland evacuate,ifnecessary.Althoughresidentssubjecttolandfallinghurricanessuchasthose livingalongtheGulfandEastCoastsoftheUSmayknowhowtopreparefortheirarrival, touristsoftenlackafamiliaritywiththeregionaswellaswiththehazardsassociatedwith hurricanes,includingfastwindsandoodingduetostormsurges,and/orheavyrainfall (Rappaport 2000 ).So,itisimportanttounderstandtourists'responsesthatinuencetheir decisionmakingwhileinalocationwhereahurricaneisforecasttomakelandfall. Thetypesofinformationusedtohelppeoplemakedecisionsduringnaturaldisasters suchashurricaneshavebeenextensivelyexplored(DashandGladwin 2007 ).Yet,not surprisingly,thisextensivebodyofresearchhasfocusedonresidentsofareasatrisk, neglectingavulnerablepopulation:touristswhoarevacationinginanareathatmightbe impactedbyahurricane.WiththeexceptionofarecentpaperbyMatyasetal.( 2011 ),the majorbodyofresearchontouristevacuationduringanaturaldisasterhasfocusedmostly onplandevelopmentandthetraitsofthemembersofthetourismindustrymorethanon consumers'reactions(Drabek 1986 1991 1993 1994 1996 ).Thus,muchremainstobe studiedinregardtothedecision-makingprocessesoftouristsduringnaturaldisasters. Researchshowsthatbothcognitiveandaffectiveresponsesinuencethedecisionmakingprocessinuncertainsituations(Loewensteinetal. 2001 ;Slovicetal. 2004 ).In general,cognitiveprocessesareseenasdeliberate,slow,mentalcalculations,while affectiveresponsesareconceptualizedasvisceral,fastreactionstotheenvironment (Loewensteinetal. 2001 ).Researcherswhohavefocusedinthecommunicationofhurricane-relatedwarningshavemostlystudiedanaudience'scognitiveprocessingofwarning messagesanditssubsequentbehavioralintention(DashandGladwin 2007 ).Ingeneral,the responseofanaudiencedependsonthreemajordimensions:(1)hurricanetraitssuchas intensityandprojectedtrack,(2)individual/householddifferences,and(3)currentconditionsorresourcestouseincaseofevacuation.Thesedimensionshavebeenlinkedto intermediateoutcomessuchasriskperceptionandnaloutcomessuchaslikelihoodor actualevacuation(Baker 1991 ;DowandCutter 2000 ;Whiteheadetal. 2000 ).Thecurrent researchfocusesontwointermediateoutcomesthatmighthelpunderstandbehaviorbetter: cognitiveandaffectiveresponses. Althoughsomeresearchersstillfocusexclusivelyoncognitiveelementsofdecision making,thereisanincreasedandrigorouslyvalidatedinterestintheroleofaffectin decisionmaking(Loewensteinetal. 2001 ;LernerandKeltner 2000 ;Slovicetal. 2004 ; Zinn 2008 ).Theresultsinmanydiverseareasofresearchsuggestthataffectcaninuence decisionmaking,insomecasesevenmorestronglythatcognition(Frijda 1986 ;Damasio 1994 ;Isen 2004 ;Izard 2007 ;LeDoux 1996 ;Loewensteinetal. 2001 ;Slovicetal. 2004 ; Zajonc 2008 ). Thecurrentstudyaidstheunderstandingofsociologicalresponsestonaturaldisasters byexploringbothcognitiveandaffectiveresponsesasintermediateoutcomesthatmight helpunderstandbetterthebehavioroftouristsafterreceivinghurricaneinformation. Matyasetal.( 2011 )foundthattouristsusetheirlocation(coastal/inland),hurricanetraits(category),knowledge,andpreviousexperiencewithhurricanes,andother 98NatHazards(2013)66:97116123

PAGE 3

tourism-relatedfactorssuchasmethodoftransportationtoevaluatetheirpersonalrisk levelsandtheirneedtoevacuate,butdidnotexplorewhetheraffectiveresponsesinuence theirevacuationlikelihood.Thiscurrentpaperutilizestheframeworkofriskasfeeling,a modelproposedbyLoewensteinetal.( 2001 )thatintegratesaffectasarelevantelementin decisionmaking,toexplorehowthetraitsofhurricanesandrelevantindividualdifferences (knowledge,personalandvicariousexperience,traveltraits)inuenceatourist'scognitive andaffectiveresponsesandtheirsubsequenteffectonevacuationdecisionmaking.Based onelementsoftherisk-as-feelingmodelthatarerelevanttothestudyofevacuation decisionsafterexposuretohurricanewarningmessages,thegeneralhypothesesare: 1.Hurricanetraits(coastoflandfall,timetolandfall,hurricanecategoryatlandfall,wind duration,andhurricaneprojections)willinuencedifferentlytourists'cognitiveand affectiveresponsestowarningmessages. 2.Tourists'knowledge,previousexperiencewithhurricanes,andtraveltraitswill inuencetourists'cognitiveandaffectiveresponsestowarningmessages. 3.Tourists'location(coastalorinland)willinteractwithhurricanetraits,knowledge, previousexperiencewithhurricanes,andtraveltraitsintheirinuenceoncognitive andaffectiveresponsestowarningmessages. 4.Tourists'cognitiveandaffectiveresponsestohurricanewarningmessageswillhavea positivestrongeffectonlikelihoodtoevacuate. WefocusoureffortsinFlorida,astatethatishosttoover80milliontouristsayear (VisitFlorida 2011 )andthestateintheUSAmostaffectedbytropicalcyclones(Elsner etal. 2004 ).Thedatausedtotestthehypotheseswerecollectedusingstatedpreference surveysthatweregivento235UStouristswhowerevisitingPinellasorOrangecountiesin Floridaduringthesummerof2009.Participantswereexposedtofourdifferentmaps (scenarios)thatdepictedahurricaneforecasttoaffecttheircurrentlocation.Foreach scenariothatparticipantswereexposedto,theyindicatedthelevelofrisk(cognitive response)andfear(affectiveresponse)thattheyperceivedandthelikelihoodtoevacuate (behavioralintention).Thehypothesesweretestedusingordered-responsemodels (McKelveyandZavonia 1975 )whereevacuationlikelihood(behavioralintention),risk (cognitiveresponse),andfear(affectiveresponse)wereusedasdependentvariablesand hurricanetraits,thetourists'experience,knowledge,andtripcharacteristicsasindependent variables.Additionally,totestwhetherrecentpersonalexperiencemighthaveaffectedthe responsesprovidedbythetourists,Chi-squaretestswereperformedaftergrouping responsesaccordingtowhethertheylivedfarfromorclosetotheGulforAtlanticcoastsof theUSorthetrackofatropicalcycloneduringtheprevioushurricaneseason. 2Relevantliterature Thedistinctionbetweendecisionmakingbasedoncognitiveandaffectiveprocesses derivesfromourunderstandingofthedualnatureofthinking.Researchattheneurological (Damasio 1994 ;LeDoux 1996 )aswellaspsychologicallevel(Frijda 1986 ;Isen 2004 ; Izard 2007 ;Zajonc 2008 )suggeststhathumanthinkingistheproductofanexperientialas wellasananalyticalsystem.AccordingtoSlovicetal.( 2004 ),theexperiential(i.e., affective)systemisorientedbypleasure-pain,itisencodedinanarrativemodality,its processingcanbequitefast,andthebehaviorbasedonthissystemismediatedbyaholistic feelingofpastexperiences.Incontrast,theanalytic(i.e.,cognitive)systemisbasedon NatHazards(2013)66:9711699123

PAGE 4

reasonandlogic,itisencodedinwordsandnumbers,usuallyisslowertobeprocessed, andthebehaviorisbasedonthejudgmentofanevent. Mostresearchershavestudiedtheeffectofriskfocusinginthecognitiveprocessingof specicevents(Zinn 2008 ).Inthesecognitive-basedmodels,thecalculationofrisk requiresthementalmultiplicationofseveritymultipliedbylikelihoodofanadverseevent tohappen(Slovicetal. 2004 ).Incontrasttothissystematicapproachtoriskanalysis, recentdevelopmentsacknowledgethatfeelings,denedasvisceralreactionslikefear (Loewensteinetal. 2001 ),playanimportantroleindecisionmaking. Thespecicmechanismsofhowandwhenaffectinuencedirectlybehaviorhavebeen presentedbymanyresearcherssuchasFinucane,etal.( 2000 ),LernerandKeltner( 2000 ), Loewensteinetal.( 2001 ),andSlovic( 1987 ).Althoughtheconceptualizationofwhatis affectmightdifferbetweentheseproposedframeworks,thereisaconsensusthatcognitive evaluationsandaffectivereactionsare(1)inuencedbythenatureoftheevent;(2) inuenceeachother;(3)bothhaveadirect,unmediatedeffectonbehavior;and(4) individualdifferencessuchasgenderorknowledge,andotherfactorssuchasprevious affectivestates(likemoodgeneratedbyanincidentalstimulus)inuencedifferentlyboth oftheseprocesses. Loewensteinetal.( 2001 )developedamodeltoexplainmixedresultsinprevious researchonrisk.Intheirframeworkcalledrisk-as-feelings,''theauthorsindicatethatin precedingresearch,subjectswhoparticipatedinriskexperimentsdidnotmodifytheirrisk calculationseventhoughtheyperceived,cognitively,achangeinprobabilitiesofthe likelihoodofanevent.Thisinsensitivitytoprobabilitiescanbeexplainedbytherisk-asfeelingframeworksinceemotionalreactionstoriskaretriggeredbymentalimagesofthe outcomesofadecision.Forexample,beingattackedbyasharkinFlorida'swatersisan extremelyunlikelyevent,yetitmightelicitastrongmentalimagethatwouldnotbeeasily modiedbyaperceptiblechangeofprobabilityoftheeventtohappen.Astheauthorsstate feelingsoffearorworryinthefaceofdecisionsunderriskoruncertaintyhaveanall-ornonecharacteristic;theymaybesensitivetothepossibilityratherthantheprobabilityof negativeconsequences''(p.276). Recentresearchonaffecthasconsistentlyfoundthatbehaviorisinuenceddirectly, withoutcognitivemediation,byaffectivestatessuchasemotion.Themainreasonforthis inuenceisthataffectivestates'mainobjectiveistomakeanindividualreadytorapidly reacttoenvironmentalchangesinordertosurvive(Zajonc 2008 ).Inameta-analysisonthe effectsofdiscreteemotionsonbehaviorssuchaschoice,reactiontime,andspecic actions,Lenchetal.( 2011 )foundnumerousrelationshipsbetweenaffectivestatesand behavioraloutcomes.Forinstance,aconsiderablenumberofresearcherscontrastedsuccessfullytheeffectofhappinessonbehaviorcomparedtootheremotionslikesadnessand anxiety. Thefeelingasriskframeworkisabletoexplainwhycognitiveandemotionalreactions touncertaineventsdiffer.Accordingtothismodel,cognitiveandaffectiveprocessesare inuenceddifferentlybyvariablesoftheeventandsituation:Whilecognitiveriskconsists ofcalculationsofseverityandprobabilityofoutcome,affect-basedriskisinuencedby vividnessoftheuncertaineventandtimeframeofthedecision.Anindividual'sforecastof howheisgoingtofeelaftertheuncertaineventoccurs,labeledanticipatoryemotioninthis model,istheresultofthementalimageryoftheexperiencethatisforecasted.Themore vividthisimageryis,thehighertheemotionalreaction.Vividnessisinuencedbyindividualdifferencesinmentalimageryability,thedescriptionoftheoutcome,andpersonal orvicariousexperience.Thetimegapbetweenthedecisionanditsoutcomealsoplayan importantdifferencebetweencognitiveandaffectiveappraisalsofrisk.Researchsuggests 100NatHazards(2013)66:97116123

PAGE 5

thatasanindividualapproachesintimeanegativeevent,thelevelsoffearandperceptions ofuncertaintyincrease.Itisimportanttonotethattheevent'sseverityandprobabilitywere notinuencedbytimeproximity(Loewensteinetal. 2001 ). Researchthathasexploredriskintourismsuggeststhattravelersmightdifferintheir processingofriskysituations.Forexample,RoehlandFesenmaier( 1992 )developedrisk dimensionsrangingfromequipment,nancial,physical,psychological,satisfaction,social, andtimethathaveconsistentlyinuencedtraveldecisionsandtravelsatisfaction(e.g., SonmezandGraefe 1998 ;PizamandFleischer 2002 ).Further,theyarguedthatonetourist mightpaymoreattentiontoonedimensionmorethanothers.Forexample,inthecontext ofhurricanerisk,onetouristmayfocusonphysicalrisks(i.e.,beinginjured),whileanother mayfocusonnancialrisk(i.e.,notgettingagoodvalueformoneyspent)forthesame traveldecision.Touristsaversetoriskarelikelytoengageinriskreducingstrategiessuch assearchingextensivelyforinformation(VogtandFesenmaier 1998 )toassistintheir travel-relateddecisions. Theroleofperceivedrisksintravel-relateddecisionmakinghasreceivedsignicant attentionoverthepastdecade(SimpsonandSiguaw 2008 ;Law 2006 ;FuchsandReichel 2011 ),withriskperceptionbeingakeyfactorthatinuencestouristsintheirtravel decisionmaking.Iftouristsperceivethatpotentialrisksmightoverweightbenets,theyare likelytocancelorshortentheirtriptoandinthedestination.ApplyingRogers'( 1983 ) protectionmotivationtheory,SonmezandGraefe( 1998 )foundthattourists'intentionto avoidaparticulardestinationrepresentsprotectivebehaviororriskavoidance.Likewise, Law( 2006 )alsofoundasignicantdifferencebetweenAsianandWesterntravelers,such thatAsiantravelersengagedingreaterriskavoidance.Additionally,tourismscholarshave alsofoundthattouristsnotonlyfeltvaryingdegreesofsafetyduringdifferenttravel activitiesandexperiences,buttheyalsoformaffectiveassociation,thatispositiveor negativefeelingassociatedwithadestination.Thisaffectiveassociationisoftenderived frompersonalexperiencesormediaexposure(Decrop 1999 ),whichinturninuencestheir travel-relateddecisions(SirakayaandWoodside 2005 ). Unfortunately,researchontourists'affectiveandcognitiveprocessesinsettingsrelated tospecicnaturaldisasterslikehurricanesisscarce.However,itisafertilegroundsince travelerswhoareinadifferent,unfamiliarenvironmentwillfeelmoreanxiety,fear,and mistrustduetotheirperceptionsofinstabilityoftheinstitutionsthataretryingtohelpthem (Zinn 2008 )aswellashaveareducedsocialnetwork.Researcherssuggestthatthenetworksofneighbors,friends,andfamilyareimportantsourcesofinformationandhelpfor decisionmakersduringhurricanes(PerryandGreen 1982 ;MiletiandSorensen 1988 ). Further,duetothedynamicnatureoftourism,thisgroupofindividualswillhavevery diverselevelsofimageryvividnessofahurricaneduetotheirheterogeneouslevelsof personalorvicariousexperiencewithhurricanes,knowledge,andmediatedorsocial sourcesofinformation(Matyasetal. 2011 ). Researchershavefoundthattheeffectofpastexperienceshasamixedeffecton evacuationrates(Baker 1991 ;Kusenbachetal. 2010 ;Lindelletal. 2005 ;Rinconetal. 2001 ;Sorensen 2000 ).Atthesametime,naturaldisastercommunicationmodelsconsider knowledgeasaveryimportantelementintheprocessingofnewinformation.Forexample, hurricane-savvyindividualstendtorelymoreontheirownpreviousexperiencesand knowledgethanontherecommendationsofgovernmentsourceswhenmakingevacuation decisions(DowandCutter 2000 ;Tayloretal. 2009 ). Experienceandknowledgeofhurricanesaredirectlylinkedtowherepeoplelive. Inhabitantsofcountiesthathadtoevacuateduetoatropicalsystemdemonstratedifferent levelsofknowledge,sourcesofinformation,andriskperceptionsthanpeoplewholivein NatHazards(2013)66:97116101123

PAGE 6

zonesthatdidnotrequireevacuation(DowandCutter 2000 ).Morespecically,individualsmightevenhavedifferentriskperceptionsbasedonthetypeofdamageexpected (stormsurgevs.wind)intheareawheretheylive(Steinetal. 2010 ).Distancefroman eventseemstoalsohaveaninuenceonaffectiveresponsessinceeventsthatarejudgedas closeareseenasconcrete,whileadistantstimulusisperceivedmoreabstractedly.It appearsthatindividualsrespondemotionallytotheprincipleofdistanceequalssafety'' (WilliamsandBargh 2008 ,p.303). Unfortunately,mostresearchershavenotusedaffectiveaswellascognitiveapproaches tounderstandprocessingofinformationrelatedtohurricanewarnings.Thisstudy'smain objectiveistoexplorehowlocation(coastalvs.inland)oftherespondentsinuencesthe relationshipbetweenhurricanetraits,tourists'knowledgeandexperiencewithhurricanes, andtriptraitswithaffectiveandcognitivevariables.Theseaimswerepursuedinthe presentstudywheredifferenthurricanescenarioswerepresentedtotouristsandtheir affectiveandcognitiveresponsesandbehavioralintentionsweremeasured. 3Surveydesign Themethodologyusedtotestthehypotheseswasastatedpreferencesurvey.Thistypeof surveypresentsaseriesofscenariostoaparticipanttoelicitresponsesthatinclude behavioralintention(inthisstudyevacuation).Sincetheobjectiveofthisstudyisto understandtheresponsesofFloridatouriststohurricanewarningmessages,vesurvey sitesinFloridawereselected.TwoofthesesiteswereinOrangeCounty,aninlandurban areaincentralFloridathathasexperiencedhurricane-forcewindsinthepast,andtheother threesiteswerelocatedonthebeachinPinellas,aGulfofMexicocoastalcountywhere pasthurricaneshavemadelandfall.Figure 1 presentsamapofthegeographiclocationsof thesurveysites.Duetotheincreaseddamageexperiencedalongthecoastwherethefastest windsandstormsurgeoccurduringlandfallandtheissuanceofmandatoryevacuationsin coastalareasduringhurricanes,weanticipatethatcoastaltouristswillindicatehigher levelsoffearandriskthaninlandtourists. InOrlando,themajorcityinOrangeCounty,304surveyswerecollectedattheFlorida MallandWyndhamBonnetCreekResort.PinellasCounty'sinterviewsiteswerethe SheratonSandKeyResort,ClearwaterBeach/Pier60,andSt.PeteBeach.Thenumberof surveyscollectedinthesethreesiteslocatedwithinviewoftheoceanwas144.The surveyswereconductedduringJulyandAugust2009priortotheformationoftherst namedstormoftheAtlanticBasinhurricaneseason. An n thsamplingprocedurewasusedtominimizepotentialbiasassociatedwiththe interceptapproachthatwasappliedinthisstudy.Afterbeingintercepted,respondentswere selectedbyscreeningpotentialparticipantsusingseveralscreeningquestions,basedonour denitionoftourists,toensuretheeligibilityofrespondents.Onlyonepersonfromeach travelpartythatwasinterceptedwaseligibletoanswerthequestionnaire.Ascreening questionwasemployedtoidentifytourists,whichisdenedbytheUSTravelAssociation asthosewhohadtravelled50milesormore,oneway,awayfromhomeorincludingoneor morenightsawayfromhome.Weveriedthedenition'sdistancebyplottingthelocation ofeachzipcodesuppliedbytherespondentwithinArcGISandcalculatingthestraight-line distancebetweenthecoordinatesofthesurveylocationandthoseofthecentroidofthe polygonenclosingthezipcode.Onlyoneadultlledoutthe27-questionquestionnaireper group.Thisquestionnaireincludedfoursections(1)tripcharacteristics(e.g.,modeof transportation,numberofcompanions),(2)afourquestiontesttoassesshurricane 102NatHazards(2013)66:97116123

PAGE 7

knowledgeaswellasquestiontoassesshurricaneexperience,(3)hurricanescenariosand statedpreferencequestions,and(4)demographicinformation. Respondents'levelofknowledgeabouthurricaneswasassessedthroughfourtrue/false questions.Thequestionswere(a)TheHurricaneSeason''inFloridaextendsfromJune1 toNovember30.(b)Itisrareforhurricane-forcewindstoaffectcitiesthatarelocated inlandawayfromthecoast.(c)ACategory1hurricanehastheleastintensityamongall hurricanes.(d)IfaHurricaneWarning''hasbeenissued,itmeansthatyoushould immediatelystartpreparingtoprotectyourselfashurricaneconditionswillbeginwithin 24h.Weusedthenumberofasubject'scorrectanswerstoindicatehurricaneknowledge. Personalorvicariousexperiencewithhurricaneswasassessedbytwoquestionsthatasked whethertherespondent(orfamily/friends)hadbeenaffectedbyahurricaneortyphoon. Wealsotestedforrecentpersonalexperiencewithatropicalcyclonetodetermine whetheratouristlivinginahurricane-pronecoastalregionoftheUSorinaninlandarea affectedbyatropicalcycloneduringtheprevious2008hurricaneseasonexhibiteddifferentresponsesthanthosewhodonotliveintheseareas.Touristswhoresideincoastal regionsarelikelytoreceivemessagesrelatedtopreparationsforhurricanelandfallsonan annualbasisandmorefrequentlyifastormapproachestheirarea(DanielsandLoggins 2007 ),sotheirsurveyresponsesmaydifferfromthoselivingoutsideofcoastalzones.For eachtouristprovidingavalidUSzipcode,theNearfunctioninArcGISwasutilizedto calculatethedistancebetweenthecentroidcoordinatesforthezipcodeandthenearest Fig.1 ResidencelocationsoftouristssurveyedinPinellasorOrangeCounty,FloridaandtheGulfand Atlanticcoastlinesandtropicalcyclonestracksfromthe2008hurricaneseasonfromwhichthedistanceto thetourists'residenceswerecalculated NatHazards(2013)66:97116103123

PAGE 8

pointalongtheUSGulforEastcoastlines.AshapelecontainingthetracksofAtlantic BasintropicalcycloneswasobtainedfromNOAACoastalServices http://csc.noaa.gov/ hurricanes/# andenteredintotheGIS.WeusedtheNearfunctionagaintomeasurethe distancebetweeneachzipcodecentroidandthenearestcyclonetrackfor2008storms.We selectedonlythisseasonasweaimedtomeasurerecentexperience,andfewrespondents arelikelytohaverelocatedtoanotherzipcodewithinthepastyear.Weestablish100km fromthecoastlineand200kmfromthestormtracktodifferentiatethetwogroups.The remnantsofHurricaneIke(2008)(Fig. 1 )broughthurricane-forcewindstotheOhio Valleyandcausedconsiderabledamageineightstateswhere28peoplediedand 2.6millionpeoplelostpower(Brownetal. 2010 );the200kmdistanceissufcientto encompassrespondentsresidingintheareasaffectedbyIke. Baseduponresultsofpreviousliterature,wedeterminedthatvehurricanetraitsshould betestedtodeterminetheircognitiveandaffectiveresponsesinFloridatourists.We developed32scenariostosystematicallymanipulatethevariablesofinterest:landfallcoast (AtlanticorGulf),timetolandfall(24or48h)category(1or4),windduration(shortor long),andtrack(offsetoroversurveysite).Thescenariosconsistedoflarge-sizedgraphics thatpresentedtheforecastedprojectionandothermanipulatedhurricanetraits(SeeFig. 2 foranexampleofascenario)andweredesignedtomimicactualhurricanewarning graphicsissuedbytheNationalHurricaneCenter(Broadetal. 2007 ).Information regardinghurricanecategoryandthedurationofhurricaneconditionswasprintedatthe Fig.2 Exampleofscenariousedinthestudy 104NatHazards(2013)66:97116123

PAGE 9

bottomofthepagesothatrespondentsdidnothavetoinferthisinformationfromthe imageitself. Eachsubjectwasexposedtofouroutofthe32scenarios,andallsubjectsviewed alternatingGulfandAtlanticcoastlandfallswithalternatingoverheadoroffsettracks. Afterviewingeachscenario,subjectsansweredthreequestionstoassessaffectiveand cognitiveresponsesaswellasbehavioralintentionmeasurements.Therstonegaugedthe feltleveloffearbyaskingHowafraiddoyoufeelonseeingthismessage?''Fearwas consideredarelevantaffectiveoutcomesinceresearcherslikeLoewensteinetal.( 2001 ) seethisemotionasanalmostinevitablereactiontouncertainevents.Thefollowing questionwasHowriskydoyouthinkitwouldbeforyoutostayinyourcurrentlocation throughthishurricanescenario?''Thisquestionwasusedasameasureofcognitiveprocessingsinceriskcanbeseenasamentalcalculationofprobabilityandconsequencesofan event(DashandGladwin 2007 ;Loewensteinetal. 2001 )Ave-pointLikert-likescale where1wasnotatalland5wasextremelywasusedforbothquestions.Thenalquestion perscenariomeasuredthelikelihoodofevacuation:Howlikelyareyoutoevacuate/leave yourcurrentlocationatthistimeandunderthissituation?''Ave-pointLikertscale, where1wasveryunlikelyand5wasverylikely,wasutilized. WehypothesizethatcoastlineoflandfallwillbeanimportantfactorforPinellasCounty tourists,butmaynotbeimportantfortouristsinOrangeCounty.However,weexpect higherratingsforcognition(risk)toresultatbothlocationswhentouristsviewscenarios withaCategory4intensity,24-hlandfalltime,trackdirectlyovertheirlocation,andalong duration.Affectiveresponsesmaynotdifferasmuchbetweenthetwochoicesforthese traitsasthosefearingahurricanemayfeelthiswayregardlessofitsintensityorwhether thetrackwillpassdirectlyoverhead. 4Proleofrespondents 434touristsansweredthequestionnaire.Inordertosimplifytheanalysis,thecurrentstudy usedexclusivelythe235responsesfromrespondentswhoindicatedthattheircountryof residencewastheUSA.132ofthequestionnairesusedwerecollectedinOrangeand103in Pinellascounties.TherstcountyisinlandandincludesthecityofOrlando,whilePinellas isacoastalcountythatincludesthecitiesofSt.PetersburgandClearwater.Thedemographicdatacollectedindicatethat46%oftherespondentsweremale,themeanagewas 43.4years(Maximum = 80,Minimum = 18,SD = 14.23),andparticipants'incomeand educationwererelativelyhigh.Intermsofaccommodationsduringthetrip,40%indicated thattheywerestayinginresorts,35%athotels,and14%withfriends.SeeTable 1 fora thoroughdescriptionoftheseresults. Table 2 presentsadetailedreportofthevariables(fear,risk,andevacuationlikelihood) usedinthemodelsdevelopedtoanswerthehypothesesofthestudyaswellastourists' hurricane-relatedknowledge,experience,andtraveltraits.Theresultsofthetablesuggest thatlikelihoodofevacuationandperceivedriskwererelativelyhigh.Whenaskedabout theirlikelihoodtoevacuateandperceivedrisk,42and44%oftherespondents,respectively,indicatedveryhighorhighlevels.However,thelevelsoffearexpressedbythe respondentswerelower:only32%oftouristsindicatedratingsof4or5fortheaffective measure.Intermsofknowledge,amajorityofrespondents,72%,answeredcorrectlythree ormoreofthefourquestionsasked(equivalenttoa75100grade).Approximately,onethirdofthetouristsindicatedpersonalexperiencewithhurricanes.Vicariousexperience, throughfriendsorfamilymembers,wasmentionedmorebytherespondents(38%).In NatHazards(2013)66:97116105123

PAGE 10

termsofthecharacteristicsofthetrip,almosthalfoftherespondentstravelledtothe locationusingtheirpersonalvehicle(48%),whileairtravelingwasthesecondmore popularoption(45%).Amajorityofrespondentsweretravelingwithatleastonecompanionwhowasyoungerthan18yearsold(63%). 5Theeffectofindividualdifferencesonriskperception,fear,andlikelihood ofevacuation Inthepreviousstudybasedonthesedata,Matyasetal.( 2011 )foundthatoneofthemost relevantinuencesonperceivedriskandevacuationlikelihoodwasdestinationlocation Table1 Demographicinformationofrespondents n = 235 Freq% Gender Male10846.0 Female12754.0 Highestlevelofeducation Highschool5623.8 Bachelordegree8636.6 Masterdegree5523.4 Other3514.9 Householdincomefor2008 \ $24,000146.0 $24,000$35,000146.0 $35,000$50,000218.9 $50,000$75,0004619.6 $75,000$100,0003916.6 $100,000$125,0004217.9 Above$125,0004519.1 Ethnicity White17274.3 Black176.7 Hispanic258.2 Asian104.1 NativeAmerican2.9 Mixed62.6 Other31.3 Residence Floridaresidents7732.8 ResidentsotherUSstates15867.2 Accommodations Resort9440.0 Hotel8335.7 Friend'shouse3314.1 Other2510.2 106NatHazards(2013)66:97116123

PAGE 11

(coastalversusinland).Thisresultthatcanbeexplainedbyresearchshowsthathurricanes haveamorenegativeeffectincoastalthaninlandsitessincestormsurgeandsustained windsarestrongerclosertotheshore(Rappaport 2000 ).Therefore,inthecurrentstudy,the ordered-responsemodels(McKelveyandZavonia 1975 )wererunseparatingthedata Table2 Responsedistribution forvariablesusedinthestudyaSubjectsindicatedforeach map(fourintotal)their evacuation,perceivedrisk,and fearassessmentsbDonotknowresponseswere notincludedintheorderedresponsemodels Freq% ResponsesafterexposuretomapsaLikelihoodofevacuation Donotknowb283.0 1 = notatall15916.9 216918.0 318619.8 414515.4 5 = verymuch25427.0 Perceivedrisk Donotknowb192.0 1 = notatall11312.0 217018.0 322223.5 420121.3 5 = verymuch22023.3 Fear Donotknowb202.1 1 = notatall19120.3 220021.3 322523.9 418519.7 5 = verymuch12012.8 Knowledgeandexperience Hurricaneknowledge 0answerscorrect93.8 1177.2 23816.2 39038.3 4answerscorrect8134.5 Personalexperience Yes7531.9 Vicariousexperience Yes9038.3 Travelcharacteristics Methodoftransportation Plane10745.5 Personalvehicle11348.1 Travelingwithpeople \ 18yearsold Yes14662.9 NatHazards(2013)66:97116107123

PAGE 12

obtainedatOrange(inland)andPinellas(coastal)counties.Astheresponsestoquestions aboutrisk,fear,andevacuationwereobtainedonaLikertscale(15),ordered-probit modelswereusedforanalysis.Thismethodology,comparedtoothertypesofregression analysis,recognizestheinherentorderingintheoutcomevariablesofinterestandallowsus tocalculatetheprobabilityofeachlevelofoutcomeasafunctionofexplanatoryfactors.In theordered-probitmodelspresentedintherestofthispaper,apositiveparameterindicates thatthecorrespondingfactorisassociatedwithhigherlevelsofperceivedrisk,fear,or likelihoodofevacuation(dependingontheoutcomebeinganalyzed),andanegative parameterindicatestheoppositeeffect. 5.1Effectofhurricanecharacteristics AccordingtoLoewensteinetal.( 2001 ),affectandcognitionvaryfromeachothersince theyareinuenceddifferentlybytheelementsofanuncertainevent.Thefollowingsetof ordered-responsemodelsexplorestheeffectsofthecharacteristicsofthehurricanesused inthescenariosonrespondents'fearandriskperceptions.Sinceoneofthehypothesisof thisstudyexpectsaninteractioneffectoflocationofthetourist(coastalvs.inland),the datacollectedinthetwositesofthesurvey,PinellasandOrangecounties,wereused separatelytocreatethemodels. InthemodelforriskperceptioninOrangeCounty,onlyonehurricanetraithada signicantimpactonrisk.Thisresultsuggeststhataprojectionthatforecastedahurricane pathdirectlyoverthelocationandastrongerhurricane(category4vs.1)wereperceivedas morerisky.ThemodelforPinellasCountyindicatesthatacategoryfourhurricane,aGulf ofMexicocoastlandfall,andatrackthatpassesoverthesiteofthesurveywereperceived asmorerisky.ThisisanexpectedresultsincePinellasislocatedonthecoastoftheGulfof MexicorenderingitmorevulnerabletoatropicalcycloneapproachingfromGulfthan fromtheAtlantic.Asinthecaseofriskperception,themodelforfearinOrangeCounty, oneparameterwassignicant.Basedonthismodel,abrieftimeofarrival(24vs.48h) triggeredastrongernegativeemotion.ForthecoastalcountyofPinellas,theresults suggestthatacategoryfourhurricaneandaGulflandfallcreatedmorefear. ThemodelspresentedinTable 3 suggestthatsubjectsintheOrangeCountysurveysite perceivedthatacategoryfourhurricaneisriskierthanacategoryone;however,asimilar inuencewasnotfoundintheirfearresponses.AsLoewensteinetal.( 2001 )argue,the characteristicsofuncertaineventsareprocesseddifferentlybythecognitiveandaffective resourcesofindividuals. Themodels'resultsalsoindicatetheimportanceoflocationtounderstandtourists' assessmentoffearandrisk.Asexpected,touristsinPinellasCountyweremoreconcerned whenahurricanewasapproachingfromtheGulfthanfromtheAtlantic.Also,itseemsthat touristsvacationinginPinellasCountyconsideredmoretraitsofthehurricanetodetermine affectiveandcognitiveperceptions.Thisresultreectsresearchthatsuggeststhatindividualswhoaremoremotivated,duetohigherperceivedvulnerability,arewillingto evaluatemoreparameterstomakeadecision(Pettyetal. 1983 ). Thedifferencesbetweentheeffectofthetraitsofthehurricaneoncognitiveand affectiveevaluationsalsoimplythatindividualsmightberiskinsensitivetosmallchanges ofanevent.Oneofthemostsalienttraitsofahurricaneisitscategory;however,this informationwasnotusedbytherespondentsofthestudytodeterminetheirfearinthe inlandcounty.ThisresultseemstofollowLoewensteinetal.( 2001 )resultssuggestingthat thestrengthofthementalimagerythatindividualscreateintheirmindswhentheythink aboutaneventmightberelativelyinsensitivetothespecictraitsoftheevent.Inother 108NatHazards(2013)66:97116123

PAGE 13

words,atouristwhoisexposedtoahurricanewarningmessagemightdevelopvivid images(i.e.,mentalimagesofdisappointedfamilymembers,outdooractivitiesinterrupted fordaysduetorain)regardlessofthespecictraitsofthetropicalcyclone.Thesemental pictureswouldcreateanaffectiveresponsethatmightnotparallelthecognitiverisk perception,whichismoresensitivetothespeciccharacteristicsoftheevent. Table3 Ordered-responsemodelsfortheeffectofhurricanecharacteristicsonperceivedriskandfear Perceivedrisk Orange(Inland)Pinellas(Coastal) Model df 55 p .003.000 Log-likelihoodmodel 214.24 196.62 Log-likelihoodequalsharesmodel 793.45 624.46 Estimate p Estimate p Parameters Timetolandfall.082.394.137.209 S.S.categoryatlandfall.357.000.576.000 Floridacoastoflandfall .081.392.546.000 Windduration .096.316.082.452 Trackpassagerelativetosite.179.050.394.000 Fear Orange(Inland)Pinellas(Coastal) Model df 55 p .163.000 Log-likelihoodmodel 204.48 187.75 Log-likelihoodequalsharesmodel 793.45 624.46 Estimate p Estimate P Parameters Timetolandfall.192.044.154.157 S.S.categoryatlandfall.133.164.402.000 Floridacoastoflandfall .130.169.324.003 Windduration .003.977 .188.082 Trackpassagerelativetosite.078.407.130.228 Valuesusedinmaps Timetolandfall1 = 24h,0 = 48h S.S.categoryatlandfall1 = category4,0 = category1 Floridacoastoflandfall1 = Gulf,0 = Atlantic Windduration1 = long,0 = short Trackpassagerelativetosite1 = through,0 = offset NatHazards(2013)66:97116109123

PAGE 14

5.2Effectoftourists'knowledge,experience,andtriptraits Researchonevacuationbehavioraswellasstudiesthathaveexploredtheaffectasrisk modelsuggestthatindividuals'levelofknowledgeandpreviousexperiencewithanevent aswellassituationalconstrainsinuencetheirresponsestoeventssuchashurricanes.In ordertotesttherelationshipbetweenthesetourists'traitsandtheirfearandriskperceptions,asecondseriesofordered-responsemodelsweredevelopedforeachcounty. AsTable 4 suggests,personalexperiencewithhurricaneshadanegativeeffectonthe fearandperceivedriskfortouristsinbothinlandandcoastallocations.Theseresults indicatethatsometouristspersonallyaffectedbyhurricanesshowedlowerperceivedlevels ofriskandfearthanthosewithoutpriorexperiencewithtropicalcyclones.Itisimportant tonoticethattheseverityofpreviouslyexperiencedhurricaneswasnotexploredinthis study.However,Matyasetal.( 2011 )foundthathavingpersonalexperiencewithhurricaneswasassociatedwithlowerratingsoflikelihoodtoevacuateaswellaslowerratings ofperceivedrisk.Thesendingscoincidewithempiricalresultsfromresearcherssuchas Rinconetal.( 2001 )andKusenbachetal.( 2010 ),whofoundanegativerelationship betweenpreviousexperienceandevacuationrates. Anotherinterestingndingisthedifferentresultsoftheroleofknowledgeofhurricanes amongtouristsinterviewedatthecoastalorinlandsites.InthecaseofOrangeCounty, knowledgeabouthurricaneshadnosignicantimpactonfearorrisk.Thisresultsuggests thatotherfactorssuchaspersonalexperiencewithprevioushurricanesplayedastronger roleinmodulatingrespondents'cognitiveandaffectiveresponsestothehurricanescenarios.Incontrast,touristsinPinellasCountyshowedastrongpositiverelationship betweenknowledgeandthecognitivejudgment,riskperception.Theseresultsshowthat inlandtouristsuselesscognitiveresourcesascomparedtocoastaltouriststodetermine theirresponsetoahurricanewarningmessage.Intheliteratureonpersuasion,aconsistent ndingsuggeststhatsubjectswithalowlevelofinvolvementwithasituationormessage useslesscognitiveresources(Pettyetal. 1983 ).Theresultsofthepresentstudyshowthat knowledgeabouthurricanesdidnothaveasignicantassociationwithleveloffearfor PinellasCountytourists.Thisresultsuggeststhatanindividual'scognitiveresourcesare relevantforthosedirectlyinharm'swayonthecoast. Theresultsofthesemodelsalsoindicatethatmethodoftransportationdidnotinuence riskandfearperceptionsexceptfortouristsinterviewedinthecoastalsitesofthesurvey. Apparently,travellingusingapersonalvehiclereducestheleveloffear.Perhapstourists whocancontroltheirvacationplans(comparedtoindividualswhoneedtocontactan airlinetochangedatesofdeparture)mighthavealessnegativeimageryofwhatwould happeniftheyhadtoevacuatethecoastalcity.Interestingly,thepresenceofcompanions lessthan18yearoldwasasignicantandpositiveinuenceonriskandfearinOrange countymodels.AccordingtoDashandGladwin( 2007 ),factorssuchaschildrenina householdinuencedifferentlythedecision-makingprocessdependingonothervariables likeresourcesavailable.Futureresearchshouldinvestigatethereasonsbehindthelackof aneffectofchildrenpresenceonriskandfearamongtouristsincoastallocations. 5.3Effectofdistanceofresidencetothecoastandprevioushurricanetracks Sinceabriefquestionnairedoesnotallowforacomprehensiveassessmentofrespondents' knowledgeandexperiencewithhurricanes,calculationsofdistanceoftheresidenceofthe respondentstothecoastandtoprevioushurricanetrackswereusedasproxiesofprevious experiencesandexposuretohurricanewarningmessages. 110NatHazards(2013)66:97116123

PAGE 15

Table4 Ordered-responsemodelsforknowledge,personalandvicariousexperience,andtraveltraitson perceivedriskandfear Perceivedrisk Orange(Inland)Pinellas(Coastal) Model df 66 p .030.031 Log-likelihoodmodel 384.12 273.63 Log-likelihoodequalsharesmodel 793.45 624.46 Estimate p Estimate p Parameters Knowledgeabouthurricanes .215.249.587.013 Personallyaffectedbyhurricane .211.037 .247.050 Family/friendsaffectedbyhurricane.006.267.010.110 Travelbyplane.029.861.014.826 Travelbypersonalvehicle.060.721 .016.806 Travelersyoungerthan18.268.011.069.530 Fear Orange(Inland)Pinellas(Coastal) Model df 66 p .000.000 Log-likelihoodmodel 288.64 281.23 Log-likelihoodequalsharesmodel 793.45 624.46 Estimate p Estimate p Parameters Knowledgeabouthurricanes .095.249.346.139 Personallyaffectedbyhurricane .446.037 .253.045 Family/friendsaffectedbyhurricane.006.237.005.376 Travelbyplane .141.393 .341.072 Travelbypersonalvehicle.005.974 .540.005 Travelersyoungerthan18.296.005.072.510 Variablevalues Knowledgeabouthurricanes100 = 4responsescorrect,0 = noresponsescorrect. Personallyaffectedbyhurricane1 = yes,0 = no Family/friendsaffectedbyhurricane1 = yes,0 = no Travelbyplane1 = yes,0 = no Travelbypersonalvehicle1 = yes,0 =no Travelersyoungerthan181 = yes,0 = no NatHazards(2013)66:97116111123

PAGE 16

Responsesfrom188touristswhoprovidedvalidUSzipcodeswereincludedinthis analysis.Inordertoanalyzethesedata,theresponsesweregroupedaccordingtotheir distancefromtheGulforAtlanticCoastsanddistancefromastormtrackfromthe2008 hurricaneseason.The2008stormtrackswereselectedforanalysisbecausetheyarefrom themostrecentseasonpriortotheadministrationofthesurveyssinceresearchofthe effectsoftimeinindividuals'memorysuggeststhatpeopletendtobemoresensitiveto morecurrentevents(BarronandErev 2003 ).Also,fewtouristsarelikelytohavemoved withinthepastyear.ThreehurricanesandthreetropicalstormsmadelandfallintheUS duringthe2008season.Figure 1 presentsresidencelocationsoftouristssurveyedin PinellasandOrangeCountyandtropicalcyclonestracksfromthe2008hurricaneseason fromwhichthedistancetothetourists'residenceswerecalculated. Themeandistancetocoastwas122km,andmeandistancetotrackwas126km. Therefore,twogroupswereformed:thosewithin100kmofthecoastortrack,andthose morethan200kmaway.Thissplitofthesamplebetweenrespondentswasdesignedto separatethosewhoaremorelikelytohavebeenexposedtohurricane-relatedmessages and,therefore,moreknowledgeable(respondentswhoseresidence'szipcodeiswithin 100kmofthecoastlineorarecentstormtrack)fromrespondentswhowouldbelesslikely tobeexposedtothistypeofmessagesandexperiences(morethan200kmfromthecoast ora2008track).ResponseswerealsoseparatedaccordingtothesurveylocationinOrange orPinellasCounty.Chi-squaretestswereperformedforfearandriskevaluation.Table 5 presentsthe p valuesfortheChi-squaretests.Thetablealsoreportswhetherthetourists residingwithin100kmhadhigherorlowerratingsthanexpectedfor p valuesthatare \ 0.100. TouristsvisitingOrlandoOrangeCountywhoresidewithin100kmoftheGulforEast AtlanticCoastsorwithin100kmofatropicalcyclonetrackfromthe2008hurricane seasonhadlowerfear,andrisk,ratingsthanthoselivingmorethan200kmfromthecoast ora2008stormtrackin2008.Whenexaminingactualversuspredictedratings,those livingclosewithin100kmindicatedmoreresponsesof1(nofearatall,notatallrisky) thanexpected.So,thissuggestsforOrlandoinlandtouriststhatwhentheyarenotas exposedtotropicalcyclonesathome,theyaremorelikelytohavesomefearanddetect somerisk.Ontheotherhand,Orlandoinlandtouristswholivenearthecoastorastorm trackindicatedlowerratingspossiblybecausetheybelievethattheyareatlessriskbeing inlandand/orcouldbelessafraidbecausetheyareusedtothesestorms(Arlikattietal. 2006 ). ResultsdifferforPinellasCountytourists.Touristsresidingwithin100kmoftheGulf orEastAtlanticCoastshadhigherratingsforfearandrisk,andthosewithin100kmofa trackfromthe2008hurricaneseasonhadhigherratingsoffearthanthoselivingmorethan 200kmawayfromthecoastorastormtrack.Thoselivingwithin100kmofthecoastora Table5 Chi-squareresultsofcoastaldistanceandhurricanetrackeffectonperceivedriskandfear OrangeCountyPinellasCounty Chi-squareWithin100kmChi-squareWithin100km Coastaldistance 9 fear0.051Lowerratings0.023Higherratings Coastaldistance 9 risk0.032Lowerratings0.084Higherratings Track2008distance 9 fear0.013Lowerratings0.059Higherratings Track2008distance 9 risk0.000Lowerratings0.456 112NatHazards(2013)66:97116123

PAGE 17

trackhadfewerratingsof1andmoreratingsof4or5thanexpected.Thoselivingwithin 100kmofthecoastoratrackmayhavemorefearbecausetheyhavemoreexperienceand astheircurrentlocationisonthebeach,sotheyrealizethatthescenariosposeathreat, regardlessofspecicslikeintensity,timetolandfall,orwhetherthecenterlineoftheTrack ForecastConepassesdirectlyovertheirlocationratherthanbeingoffset. 5.4Relationshipofcognitiveandaffectiveresponsestoevacuationlikelihood Thestudyofhowcognitiveandaffectiveresponsesareinuencedbyindividualand environmentaltraitsrequirestheinclusionofalinktoabehavior-relatedtoadecisionmakingprocess.Inthisstudy,twoordered-responsemodels,onepercounty,designedto testtheeffectoffearandperceivedriskonevacuationlikelihoodwereanalyzed.These modelssuggestthattheserelationshipsareconsistentinthetwosamplingsites(See Table 6 ).Forbothinlandandcoastalsurveylocations,bothparametersweresignicantat a = 0.0001.Interestingly,theperceivedriskparametersonbothmodelsareapparently strongerthanfear'sparameters.Thisresultmightbeexplainedbyresearchthathasconsistentlyfoundthatindividualswhoneedtomakeimportantdecisionsarehighlymotivated toarrivetoacorrectdecisionand,therefore,relymoreoncognitivethanaffective information(Pettyetal. 1983 ;SvensonandSalo 2007 ). 6Conclusionsandfutureresearch Landfallingtropicalcyclonescanimpactsocietybefore,during,andafterthestorm.This study'skeyobjectivewastoaddressthepaucityofresearchonhowtouristsrespond cognitivelyandeffectivelytohurricanewarningmessages.Morespecically,theinuence offear,anaffectiveresponse,andriskperception,acognitiveresponse,onevacuation likelihoodoftouristsintwodifferentdestinations(coastalandinland)wasexplored.The resultssuggestthatevacuationlikelihoodisstronglyinuencedbycognitiveandaffective responsesamonginlandaswellascoastaltourists.Atthesametime,fearandriskperceptionwereinuenceddifferentlybyhurricanetraits,andindividualdifferencesamong touristslikeknowledgeandpreviewsorvicariousexperiencewithhurricanes. Table6 Ordered-responsemodelsfortheeffectofriskperceptionandfearonevacuationlikelihood EvacuationLikelihood Orange(Inland)Pinellas(Coastal) Model df 22 p .0001.0001 Log-likelihoodmodel 203.30 161.05 Log-likelihoodequalsharesmodel 793.45 624.45 Estimate p Estimate p Parameters Perceivedrisk.810.0001.04.000 Fear.385.000.177.003 NatHazards(2013)66:97116113123

PAGE 18

Eventhoughthisresearchprovidesaninitialsteptowardunderstandinghowaffective andcognitiveresponsestohurricanewarningmessagesareutilizedbytouriststodetermine theirlikelihoodofevacuation,otherimportantissuesstillremain.Forexample,itis importanttonotethatinthisstudy,simpledichotomical(yes/no)questionswereaskedin regardtoexperiencewithhurricanes.Previousexperiencewithhurricanesisacomplex issuesinceresearchershavefoundaverydiversenumberofcognitiveandbehavioral consequences.Forexample,DowandCutter( 1998 )foundthattheeffectsoffalsehurricanealarmsamongSouthCarolinaresidentsdidnotchangetheirevacuationdecisions(a behavioralmeasure),butreducedthecondenceinofcialsourcesofinformation(a cognitivemeasure).Recently,Dillonetal.( 2011 )suggestedthatpeoplewhowerenot affected,becauseofchance,byanegativeeventsuchasahurricanetendtoreducetheir perceptionsofrisk.Also,DashandMorrow( 2001 )foundthatavicariousexperience throughmediamightinuencemorebehavioralintentionthanapreviouspersonal experience. Onepossiblelimitationofthisstudyisself-selectionofitsparticipants.Thereisthe possibilitythatindividualswhoaremoreworriedorknowledgeableabouthurricanesmight avoidFlorida,aswellasotherhurricane-pronedestinations,duringthemonthswhen tropicalcyclonesaremorefrequent.FutureresearchshouldassessknowledgeandattitudinaldifferencesbetweencurrentandpotentialtouristsofAtlanticandGulfcoaststatesor comparetheresponsesoftouriststhatvisitthesestatesindifferentmonthsoftheyear. Anotherpossiblefuturestepwouldbetoexaminehowtouristsobtaininformationwhen anemergencypresentsitself.MiletiandSorensen( 1988 )suggestthatmedia,emergency ofcials,andanindividual'ssocialnetwork(friends,neighbors,andfamilymembers)are importantsourcesofinformationduringevacuationdecisionmaking.Theprocessof informationgatheringandsharingofmultiplesourceshelpsindividualstodevelopa personalizedandrelevantriskassessment.Eachofthesesourceswouldhavedifferent effectsonbehaviorsincethedeliveryofthemessagevariesintermsofmodality(visualor verbal),hencevividness,(Burnsideetal. 2007 )andbias(Slovic 1987 ).Theseissuesareof extremeimportancesincevividnessofaneventstronglyinuencesaffectiveriskevaluations(Loewensteinetal. 2001 )andtouristsinanon-familiarenvironmentdonotcount withthelargesocialnetworkthattheymightuseattheirresidence(PhillipsandMorrow 2007 ).Acknowledgments WethanktheEricFriedheimFoundationfortheirnancialcontributionthatallowed ustoconductthisstudy.Wealsothanktheteamofstudentassistantswhohelpedtoadministerthesurveys.ReferencesArlikattiS,LindellMK,PraterCS,ZhangY(2006)Riskareaaccuracyandhurricaneevacuationexpectationsofcoastalresidents.EnvironBehav38:226247 BakerEJ(1991)Hurricaneevacuationbehavior.IntJMassEmergDis9(2):287310 BarronG,ErevI(2003)Feedback-baseddecisionsandtheirlimitedcorrespondencetodescription-based decisions.JBehavlDecisMaking16:215233 BroadK,LeiserowitzA,WeinkleJ,SteketeeM(2007)Misinterpretationsoftheconeofuncertainty''in Floridaduringthe2004hurricaneseason.BullAmMeteorolSoc88(5):651667.doi: 10.1175/bams88-5-651 BrownDP,BevenJL,FranklinES,BlakeES(2010)Atlantichurricaneseasonof2008.MonWeatherRev 138(5):19752001 BurnsideR,MillerDS,RiveraJD(2007)Theimpactofinformationandriskperceptiononthehurricane evacuationdecision-makingofgreaterNewOrleansresidents.SociolSpect27(6):727740.doi: 10.1080/02732170701534226 114NatHazards(2013)66:97116123

PAGE 19

DamasioAR(1994)Descartes'error:emotion,reason,andthehumanbrain.Putnam,NewYork DanielsRJ,LogginsGM(2007)Conceptualizingcontinuouscoverage:astrategicmodelforwall-to-wall localtelevisionweatherbroadcasts.JApplCommunRes35(1):4866 DashN,GladwinH(2007)Evacuationdecisionmakingandbehavioralresponses:individualandhousehold. NatHazardsRev8(3):6977 DashN,MorrowBH(2001)Returndelaysandevacuationordercompliance:thecaseofHurricaneGeorges andtheFloridaKeys.EnvironHazards2:119128 DecropA(1999)Tourists'decision-makingandbehaviorprocesses.In:PizamA,MansfeldY(eds) Consumerbehaviorintravelandtourism.HaworthPress,London,pp103133 DillonRL,TinsleyCH,CroninM(2011)Whynear-misseventscandecreaseanindividual'sprotective responsetohurricanes.RiskAnal31(3):440449 DowK,CutterSL(1998)Cryingwolf:repeatresponsestohurricaneevacuationorders.CoastManage 26:237252 DowK,CutterSL(2000)Publicordersandpersonalopinions:householdstrategiesforhurricanerisk assessment.GlobalEnvironChangeBEnvironHazards2(4):143155 DrabekTE(1986)Humansystemresponsestodisaster.Springer,NewYork DrabekTE(1991)Anticipatingorganizationalevacuations:disasterplanningbymanagersoftourist-orientedprivaterms.IntJMassEmergDisasters9:219245 DrabekTE(1993)Variationsindisasterevacuationbehavior:publicresponsesversusprivatesector executivedecision-makingprocesses.Disasters16:104118 DrabekTE(1994)Disasterevacuationandthetouristindustry.In:Environmentandbehaviormonograph no.57.NaturalHazardsResearchandApplicationInformationCenter,Boulder,CO,p282 DrabekTE(1996)Disasterevacuationbehavior:touristsandothertransients.In:Environmentandbehavior monographno.58.NaturalHazardsResearchandApplicationInformationCenter,Boulder,CO,p354 ElsnerJB,NiuXF,JaggerTH(2004)DetectingshiftsinhurricaneratesusingaMarkovchainMonteCarlo approach.JClim17:26522666 FinucaneML,AlhakamiA,SlovicP,JohnsonSM(2000)Theaffectheuristicinjudgmentsofrisksand benets.JBehavDecisMak13(1):117.doi: 10.1002/(sici)1099-0771(200001/03)13:1 \ 1:aidbdm333 [ 3.0.co;2-s FrijdaNH(1986)Theemotions.CambridgeUniversityPress,Cambridge FuchsG,ReichelA(2011)Anexploratoryinquiryintodestinationriskperceptionsandriskreduction strategiesofrsttimevs.repeatvisitorstoahighlyvolatiledestination.TourManag32:266276 IsenAM(2004)Someperspectivesonpositivefeelingsandemotions:positiveaffectfacilitatesthinkingand problemsolving.In:MansteadASR,FrijdaN,FischerA(eds)Feelingsandemotions:theAmsterdam symposium.Studiesinemotionandsocialinteraction.CambridgeUniversityPress,NewYork, pp263281 IzardCE(2007)Basicemotions,naturalkinds,emotionschemas,andanewparadigm.PerspectPsycholSci 2(3):260280.doi: 10.1111/j.1745-6916.2007.00044.x KusenbachM,SimmsJL,TobinGA(2010)Disastervulnerabilityandevacuationreadiness:coastalmobile homeresidentsinFlorida.NatHazards52(1):7995.doi: 10.1007/s11069-009-9358-3 LawR(2006)Theperceivedimpactofrisksontraveldecisions.IntJTourRes8(4):289300 LeDouxJE(1996)Theemotionalbrain.Simon&Schuster,NewYork LenchHC,FloresSA,BenchSW(2011)Discreteemotionspredictchangesincognition,judgment, experience,behavior,andphysiology:ameta-analysisofexperimentalemotionelicitations.Psychol Bull(Suppl).doi: 10.1037/a0024244 LernerJS,KeltnerD(2000)Beyondvalence:towardamodelofemotion-specicinuencesonjudgement andchoice.CognEmot14(4):473493.doi: 10.1080/026999300402763 LindellMK,LuJC,PraterCS(2005)HouseholddecisionmakingandevacuationinresponsetoHurricane Lili.NatHazardsRev6:171179 LoewensteinGF,WeberEU,HseeCK,WelchN(2001)Riskasfeelings.PsycholBull127(2):267286.doi: 10.1037/0033-2909.127.2.267 MatyasC,SrinivasanS,CahyantoI,ThapaB,Pennington-GrayL,VillegasJ(2011)Riskperceptionand evacuationdecisionsofFloridatouristsunderhurricanethreats:astatedpreferenceanalysis.Nat Hazards59(2):871890.doi: 10.1007/s11069-011-9801-0 McKelveyRD,ZavoniaW(1975)Astatisticalmodelfortheanalysisofordinalleveldependentvariables. JMathSociol4:103120 MiletiDS,SorensenJH(1988)Planningandimplementingwarningsystems.In:LystadM(ed)Mental healthresponsetomassemergencies:theoryandpractice.Brunner/Mazelpsychosocialstressseries; no.12.Brunner/Mazel,Philadelphia,PA,USA,pp321345 NatHazards(2013)66:97116115123

PAGE 20

PerryRW,GreenMR(1982)TheRoleofethnicityintheemergencydecision-makingprocess.SociolInq 52(4):306334 PettyRE,CacioppoJT,SchumannDW(1983)Centralandperipheralroutestoadvertisingeffectiveness:the moderatingroleofinvolvement.JConsumerRes10:134148 PhillipsBD,MorrowBH(2007)Socialscienceresearchneeds:focusonvulnerablepopulations,forecasting, andwarnings.NatHazardsRev8(3):6168 PizamA,FleischerA(2002)Severityversusfrequencyofactsofterrorism:whichhasalargerimpacton tourismdemand?JTravelRes40(3):337339 RappaportEN(2000)LossoflifeintheUnitedStatesassociatedwithrecentAtlantictropicalcyclones.Bull AmMeteorolSoc81:20652073 RappaportENetal(2009)AdvancesandchallengesattheNationalHurricaneCenter.WeatherForecast 24:395419 RinconE,LinaresMYR,GreenbergB(2001)Effectofpreviousexperienceofahurricaneonpreparedness forfuturehurricanes.AmJEmergMed19(4):276279 RoehlW,FesenmaierD(1992)Riskperceptionsandpleasuretravel:anexploratoryanalysis.JTravelRes 30(4):1726 RogersRW(1983)Cognitiveandphysiologicalprocessinfearappealsandattitudechange:arevisedtheory ofprotectionmotivation.In:CacioppoJ,PettyR(eds)Socialpsychophysiology.Guilford,NewYork, pp153174 SimpsonPM,SiguawJA(2008)Perceivedtravelrisks:thetravelerperspectiveandmanageability.IntJ TourRes10:315327 SirakayaE,WoodsideAG(2005)Buildingandtestingtheoriesofdecisionmakingbytravelers.Tour Manage26:815832 SlovicP(1987)Perceptionofrisk.Science236(4799):280285 SlovicP,FinucaneML,PetersE,MacGregorDG(2004)Riskasanalysisandriskasfeelings:somethoughts aboutaffect,reason,risk,andrationality.RiskAnalIntJ24(2):311322.doi: 10.1111/j.0272-4332. 2004.00433.x SonmezS,GraefeA(1998)Determiningfuturetravelbehaviorfrompasttravelexperienceandperceptions ofriskandsafety.JTravelRes37(2):171177 SorensenJH(2000)Hazardwarningsystems:reviewof20yearsofprogress.NatHazardsRev 1(2):119125 SteinRM,Duenas-OsorioL,SubramanianD(2010)Whoevacuateswhenhurricanesapproach?Theroleof risk,information,andlocation.SocSciQ91(3):816834 SvensonO,SaloI(2007)Mentalrepresentationsofimportantreal-lifedecisions.EurJOperRes 177(3):13531362.doi: 10.1016/j.ejor.2005.04.007 TaylorK,PriestS,SiscoHF,BanningS,CampbellK(2009)ReadinghurricaneKatrina:information sourcesanddecision-makinginresponsetoanaturaldisaster.SocEpistemol23(3/4):361380.doi: 10.1080/02691720903374034 VisitFlorida(2011)20102011annualreport.VisitFlorida,Talahassee,FL VogtCA,FesenmaierDR(1998)Expandingthefunctionalinformationsearchmodel.AnnTourRes 25:551578 WhiteheadJC,EdwardsB,VanWilligenM,MaioloJR,WilsonK,SmithKT(2000)Headingforhigher ground:factorsaffectingrealandhypotheticalhurricaneevacuationbehavior.EnvironHazards 2:133142 WilliamsLE,BarghJA(2008)Keepingone'sdistance:theinuenceofspatialdistancecuesonaffectand evaluation.PsycholSci19(3):302308.doi: 10.1111/j.1467-9280.2008.02084.x ZajoncRB(2008)Feelingandthinking:preferencesneednoinferences.In:FazioRH,PettyRE(eds) Attitudes:theirstructure,function,andconsequences.Keyreadingsinsocialpsychology.Psychology Press,NewYork,pp143168 ZinnJO(2008)Headingintotheunknown:everydaystrategiesformanagingriskanduncertainty.Health RiskSoc10(5):439450.doi: 10.1080/13698570802380891 116NatHazards(2013)66:97116123