CREATE Research Archive

Published Articles & Papers

5-2011

Scenario Simulation Group Reactions to theAftermath of the Great ShakeOut Magnitude 7.8EarthquakeHeather RosoffUniversity of Southern California, rosoff@usc.edu

Richard JohnUniversity of Southern California, richardj@usc.edu

William J. BurnsCal State San Marcos, bburns@csusm.edu

Isaac MayaUniversity of Southern California, imaya@usc.edu

Follow this and additional works at: http://research.create.usc.edu/published_papersPart of the Experimental Analysis of Behavior Commons, and the Industrial and Organizational

Psychology Commons

This Article is brought to you for free and open access by CREATE Research Archive. It has been accepted for inclusion in Published Articles & Papersby an authorized administrator of CREATE Research Archive. For more information, please contact gribben@usc.edu.

Recommended CitationRosoff, Heather; John, Richard; Burns, William J.; and Maya, Isaac, "Scenario Simulation Group Reactions to the Aftermath of theGreat ShakeOut Magnitude 7.8 Earthquake" (2011). Published Articles & Papers. Paper 176.http://research.create.usc.edu/published_papers/176

Scenario Simulation Group Reactions tothe Aftermath of the Great ShakeOutMagnitude 7.8 Earthquake

Heather Rosoff,a) Richard John,b) William J. Burns,c) and Isaac Mayad)

This paper demonstrates an innovative approach for learning about earth-quake victims’ behavioral responses to an emergency situation in the immediateaftermath of an earthquake. Researchers developed a scenario following a mag-nitude 7.8 earthquake that leads to escalating complications described over eightepisodes. Subjects were assigned to scenario simulation groups (SSG) andinstructed to discuss how they would cope with problems as if they were experi-encing the scenario. Subjects first discussed their reactions and potential deci-sions they might have to make as a group. They were then asked to individuallyrecord their behavioral intentions, cognitive reactions (concern) and emotionalstate (fear) in a survey instrument. Subjects’ responses were tracked over theeight episodes of the scenario. The SSG methodology yielded a more realisticunderstanding of how a respondent’s reactions and behavior change in the im-mediate aftermath of an earthquake. The implications of the SSG approach ondisaster preparedness and response are discussed. [DOI: 10.1193/1.3574450]

INTRODUCTION

Earthquake mitigation, response, and recovery have been an important topic of scientificinquiry as researchers seek to improve disaster management. Researchers have identifiedand addressed challenges posed by earthquakes that span all five stages of disaster—planning, crisis communication, response, relief, and recovery. Some studies have assessedhow earthquake planning challenges are greatly influenced by the extent of involvement oflocal, state, and federal organizations (May 1991, Berke and Beatley 1992, Godschalk et al.1999, Nelson and French 2002) and the general public in at-risk seismic regions (Kartezand Lindell 1990, Pearce 2004). Other studies have focused on the immediate aftermath ofan earthquake, where the effectiveness of coordination among emergency responders (bothlocal and national) in providing response services and restoring order is of greatest concern(Drabek 1985, Tierney 1985, Fiedrich et al. 2000; Saab et al. 2008). Lastly, research hasfocused on whether local victims of disaster are amply supplied with aid, services, and

a)University of Southern California, National Center for Risk and Economic Analysis of Terrorism Events, 3710McClintock, RTH 322, Los Angeles, CA 90089-2902

b)University of Southern California, Department of Psychology MC-1061, 3620 South McClintock Ave., SGM621, Los Angeles, CA 90089-1061

c)Decision Research, 1201 Oak Street, Suite 200, Eugene, OR 97401d)University of Southern California, National Center for Risk and Economic Analysis of Terrorism Events, 3710

McClintock, RTH 308, Los Angeles, CA 90089-2902

597

Earthquake Spectra, Volume 27, No. 2, pages 597–614, May 2011; VC 2011, Earthquake Engineering Research Institute

goods (Berke et al. 1993, Fiedrich et al. 2000) and how these disaster cost consequences aredistributed across the nation (Brookshire et al. 1997, Okuyama and Chang 2004).

Disaster response officials rely on study findings to guide the implementation of techni-cal, economic, and political improvements to earthquake preparedness and response efforts.Unfortunately, these programs have not provided solutions that optimize disaster responseand maintain ongoing readiness of at-risk communities. The remedies proposed have beenineffectual, in part because they fail to account for variations in human behavioral responseand decision making during crisis situations.

The decision making of emergency responders is instrumental to the successful imple-mentation of the preparedness and response infrastructure developed by researchers and pol-icy makers. Often, in the aftermath of a catastrophic earthquake, the technical and resourceassets of policy programs are temporarily unavailable or destroyed. Human emotion, prob-lem solving, and behavior then become paramount in determining loss of life. It is the stra-tegic thinking of first responders, the managerial skills of local officials, and the generalreadiness of response officials, volunteers, and survivors that collectively determine theeffectiveness and speed of disaster recovery efforts (Kreps and Bosworth 2006).

Furthermore, the consideration of human behavior extends beyond that which isrequired for disaster response. Following an earthquake, situations arise that researchersand response officials cannot anticipate. Delays in the delivery of medical supplies orthe inability to seek help (e.g., due to being trapped in a building or communicationstower damage) are among some of the reasons why victims may need to rely on theirown resourcefulness in the first hours or days of an emergency. The ingenuity and de-cision making of victims is an important part of understanding what to expect and howto manage disaster situations.

Social scientists’ investigations into the behavioral aspects of earthquakes have focused onthe relief and recovery stage of the disaster event, with much of the research and policy-mak-ing emphasis placed on treating trauma and post-traumatic stress (Goejian et al. 2000, Yehuda2002). More recently, integrating behavioral science research into preparedness and responseefforts has been proposed for improving earthquake management in all the major stages of adisaster. Researchers have since studied how experience with earthquakes, level of education,and personal earthquake preparedness impact disaster response for residents of seismic riskzones (Eisenman et al. 2006, Heller et al. 2006) and influence perceptions of earthquake risk(defined in terms of event likelihood and consequence) (Slovic et al. 1981, Palm and Hodgson1992, Shaw et al. 2004). Following an earthquake, researchers have directed their efforts to-ward evaluating the public response to hazard information and risk communication (Glass2001, Eisenman et al. 2007); the public response and communication of both victims and non-victims in the weeks and months postdisaster (Mileti and Fitzpatrick 1992, Glass 2001); andthe victim’s household, workplace, and personal preparedness adjustments (Palm and Hodg-son 1992, Lindell and Perry 2000).

Less research attention has been paid to individual choices and social networking bothduring and in the immediate aftermath of earthquakes. Little is known about the actionstaken by victims during and immediately following shaking, how they would attempt tocommunicate with emergency responders, or what decisions they might make when

ROSOFF ETAL.598

determining how to remove themselves and others from harm’s way. Research on socialnetworking during and immediately following a disaster has primarily addressed theimprovisations and coordination undertaken by emergency responders (Tierney 2002,Wachtendorf and Kendra 2005, Kendra and Wachtendorf 2006, Drabek 2010). Research onevent victims has mainly focused on their altruistic nature toward helping themselves andothers, and how this emotional and behavioral response runs a quick course over the firstfew days following an event (Dynes 1994, Tierney et al. 2001, Kreps and Bosworth 2006).Knowledge of whether the same innovativeness of emergency responders and philanthropicnature of victims holds true for people struggling through emergency situations in the im-mediate aftermath of a disaster is unknown.

We propose that there is more to be learned about the individual responses of earthquakevictims and social networking among them in the immediate hours following a disaster event.In this paper, we describe an innovative approach for studying how a person’s individual reac-tions and a group’s collective reactions evolve during an escalating emergency following amagnitude 7.8 earthquake. The unfolding scenario is presented in discrete episodes, allowingus to track responses over time. Of particular interest is characterizing changes in emotions,concerns, and actions as the scenario unfolds over a several-hour period.

The impetus for this study was California’s 2008 Golden Guardian Emergency Prepar-edness Exercise—an earthquake simulation event designed to test the state’s ability to pre-vent, respond to, and recover from a terrorist attack or catastrophic natural disaster. Wedesigned a simulation exercise that narrates through an escalating emergency on a collegecampus in the aftermath of a 7.8 earthquake. The scenario progresses over eight delineatedepisodes. Subjects were randomly assigned to one of two scenario simulation groups(SSGs) of college students to keep the group size to about 12 students each. All subjectswere then instructed to imagine they were students in a classroom when the earthquakeoccured and told to role-play their “real-time” reactions to each disaster episode. Realisticnotions about public response and needs are captured by how subjects’ actions, fears, andconcerns evolve as they interact with other SSG subjects and struggle for survival throughthe crisis. More specifically, we explore how the study design might provide additionalinsight into the following research questions:

1. What are college students’ immediate affective, cognitive, and behavioral reactionsto the simulated magnitude 7.8 earthquake?

2. How do concerns change during an escalating emergency following an earthquake?

3. How does fear change during an escalating emergency following an earthquake?

4. How does hope for rescue change during an escalating emergency following anearthquake?

5. What actions are taken and decisions made to protect oneself and others during anescalating emergency following an earthquake?

6. Do males and females differ in their pattern of self-reported fear ratings over time?

The next section describes the simulated earthquake scenario and the methodology usedfor assessing subjects’ behavioral, emotional, and cognitive responses to the event. We thenreport the findings from two separate SSGs, and finally, we summarize our conclusions andexamine the benefits of the SSG approach for informing policy decision makers.

SCENARIO SIMULATION GROUP REACTIONS 599

METHODOLOGY

PARTICIPANTS

For this exercise, we were interested in demonstrating the applicability of the SSGmethodology using a small sample of subjects likely found in a nonresidential building.University students are often in classrooms or office buildings throughout their academictenure, and thus were selected as an illustrative sample. To recruit subjects, flyers wereposted and handed out around the University of Southern California campus to under-graduate and graduate students. The flyer stated that researchers were seeking volunteersto participate in the simulated earthquake scenario study. The flyer also explained thatthe purpose of the study was to learn more about social interactions during an emer-gency situation and provided information about the time, location, and compensationprovided for their participation. Subjects were compensated $10 per hour for a two-hoursession (for a total of $20) on the day of the earthquake simulation exercise. They alsowere given a chance to win one of three Apple iPods. A total of 23 students (70%female) participated in the exercise. This study is intended to be a demonstration using asample of convenience of university students to explore the usefulness of the SSG meth-odology. For any application of the SSG methodology, to investigate likely behaviorsfollowing an earthquake (or other disaster event) a population of interest would be iden-tified and a representative sample selected.

STUDY DESIGN

Earthquake Scenario

The scenario opens with a shaking video simulating the occurrence of a magnitude 7.8earthquake with an epicenter on the southern end of the San Andreas Fault. The video is fol-lowed by a narrative describing an emergency situation that escalates over eight episodes.Episode One opens with the earthquake hitting shortly after students have arrived and set-tled in for their morning class. Following the earthquake, students discover their professoris seriously injured and they are unable to call for help as cell phone service is down. OverEpisodes Two through Four the situation continues to worsen for the class. The students aretrapped in a room with no power, no windows and a door that is forced shut by fallen debrison the other side. Some students also start to smell smoke, which creates concern about afire in the building. To further complicate matters, the professor’s wounds are so severe thathe loses consciousness. In Episodes Five and Six, the students manage to pry the door tothe hallway open. One of the smaller students is able to slip through the door and survey thesituation outside while other students continue to work on widening the door opening. Dur-ing this time the building fire alarms are blaring and the smell of smoke grows stronger.There is an escalating urgency to evacuate the building as soon as possible. In EpisodesSeven and Eight the class makes its way out of the building, while grappling with the chal-lenges of safely removing their professor and deciding how, if at all, to help studentstrapped inside an elevator. The scenario closes with the students about to exit the buildingwhen they learn of a chlorine plume heading toward campus from a truck overturned on anearby freeway. A summary of the earthquake simulation scenario by episode is providedin Table 1.

ROSOFF ETAL.600

EARTHQUAKE SIMULATION SCENARIO SUMMARY

MEASURES

Pre-measure

Prior to participating in the SSG earthquake exercise, subjects completed a pre-measurethat included demographic questions about sex, age, geographic region of origin, year inschool, and academic area of study. Table 2 provides a summary of the demographic data forthe combined SSG sample. (Note: Percentages do not always sum to 100 due to missing data.)

SAMPLE DEMOGRAPHIC DATA

Survey Instrument

To catalog the SSG reactions to the earthquake scenario, subjects completed a shortround of survey questions after each of the eight episodes. Subjects were told they would be

Table 1. Earthquake simulation scenario summary by episode

Episode Summary

OpeningEvent At 9:57 a.m. a 7.8 earthquake rips through Los Angeles for 2 minutes.

1 Its now 10:01 a.m. and the students are regrouping after what just occurred. Some try todial out, but there is no cell reception. Others are surveying the damage and discover theirprofessor was seriously injured by the ceiling projector that had fallen to the ground. Thepower is out, leaving the windowless classroom in darkness.

2 At 10:05 a.m. the students are sitting in the darkness listening to the noises around them. Onestudent feels her way to the door to discover that it is jammed, leaving the students and theprofessor trapped inside. Fire alarms also have started going off throughout the building.

3 Stuck in the classroom, students are left listening to the screaming and bustling from out-side the classroom door. The professor is now unconscious and it’s up to the students toget him medical attention immediately.

4 Now 25 minutes after the earthquake (10:22 a.m.), the sound of alarms and smell ofsmoke have students concerned there is a fire in the building, placing more pressure onfinding a way out of the building.

5 The doorway becomes more manageable to open. One of the smaller students is able toslip through the opening and go survey the damage outside. She reports of serious wreck-age outside and confirms suspicions of a fire in the building.

6 By 10:45 a.m. some of the larger students make significant progress in getting the dooropen. The smell of smoke is getting stronger and visibly thicker. There is a sense of ur-gency among the students to evacuate from the classroom.

7 By 10:55 the students have evacuated the classroom carrying the professor out with them.While heading for the building exit, they come across a group of students trapped in theelevator and are uncertain of how, if at all, to help them.

8 As students approach the building exit, they come to a stop when they see a crowd of peo-ple lingering by the doors. They are informed there is rumor of a chlorine plume sittingover the University’s campus. Uncertain about the consequences of the cloud’s toxins,the students are unsure about what to do.

SCENARIO SIMULATION GROUP REACTIONS 601

responding to questions asking them to describe their concerns and intended actions, andto rate their fear level as though they were experiencing the simulated earthquake scenario.Table 3 lists the survey questions asked in sequential order and the scales, when applicable,used for each.

SURVEY QUESTIONS AND RATING SCALES

After watching an earthquake shaking video and learning that the simulation involved amagnitude 7.8 earthquake, SSG subjects were presented with eight questions related to theirimmediate reactions to the situation. The first four questions were open-ended and focusedon immediate actions and information gathering once the shaking ended. More specifically,they were asked about what emergency action they would take, what information theywould seek, whom they would try to contact, and what means they would use to contactsomeone. The remaining four questions were close-ended. Subjects rated their fear level,the degree to which a large earthquake would disrupt their daily life, and how effectivelythey would respond to an earthquake. Subjects also estimated the likelihood of an earth-quake of the same magnitude occurring in the next 30 years.

Following each of the eight episodes, the SSG subjects engaged in a discussion of theirconcerns, fears, and possible actions. After each brief discussion, they responded to thesame three survey questions. The first two were open-ended questions designed to captureindividual priorities for actions related to the emergency situation and individual concernsrelated to each episode. The third question was a self report of fear level related to theepisode.

Table 2. Sample background and demographic breakdown

Sample Characteristics: N¼ 23

Demographic Variable Variable Response Cateogy Percentage of Sample

Age Median¼ 21 (Range¼ 18–25) NA

Sex Female 69.6%

Male 30.4%

Birthplace non-US 47.8%

US 43.5%

Race Asian 43.5%

Caucasian 30.4%

Black, Hispanic, Other 17.4%

Status Frosh=Soph 30.4%

Jr=Sr 34.8%

Grad Student 26.1%

Major Social Science & Humanities 30.4%

Business=Management=Policy 26.1%

Science and Engineering 30.4%

Undeclared 4.3%

ROSOFF ETAL.602

The survey closed with two questions encouraging subjects to reflect upon the scenariosimulation exercise and experience in the SSG. The first asked subjects to think about howthey felt without access to functional technology and to select what technological devicethey would like to have at their disposal at the current moment. The second question askedwhat lessons, if any, the subject learned from the exercise.

PROCEDURE

On 24 September 2008, (the same day as the 2008 California Golden Guardian Exer-cise) subjects reported for participation in a two-hour earthquake simulation exercise. Uponarrival, subjects were assigned to one of two SSGs (12 subjects each)—the rationale beingthat smaller groups would foster more active participation in the role-play desired for thesimulation. Once in their designated classrooms, SSG subjects were asked to watch anearthquake video. The video contained a sequence of segments depicting the shaking thatoccurs during earthquakes and the subsequent range of consequences that might ensue.

Following the video, each SSG moderator read the episode narrative introduced in theStudy Design section using a PowerPoint presentation and facilitated discussion among the12 SSG subjects. The moderator read the contents of each episode while subjects followed

Table 3. List of survey questions and rating scales in sequential order

Episode Survey Question Scale

PreliminaryQuestions

Immediate emergency action Open-ended

Immediately attempt to contact Open-ended

Immediate modes of attempted communication Open-ended

Immediate information sought Open-ended

Fear of magnitude 7.8 earthquake 1 (not afraid) - 5 (very afraid)

7.8 earthquake likelihood in the next 30 years 0% (no likelihood) - 100% (verylikely)

Disruption of life from magnitude 7.8earthquake

1 (not at all) - 5 (a great deal)

Ability to respond effectively to the magnitude7.8 earthquake

1 (very low) - 5 (very high)

1 through 8 Actions taken (by you and your “classmates”) toprotect each other and seek help

Open-ended

Fear caused by the current post-earthquakesituation

1 (not afraid) - 5 (very afraid)

Biggest concern right now Open-ended

5 Siren influence on feeling hopeful about rescue 1 (no hope) - 5 (very hopeful)

7 Exiting classroom influence on feeling hopefulabout rescue

1 (no hope) - 5 (very hopeful)

8 Exiting building influence on feeling hopefulabout rescue

1 (no hope) - 5 (very hopeful)

8 Feelings about no technology access anddesired device at disposal

Open-ended

8 Reflections on exercise Open-ended

SCENARIO SIMULATION GROUP REACTIONS 603

on the PowerPoint display and then instructed them to discuss the episode events as if theywere actual victims of the earthquake. Following the group discussion, subjects weredirected to individually report intended actions, concerns, and fears related to the latest epi-sode of the simulated earthquake scenario. Group discussion followed by individual surveyswas used to facilitate both learning and thought about the decision problems and to helpfully immerse respondents into the role-playing aspect of the exercise.

SSG moderators allocated a total of 9 minutes for the execution of each episode; 1 mi-nute for reading the scenario, 6 minutes for group discussion, and 2 minutes for surveycompletion. This process was repeated over eight times (72 minutes total) for each of thescenario simulation episodes.

CODING

Responses from all open-ended questions were grouped into discrete categories by aresearch assistant, blind to any of the study hypotheses. These categories were defined to beexhaustive, but not mutually exclusive. Two other research assistants, also blind to researchhypotheses, coded all responses to the open-ended questions using these independentlydefined response categories. All disagreements between the two coders were resolvedthrough a consensus discussion.

RESULTS

IMMEDIATE RESPONSE TO SHAKE

Immediate responses from both SSGs (total combined N¼ 23) to the initial earthquakeshaking video are summarized in Table 4. By almost a 7 to 1 margin, respondents favored aduck and cover action over running out of the building. Over 90% of the respondents indicatedthat they would first attempt to contact family, and only about a quarter indicated they wouldfirst attempt to contact emergency services. Every respondent indicated that they would attemptto use their cell phone immediately after the shaking stopped; other modes of communicationincluded email and in-person, each about 20%. When asked what information they mostwanted immediately following the shaking, about 41% of the respondents said they wanted toknow about the location and extent of damage caused by the earthquake and about a third ofthe respondents indicated they would seek information about each of the following: fate of fam-ily and friends and advice from official emergency response officials. Relatively few respond-ents indicated interest in the medical needs of those nearby (18%), practical information aboutthe loss of water, power, and transportation infrastructure (14%), location of evacuation centers(14%), risk of aftershocks (9%), and casualties resulting from the quake (5%). Surprisingly,14% of respondents indicated they would not seek any information (or did not know).

IMMEDIATE RESPONSE TO THE MAGNITUDE 7.8 EARTHQUAKE VIDEO

Overall, subjects indicated that they were very afraid of the magnitude 7.8 earthquakedescribed in the video. Mean fear ratings were 4.3 on a 1 to 5 scale, with over half of therespondents indicating maximum fear (5) and about three-quarters of the respondents indi-cating fear levels of either 4 or 5. The overwhelming majority of respondents (nearly 90%)indicated a belief that a magnitude 7.8 earthquake in the next 30 years is more likely than

ROSOFF ETAL.604

not, and an overall median response probability of 0.70. Respondents also agreed that amagnitude 7.8 earthquake would disrupt their life a great deal; mean rating of 4.3, with overhalf indicating maximum disruption (5) and the vast majority (87%) rating daily life disrup-tion as either a 4 or 5. The largest variability in responses to the earthquake was related tosubjects’ ability to effectively respond to such an event, with responses evenly distributedacross the 5 point scale (mean¼ 2.8).

RESPONSE TO ESCALATING SCENARIO

Respondents indicated their major concerns following each of the eight unfolding epi-sodes of the simulated earthquake scenario. These open-ended responses were coded intosix categories: (1) Lack of preparation, (2) Family and friends, (3) Escape (evacuation), (4)Aiding the injured, (5) Immediate dangers to self, and (6) Calling for rescue. Percentages ofrespondents indicating concerns for each of these categories are plotted across all eight epi-sodes in Figure 1. Concerns about lack of preparation and about family and friends werecommon during Episode One, but extremely rare or nonexistent after Episode One. The

Table 4. List of responses about immediate reactions

Survey Question Behavioral Intent % of Sample

Immediate emergency action Duck and Cover 86.4%

Leave Building 13.6%

Immediately attempted tocontact

Emergency Response 27.3%

Family 90.9%

Friends 54.5%

Immediate modes of attempt tocommunication

Cell Phone (voice & text) 100.0%

In Person 18.2%

Email 22.7%

Immediate information sought Location and extent of damage 40.9%

Well-being of family and friends 31.8%

Advice from ER officials & radio 31.8%

Medical needs of those near-by 18.2%

Casualties (Deaths and Injuries) 4.5%

Infrastructure failures, i.e., water, power,transportation, communications

13.6%

None or Do Not Know 13.6%

Quake Details (Location and Magnitude) 9.1%

Risk of aftershocks 9.1%

Location of evacuation centers 13.6%

Survey Question Scale Mean Response

Fear of magnitude 7.8 earthquake 1 (not afraid) - 5 (very afraid) 4.27

Likelihood of magnitude 7.8 earthquake 0% (no likelihood) - 100% (very likely) 70%

Disruption of life from magnitude 7.8 1 (not at all) - 5 (a great deal) 4.32

Ability to effectively respond 1 (very low) - 5 (very high) 2.82

Note: Percentages sum to greater than 100% since some indicated multiple contacts.

SCENARIO SIMULATION GROUP REACTIONS 605

dominant concern after Episode One was about finding an escape route out of the damagedand burning building. Providing immediate medical assistance to the injured, including theprofessor, was subordinate to finding an escape, except in the first and seventh episodes.The lack of cell phone service extinguished concerns about calling for help, although therewere some initial concerns related to signaling to others on the ground through the win-dows. As the situation increased in severity over the eight episodes, respondents’ concernsabout the immediate danger to their own safety also increased steadily. As the situationworsened, concerns shifted to escape and danger to self; however, concerns about theinjured remained a focus for a subgroup of the respondents.

MAJOR CONCERNS BY EPISODE

Respondents also indicated actions they would take in response to each episode of theunfolding earthquake scenario. These open-ended responses were coded using the followingthree categories: (1) Escape/evacuation, (2) Calling for help, and (3) Sheltering in place andproviding medical assistance to the injured. Percentages of respondents indicating actionsfor each of these categories are plotted across all eight episodes in Figure 2. Respondentsinitially focused on calling for help, but, as the situation deteriorated, there was an increas-ing tendency toward actions related to escape. Intended actions related to sheltering fromimmediate dangers and assisting others persisted throughout the eight episodes. Overall,intended actions varied substantially by episode. As one might expect, respondents’ actionswere quite sensitive to the unfolding circumstances of the simulation scenario.

Figure 1. Percentage of concerns about preparation, family and friends, aiding injured, evacua-tion, immediate dangers, and calling for help by episode. Note: Percentages sum to greater than100% since some indicated multiple contacts.

ROSOFF ETAL.606

MAJOR ACTIONS BY EPISODE

Respondents also provided fear ratings for all eight episodes; these are plotted sepa-rately for male and female respondents in Figure 3. Across all eight scenarios, females indi-cated higher levels of fear than males. A mixed model 2 X 8 ANOVA indicates that fearreported by females is higher than that for males, F(1,21)¼ 5.5, p< .05. The partial eta-squared (.21) indicates that about a fifth of the variance in fear ratings can be predicted bythe sex of the respondent. There is also substantial variation in fear across scenarios. Fearlevels vary significantly by episode, F (7,147)¼ 2.1, p< .05; about 9% of the variance infear can be attributed to changing circumstances across episodes. Fears for female respond-ents peak when fire is confirmed in Episode Four, but decreases until rumor of an approach-ing chlorine gas plume in Episode Eight. Fears for male respondents do not peak at EpisodeFour, but do show a steady increase over the earthquake simulation scenario. Fears for bothmales and females dip sharply when an escape route from the building is finally realized inEpisode Seven, followed by an immediate rebound in fear level in response to the gasplume rumor introduced in Episode Eight. The test for an episode by sex interaction failedto reach significance, F (7,147)< 1.0.

FEAR BY GENDER BY EPISODE

In addition, respondents provided ratings for hope for rescue for three episodes: Five,Seven and Eight. These are plotted separately for male and female respondents in Figure 4.There was significant change in hope for rescue over time, F (2,35)¼ 3.96, p< .05; about

Figure 2. Percentage of indicated actions related to evacuation, seeking help and sheltering byepisode. Note: Percentages sum to greater than 100% since some indicated multiple contacts.

SCENARIO SIMULATION GROUP REACTIONS 607

16% in the variances of hope for rescue can be attributed to change across episodes. Hopefor rescue peaked in Episode Seven when students identified a way to exit the classroomand building. However, in Episode Eight hope decreased when they reached the buildingexit and learned of rumors about an approaching chlorine gas plume outside. Hope for res-cue in Episode Eight was still higher than in Episode Five when students still were assessingtheir ability to exit the classroom. The test for sex (F (1,21)< 1.0) and sex by episode inter-action failed to reach significance, F (2,35)< 1.0.

HOPE BY GENDER BY EPISODE

REFLECTIONS ON THE SIMULATION

Upon completion of the earthquake simulation exercise, respondents were asked toreflect on the experience just completed. When asked to identify a single most desired tech-nological device during the simulation, 70% indicated a working cell phone and about aquarter of the respondents indicated an oxygen tank and mask. Responses to the open-endedquestion to identify “take-away” lessons from the simulation were quite varied. Theseresponses were coded into four categories: (1) Importance of preparation (57%), (2) Impor-tance of teamwork (35%), (3) Importance of staying calm (26%), and (4) Scariness of thesimulated 7.8 earthquake (17%).

DISCUSSION

We have introduced an innovative approach to the study of public reactions to a simu-lated life-threatening scenario following a devastating earthquake. The SSG approach

Figure 3. Fear ratings by episode for females and males.

ROSOFF ETAL.608

allowed subjects to engage in discussions about an earthquake scenario followed by individ-ual responses to a survey. This design uses written vignettes, role playing, and group discus-sion to emotionally immerse subjects into the earthquake scenario. Of particular interest toresearchers was capturing the dynamic nature of subjects’ reactions, including concerns,fears, and intended actions as an escalating emergency situation unfolds in the hours follow-ing the earthquake over eight discrete episodes.

Collecting data in the immediate aftermath of an event is very difficult. Researchersfrom different disciplines have explored various study designs for generating awareness ofthe public’s cognitive and emotional response to both natural and man-made disaster situa-tions. Psychologists have worked primarily within a psychometric paradigm to explore thevarious factors of risk perception and how these are related to attitudes and intended behav-iors in particular contexts (Engelberg and Sjoberg 2005, Sjoberg 2002, Alhakami andSlovic 1994, Kraus and Slovic 1988). A limitation of this approach is that the research istypically conducted in a lab setting which tends to be a more dispassionate context, and thestudies rarely assess how respondents react to an evolving disaster event. Similarly, sociolo-gists have worked primarily within a field research paradigm, conducting surveys and inter-views with disaster survivors (Tierney 2008, 2007, 1989). The drawback to this approach isthat researchers must rely on the memories of their respondents, who are also susceptible tosocial desirability biases. The SSG methodology is an alternative approach based on emo-tionally evocative scenarios and group discussion.

An ongoing challenge to the SSG design is whether subjects are reacting in the studysetting as they would in the real situation. While this can never be known with certainty,researchers expect to obtain a more realistic understanding of how a respondent’s emotions

Figure 4. Hope ratings by episode for females and males.

SCENARIO SIMULATION GROUP REACTIONS 609

and behavior change as the earthquake scenario develops. Furthermore, the change in reac-tions over time can also be explored, as the scenario is broken into eight discrete stages. Assuch, researchers might learn if responses monotonically increase in magnitude over theentire simulation, reach a plateau, or reach a maximum and then subside. Also, they mightassess how the patterns of change in respondents’ emotional and behavioral responses differas a function of the manipulation of key details about the simulated scenario (e.g., commu-nication with emergency officials possible or not), or covary with relevant demographic var-iables (e.g., sex) or personality variables (e.g., locus of control).

The results presented in this paper are a demonstration of the type of information andinsights possible from using the SSG simulation approach to track the dynamic emotionaland behavioral responses to an earthquake. Findings suggest that there are variations in theconcerns of respondents over time. Initially respondents were preoccupied with the safetyof family and friends and their lack of preparation. However, as time progressed their focusshifted towards mitigating the danger to themselves and identifying an escape route out ofthe classroom building. When respondents were asked about their intended actions, per-ceived actions were quite sensitive to the evolving situation of the simulation scenario.Respondents initially focused on calling for help, but as the situation deteriorated, there wasan increasing focus on actions related to escape.

What remains unanswered is how respondents’ perceptions and decisions vary when dif-ferent factors critical to the assessment of an earthquake are considered. This can be addressedin future studies where important variables, such as size of the earthquake or level ofemergency response, are systemically manipulated for different (independent) SSGs. Bymanipulating selected components of the scenario with the SSG methodology, researcherscan provide a strong justification for a causal link between related variables. Such an experi-mental design, involving the manipulation of earthquake scenario features, allows researchersto identify characteristics that influence potential earthquake victims’ emotional reactions andbehavioral decision making. This understanding of public concerns, fears, and actions is criti-cal to the development of effective policies for earthquake emergency management.

Results for the SSG-tested sample also indicated that females consistently experiencedhigher levels of fear than males across all eight episodes. This finding is consistent withgender risk perception research findings, which have shown that risk tends to be judgedlower by males than by females (Flynn et al. 1994). Females are believed to be more vulner-able to natural disasters than males, and thus more sensitive to risk (Slovic 1997, Grangerand Hayne 2001). In the same vein, because of physical differences, males have a greaterillusion of control over their ability to respond a disaster situation (Rosen 2005). Othersposit that such sex differences are apparent from an early age, as females have higher per-ceived risk of injury from play and are more able to identify risks (Morrongiello and Rennie1998). This study replicates this sex difference using a dynamic and emotionally evocativescenario simulation.

Other studies have been conducted to evaluate the effect of different subgroup perspec-tives on perceived risk. Researchers have found that respondents’ proximity to a terroristattack (location) traditionally has been associated with increased perceived risk by thosewho are in the immediate vicinity or are strongly linked through a personal story (Fischoff2003). Also, when looking at education and income, researchers have found that a more

ROSOFF ETAL.610

educated person is capable of understanding new information more easily, and thereforefeel a higher degree of “control” in a disaster situation (Ho et al. 2008). Additional studiesapplying the SSG group design using specific demographic groups (e.g., older adults, moth-ers with children, etc.) would test whether the same aforementioned conclusions hold truein the aftermath of a devastating earthquake, as well as provide insight into other potentialinfluential variables for predicting public response.

Lastly, findings from the SSG-tested sample showed that respondents’ hope for rescueincreased as the scenario changed over time. This finding supports Snyder’s definition ofhope as the perceived capability to achieve desired goals (Snyder 2002). By this definition,respondents’ perception of how successful they were being in evacuating the building wasreflected in how they evaluated hope for rescue. Accordingly, hope for rescue was higher asstudents made progress in securing their safety (e.g., exiting the classroom), and lowerwhen they were confronted by challenges (e.g., rumors of the gas plume outside). The SSGstudy design could be extended in future research to consider how individual experienceswith disaster situations and safety in terms of feeling vulnerable and protected might influ-ence the respondents’ assessment of hope for rescue.

A major challenge for policy analysts is understanding how risks are perceived andresponded to by individuals. The application of the SSG approach is intended to providepolicy makers and emergency responders with information critical for (1) Providing insightinto how earthquake victims respond and cope in the aftermath of an earthquake and (2)Contributing to the development of effective organizational policies related to earthquakeemergency response.

ACKNOWLEDGMENTS

This research was supported by the National Science Foundation under grant numbersSES-0728934 and SES-0901036. It was also supported by the U. S. Department of Home-land Security through the National Center for Risk and Economic Analysis of TerrorismEvents under grant number 2007-ST-061-000001. However, any opinions, findings, conclu-sions, and recommendations in this document are those of the author and do not necessarilyreflect views of the National Science Foundation or the U. S. Department of Homeland Se-curity. We would like to thank Kelly Buccola, Jaime Carias, Sarah Hammond, SabrinaFeeley, Kimberly Lewkowitz, Shelly Christine McArdle, and Robert Siko for their supportin the planning and execution of the earthquake simulation exercise.

REFERENCES

Alhakami A. S., and Slovic P., 1994. A psychological study of the inverse relationship betweenperceived risk and perceived benefit, Risk Analysis 14, 1085–1096.

Berke, P. R., and Beatley, T., 1992. Planning for earthquakes: Risk, politics and policy, JohnsHopkins University Press, Baltimore, MD, 232 pp.

Berke, P. R., Kartez, J., and Wenger, D., 1993. Recovery after disaster: Achieving sustainabledevelopment, mitigation and equity, Disasters 17, 93–109.

Brookshire, D., Chang, S., Cochrane, H., Olson, R., Rose, A., and Steenson, J., 1997. Directand indirect economic losses from earthquake damage, Earthquake Spectra 13, 683–701.

SCENARIO SIMULATION GROUP REACTIONS 611

Drabek, T. E., 1985. Managing the emergency response: A challenge for public administration,Public Administration Review 45, 85–92.

Drabek, T. E., 2010. The Human Side of Disaster, CRC Press, Boca Raton, FL, 288 pp.

Dynes, R., 1994. Situational Altruism: Toward an Explanation of Pathologies in Disaster Assistance,Preliminary Paper No. 201, University of Delaware, Disaster Research Center, Newark, DE.

Eisenman, D., Wold, C., Fielding, J. Long, A., Setodji, C., Hickey, S., and Gelberg, L., 2006.Differences in individual-level terrorism preparedness in Los Angeles County, AmericanJournal of Preventive Medicine 30, 1–6.

Engelberg E., and Sjoberg L., 2005. Perceived reality of visually mediated hazards and beliefsabout risk, Applied Cognitive Psychology 19, 899–912.

Fiedrich, F., Gehbauer, F., and Rickers, U., 2000. Optimized resource allocation for emergencyresponse after earthquake disasters, Safety Science 35, 41–57.

Fischoff, N., Gonzales, R., Small, D., and Lerner J., 2003. Judged terror risk and proximity tothe World Trade Center, The Journal of Risk and Uncertainty 26, 137–151.

Flynn, J., Slovic, P., and Mertz, C. K., 1994. Gender, race, and perception of environmentalhealth risks, Risk Analysis 14, 1101–1108.

Glass, T. A., 2001. Understanding public response to disasters, Public Health Reports 116, 69–73.

Godschalk, D. R., Beatley, T., Berke, P., Brower, D., and Kaiser, E., 1999. Natural Hazard Mit-igation: Recasting Disaster Policy and Planning, Island Press, Washington D.C., 550 pp.

Goejian, A. K., Steinberg, A. M., Najarian, L. M., Fairbanks, L. A., Tashjian, M., and Pynoos,R. S., 2000. Prospective study of posttraumatic stress, anxiety, and depressive reactions afterearthquake and political violence, American Journal of Psychiatry 157, 895–911.

Granger, K., and Hayne, M. (eds.), 2001. Natural Hazards and the Risk They Pose to South-East Queensland, Australian Geological Survey Organisation, Commonwealth Governmentof Australia, Canberra, Australia.

Ho, M., Shaw, D., Lin, S., and Chiu, Y., 2008. How do disaster characteristics influence riskperception? Risk Analysis 28, 635–643.

Heller, K., Alexander, D., Gatz, M., Knight, B., and Rose, T., 2006. Social and personal factorsas predictors of earthquake preparation: The role of support provision, network discussion,negative affect, age, and education, Journal of Applied Social Psychology 35, 399–422.

Kartez, J., and Lindell, M. K., 1990. Adaptive planning for community disaster response, Citiesand Disaster: North American Studies in Emergency Management, Sylves, R., and Waugh,W. (eds.), Charles C. Thomas, Springfield, IL, 5–31.

Kendra, J. M., and Wachtendorf, T, 2006. Community innovation and disasters, Handbook ofDisaster Research, Rodriguez, H., Quarantelli, E. L., and Dynes, R. R. (eds.), Springer Ver-lag, New York, 316–334

Kraus, N. N., and Slovic P., 1988. Taxonomic analysis of perceived risk: Modeling individualand group perceptions within homogeneous hazard domains, Risk Analysis 8, 435–455.

Kreps, G. A., and Bosworth, S. L., 2006. Organizational adaptation to disaster, Handbook ofDisaster Research, Rodriguez, H., Quarantelli, E. L., and Dynes, R. R. (eds.), Springer Ver-lag, New York, 297–315.

Lindell, M. K., and Perry, R. W., 2000. Household adjustment to earthquake hazard, Environ-ment and Behavior 32, 461–501.

ROSOFF ETAL.612

May, P. J., 1991. Addressing public risks: Federal earthquake policy design, Journal of PolicyAnalysis and Management 10, 263–285.

Mileti, D., and Fitzpatrick, C., 1992. The causal sequence of risk communication in the Park-field earthquake prediction experiment. Risk Analysis 12, 393–400.

Morrongiello, B. A. and Rennie, H., 1998. Why do boys engage in more risk taking than girls?The role of attributions, beliefs, and risk appraisals, Journal of Pediatric Psychology 23, 33–43.

Nelson, A., and French, S., 2002. Plan quality and mitigating damage from natural disasters: Acase study of the northridge earthquake with planning policy considerations, Journal of theAmerican Planning Association 68, 194–207.

Okuyama, Y., and Chang, S. (eds.), 2004. Modeling Spatial and Economic Impacts of Disas-ters: Advances in Spatial Science, Springer Verlag, New York, NY, 323 pp.

Palen, L., and Liu, S., 2007. Citizen communications in crisis: anticipating a future of ICT-sup-ported public participation, Special Interest Group on Computer-Human Interaction (SIG-CHI) Conference on Human Factors in Computing Systems, April 28–May 3, 2007, SanJose, CA.

Palm, R. and Hodgson, M. E., 1992. After a California Earthquake: Attitude and behaviorchange, University of Chicago Press, Chicago, IL, 130 pp.

Paulhus, D., 1983. Sphere-specific measures of perceived control Journal of Personality andSocial Psychology 44, 1253–1265.

Pearce, L., 2004. Disaster management and community planning, and public participation: Howto achieve sustainable hazard mitigation Natural Hazards 28, 211–228.

Rosen, J., 2004. The Naked Crowd: Reclaiming security and freedom in an anxious age, Ran-dom House Publishing Group, New York, NY, 272 pp.

Saab, D. J., Maldonado, E., Orendovici, R., Tchouakeu, L. M., van Gorp, A., Zhao, K., Mait-land, C., Tapia, A. H., 2008. Building global bridges: Coordination bodies for improved in-formation sharing among humanitarian relief agencies, 5th International ISCRAMConference, May 4–7, 2008, Washington, D.C.

Shaw, R., Kobayashi, K., and Kobayashi, M., 2004. Linking experience, education, perceptionand earthquake preparedness, Disaster Prevention and Management 13, 39–49.

Slovic, P., 1997. Trust, emotion, sex, politics and science: Surveying the risk-assessment battle-field, Environment, Ethics and Behavior, Bazerman, M., Messick, D., Tenbrunsel, A., andWade-Benzoni K. (eds.), New Lexington Press, San Francisco, CA, 277–313.

Slovic, P., Fischoff, B., Lichtenstein, S., and Roe, F. J. C., 1981. Psychological factors andsocial implications, Proceedings of the Royal Society of London, Series A, Mathematical andPhysical and Engineering Sciences 376, 17–34.

Snyder, C. R., 2002. Hope theory: Rainbows in the mind Psychological Inquiry 13, 249–275.

Sjoberg, L., 2002. Are received risk perception models alive and well? Risk Analysis 22, 665–669.

Tierney, K. J., 2008. Hurricane in New Orleans? Who knew? Anticipating Katrina and its dev-astation, Sociological Inquiry. Special Issue: Twenty-first century disasters: Why should wecare about worst cases? 78, 179–183.

Tierney, K. J., 2007. From the margins to the mainstream? Disaster research at the crossroads,Annual Review of Sociology 33, 503–525.

Tierney, K. J., Lindell, M. K., and Perry, R. W., 2001. Facing the Unexpected: disaster prepar-edness and response in the United States, Joseph Henry Press, Washington, D.C., 301 pp.

SCENARIO SIMULATION GROUP REACTIONS 613

Tierney, K. J., 2002. Lessons learned from research on group and organizational responses todisasters, Countering Terrorism: Lessons Learned from Natural and Technological Disas-ters, February 28–March 1, 2002, National Academy of Sciences, Washington, D.C.

Tierney, K. J., 1989. The social and community contexts of disaster, Psychosocial Aspects ofDisaster, Gist, R. and, Lubin, B., (eds.), John Wiley & Sons, Oxford, England, 11–39.

Tierney, K. J., 1985. Emergency medical preparedness and response in disasters: The need forinterorganizational coordination, Public Administration Review 45, 77–84.

Wachtendorf, T., and Kendra, J. M., 2005. A typology of organizational response to disasters,Presentation to the American Sociological Association, August 14, 2005, Philadelphia, PA.

Yehuda, R., 2002. Post-traumatic stress disorder, New England Journal of Medicine 346, 108–114.

(Received 25 January 2010; accepted 13 November 2010)

ROSOFF ETAL.614