programs, training, and traditions that minimize the ...... B. A mechanic is protected by a heavy bench used in ... out of the car and knock on the front door of the.
PLANNING, WARNING, EVACUATION, AND SEARCH AND RESCUE: A REVIEW OF THE SOCIAL SCIENCE RESEARCH LITERATURE
B. E. Department Texas A&M College Station,
Aguirre of Sociology University Texas, USA, 77843
PLANNING, WARNING, EVACUATION, AND SEARCH AND RESCUE: A REVIEW OF THE SOCIAL SCIENCE RESEARCH LITERATURE ABSTRACT The focus of this review is on planning, warnings, evacuations, and search and rescue activities that are designed to minimize the deleterious effects of disasters on populations. The review discusses the points of consensus among specialists in these disaster-related activities, and it illustrates them with my own research in Puerto Rico and Mexico. The review disccusses present-day lack of interest in planning for disasters and the determinants of public response to warnings. A third section presents the prevailing definition of evacuation, types of evacuation and a list of its determinants, as well as the two major extant conceptual frameworks of evacuation. The fourth and final section of this review presents what is known about search and rescue activities in disasters, to include the importance of volunteers and emergent group activities, the importance of the timing of rescue, and the policy implications of these patterns.
2
PLANNING, WARNING, EVACUATION, AND SEARCH AND RESCUE: A REVIEW OF THE SOCIAL SCIENCE RESEARCH LITERATURE This review of the social science research literature on planning, warning, evacuation, and search and rescue assumes that disasters are failures of social systems (Dynes, 1993). As such, its emphasis is not on physical agents but on what people do together to respond in an organized fashion to the crisis of disasters, and on the continuities between pre and post disaster social organizations. Its focus is on disaster preparedness, or on the programs, training, and traditions that minimize the effects of disasters that will occur (Gillespie and Streeter, 1987). Planning, warnings, evacuations, and search and rescue are processes designed to minimize the deleterious effects of disasters on populations. While limitations of space and time preclude the treatment of other preparedness issues such as the level of awareness of hazards of populations, the four topics presented in this review cover most of the empirical work on the subject of disaster preparedness in the social sciences today, although perhaps some important contributions may have been omitted by error. The most vigorous research tradition is on warnings, followed by research on evacuation. Search and rescue and planning run a distant third in the amount of attention they have received so far. Two caveats are in order. Search and rescue activities have until now being conceptualized as part of disaster response. I would like to suggest ways in which they should be seen as presenting legitimate disaster preparedness planning and program issues. Moreover, while it can be argued that there is a chronology to the four types of activities reviewed in this paper, with planning preceding warnings which are then followed by evacuation behavior and search and rescue efforts, in fact this chronological assumption is not made in this paper. More complex interconnections among the four processes take place. PLANNING. Despite its obvious relevance to preparedness activities, planning for disasters has not received a great deal of research attention in the social sciences. Disaster planning is a rather unpopular activity for most Americans. This is the case in spite of the importance of disaster planning and its complexity, made clear, for example, by Vogt and Sorensen's (1992) review of the many planning issues related to evacuations (see below). Quarantelli (1991; see also Britton, 1987; Burby and French, 1981) has recently summarized these patterns. He points out that: 1. It is very difficult to get individuals and families to become interested and concerned about disasters before they happen. Most people are involved in the here and now and are uninterested in low probability, future disaster events. Moreover, they see planning as the responsibility of the state, not of themselves. The exception to this disinterest in planning occur in communities with a recurrent and serious disaster threat
3
(e.g., Perry and Green, 1983), and among emergent groups that become active proponents of disaster preparedness and mitigation measures in their communities. The proportion of populations that engage in some sort of preparatory activity increases with the frequency of hazards (Drabek, 1986, 24). After extensive review of the social science literature, Drabek (1986, 26-28; see also Palm, 1981; 1987) write that, after disasters, families in impacted communities express interest in disaster planning but very little effective planning gets done. 2. The few complex organizations that plan for disasters often do not do a good job of it. With the exception of organizations formally charged to respond to crises, such as police, fire departments, and medical organizations, most private and public organizations in the United States do not plan for disasters. In the US the greater the size of the organization the greater the presence of disaster planning (Drabek, 1986. However, the plans that are done are rather limited, assume that the crises will be of the most severe type imaginable, and that the effects of the disasters will impact others and not the organizations themselves. Most planning succumbs to the technological bias, with much emphasis put on having sophisticated technology such as computers to respond to the crisis rather than the development of the social organization required for successful operations. Moreover, those organizations that have developed disaster plans tend to think of disasters as accidents, ignoring that disasters are much more complex than accidents, for they force the organizations to interact intensively with many and different groups from both the public and private sectors, in situations where the organization has to relinquish parts of its autonomy and apply different performance standards, and where the organizations may have suffered losses of their own resources. In short, organizational planning is often done incorrectly. 3. Finally, Quarantelli points out that communities give very low priority to disaster planning. Most communities in the US do not spend a great deal of time, money, or other resources in planning for disasters. It is very seldom that local governments attempt to educate the public to the hazards that threaten them. In the US whatever is done is to a large extent the result of the activities of the federal government, through federal matching grants and planning grants. Community planning for disasters is rendered difficult by the divisions, disputes, and conflicts that punctuate community life. Such divisions are often expressed in conflicts among the various organizations charged to respond to disasters, such as the police, army, emergency medical organizations, and fire departments. Such conflicts make it very difficult for organizations to cooperate on a community-wide disaster plan, for the plan requires their cooperation and trust, and a willingness to forego some degree of organizational autonomy. Drabek (1986) points out that in the US the local emergency management offices, charged to develop and implement disaster planning, are often hampered by uncertainty regarding their authority, task domains, and public support. Successful disaster planning by local
4
civil defense offices are a product of their previous experiences in handling disasters. Their success is also a function of the willingness of local governments to make civil defense a legitimate and important organization, the centrality of the organization in the structure of local government, and the extent to which the office is a useful source of information to the other agencies in the community. Their success is also determined in part by the ability of the civil defense director to develop relationships with key officials in key disaster-relevant organizations in and out of their communities, and by the resources at their disposal. Effective community emergency managers are recognized nowadays as perhaps the most important force facilitating disaster planning and other preparedness and mitigation activities. Effective coordinators make it a habit to visit other officials on a regular basis, feel comfortable working with them, and are in agreement with the general goals of their organizations. Most effective community disaster managers have a regular, comprehensive preparedness program, have an updated program with defined duties and responsibilities, have good communication equipment at their disposal, and spend a great deal of their time educating the public through mass media and other means. Professors E. L. Quarantelli and R. R. Dynes have argued that planning for a disaster should be differentiated from managing a disaster. Such planning is an ongoing process and not a product. It is based on likely events, not worst possible cases. It aims for appropriate, not necessarily speedy responses. It is based upon accurate knowledge of disaster behavior and on patterns of everyday routines. And it avoids command and control structures. Reflecting these insights, Wenger, James, and Faupel (1980) have identified the most common errors made by planning officials. These are: 1. To see disaster planning as a product, not a process. 2. to think of disaster planning as isolated from the day-to-day planning process. 3. to see the plan disconnected from the the behavior of collectivities during disasters, so that little attention is given to public response to disasters. 4. to be unclear as to what constitutes an emergency and who is responsible to declare the emergency. 5. to fail to make available and to distribute disaster relevant information about the plan to all relevant community organizations. 6. to fail to create a command center to respond to the crisis. 7. to fail to take into consideration the problem of interorganizational cooperation in times of crisis.
5
8. for the plan to remain a paper plan, so that it is not rehearsed. Faupel (1987, p. 206-207) argues that disaster planning has three distinct analytical characteristics that can be conceptualized using the theoretical paradigm known as sociological human ecology. These aspects are the development of the formalized disaster plan, the constant updating and testing of the plan, and the formalized plan itself. He points out that the formal plan is a microcosm of the ecological complex, for it incorporates assumptions about the population, environment, social organization, culture, and technological dimensions of communities. While largely ignored at present, I find Faupel's attempt to conceptualize disaster planning in terms of human ecological processes potentially a very fruitful approach to planning. WARNINGS. There is considerable consensus among social scientists that it is useful to conceptualize the public's response to warnings as a function of the physical environment, population, technology, social relations, and culture. Moreover, it is necessary to differentiate between the warning message and the system which produces and responds to it. Warning systems are complex. Such systems include the individuals or organizations that detect the hazard and communicate the information to threatened populations, others who issue and transmit the warning, and yet still others who receive the warning, interpret it, and hopefully act out specific protective behavior. Quarantelli refers to this behavioral response to disaster warnings as the specific "adjustive behavioral outcome of the reaction pattern." Eliciting protective responses constitutes the primary goal of warning systems. Another point on which there is consensus among specialists is that warning is not the linear transmission of a message but the result of a complex interaction of physical, technological and social systems. The complexity of warning systems derives from the interdependence among these components and the intereffects that characteristically take place among them. It is inappropriate to think of parts of warning systems in isolation from each other. Thus, the optimum design and implementation of warning systems require a generalized openness of disaster preparedness agencies to system-generated demands for change. Warning messages often have a scientifically grounded content. However, such content is insufficient to make them effective in protecting threatened populations. Warning messsages and the systems that produce them must also consider people's perception and interpretation. To be effective, warnings must take into consideration people's reactions to warnings. Thus, considerable knowledge of local populations is required.
6
Consideration of the interactive, interpretative nature of warning systems means that warning messages need to adjust and change in response to historical events during which people reacted to warnings. Thus, the designers of warning systems need to allow for the opportunity of programmatic learning. Imbalances and lags in the production and distribution of warnings account for a considerable proportion of present-day failures of warnings in the US. Numerous investigators have shown that people's reactions in disaster situations (both reactive and proactive) cannot be meaningfully described as panicky (i.e., wild flight, hysterical breakdown, irrational or impulsive behavior). Instead, their reactions to warnings are affected by how they define the situations in which they find themselves. People have their own definition of the situation which incorporates the person's identity and previous history. This personal history shapes the meanings of social objects and symbols, and is impacted by the observed or imputed actions and reactions of other people perceived as "significant others." It follows that upon receiving warnings people try to evaluate the message not only in terms of their own previous experiences, values, ideologies, and personal goals but also by seeking information and confirmation of the warning through observing changes in their surrounding, and by inquiring how others, including authorities, perceive the risk (Sorensen and Mileti, 1989). Social science research has shown that in disasters different categories of people at risk have different definitions of the situation, different perceptions of risk, and different reactions to warnings. For example, previous experience with disaster make people more receptive to warnings and to the need for protective action, and actual or potential victims with relatives in the area threatened by the disaster receive more offers of assistance than their counterparts. Similarly, people have difficulty believing the warning message when they cannot see or hear the hazard. Sequentially, Mileti and Sorensen summarize some of the more important aspects of this process as involving hearing the warning, understanding its contents, believing that the warning is accurate and credible, personalizing its message to the person's life situation, confirming the fact that others are responding to the warning, and responding by taking proctective action. Rogers and Sorensen (1989), in their research on warnings and response in chemical accidents, report that response time is in part determined by the timing of the warning, the content of the warning message, and the source of the information presented in it. Reactions to warnings are an outcome of this complex sociopsychological process. They range from inertia--the explicit or implicit refusal or denial to grant validity to the announcement of the risk and the need for protective action--to the complete, conscious acceptance of the risk and the need for protection. Between these two opposites are many stages, in which people attempt to provide, with varying degrees of explicitness, purposefulness, and success, independent evaluation of the risk and assessment of their options. The important
7
thing to keep in mind is that such reactions are social products, impacted not only by the person's subjective interpretation of the situation at hand but also by the qualities of the warning message. Recently, Sorensen (1993) identified the social processes that impact people's responses to warnings in the US and the extent of empirical support associated with each of these predictions. TABLE 1 ABOUT HERE Drabek identifies three important qualities of messages: content, source, and number. Sorensen (1993) mentions that a public warning message should include information about the hazard or risk, location, what the person should do, the time constraints involved, and the source of the message. Specific messages are more effective in producing protective response among people at risk of the hazard. Warnings should be given in the language of the population at risk of the hazard. Specific warnings should include general information about who should do what, when, how, with whom, why, and with what consequences. Messages need to be internally consistent regarding such matters as the history, origin, nature, and seriousness of the risk. For example, they might include brief coverage of such topics as the causes, probability of occurrence, probable effects, and probable place, time, and length of impact of the hazard. Most crucially, warnings should include the effective reaction choices open to individuals; they should offer clear guidance as to what people should do to prevent or minimize the risk. People react differently to the warnings they receive from different sources such as police, fire departments, mass media, and primary groups. Locally, each source has a different level of credibility which affects people's reactions to the warnings. The probability of eliciting proper responses from the public is maximized if the issuer is perceived as knowledgeable and has the public's trust. Numerous, consistent, timely, and complete messages have more favorable effects in bringing about appropriate protective actions than few, inconsistent, inaccurate, or inconsistent messages. Moreover, there should be congruence between the content, context, and the tone of the warning message. Effective warnings should have the appropriate emotive character or "envelope." Emotion and rationality are not competing dimensions of social life. Warnings should have a metalanguage of constrained emotion which, while not distorting their contents, enhances their ability to bring about protective responses. Effective warnings evince emotionality, to mirror the natural tendency of people to react to crisis in those terms. There have been instances of disasters in the US in which people did not respond to warnings of severe weather and continued to behave under the assumptions of normalcy because the appropriate emotive envelope of the message was absent. Another problem occurs in situations when, after the message of warning is presented, its source reverts to its customary or habitual routine. Thus, warnings especially about an impending, quick-onset hazard should
8
be presented continuously by the mass media and in a crisis context for the duration of the threat, rather than being presented in a normal programming context which is only momentarily interrupted. Testing the effectiveness of warnings should not only include hardware, evaluation of decision making structures, and the outcome of various diffusion methods, but should also include the population at risk. Effective warning systems help people develop a tradition, a custom, a collective acceptance, of hazards in their lives, so that testing and modification of warning systems becomes a significant part of their lives. Thus, while unexplored so far, my guess is much can be learned from the study of warning systems in communities with strong disaster subcultures. The study of warnings cannot be separated from people's awareness of risk, of the extent of their awareness of the implications for disaster preparedness of what they do or fail to do. While people need to receive instructions as to how best to prepare for different disaster hazards through public boards, schools, and telephone directories, for example, they also need to be encouraged to change their attitudes towards specific hazards, to become sensitive of environmental cues, such as weather signs of impending tornadoes (National Weather Service; Doswell and Ostby, 1982), to notice their landscapes for places where they might find refuge in rapid onset hazards, and to build shelters in their homes and adopt other preparedness measures. An example of a warning program that has had mixed success in the U.S. is the weather warning system currently used by the National Weather Service (NWS) to produce tornado watches and warning messages. This warning program needs to be reevaluated in light of the social science knowledge about warnings already available, for the two products, especially tornado watches, violate many of the most important characteristics of effective warnings.. Changes in the system will require laboratory and field experimentation of human reaction to weather warnings. Such changes become particularly pressing as detailed weather information becomes available in the US from improvements in weather technology such as NEXRAD (next generation weather radar) and the integration of information from different weather monitoring systems. The foundation of effective warning systems is the behavior of consumers. The challenge in the US and throughout the world is to reshape public perceptions, attitudes, and customs towards the acceptance of individual responsibility for disaster preparedness and away from a unidirectional educational approach which makes people passive recipients of the information generated by experts and the services of government bureaucracies. Such passive-type programs were inventoried by Sorensen and Mileti. They take many forms: giving practical instructions or scientific information; engaging in norm-oriented (social pressure) communication; modeling behavior through the use of admired public figures who serve as models for
9
appropriate response; arousing fear and using prompts; and teaching appropriate protective response through participation in an already elaborated program. Despite their popularity, the effectiveness of these programs is limited. It is instructive to quote Sorensen and Mileti on this point: "Overall (there is) no conclusive evidence that people are more prepared and protected as a result of information programs. Furthermore (the evidence) is inconclusive about how programs could be improved to produce higher levels of protection. Finally, the experience gained with one program at a single location, even if evaluated, may not be useful in designing protective action schemes for different locations or for the entire country (Sorensen and Mileti, 1987a: 225)." Perhaps this overall evaluation of existing programs is too pessimistic, for there may be useful elements of this "planning from above" approach. Nevertheless, my opinion is that local citizenry can play an important part in planning, creating, testing, and modifying warning systems. The principles of effective planning previously summarized elsewhere in this paper coincide with my view that the participation of the residents of local areas in the creation, implementation, and change of warning systems is essential. People who are potential victims must also become active creators and guarantors of their own welfare (Mileti, Farhar, Fitzpatrick, 1990). Effective national warning systems are diverse. What works in one place might not necessarily work in other places. Some of the elements of warning systems which can be made to vary are: (1) mass public participation and education; (2) coverage of special populations and places like the elderly, school children (their protectiveness training is routine in schools in tornado-prone communities in the US), hospitals, and trailer parks; (3) variations in the sources and content of the messages and their means of dissemination (e.g., sirens/alarms, modulated power lines, aircraft, VHF-FM Tone Alert Warnings or pagers triggering home receivers, telephone automatic dialers, radio and TV announcement, cable override, personal notification through police and firemen's loudspeakers and sirens alerting neighbors and persons in stores, markets, or theaters, citizen band radios, door to door volunteers, taxi companies, whistles, flares, electronic display boards in major traffic routes); and (4) the extent to which these different means of dissemination of warnings are used. The consensus among specialists is that some degree of redundancy is appropriate, for a multiplicity of channels insures that the warning system will work before, during, and after primary impact of the hazard, and reduces the problem which is created through dependence on one channel or on key personnel (Roger and Sorensen, 1991). EVACUATION. Quarantelli, reviewing the literature on evacuations, define them as "mass physical movements of people, of a temporary nature, that collectively emerge in coping with community threats, damages, or
10
disruptions (p.10)." This definition is useful for our purposes, for it excludes permanent or semi-permanent relocations such as war population movements (Moses, Rosenfeld, and Moses-Hrushovski, 1987), and the evacuations of buildings (Pauls, 1980; Weinroth, 1989). It does not exclude, however, the recent use of high rise buildings for vertical evacuation (Ruch et al., 1991) . Quarantelli's definition makes it clear that evacuations involve a community context and a multiplicity of behavioral sequences occurring more or less cotemporally. Evacuations are defined as the round-trip movement of throngs of people. They are one of many behavioral outcomes engaged by people faced with imminent danger. Evacuation behavior is not instinctive or automatic. Rather, it is contingent, socially constructed behavior. Aguirre (1983) uses relative distance, voluntarism, and permanence of population movements to differentiate evacuation from migration. In evacuations the movement of people away from danger is accomplished without a great deal of difficulty (Quarantelli, 1980); evacuations are usually orderly from the perspective of the evacuees, and are generally effective in removing people from danger. Instead, Quarantelli concluded that the problems that occur in evacuations take place before and after people's movement occur. The organizational and interorganizational planning required to begin evacuations is often poor and not based on knowledge of the actual problems evacuees face during transit, after arrival at places of destination, and during their return to their places of origin. Evacuations present distinct planning issues that should be seen as separate from warning. R. Perry and his colleagues, in their studies of evacuation, subscribe in general terms to Quarantelli's definition of evacuation, although they help clarify it through the identification of four subtypes of evacuation. These subtypes are preventive, protective, rescue, and reconstructive evacuations. They identify two factors that produce different evacuations: the first is the amount of time available between the evacuation and the time of impact of the hazard; the second is the amount of time evacuees remain away from their places of origin. The occurrence of pre-impact evacuation, done to minimize loss of life and property, depends on the type of hazard and detection system available in a community. Some hazards, such as hurricanes, allow many hours of preparation before impact. Likewise, communities differ in their ability to detect the threat and issue warning to the population. Post-impact evacuation is part of community reconstruction, involving search and rescue efforts and the long term relocation of survivors while the reconstruction of their communities occur. Perry and his colleagues also differentiate short-term from long-term evacuations. In short term evacuations shelters are used to house evacuees. These evacuations require fewer and less complex arrangement for the provision of food, medical attention, and other essentials. In contrast,
11
long term evacuation requires complex planning and social control. Quarantelli's definition is most appropriate for preventive evacuations (pre-impact, short-term). This is the type of evacuation that has received most social science research attention. Its counterpart, protective evacuations (pre-impact, long-term), or the removal of potential victims before impact, received some attention from U.S. civil defense analysts. The threat of nuclear war or accidents (Johnson, 1986; Tweedie et al., 1986) provided the impetus for studies that considered the logistics of massive, region-wide human evacuations. Rescue evacuation (post-impact, short-term), also known as search and rescue activities, see below, is increasingly gaining importance as a type of evacuation, as chemical accidents become more common throughout the world (Rogers and Sorensen, 1989). Finally, reconstructive evacuation (post-impact, long-term) the other remaining subtype, is usually studied in wartime relocation of populations. Recently, however, nuclear reactor accidents have generated studies of reconstructive evacuations (Johnson, 1986). Ronald Perry and his colleagues also emphasize the importance of voluntary versus coercive compliance structures for disaster planning. Coercive evacuation involves the forceful removal, by the military or police, of potential victims. The use of coercion complicates the logistics of evacuating populations. It creates added demands on the complex organizations dealing with the crisis. Brabek (1986) has inventoried available social science knowledge regarding evacuations. Drabek groups most empirical generalizations about evacuations into individual, group, complex organizational, community, societal, and international categories. As Drabek points out, many of these empirical generalizations do not have crosscultural and international generalizability. Other generalizations are based on one or two studies in the US. Clearly, they need to be tested and hopefully replicated in future studies of evacuations elsewhere throughout the world. I now reproduce some of these empirical propositions, albeit in simplified and somewhat changed form: 1. Adequate warning with sufficient lead time increases the probability of evacuation. The previous section of this report reviewed the literature on warnings, and tried to indicate the agreement that exists regarding effective warning systems. 2. The probability of evacuation increases a) if the the threat is seen as real and immediate, b) if the risk is personalized, c) if the person has a way of responding to the threat, i.e., a personal plan, d) to the extent that the person maintains vigilance over changing threatening conditions, and e) with increasing contacts with community organizations responding to the crisis.
12
3. Previous experiences with disasters increases the probability of evacuation, although not consistently across warning and disaster experiences. 4. The greater the person's perception of vulnerability the greater is the probability of evacuation, and such perceptions are impacted by site characteristics such as proximity to projected impact area, location, age of buildings, etc. 5. Recent migrants are more prone to evacuate than long-term residents. 6. Age is positively associated with the probability of evacuation. 7.
Women are more prone to evacuate than men.
8. The probability of evacuation increases if the family members are together and available to evacuate, and if absent family members are accounted for before leaving. 9. Evacuation behavior is a primary group activity rather than an individual-level behavior. 10. The probability of evacuation increases if neighbors and significant others evacuate. 11. The majority of families leave in their own means of transportation. 12. Poorer families evacuate under official programs more often than families from higher social classes. 13. The probability of evacuation increases if families have members in high risk categories, such as elderly persons and young children. 14. Evacuees prefer to go to the homes of relatives and friends rather than to official shelters. 15. Evacuees decide by themselves when it is appropriate for them to go back home, and such decisions may conflict with official views of risk. 16. Institutionalized populations present serious difficulties to organizations attempting to evacuate them. 17. Organizations have more success in carrying out evacuation tasks when their members are familiar with the details of their evacuation plan. 18. There is often misunderstanding among organizations involved in evacuations regarding what each is doing and the scope of the problems. For example, there is the tendency to overestimate the number of people needing shelter, and officials are often unwilling
13
to order evacuations because of their perceptions regarding the legal and political effects of their acts. Emergency operation centers usually increase the coordination and overall efficiency of these organizations. As acknowledged by Quarantelli and Drabek, among other scholars who have considered the problem, the state of social science knowledge of evacuation behavior is rudimentary. It is true that, as indicated, there are some empirical propositions for which there are limited research support. However, empirical propositions, no matter how many and how well documented, do not make up for the present-day dearth of deductive theory in research on evacuation which would allow the testing of falsifiable predictions. Rather than theories, what we have today are conceptual frameworks that have been used to order in some logical and temporal fashion the more obvious conditions and processes affecting evacuations. Two of these conceptual approaches to evacuation are Quarantelli's and Perry and Mushkatel's. Quarantelli's Conceptual Framework. The five components of Quarantelli's model of evacuation are: the community context, threat conditions, social processes, patterns of behavior, and consequences for preparedness. The community. Quarantelli argues that the community provides a context for the disaster threat or impact. This context is the capability of the community to deal with emergencies. It includes resources, social links, and social climate. Communities have conceptual and material resources that can be used to solve the problems caused by disasters. Social links refer to the extent of integration, or the closeness and ability to work together of families, groups, and community organizations. Social integration is a variable condition of communities. An example of high integration would be communities where the police, army, fire, water, and telephone companies have a history of working together to respond to community crises and have developed and implemented a collective disaster response plan. Such plans facilitate the communication, coordination, and decision-making necessary for evacuation. Quarantelli's general point is that the social links among social actors in a community tend to cluster in space and time, so that often the resulting clusters of people, groups, and organizations do not interact among themselves. Communication, coordination, and decisionmaking is thus rendered less effective. The final element of the community is its social climate, a residual category in the framework. As such it is quite broad, including social, political, economic, historical, and psychological factors that affect resources and social linkages. Threat. The second general dimension of Quarantelli's evacuation model is the threat, or impact of a disaster agent. The threat creates a special situation in the community. This special situation or context is affected by the characteristics of the
14
disaster agent, situational factors, and the individual and collective definition of the disaster and the response to it. Disasters agents vary in their frequency, predictability, duration, scope of impact, destructiveness, speed of onset, and length of possible forewarning. These characteristics impact the evacuation preparations and implementations that are possible. Situational factors are unique to the communities, such as the time of day of the disaster or season of the year, especially if there are seasonal fluctuations in their populations. The final element in the threat is the definition of the situation, or how the threat comes to be understood by the impacted population. The general point Quarantelli makes is that people, and evacuees among them, through whatever means, develop collective beliefs of what is happening in their communities and then act on the basis of these beliefs. Social Processes. The community context and the conditions associated with the threat in turn create responses, what Quarantelli refers to as social processes. He identifies four social processes: communication, decision-making, coordination, and task manifestation. Communication refers to the means and channels used in information flow and the content of the messages that are transmitted. Decision-making refers to the evaluation and use of information to make a choice, such as ordering an evacuation or searching for a missing family member before evacuating. Coordination refers to the relative presence of joint activities, and includes conflict. Finally, task manifestation refers to the specific behavior that people and organizations enact as they bring about an evacuation, from readying transportation equipment to restructuring transportation routes to make the mass movement possible. The general point Quarantelli makes is that communication is necessary for decision-making. Decision-making can lead to coordination of efforts. And coordination can bring about task manifestation. Behavior. The above-mentioned social processes produce patterns of behavior, such as warning, evacuation, shelter, and the eventual return of the population. In turn, these behaviors may have long term impact on the community's resources, social linkages, and social climate. Warning behavior and response is logically prior to evacuation, for as we have defined it, evacuation is the actual physical flight behavior. Part of the model is sheltering, which includes the behavior of evacuees at their place of destination. The analysis of behavior also includes the study of return behavior, or what happens during the trip back to the community and the period of readjustment. Consequences. Quarantelli argues that after the evacuation is over and people have returned to their places of destination the community may experience substantial and permanent changes in its resources and their use, the social links available for future crises,
15
and its social climate. In turn, these changes may help communities become better prepared to respond to future crises. Thus, the model is recursive rather than linear in its causal assumptions. Perry and Mushkatel's Conceptual Framework. Perry and Mushkatel's (1984) conceptual model is developed in path diagram form. they identify four sets of most proximate predictors of evacuation behavior. These sets are the family context, the adaptive plan, the sense of personal risk, and the warning belief. Warning belief is the determination that a real environmental threat exists. If such a determination is made, people assess, by noting their proximity to the impacted area and other personal vulnerabilities, their level of personal risk. If people feel that both the warning and the threat are real and that they are in danger, they begin to make a plan to protect themselves, or to activate previously formulated plans of action. The model includes the family context, to incorporate the well-established empirical generalization that most people will not evacuate unless their family members are accounted for. Perry and Mushkatel, both in the path diagram and in the text of their monograph, also identify less proximate predictors that impact the four aforementioned variables. They write that family context is influenced by kin relationships. Warning content, or the extent to which the warning message is properly constructed, influences both the warning belief and the sense of personal risk of people. Even further removed from the causal explanation of evacuation behavior are people's kin relationships (frequency and nature of contacts) and community involvement (nature and frequency of contacts with friends, neighbors, participation in voluntary associations), which influence the warning content of potential evacuees. In contrast to what was the case when Quarantelli did his influential review of the literature in 1980 (Quarantelli, 1980), evacuation research has blossomed in the ensuing years. A number of examples attest of this new-found popularity. Aguirre (1991) tested some generalizations regarding evacuation, using information from Cancun, Mexico, in the aftermath of Hurricane Gilbert. He shows that, as in the US, the majority of the evacuees found shelter in the homes of friend, neighbors, and relatives and were gone from their homes one week or less, and that calculations of personal risks were the best predictors of evacuation behavior. He also shows that socio-demographic variables such as the size of the household, and the gender, age, and marital status of the respondents were not very useful predictors of the behavior. Baker (1993; 1991; 1990; see also Southwood and Chin, 1987) is also a recent example of research efforts by geographers and city and traffic planners to solve the problems attending the protective evacuation of populations threatened by massive hurricanes and floodings. Another example that research on evacuation is thriving is the recent August, 1991 special issue of the International Journal of Mass Emergencies and Disasters
16
(IJMED) devoted to the topic. The articles in this special issue are excellent and should be carefully studied by anyone interested in the topic of evacuation. The special issue includes Barbara M. Vogt's research on the issues attending nursing home evacuations, D. Gillespie and S. Murty's assessment of organizations involved in evacuations, S. L. Cutter's analysis of evacuations produced by chemical accidents, C. L. Streeter's argument regarding redundancy in social systems and its import to disaster planning and programs, and Fitzpatrick and Mileti's article on warning and warning systems. It is a study of the characteristics of warning messages that have the effect of bringing about the perception that evacuation is warranted. These characteristics are the source, consistency, accuracy, clarity, certainty, sufficiency, guidance, frequency, and channel of the warning message. The IJMED also includes John Sorensen's study of the factors affecting the timing of evacuation departures,in which he concludes that the personalization of the warning is the only variable impacting the timing of flight, as well as Robert Stalling's study on the problem of ending evacuations. Stalling contrasts evacuations in natural events and manmade accidents. He points out that evacuations caused by chemical spills occur more often in a context of community conflict and distrust of authorities, so that authorities' all-clear messages often are not believed by the evacuees. T. Drabek's article is yet another important contribution in the special issue. Drabek extends our knowledge of evacuation by considering the evacuation readiness of corporations in the U.S. tourist industry. SEARCH AND RESCUE. During close to forty years disaster researchers (Fritz and Marks, 1954; Form and Nosow, 1958; Fritz, 1961; Barton, 1969; Drabek et al., 1981; Durkin, 1989; Krimgold, 1988, 1989; Wenger, 1987) have endeavored to understand what accounts for the relative success of search and rescue (SAR) activities in disasters, to include factors such as the nature of structural and nonstructural damage to the built environment (Culver et al., 1975; Hart, 1976; Anagnostopoulos and Whitman, 1977; Hasselman et al., 1980; Tiedemann, 1989; Stubbs et al., 1989; Lechat, 1989), the epidemiology of SAR events (Lechat, 1976; Glass et al., 1977; 1979; Lechat et al., 1985), and the effectiveness of medical services (Quarantelli, 1983). Recently, I have argued (Aguirre et al., forthcoming) for the importance of social organization in search and rescue in disasters. Wenger (1990, and literature cited therein) has summarized the consensus among SAR specialists to the
17
effect that volunteer and emergent group response is massive and that the initial activity is accomplished by volunteers and emergent groups. There is also agreement among specialists that the SAR behavior of these volunteers is of crucial importance because the chances of live rescue decrease rapidly after the initial "golden" hour. Buried and entrapped victims are likely to suffer from injuries that require life-sustaining intervention including compromised access to air, severe loss of blood and body fluid, crushing injury, and internal damage to essential organ systems. Professional SAR teams, despite the massive attention they usually receive from the mass media (Quarantelli, 1991), most often arrive too late to rescue alive significant proportions of victims. This is due in large part to the particular nature of the socio-geography of disasters in which professional SAR teams are especially hampered by problems of access, victim identification, inadequate resources, the breakdown of normal operational parameters, and the magnitude of the events. Finally, there is consensus that oftentime the integration and coordination of volunteers and SAR professionals is difficult, due in part to disagreement over rescue strategy, ambiguous authority relationships, and conflicts among independent agencies. EXTENDED EXAMPLE OF SAR. I would like to illustrate many of these principles with examples from research on the search and rescue activities (SAR) in Guadalajara, Mexico, in the aftermath of the 22 April 1992 gasoline explosion. We studied the SAR that occurred in the neighborhood of Analco, Guadalajara, Mexico, impacted by the gasoline explosion in which approximately 300 people died and 1120 were injured. Analco is one of the oldest neighborhoods of Guadalajara. It is part of the Sector Reforma. Analco has a very geographically-stable population. Many of the resident families have lived in the neighborhood for many years. SAR activities were dominated by the collective behavior of the neighbors of Analco. Most of the people that were rescued alive in the aftermath of the tragedy were rescued by these volunteers during the first phase of the postimpact emergency period (Form and Nosow, 1958). In trying to find out how they did it, we were impressed by the importance of selected social organizational features of the search and rescue activities that occurred, and by the linkage between institutionalized and emergent types of social organizations. It is not possible to understand search and rescue activities in Guadalajara in the aftermath of the explosion without considering the community's culture and social relations. The search and rescue processes occurred in a context of human solidarity, what can be called a "sociedad solidaria," in which, because of the culture and social history of the community, there is a much greater probability than in similar-sized cities in North American and Western European countries for people to know and to relate to others as full persons, and in which social relations assume greater immediacy and intimacy. The social formations of the peer group, the extended family, the neighborhood, and the Catholic Church constitute viable and active social networks in the lives of people. The social identitities derived
18
from these social formations impacted the search and rescue activities we studied. The importance of these social identities in people's participation in SAR is magnified by the absence of official disaster programs and plans, for as Quarantelli (1993) has argued, in comparison to the U.S. and Western European countries, Mexico has fewer disaster preparedness and responding organizations. The dearth of this type of organization is acute at the local level. Building upon the work of Drabek et al. (1981) and others, Olson and Olson (1987) documented that most lives are saved and victims rescued during this immediate postimpact period. Thus, it is this phase in SAR that we were most interested in studying. There are clear indications that the organization of SAR activities changed as the day progressed and the agencies responsible for crisis intervention established control over the SAR process. As it is typical in community disasters, a control and command center in charge of the societal response was formed in the evening of the day of the explosion. We met with forty three victims that had been buried alive by the explosion throughout the impacted area, and with twenty two volunteers who had participated in the direct rescue phase. They reported on their own experience during SAR and the experience of victims and rescuers near them. Separate interviews were also conducted with six neighbors who had participated in search and rescue activities immediately after the explosion and had subsequently formed themselves, in conjunction with some of their neighbors, into a search and rescue voluntary association. In addition, we also interviewed 5 Red Cross paramedics who had participated in the search and rescue activities. Multiple activities were undertaken by supporting volunteers throughout the broader community. Direct SAR activities within the impact area generated the need for services, tools, food and other commodities. Two systems emerged to resolve these needs. First, a communication system was created that involved the functional transformation of privately owned radio stations in Guadalajara into search and rescue communication systems serving the public. Second, volunteers established and operated a nascent transportation system that provided needed commodities donated by the public. Numerous students of mass media response during the emergency period of disasters have noted that mass media are often transformed into personal media (Waxman, 1973; Scanlon et al., 1985; Wenger, 1985; Wenger and Quarantelli, 1989). In effect, they alter their normal functioning and serve to transmit personal messages, relay personal information, and request information from specific individuals and groups. As such, they take on a new, emergent role in the emergency response system. This pattern was observed in Guadalajara. There are two major radio station corporations in the city. We obtained information from one of them. Immediately after the explosion they ceased normal
19
operations, and for two days helped structure the societal response to the explosion. The corporation owns five radio stations throughout the city. In what is a rather unusual pattern of response for U.S. corporations of its type (Quarantelli, personal communication), they reformed the stations into a network that functioned during the first two days of the emergency period. Moreover, their transmission power was supplemented by seven radio field stations donated temporarily by a major local manufacturing firm. These seven field stations were positioned in key places throughout the city such as the Red Cross, the morque, the local stadium were the homeless had congregated, and the hospital where many of the victims were receiving treatment. The field radio stations were manned by volunteers under the supervision of professional staff of the corporation. They helped the public in finding the whereabouts of lost persons and of needed supplies. Initially, before the system was supplemented with the loaned radio equipment, it was used to satisfy the immediate needs for tools, supplies, and food of the people manning the boundary areas and conducting SAR in the impacted city blocks of Analco. People in the impacted area would call the stations relaying their direct needs. The stations would then request the voluntary donation of the needed resources from people in the rest of the city. Once their requests were satisfied, the stations would communicate it to the rest of the population in an effort to minimize needless duplication. Later on, the expanded radio communication system was used to help locate missing persons. In this capacity the network was made more effective by the use of a computer program that compared alphabeticallyordered lists of names of victims kept in the various locations in the city involved in the emergency. A second social organizational emergence in the city were the groups of transportation volunteers who distributed food, tools, and other needed resources to people in the boundaries and the impacted areas. We know very little about the organization of this category of people involved in transporting goods and services. Reportedly, as it occurred during the mass assault phase of the 1985 Mexico City earthquake (Dynes et al., 1990), many of these volunteers were young university students with their own means of transportation (c.f., Arreola et al., 1986). College students were in Easter recess. Apparently, they were recalled by their local colleges and organized by the colleges to participate in the emergency phase of the societal response to the explosion, in conjunction with the activities of the radio stations. The filter (or boundary) area was dominated by the army and police. Reportedly the army, reflecting its importance in Mexico's national disaster planning, was one of the formal organizations (the others were the Mexican Red Cross, the Green Cross, and the Guadalajara Police and Fire Departments), that deployed most quickly into the impacted area, setting up perimeter controls in many of the impacted city blocks before the first hour after the explosion. Testimony from the neighbors in the impacted area shows that the army and the other formal organizations were very well received by them, that they
20
were able to work together with little friction, and that their search and rescue efforts were very much appreciated by the citizenry. FIGURE ONE ABOUT HERE The very narrow and mostly rectilinear pattern of destruction is despicted in Figure One. In this area the blast destroyed the streets, sewer and potable water systems, more than 1,000 houses, and an unknown number of businesses. It extends for more than nine kilometers, from the intersection of Gantes and 20 de Noviembre at the northwest end segment of the line to the intersection of Yunque and Calzada L. Cardenas, in the extreme southeast segment of the line. People were killed or injured either on these city streets or inside the homes lining these streets (c.f., Alexander, 1989). Inside the homes, most of the people who were killed or injured were in the rooms closest to the streets where the explosion occurred. Cultural Influences. The explosion occurred at approximately 10.07 a.m. This time of day (Lomnitz, 1970) determined to some extent the different risks of victimization of population categories. An important cultural factor determining the gender and age of the victims of the explosion are the rituals associated with eating. The custom in Guadalajara is to eat supper around 9 p.m. Many families leave the cleaning of the kitchen and dining areas for the morning after. The customary cleaning sequence in the homes is kitchen, living room, and bathrooms and bedrooms last. This is done since typically in Analco the living rooms and kitchens are the areas of the homes where visitors have more visual access, and are thus straightened out first. People in social categories that customarily clean the homes, such as adult women, were protected to the extent that the kitchen and dining spaces were in the back of the homes, away from the street and thus away from the source of the blast. Reportedly, this is the use of space that predominates in the destroyed houses. Another cultural factor that determined the age composition of the victims was that the gas explosion occurred during the Easter vacation so that school-age children were at home. Many were playing in the streets when the explosion occurred. This accounts for the many children victimized by the blast. A third cultural pattern determining victimization was the use of space in the streets. Some of the streets impacted by the explosion were part of intercity bus routes. Small restaurants fronting these city blocks serve breakfast for travelers. Many of their clients became victims of the explosion. Pre-Impact Behavior. Besides these cultural mechanisms that dictated the demographic characteristics of the victims, the chance movement of people inside and outside their homes also affected their exposure to the effects of the explosion. An example is of a single mother and her baby. She leaves the baby in her crib to go to the kitchen to fetch food for the baby. As she is crossing the interior patio of her home the explosion
21
occurs and her baby is killed. unharmed.
She escapes physically
For at least 12 hours prior to impact, neighbors were aware that they were in danger of a gasoline explosion. The day before the explosion the Fire Department cooperated with the Transit Police to cordon off part of the area (corner of Gante and Analco) that eventually exploded. At least one radio station had deployed personnel to the area endangered by the gasoline spill hours prior to the explosion. Red Cross and Fire Department personnel had been deployed the night prior to the explosion to the city blocks that were eventually destroyed. During their deployment they confirmed the presence of pressurized gasoline vapor columns spewing out of drainage manholes and of housewives who complained to them of the presence of gasoline in their toilets. The Red Cross treated a police officer, a fireman, and a worker from the Mexican petroleum corporation (PEMEX) for gasoline inhalation. The day of the explosion a local newspaper had published the news of the presence of gasoline in the drainage system in its morning edition. The existence of these unofficial but nevertheless important cues of imminent danger probably meant that an unknown number of neighbors evacuated their homes prior to the explosion. It is not know the extent to which such evacuations occurred and were effective in protecting lives, or the social and demographic characteristics of the families that left their homes. Behavior After the Explosion. The explosion affecting the entire area occurred very quickly, almost simultaneously. The only indication of warning we obtained is of people looking down the streets and seeing a rapidly disintegrating landscape advance towards them. Those who survived turned away from the center of the street where the drainage pipe that blew up was located. Apparently the explosion of the pipeline was not simultaneous. Rather, it was nearly simultaneous throughout the city blocks that were destroyed. The noise of the explosion has been described as a very loud hissing sound. Corraborating the findings of Norris Johnson's (1987a; 1987b; 1988; see also Baker, 1960) research on the behavior of people in extreme situations, the behavior of the victims that we talked to was marked by the continuation of preexisting motivational, normative, and value orientations. Victims, under the very difficult conditions of being buried alive, often in imminent danger of death, continued to be social beings. As we will show, their actions during entrapment showed the constraint generated by their membership in primary groups and other meaningful social categories. Victims acted cooperatively during entrapment. They provided information to potential rescuers about other people in the rubble and thus assisted in increasing their chances of surviving the explosion. Many examples of these patterns are available from the record: A. A man and his two nephews are having breakfast in their home. The explosion buries them alive. The man reports experiencing a great amount of
22
difficulty breathing. He can hear his two nephews near him in the rubble. He talks to them and synchronizes their scream for help at his count of 3. Eventually, people hear them and save them. B. A mechanic is protected by a heavy bench used in the shop that lands on top of metal engines that were being repaired. The bench and engines provide him a cavity in the debris. Within the cavity thus formed he is buried from his waist down. There are three other people buried with him and he hears two of them that are in very close physical proximity talk to each other. The rubble is very unstable and the men agree that they must tell the searchers to pull them out simultaneously, for otherwise the one who is left behind will be seriously injured by the resulting collapse of the pile of debris in which they find themselves. C. A mother and her two children drive up to her friend's house. She parks her car in the side of the street opposite the house. As she gets ready to get out of the car and lock it, the children rush out of the car and knock on the front door of the friend's house. The explosion dislodges the front door amd one side of it falls on top of the car parked in front of the house, providing a protecting space for the two children below it. Soon afterwards, the brother helps his younger sister escape from their entrapment. In turn, his sister calls attention to rescuers nearby about where her brother and mother are entrapped. The rescuers then start digging up the mother. The mother can hear through the rubble and tells the volunteers that she is alright and first to rescue her son. D. Neighbors, in conjunction with Army and Red Cross personnel, begin to look for a mother and her six children that lived in a one-room apartment in a multi-family building (casa de vecindad). When they find the woman in the rubble she is embracing three of her children. A table protected them from getting hurt by pieces of the roof and walls, and they are alive. The mother then tells the rescuers that her other children are still buried in the room. They continue to search, and twenty minutes later find them. Two of the children are rescued alive. However, her five year old daughter is dead of massive head injuries. E. Rescuers hear through the rubble a man calling for help. He has been protected by a heavy slab of concrete, and they rescue him alive. The man then tells the rescuers about his three children still in the rubble. The rescuers eventually find them, but the young boys are crushed to death. F. A woman owns a small restaurant. She is serving breakfast to four men with whom she has a relationship as clients. She is the first to be rescued from the rubble of her restaurant and immediately reports to her rescuers the last location of the entrapped clients in it, thus facilitating their rescue.
23
G. A young adult shares a bedroom with his younger brother. They are asleep when the explosion occurs and are buried alive. He reports feeling initially confused, but soon begins to struggle to free himself from the debris and succeeds in doing so after ten minutes or so. Just as he frees himself his other brother arrives and he is able to alert him that their brother is still trapped. He is injured and cannot participate in the rescue, but directs it by telling the rescuers where the brother's bed was in the room. The respondent refuses to leave for the hospital until his own brother is rescued alive, for initially the rescuers ignored his directions and searched for the brother in the wrong side of the room. Even those victims who were alone at the time of the explosion continued to be social beings engaging in imaginary social interaction with significant others. A victim reports conversing with the Virgin of Talpa, the saint of his devotion, and with many of his dead relatives during entrapment, He reports meeting his dead father, until then unknown to him. A mother, also buried in the rubble, reports thanking God for her good luck; her children were visiting relatives outside the city, away from the explosion. Victims were able to hear, despite being buried often in one or two meters of rubble, what people on the surface were saying and doing. It was another way they had of maintaining social ties to the world around them. As shown by some of these examples, many of the victims actively participated in increasing their chances of survival and in their rescue. Some victims mentioned that they moved their bodies ever so slowly to create more space for themselves among the rubble that trapped them. Others called attention to themselves to help rescuers know their location. Perhaps one of the most dramatic cases we encountered is that of a man who tells his son to try to move his arm towards the surface of the earth to help him locate him. He finally rescues the child and as he is moving out of the location of the rescue he turns and sees the arm of his son's friend protrude from the ground. He returns, calls others to continue helping him, and the friend is saved. Response Time and Composition of the Initial Rescuers. Similar to the findings of research in other major disasters (e.g., De Bruycker et al., 1985; Durkin, 1988; Durkin et al., 1989; Noji et al., 1990) and accidents (Quon and Laube, 1991) documenting the importance of quick response in saving lives, the limited evidence that we have shows that most of the victims in Guadalajara that were rescued alive were rescued during the first two hours immediately after the explosion. And as in other disasters (e.g., Lechat, 1976; Drabek et al., 1981; Abrams, 1989), they were rescued by their neighbors, kinfolk, and after the first hour, by Mexican Army, Guadalajara Police and Fire, and Red and Green Cross personnel. Very few people were pulled out of the earth alive after the first two hours. A man was extricated alive 8
24
days after the explosion, but his rescue was an unusual event. Indeed, none of the 43 victims we contacted reported being trapped more than two hours. The testimony of the four rescuers that could remember the figures corroborates the victims's experiences. Among them they found seven dead and thirty two persons alive before noon. From noon until nightfall they report finding one person alive and thirty seven dead. Records from the Mexican Red Cross in Guadalajara indicate that in the first two and a half hour after the explosion 265 victims were admitted to their facility. 49 were dead on arrival (31 males, 18 females) and 3 died after arrival. The Green Cross reports finding, near their field operation center, 5 victims after 7 p.m., all dead. While incomplete, this evidence is in agreement with what is known from other disasters about the importance of quick extraction for rescuing people alive. Reportedly, SAR dog teams arrived from Mexico City 26 hours after impact, helping find corpses throughout the impact area. A local SAR dog team was more successful, for it movilized three hours after the explosion and found 2 live and 3 dead victims. Emergent Organizational Patterns of SAR Groups. During the first phase of the search and rescue activities most of the rescuers were the neighbors, associates, and relatives of the victims, as well as personnel from the Mexican Army, Red and Green Cross, and Guadalajara Fire and Police departments, the agencies that had sustained and dispersed involvement in the societal response to the explosion. We were very interested in understanding the division of labor, leadership structure, and role relationships characterizing the groups they formed. The evidence shows that the degree of formalization in these search and rescue groups was minimal. Apparently, high levels of formalization were neither necessary to achieve the tasks at hand nor possible under the typical circumstances. Initially, the primary social formation that carried out search and rescue activities were the people residing in the neighborhoods impacted by the explosion. This pattern is similar to the immediate reactions of people impacted by disasters elsewhere in the world (Quarantelli, 1988). Neighbors grouped themselves by city block and cooperated with each other in searching and rescuing victims of the explosion. Neighbors, friends, and relatives had privileged information about the customary activities, habits, and probable whereabouts of known or potential victims and of the layout of their residences. This knowledge was of paramount importance in the search and rescue process, and is an important advantage these emergent groups have over formal SAR organizations. It allowed its possessors to act as keynoters who distributed the volunteers and personnel from service agencies that quickly arrived from outside their city block into the various searches that were being conducted in it. The division of the neighbors into city blocks and the ability of the resulting SAR groups to distribute subsequent volunteers into their SAR activities was a
25
very important form of division of labor that emerged to respond to the crisis. It was an emergence characterized to varying degrees by both new social norms and new social relationships (Weller and Quarantelli, 1974). The rudimentary divisions of labor that emerged among SAR groups included a number of ephemeral roles. For example, in Guadalajara, as is true of other cities in Mexico and the developing world, houses have their own gas tanks which are used for cooking. Immediately after the explosion, the potential leaks in these tanks presented a real threat to the searchers. To deal with the threat some of the searchers volunteered and would enter the homes throughout their city blocks to disconnect the tanks. Yet another rough division of labor occurred in the control of pedestrian traffic into and out of the impacted city blocks. Some of the neighbors, in conjunction with social control personnel, took over this responsibility, which at time involved challenging people they did not recognize as neighbors, requiring them to justify their presence and insisting that they would leave the premises if they could not state a legitimate purpose for being there. Still another example of division of labor centered on the actual search activity. The removal of rubble necessitates both the picking up of the pieces of debris as well as their transportation away from the places in which the digging is taking place. These tasks required two types of workers acting in close coordination. Characteristically, there would be three of four people picking up the pieces. And surrounding them would be other rescuers ordered in lines, usually away from the center in each of the four cardinal directions. These people actually moved the piece of rubble, passing it hand to hand away from the dig. During the Mexico City 1985 earthquake response, most of the volunteers removing the rubble were males (Dynes et al., 1990, 86-90). Our impression is that a similar pattern occurred in Guadalajara, although we do not have the survey information to determine the gender composition of the volunteer searchers in Guadalajara during the immediate post-explosion phase. This rough division of labor existed in places where medical personnel was not available. Thus, as is typical of disasters elsewhere, during the first 45 minutes or so after the explosion the people at the center of these search and rescue formations also extracted the victims, and they, or other volunteers, would transport them in private automobiles to places where the victims could receive medical treatment. However, once Red and Green Cross and other medical personnel arrived at the scene of the disaster, they ceased doing so, and paramedics would carry out the actual removal and transportation of the victims. During the initial phase of the response simple, hand held, small mechanical tools were most effective in helping people do their rescue work (e.g., Abrams, 1989). Lechat (1989) also reports that almost 97 percent of the injured victims trapped by the 1980 earthquake in Italy and evacuated to medical centers were rescued with bare hands, shovels, and ladders. In Guadalajara, some of these simple tools were heavy gloves to protect the hands, ropes to wrap around searchers as they entered
26
particularly dangerous places; lumber, especially 4 by 4 and 2 by 6 pieces with which to construct temporary retainers needed in some of the rescue sites; Wire cutters, small hydraulic jacks with wheels, used to remove heavier pieces of concrete, rock, and steel columns, metal buckets used to remove sand and loose earth, metal bars used to remove heavy objects, and heavy hoes used to remove debris and commonly used in agricultural work. Many of the respondents that had participated in the search thought that hydraulic arms or small excavators, mostly unavailable during the initial response phase, would have been very helpful in the SAR work. Upon arrival at the site, formal rescuers used the knowledge of neighbors to locate victims and made use of the volunteer manpower to remove debris and help in all phases of the rescue efforts, while volunteers relied in their specialized knowledge of extrication and transport of the victims. Paramedics were heavily engaged in victim transport, and cooperated with volunteers and personnel from other agencies in the SAR activities. Some of these volunteers were medical doctors and nurses who very rapidly joined the rescue efforts at the site of the explosion. Thus, the organization of SAR activities and their location at the site of the explosion changed rapidly, for it was supplemented by the efforts of formal organizations. For example, the Red and Green Cross teams began to organize civilians into SAR groups of 20 persons or so, and these teams were augmented by military, fire, and police personnel. The excellent integration of the Mexican Red and Green Cross personnel with the SAR volunteers was facilitated by the semi-formal organizational structure of these agencies. In contrast to EMS organizations in the US, these organizations are heavily manned by volunteers who often have limited training and few resources for on-site medical treatment. Many of these paramedics are local citizens; in Mexico, the name paramedic itself does not indicate the para-professional status and training that it does in the US. Thus, the social distance between paramedics and informal groups was considerably less than it would have been in the US, and led to greater degrees of cooperation and less conflict among the rescuers. Paralleling the response of volunteers, the response of official agencies involved in SAR was dominated by intra and inter-organizational normative and role emergence. All agencies were augmented by volunteers. The Fire and Police deparments also received sustantial assistance from neighboring departments. The impacted city blocks were divided among them. Thus, the Red and Green Cross sectored off the impacted area for systematic search and extrication of victims, while the Army, Police, and Fire departments similarly sectioned off the same city streets, creating overlapping jurisdictions with the Red and Green Cross organizations. In the absence of an interorganization plan, these geographical subdivisions helped the agencies assign responsibility for the SAR effort and distribute the needed resources.
27
This division of labor among the various agencies was unplanned prior to the explosion, for while Guadalajara had a written disaster plan it had not been implemented in the past. Interagency cooperation and coordination emerged from the bottom up rather than from the top down. Feedback from the teams in the field guided the relationships between agencies in their response rather than pre-established, planned procedures providing the context of coordination. Reminiscent of Form and Nosow's (1958; see also Quarantelli, 1988) findings regarding the importance of family and gender roles on disaster-relevant activities of volunteers, the evidence we have show that the concerns of people engaged in search and rescue was first for their kin, second for their immediate neighbors and other nearby residents, and finally for residents of nearby blocks; people farther removed from their spheres of everyday interaction. People did not participate in the search and rescue efforts at random. Instead, their participation was a function of the strength of their preexisting social linkages and interdependencies with the victims and fellow rescuers. Their search and rescue efforts were part of a stream of ongoing social relations in which people participated, and from which their activities on behalf of their relatives, friends, acquaintances, or even strangers obtained meaning. The rescuers prioritized life; all human life was precious for them but the lives of those socially closest to them was deemed more important. One of the most important findings of the field work is that the chances of people surviving the blast were directly proportional to the presence among the searchers of a person or persons who cared for the victim and who knew the victim's likely location at the time of the blast (e.g., Barton, 1969, p. 129; Abrams, 1989; Mileti and O'Brien, 1991). The accuracy of the information they provided about the likely location in the rubble of the persons presumed buried by the explosion very importantly determined the length of their entrapment and the relative effectiveness of SAR efforts in rescuing victims alive. In Guadalajara, these people often were kin or neighbors of the victims, although at times they were people who worked in the same establishment, or who had religious or other stable and primary relations with the victim. This pattern is reminiscent of the findings of research that indicate that family context is an important determinant of the death rate of victims trapped by earthquakes (Lechat, 1989). In an important pattern, these significant others acted as proxies for the victims. The victims were assumed to be in the rubble and could not act to direct their own search and rescue, and so their proxies did it for them. They served to remind the neighbors and others in the immediate environment that the victim was missing and that they had an ongoing pattern of reciprocity with the victim and the proxy which at this critical juncture demanded their attention and enactment. The proxy actors injected meaning into the search, helping define the situation in terms of their own priorities. Moreover, many of the proxies constituted the nucleus around which search and rescue groups formed. Some examples will help elucidate the general pattern.
28
A. Residents of a "casa de vecindad" alert others and begin searching for their neighbors, an elderly couple buried in their one-room apartment. They eventually find them in a semi-conscious state after more than two hours of entrapment in more than one meter of rubble, and are able to pull them out alive. B. A lawyer is a life-long resident of the neighborhood. He hears the explosion and runs from his house, two city blocks away from the impacted area, to the house of his father-in-law, where his wife and children are visiting. Once he gets there he verifies that no one is severely hurt. However, he is informed by his wife that his nephew is missing. She tells him that the boy had left to buy candy at the corner store. The store, part of the landscape, is gone. Nevertheless, he remembers that there was a tree in front of the store, finds the remaining tree stump, and from its position reconstructs in his imagination the most likely place where his nephew is buried. He begins to dig through the rubble. Others see him, ask him who is missing, and begin to work with him. Eventually, they find the boy and are able to rescue him alive. Immediately thereafter, his wife arrives and tells him that his daughter and niece are also missing. He asks his nephew to remember where he saw the girls last. Eventually, the child tells him that they had told him that they were going to play an electronic game at the nearby parlor. Soon, neighbors and other friends of the man learn through word of mouth that the girls are missing, approach him, and ask him where he wants them to search for them. He begins to distribute them along the route which the girls were presumed to have taken. He uses the remains of housing plumbing exposed by the explosion to reconstruct the location of buildings and other elements in the now transformed landscape. All during the day he receives information through word of mouth about other families in the neighborhood known to him that are also searching for missing relatives, but he cannot join others helping them because of his own troubles. Six hours later all hope is lost that the girls are still alive and he and his wife are interviewed by a television crew covering the disaster. It is only then that his daughter learns that she is presumed dead and sends a message with another friend that she and her cousin have survived the explosion and have found shelter in the home of yet another friend of the family. C. A man, born and raised in the neighborhood, is at work as a security officer in a major retail store away from his house situated in one of the city blocks which would be destroyed by the explosion. Before going to work he has been concerned with the gasoline threat, and now makes a telephone call to his home. As he is talking to his mother the explosion occurs. He rushes home in uniform and becomes the leader in organizing the search and rescue activities of the neighbors in his city block. All the time, his little son is
29
missing. The boy was on the street at the time of the explosion, and it is unclear to anyone the location where he could be buried. His father continues to search for him even as he coordinates the search for others in the block. Eventually, after more than 24 hours have passed, he and the friends that have helped him in the search are able to locate and rescue his remains. In the interim, he and his friends searching for his son successfully block the entry into their city block of heavy earth removing equipment for fear that it will accidently mutilate and kill his son. D. An engineer is a lifelong resident and a member of a large extended family unit in the neighborhood. He rushes to his home immediately after the explosion. His mother has escaped uninjured. One of his cousins tells him that yet a third cousin is unaccounted for and presumed buried in the rubble. He then joins the entire family in searching for the missing member. Eventually, after a couple of hours of searching for him they find the missing cousin helping neighbors in an adjacent block. Once he learns that his cousin is safe he joins one of the human chains removing rubble from a site in which a search is taking place. E. An automobile mechanic is talking to two of his friends in the shop where he works. His twenty one year old son leaves the group to retrieve a tool in his nearby house. The explosion occurs when the son is inside the house, and he returns to where he left his father, only to find that the shop has disappeared. Since he knew where his father and friends were immediately before the explosion, he begins to dig for them. He is joined by other neighbors. Eventually, he and his father are able to talk to one another. However, the son cannot determine the location of his father in the pile ot rubble. By accident, he finds a carburetor hose his father had bought that morning, and he begins to use it to probe the pile of rubble for his father. From below, the father eventually sees the light of day through the debris, and begins to direct the probe. After some misses, he grabs a hold of the hose. The son then uses the hose to direct the rescue effort, and the father is rescued alive. F. A group of Red Cross paramedics joins others in digging for a man who is presumed buried inside his automobile. They help dig the car out of the rubble but the man is not found in it. As the excavation is almost completed someone removes a heavy slab of concrete, and to everyone's surprise a dog emerges from the small cavity thus revealed. The dog rushes to a puddle of water nearby, and after dipping in it returns to the pile of rubble and begins to dig. The men then begin to dig where the dog indicated, and eventually they succeed in finding the dead man. Afterwards, neighbors identify the dead man and Rambo, his dog. Another example also shows, by its absence, the importance, in rescuing people alive, of quick
30
identification of their likely position in the rubble. In this case, the mother is at work and the father and his three children are buried by the explosion in the rubble of their two-story home. The father is able to escape and calls for help in locating his missing children. A SAR group is formed. It is led by an army officer and composed of Army, Red Cross, and Green Cross personnel as well as neighbors. Unfortunately, however, their work is hampered by the man's inability to give them an accurate identification of the location of the children in the house at the time of the explosion; the father is disoriented and is unable to give accurate information to the rescuers, telling them to dig in the wrong place. Eventually, after more than one and a half hours of trial and error by groups of people digging in various parts of the rubble, the children are found. They had been sleeping in the same bed, and are now found one on top of the other. The third, at the bottom of the pile, is dead. Her two brothers survive. Implications. These patterns show the importance of pre-existing social organization in the formation of emergent social organization. The fact that people grouped themselves by city block or neighborhood subunits in carrying out SAR activities extends social science understandings regarding the structure of emergent, volunteer search and rescue groups. Perhaps more significantly, the observation that the chance of being rescued alive were directly proportional to the presence among the searchers of a person or persons who cared for the victim and who acted as social proxies for them, is a powerful testimony to the influence of social integration upon physical well being. These results document once again the validity of the principle of continuity advocated by Quarantelli and Dynes (1977). The empirical generalization supported in this and other studies of SAR in disasters, to the effect that most victims are saved by volunteers rather than by professional rescuers, is at first sight an interesting conundrum, for the greatest effectiveness is achieved with the least amount of technical know-how. The conundrum is solved once the social organization of the emergent collectivities of people involved in the first phase of the SAR activity is understood. Pre-existing social relations provide the foundations for emergent social relations and organizations in the immediate aftermath of disasters. This emergent social organization is the most important tool for saving lives. There is considerable urgency in understanding such emergent organizations so as to plan for it, making them part of disaster preparedness measures. Zurcher (1968) analyzed an emergent ad hoc group of volunteers that formed 36 hours after the massive impact of the 1966 Topeka tornado. A number of the characteristics of Zurcher's model also apply to the present case. Thus, our respondents experienced a sudden and unplanned disruption to their social world, felt the need to act to reasert their control of the situation, took on new roles, and constituted ephemeral, short-lived SAR groups that experienced rapid albeit limited conventionalization. In contrast to the Topeka SAR emergent group, however, and replicating the findings of research on collective behavior, in Guadalajara the
31
cohesion of the SAR groups and relationships we studied often came from preexisting social identities rather than from the shared need of their members to "restructure activities." Preexisting social identities and relationships oftentime determined the occurrence and type of participation in the search and rescue activities that took place. There are important continuities between institutionalized social life and the collective behavior of SAR activities. The emergent noninstitutionalized nature of volunteer SAR groups provides a vivid example that such collectivities do not emerge from a vacuum. Not only are there always elements of the traditional social structure embedded within collective behavior entities, but their emergent division of labor, role structure, activities, and effectiveness, are also dependent upon prior social relationships and forms of organization. Dynes and Quarantelli (1980) identified four types of disaster volunteers, what they term organizational volunteers, group volunteers, volunteers in expanded roles, and volunteers in new roles. Clearly, as Dynes (1970) theorized years ago, in the aftermath of the Guadalajara explosion there was a great deal of group emergence involving extending, expanding, and emergent organizations. We have shown that preexisting networks of human relationships were used to alleviate novel and unexpected collective problems that demanded immediate attention. Thus, while it is probable that volunteers representing all of these types could be found participating in the SAR during the days following the explosion, during the period immediately following the explosion the last two types of volunteer appeared to us to have predominated in the SAR efforts, albeit not in as clearly differentiated a form as represented in the analytical types. Instead, people expanded their sense of responsibility towards each other, and often they did so by becoming members of new emergent groups which carried out SAR activities. Second, our findings once again seriously question the validity of "breakdown" models of social organizational patterns in disaster. As is true in other disasters, television reports of the Guadalajara disaster depicted throngs of people moving seemingly at random at the sites destroyed by the explosion. From these depictions it is relatively easy to assume that the people were disoriented immediately after the explosion and had lost their ability to enact social roles. Indeed, in light of the marked absence of formal planning and organization, it might have been expected that widespread confusion, lack of coordination, and civil panic had occurred. Our data suggest a considerably different picture, one in which naturally occurring social networks provided an effective and fluid framework for the formation of a relatively successful search and rescue effort. The social organizational patterns of the SAR activities that occurred in Guadalajara reaffirm once again our understanding regarding collective behavior at the sites of disasters. As opposed to breakdown, panic, or antisocial behavior, the seeming disorganization and aimless movement of people was the result of persons acting individually or as part of collectivities who were
32
trying to accomplish multiple individual and collective goals in a reduced and relatively well defined space under important felt time constraints (c.f., Fritz and Mathewson, 1957). Creative problem-solving and rationality were much more appropriate descriptions of their actions than panic (Aroni and Durkin, no date, p. 30). Third, we have observed the important role played by a person's imagination in survival--whether it is to reconstruct a destroyed area, to adopt tools for new uses, or to forget pain and fear by talking to the saints or to dead relatives. Fourth, these findings provide a comparison with what has been reported about the search and rescue efforts in the 1985 Mexico City earthquake. In describing the initial search and rescue effort following the earthquake, Dynes, Quarantelli, and Wenger (1990) describe patterns of activities that are very similar to what was observed in Guadalajara. In both disasters the initial search and rescue activities were undertaken by volunteers and emergent citizen groups. The vast majority of those who were rescued and survived were extricated within the first two days. Pre-existing social relationships, such as neighborhood and work place relationships, served as a basis for the emergence of new SAR groups. The major difference in the search and rescue efforts in the two disasters appears to focus upon the formal established organization, not the volunteers. In the Mexico City earthquake it took about three days for the formal, professional rescue and response organizations to become operational. For the first two days there was little involvement and no overall coordination by formal organizations. In Guadalajara the formal units were involved much more rapidly, perhaps because the impact area was more concentrated, communication and transportation lifelines were not severely affected, damage assessment was facilitated by the ecology of the impact zone, and considerably more national attention had been given to emergency planning in the aftermath of the 1985 Mexico City disaster. Fifth, there is a pressing need in the planning for disasters to reestructure the societal reaction to massive disasters so as to give more weight to the emerging social organizations that volunteers create in the aftermath of disasters (Forrest, 1978; Drabek, 1987). This social organizational emergence should be seen as the most important societal resource available in the immediate aftermath of disasters. Disaster planners should facilitate the work of search and rescue collectivities. The recognition of its importance should help reevaluate the function of volunteers, which until now have been commonly seen as appendages to the work of official agencies or as impediments to the work of these agencies. Part of the needed reestructuring should involve an international program of public education. For example, the cities of Los Angeles and Oakland, California, USA, have begun extensive public education and training
33
programs to prepare citizens and neighborhood groups to undertake post-impact tasks, including search and rescue, and international programs based on these and similar efforts are needed. There is a clear need to provide populations throughout the world, perhaps through cartoons or other graphic despictions, with guidance for the appropriate behavior of potential victims, and in the principles of sound SAR activities, especially in the excavation of unstable and demolished structures. Victims need to know what behaviors increase their chances of being rescued alive, and searchers need to know how and whom and what to ask (and search). Not all SAR efforts in Guadalajara benefitted the victims. We learned of one instance in which volunteers climbed a pile of rubble in which a woman was trapped. Their collective weight collapsed the internal cavity that had protected her, killing her. Much thought should be given to organizing an SAR educational campaign that would prove effective in preparing people to act collectively on their own behalf in case of massive disasters. There is also an international need to rethink public investment and use of disaster preparedness technology, to give greater importance to the purchase and effective distribution of hand-held tools used in SAR by emergency management agencies and volunteers, and to the pre-positioning, rapid deployment, and use in communities of appropriate light-weight, easily maneuverable machines such as hydraulic arms, jacks, and drills that could be easily integrated into the work of volunteer SAR groups. Currently, considerable planning has been undertaken at the national level in the US to create teams of mobile, stand-by teams of professional rescue personnel. While they obviously have an important role to play in difficult extrications and complex operations, professional SAR teams face the barrier of time and the lack of localized knowledge so critical for successful SAR operations. I would argue for increased attention to local and citizen training and capabilities. For example, one of those areas of capabilities involve rescue technology. We have found that simple hand tools and other "low tech" implementations were judged most effective by those who rescued alive most of the victims. The policy implication is obvious. Greater importance should be given to the purchase and distribution of handheld tools used in SAR, and to the use and rapid deployment of appropriate light weight easily maneuverable machines, such as hydraulic arms. CONCLUSION. The foregoing review of the planning, warning, evacuation, and search and rescue literature is not exhaustive. Rather, it is intended to convey the range of social science work that has taken place mostly in the US in each of these preparedness topics. The findings presented herein reflect primarly US culture and social organizations, and I have not reviewed the social conditions of implementation of social science knowledge about these topics conveyed in the preceeding pages elsewhere in the world. This review has not attempted to answer the problem that disaster planners in developing and Eastern European countries must ask themselves,
34
namely, how can they use this information in their own work. It is generally understood that the effective transfer of knowledge from one society to another is a delicate and complex matter. And so it is in the area of disaster studies. Thus, I would like to conclude this paper with the thought that implementation is very much linked to the social and cultural matrix in which disaster planners operate (Mileti and Sorensen, 1989b). My work on disaster preparedness and mitigation in Puerto Rico (Aguirre and Bush, 1992; see also Ali, 1992) illustrates this issue quite well, for the failure of disaster programs in the island is very much explained by their sociocultural inappropriatedness. In Puerto Rico disaster programs are based on conceptual schemes developed in North America. The programs are organized in accordance with administrative and sociocultural assumptions appropriate to the US and other core societies. When some of the programs, such as the one protecting the Puerto Rican shoreline, fail, the explanation for their failure is couched in terms of cultural resistance to modernization. The more mundane fact that the planners do not consider the effects of the social organization of Puerto Rican society on the likelihood of success of their disaster programs is not seen as relevant! Based on these experiences and on the results of other research efforts, and at the risk of sounding normative, I would like to venture the following observations. Planners should resist the impulse to use programs and technology that do not have a local support base. Thus, instead of depending on sophisticated stateof-the-art technology they should prefer, to the extent that it is available, the use of simpler, reliable, and proven native techniques. Planners need to become good ethnographers of previous local responses to disaster events. They need to know the history of community responses to disasters, the successes of previous plans, warnings, evacuations, and search and rescue efforts, and the political dynamics existing among the significant local institutions, neighborhoods, ethnic groups, social classes, and leaders. The importance of this type of knowledge for planners is a central theme in Drabek's excellent primer for North American disaster planners. It is an important principle for disaster planners elsewhere as well, worth repeating. Planners need to know how the people of their communities responded to the challenge of disasters. They also need to know the power structure of their communities, so as to organize constituencies to carry out the disaster agenda (e.g., Seitz and Davis, 1984; Tierney, 1992). Successful planning requires knowledge of the history of previous disaster events so as to derive improved solutions that, given the political realities of communities, have a good chance of implementation. There is a tremendous amount of innovation and human inventiveness in the societal responses to disasters, and some of that human creativity is worth preserving. Planners should try to
35
institutionalize the best of this ethno-knowledge, so as to make them available for future crises. Planners can derive from their studies of past disaster experiences an understanding of the social organization implicated in those experiences and the sort of planning that is feasable and has a chance of doing some good for their communities. In their work, planners will need to build consensus, sharing their tentative answers with local key informants before attempting to put them into practice, obtaining support through a multiplicity of strategies and tactics from the key players in the communities. Only from such ethnographic basis and from an understanding of local politics will planners be able to evaluate the extant principles of good planning, such as D. Wenger's directive that they should not plan for the worst possible case scenario but rather that they should plan for the average generic disaster event. To repeat, the problem with these directives, as with the knowledge base presented in this paper, is that most of them have been derived from studies in the US and that many of them assume the presence of interventionist, public-service oriented government bureaucracies and programs. Their generalizability to other countries is still not known, and should be an important goal of future research. In sum, my view is that the improvement of disaster preparedness programs in much of the world must come from below, not from the national governments but from the experiences of the people, as such experiences are assessed and researched by planners. At this stage the disaster community should follow the lead of the Latin American Catholic Church in its decision to build "comunidades de base" as the nucleus from which to improve the lot of the people. These base communities generated social power for the Church and created for it the resources to reshape the world to better reflect Christian ideals. The creation of base communities was not a chance event. Instead, it resulted from a thorough analysis by the Church of the society and culture of Latin America and Central America in the second half of the 20th Century (Smith, 1991). Similarly, disaster planners must build their own local constituency and their own source of political support from which to effect social change.
36
REFERENCES Abernethy, James, D. Schmiedeke, H. Ferrero, U. Koehler. 1982. Tornado Protection. Selecting and Designing Safe Areas in Buildings. Southfield, Michigan: School of Architecture, Lawrence Institute of Technology. Abrams, J.I., and the Disaster Reanimatology Study Group. 1989. "Detection and Extrication in the Armenian Earthquake." Pp. 443-451 in Proceedings of the International Workshop on Earthquake Epidemiology for Mitigation and Response. Baltimore, MD: Johns Hopkins University. Aguirre, B. E. and D. Bush. 1992. "Disaster Programs as Technology Transfers." International Journal of Mass Emergencies and Disasters, Vol. 10 (1): 161-178. Aguirre, B. E., D. Wenger, T. A. Glass, M. Diaz-Murillo, G. Vigo. (forthcoming). "The Social Organization of Search and Rescue: Evidence from the Guadalajara Gasoline Explosion." International Journal of Mass Emergencies and Disasters. Aguirre, B. E. 1983. "Evacuation as Population Mobility." International Journal of Mass Emergencies and Disasters, Vol. 1 (3): 415-438. Alexander, David. 1989. Epidemiology." Pp. 82-94 International Workshop on Mitigation and Response. University.
"Spatial Aspects of Earthquake in Proceedings of the Earthquake Epidemiology for Baltimore, MD: Johns Hopkins
Ali, Lynne. "Symbolic Planning and Disaster Preparedness in Developing Countries: The Presbyterian Church in Vanuatu." International Journal of Mass Emergencies and Disasters, Vol. 10 (2): 293-314. Anagnostopoulos, S.A. and Whitman, R.V. 1977. "On Human Loss Prediction in Buildings During Earthquakes." Pp. 671-676, (Vol. 1) in Proceedings of the Sixth World Conference on Earthquake Engineering, New Delhi. Applied Technology Council. 1988. "Earthquake Damaged Buildings: An Overview of Heavy Debris and Victim Extrication." Earthquake Hazards Reduction Series 43. Washington, D.C.: Federal Emergency Management Agency, Document No. 158. Aroni, Samuel and Durkin, Michael E. "Injuries and Occupant Behavior in Earthquakes." Los Angeles, California: UCLA Graduate School of Architecture and Urban Planning. Unpublished manuscript. Arreola, Alvaro, Jose, Georgette, Luna, Matilde, and Tirado, Ricardo. 1986. "Memoria: Los Primeros Ocho
37
Dias." Revista Mexicana de Sociologia, Vol. 48 (2): 105-120. Barton, Allen H. 1969. Communities in Disaster: A Sociological Analysis of Collective Stress Situations. Garden City, New York: Doubleday and Company Inc. Baker, G.W. 1956. Individual and Group Behavior in a Coal Mine Disaster. New York, New York: National Academy ofSciences, National Research Council Disaster Research Group, Publication No. 834. Baker, Jay. 1990. "Evacuation Decision Making and Public Response in Hurricane Hugo In South Carolina." Quick Response Research Report No. 90, Natural Hazards Center, University of Colorado at Boulder. Baker, Jay. 1991. "Hurricane Evacuation Behavior." International Journal of Mass Emergencies and Disasters, Vol. 9 (2): 287-310. Baker, Jay. 1993. "Emergency Preparedness and Public Response in Southeast Florida in Andrew." Pp. 19-30, L. S. Tait, editor. Lessons of Hurricane Andrew. Tallahassee, Florida: national Hurricane Conference. Britton, N. R. 1987. "Towards A Reconceptualization of Disaster for the Enhancement of Social Preparedness." Pp. 31-55 in R. R. Dynes, B. De Marchi, and C. Pelanda, editors. Sociology of Disasters. Milano, Italy: Franco Angeli. Burby, R. J. and S. P. French. Floods." APA Journal 289-300.
1987.
"Coping with
Cowley, Adams R. 1982. "Mass Casualty Needs." Mass Casualties: A Lessons Learned Approach. Pp. 141-146 in Proceedings of the First International Assembly on Emergency Medical Services. Culver, C.G., Lew, H.S., Hart, G.C. and Pinkham, C.W. 1975. "Natural Hazards Evaluation of Existing Buildings." NBS Building Science Series Report No. 61. Washington, D.C.: National Bureau of Standards, Institute for Applied Technology, Center for Building Technology. De Bruycker, Marc, Greco D., and Lechat, M. F. 1985. "The 1980 Earthquake in Southern Italy-Morbidity and Mortality." International Journal of Epidemiology. Vol. 14 (1): 113-117. Drabek, T. E. 1987. "Emergent Structures." Pp. 259290, R.R. Dynes, B. De Marchi, C. Pelanda, editors. Sociology of Disasters. Milano, Italy: International Sociological Association Research Committee on Disasters.
38
Drabek, T.E. 1986. Human System Responses to Disasters. An Inventory of Sociological Findings. London: SpringerVerlag. Drabek, Thomas E., Tamminga, Harriet L., Kilijanek, Thomas S. and Adams, Christopher R. 1978. "The Sociology of SAR." Paper presented at the 10th Annual Meeting of the National Association for Search and Rescue, Albuquerque, NM. Drabek, Thomas E., Tamminga, Harriet L., Kilijanek, Thomas S. and Adams, Christopher R. 1981. Managing Multiorganizational Emergency Responses: Emergent Search and Rescue Networks in Natural Disasters and Remote Area Settings. Boulder, Colorado: The University of Colorado, Institute of Behavioral Science. Doswell III, Charles A. and Frederick P. Ostby. 1982. A Kansas City Area Guide to Severe Thunderstorms and Tornadoes. Kansas City, Missouri: National Severe Storms Forecast Center. Durkin, Michael E., and Murakami, Hitomi Ohashi. 1988. "Casualties, Survival and Entrapment in Heavily Damaged Buildings." Pp. 2-8 (vol. 3) in Proceedings of the Ninth World Conference on Earthquake Engineering, Tokoyo-Kyoto, Japan, August 2-9. Durkin, M.E. 1989. "The Role of the Physical Setting in Earthquake Injuries: The Mexico Experience." Proceedings from the Third U.S.-Mexico Workshop on 1985 Mexico Earthquake Research, March 16-18. Dynes, R. R. 1993. "Conceptualizing Disaster in Ways Productive for Social Science Research." Paper presented at the Workshop Socioeconomic Aspects of Disaster in Central America, San Jose, Costa Rica, January 21-23. Dynes, R. R. and Quarantelli, E.L. 1980. Helping Behavior in Large Scale Disasters. Newark, Delaware: University of Delaware, Disaster Research Center, article # 132. Dynes, R.R., Quarantelli, E.L., and Wenger, D. 1990. Individual and Organizational Response to the 1985 Earthquake in Mexico City, Mexico. Newark, Delaware: University of Delaware, Disaster Research Center, Book and Monograph Series # 24. Faupel, C. 1987. "Human Ecology: Contributions to Research and Policy Formation," in Sociology of Disasters, R.R. Dynes, B. De Marchi, C. Pelanda, (eds.). Milano, Italy: Granco Angeli Libri. Forrest, T. R. 1978. "Group Emergence in Disasters." Pp. 106-125 in E. L. Quarantelli, editor. Disasters: theory and research. Beverly Hills, California: Sage.
39
Form, William H., and Nosow, Sigmund. 1958. Disaster. New York: Harper and Brothers.
Community in
Fritz, Charles E. 1961. "Disasters," in Contemporary Social Problems, Robert K. Merton and Robert A. Nisbet (eds.). New York: Harcourt. Fritz, Charles E., and Marks, Eli S. 1954. "The NORC Studies of Human Behavior in Disaster." The Journal of Social Issues, Vol. 10 (3): 26-41. Glass, Roger I., Urrutia, J., Sibony, S., Smith, H., Garcia, B., and Rizzo, L. 1977. "Earthquake Injuries Related to Housing in a Guatemalan Village." Science, Vol. 197: 638-643. Gillespie, D. F. and C. L. Streeter. 1987. "Conceptualizing and Measuring Disaster Preparedness." International Journal of Mass Emergencies and Disasters, Vol. 5 (2): 155-176. Glass, Roger I., O'Hare, Phillip and Conrad, J. Lyle. 1979. "Health Consequences of the Snow Disaster in Massachusetts, February 6, 1978." American Journal of Public Health. Vol. 69 (October): 1047-49. Goodman, L.A. 1961. "Snowball Sampling." Mathematics and Statistics. Vol. 32: 148.
Annals of
Hart, G.C. 1976. Natural Hazards: Tornado, Hurricane, Severe Wind Loss Models. Redondo Beach, CA.: J.H. Wiggins Company. Hasselman, T.K., Equchi, R.T. and Wiggins, J.H. 1980. Assessment of Damageability for Existing Buildings in a Natural Hazards Environment. Technical Report No.801332-1, September. Redondo Beach, CA. Holland, Constance J. 1989. "Effective Utilization of Victim Volunteers in the Emergency Response." Paper presented at Workshop on Earthquake Injury Epidemiology. Baltimore, MD: Johns Hopkins University, May 16. Johnson, Jr., H. J. 1986. "Predicting Nuclear Reactor Emergency Evacuation Behavior." Energy. Vol. 11 (9): 861-868. Johnson, Norris R. 1987a. "Panic at 'The Who Concert Stampede': An Empirical Assessment." Social Problems. Vol. 34 (4): 362-373. Johnson, Norris R. 1987b. "Panic and the Breakdown of Social Order: Popular Myth, Social Theory, Empirical Evidence." Sociological Focus. Vol. 20 (3): 171-183. Johnson, Norris R. 1988 "Fire in a Crowded Theater: A Descriptive Investigation of the Emergence of Panic." International Journal of Mass Emergencies and Disasters. Vol. 6 (1): 7-26.
40
Kreps, G. 1978. "The Organization of Disaster Response." Disasters: Theory and Research. E.L. Quarantelli, editor. Beverly Hills, California: Sage. Krimgold, Frederick. 1988. "Search and Rescue in Collapsed Reinforced Concrete Buildings." Proceedings of the Ninth World Conference on Earthquake Engineering. Vol. 3, Tokoyo-Kyoto, Japan, August 2-9, 1988. Krimgold, Frederick. 1989. "Search and Rescue in Collapsed Buildings." Proceedings from the Third U.S.Mexico Workshop on 1985 Mexico Earthquake Research. Mexico City, March 16-18. Lavell, Allan. 1991. Prevention and Mitigation of Disasters in Central America and Panama: A Pending Task. San Jose, Costa Rica: Confederation of Central American Universities, Central American Research Programme. Lechat, Michael F. 1976. "The Epidemiology of Disasters." Section of Epidemiology and Community Medicine. Proceedings of the Royal Society of Medicine. Vol. 69 (June): 421-426. Lechat, Michael F. 1989. "Corporal Damage as Related to Building Structure and Design. The Need for an International Survey," unpublished manuscript. Lomnitz, Cinna. 1970. "Casualties and Behavior of Populations During Earthquakes." Bulletin of the Seismological Society of America. Vol. 60 (4): 13091313. Quarantelli, E.L. 1980. Final Project Report. Evacuation Behavior and Problems: Findings and Implications from the Research Literature. Columbus, Ohio: Ohio State University, Disaster Research Center, Publication No. 27. Quarantelli, E.L. 1983. Delivery of Emergency Medical Services in Disasters: Assumptions and Realities. New York: Irvington Publishers, 1983. Quarantelli, E.L. 1983. "Perceptions and Reactions to Emergency Warnings of Sudden Hazards." DRC article no. 173, reprinted from Ekistics. Vol. 309 (Nov.-Dec.): 511515. Quarantelli, E. L. 1984. "People's Reactions to Emergency Warnings." Disaster Research Center Article no. 170, reprinted from Transcript of the Dam Safety Conference, Charleston, West Virginia, August 23-24, pp. 177-187. Charleston, West Virginia: Department of Natural Resources. Quarantelli, E.L. 1988. "The NORC Research on the Arkansas Tornado: A Fountainhead Study." International Journal of Mass Emergencies and Disasters. Vol. 6 (3): 283-310.
41
Quarantelli, E.L. 1991. Lessons from Research: Findings on Mass Communication System Behavior in the Pre, Trans, and Postimpact Periods. Newark, Delaware: University of Delaware, Disaster Research Center, article # 227. Quarantelli, E.L., and R. R. Dynes. 1977. "Response to Social Crisis and Disasters." Annual Review of Sociology. Vol. 3: 23-49. Quon, Tony K., and Laube, A. 1991 "Do Faster Rescues Save More Lives?" Risk Analysis. Vol. 11, No. 2: 291301. Mileti, Dennis S. and John H. Sorensen. 1987. "Natural hazards and precautionary behavior." Pp. 189-207 in Neil D. Weinstein, ed., Taking Care. Understanding and Encouraging Self-Protective Behavior. Cambridge: Cambridge University Press. Mileti, Dennis S. and J. H. Sorensen. 1989. "La Eficacia de los Sistemas de Alarma en America Latina." Pp. 187-198 in Bruno Lima, editor. Consecuencias Psicosociales de los Desastres: La experiencia latinoamericana. Venezuela: Universidad Simon Bolivar, Departamento de Psiquiatria.. Mileti, Dennis S., B. C. Farhar, C. Fitzpatrick. 1990. "How to Issue and Manage Public Earthquake Risk Information: Lessons from the Parkfield Earthquake Prediction Experiment. Fort Collins, Colorado: hazards Assessment Laboratory, Colorado State University. Mileti, Dennis S. and P. W. O'Brien. 1991. Public Response to the Loma Prieta Earthquake Emergency and Aftershock Warnings: Findings and Lessons. Fort Collins, Colorado: hazards Assessment Laboratory, Colorado State University. Moses, Rafael, J.M. Rosenfeld, R. Moses-Hrushovski. 1987. "Facing the Threat of Removal: Lessons from the Evacuation of Ofira." The Journal of Applied Behavioral Science. Vol. 23 (1): 53-71. Noji, E. K., Kelen, G. D., Armenian, H. K., Oganessian, A., Jones, N. P., and Silverton, K. T. 1990. "The 1988 Earthquake in Soviet Armenia: A Case Study." Annals of Emergency Medicine. vol. 19 (8): 891-897. Olson, R. and Olson, R. A. 1987. "Urban Heavy Rescue." Earthquake Spectra. Vol. 3 (4): 645-658. Palm, Risa. 1981. Real Estate Agents and the Special Studies Zones Disclosure: The Response of California Home Buyers to Earthquake Hazards Information. Institute of Behavioral Science, The University of Colorado. Palm, Risa. 1987. "Pre-Disaster Planning: The Response of Residential Real-Estate Developers to Special Studies
42
Zones." International Journal of Mass Emergencies and Disasters. Vol. 5 (1): 95-102. Pauls, J.L. 1980. "Building Evacuation: Research Findings and Recommendations." Pp. 227-250 in D. Canter, editor, Fires and Human Behavior. New York: John Wiley and Sons, Inc. Perry, R.W., M.K. Lindell, M.R. Greene. 1981. Evacuation Planning in Emergency Management. Lexington, Massachusetts: Lexington Books. Perry, R.W. and M. R. Greene. 1983. Citizen Response to Volcanic Eruptions: The Case of Mt. St. Helens. New York: Irvington Publishers, Inc. Perry, R.W. and A.H. Mushkatel. 1984. Disaster Management: Warning Response and Community Relocation. Westport, Connecticut: Quorum Books. Peters, Brian E. 1988. Natural Hazard Deaths in 1987 in the United States. A Summary. Fort Worth, Texas: NWS Southern Region Headquarters, Fort Worth, TX. Rogers, G.O., J. H. Sorensen. 1989. "Warning and Response in Two Hazardous Materials Transportation Accidents in the U.S." Journal of Hazardous Materials. Vol. 22 (1): 57-74. Rogers, G.O. and J. H. Sorensen. 1991. "Diffusion of Emergency Warning: Comparing Empirical and Simulation Results." Pp. 117-134 in C. Zervos, editor, Risk Analysis. New York: Plenum Press. Ruch, C., H.C. Miller, M. Haflich, N.M. Farber, P. R. Burke, N. Stubbs. 1991. The Feasibility of Vertical Evacuation. Boulder: University of Colorado, Institute of Behavioral Science. Scanlon, J., Alldred, A., Farrell, A., and Prawzick, A. 1985. "Coping with the Media in Disasters: Some Predictable Problems." Public Administration Review. Vol. 45: 123-133. Seitz, S. T. and M. Davis. 1984. "The Political Matrix of Natural Disasters: Africa and Latin America." International Journal of Mass Emergencies and Disasters, Vol. 2 (3): 231-250. Smith, C. 1991. The Emergence of Liberation Theology: Radical Religion and Social Movement Theory. Chicago: University of Chicago Press. Sorensen, John H. 1993. "Warning Systems and the Public Warning Response." Paper presented at the Workshop Socioeconomic Aspects of Disaster in Central America, San Jose, Costa Rica, January 21-23.
43
Sorensen, John and Dennis S. Mileti. 1987a. "Programs that encourage the adoption of precautions against natural hazards: review and evaluation." Pp. 208-230 in Neil D. Weinstein, ed., Taking Care. Understanding and Encouraging Self-Protective Behavior. Cambridge: Cambridge University Press. Sorensen, John and Dennis S. Mileti. 1987b. "DecisionMaking Uncertainties in Emergency Warning System Organizations." International Journal of Mass Emergencies and Disasters. Vol 5 (1): 33-61. Sorensen, John H. and Dennis S. Mileti. 1988. Warning and Evacuation: Answering Some Basic Questions. Industrial Case Quarterly, Vol. 2, 195-209. Southworth, F. and S. M. Chin. 1987. "Network Evacuation Modelling for Flooding as a result of dam failure." Environment and Planning, Vol. 19, 1543-1558. Stubbs, Norris, Sikorsky, C., Lipnick, P., and Lombard, P. 1989. "Simulation of Fatalities in Structures Subjected to Hurricane Loading," Pp. 88-93 in Simulation in Emergency Management and Technology. San Diego, CA: Simulation Councils, Inc. Tiedemann, Herbert. 1989. "Casualties as a function of Building Quality and Earthquake Intensity," Paper presented at Workshop on Earthquake Injury Epidemiology. Baltimore, MD.: Johns Hopkins University, May 16. Tierney, K. J. 1992. "Politics, Economics, and Hazards." Unpublished manuscript. Tweedie, S. W., J. R. Rowland, S. J. Walsh, R. P. RHoten, P. I. Hagle. 1986. "A Methodology for Estimating Emergency Evacuation Times." The Social Science Journal, Vol. 23 (2): 189-204. Ville de Goyet, Claude de. 1982. "Emergency Health Management Following Natural Disasters in Latin America." Pp. 185-192 in Mass Casualties: A Lessons Learned Approach. Proceedings of the First International Assembly on Emergency Medical Services. Vogt, B. and J. Sorensen. 1992. Evacuation Research: A Reassessment. Oak Ridge, Tennessee: Oak Ridge National Laboratory. Wallace, Anthony F.C. 1956. Tornadoes in Worcester: An Exploratory Study of Individual and Community Behavior in Extreme Situations. Washington, D.C.: National Academy of Sciences, National Research Council. Waxman, Jerry. 1973. "Local Broadcast Gatekeeping During Natural Disaster." Journalism Quarterly. Vol. 50: 71-758.
44
Weinroth, J. 1989. "A Model for the Management of Building Evacuation." Simulation. Vol. 53 (September):111-119. Weller, J.M., and Quarantelli. E.L. 1973. "Neglected Characteristics of Collective Behavior." American Journal of Sociology. Vol. 79: 665-683. Wenger, Dennis E., T. F. James, C. F. Faupel. 1980. Disaster Beliefs and Emergency Planning. Newark, Delaware: Disaster Research Center, University of Delaware. Wenger, Dennis E. 1985. "Mass Media and Disasters." College Station, Texas: Hazard Reduction and Recovery Center, Preliminary paper. Wenger, Dennis E. 1987. "Patterns of Participation in Collective Action in a Consensus Crisis: The Mexico City Earthquake of 1985." Paper presented at the Annual Meeting of the American Sociological Association, Chicago, Illinois. Wenger, Dennis E. 1990. "Volunteer and Organizational Search and Rescue Activities Following the Loma Prieta Earthquake: An Integrated Emergency and Sociological Analysis." College Station, Texas: Texas A&M University, Hazard Reduction and Recovery Center. Wenger, Dennis E., and James, Thomas F. 1991. "The Convergence of Volunteers in a Consensus Crisis: The Case of the 1985 Mexico City Earthquake." College Station, Texas: Texas A&M University, Hazard Reduction and Recovery Center. Wenger, Dennis E. and Quarantelli, E. L. 1989. Local Mass Media Operations, Problems, and Products of Disasters. Newark, Delaware: University of Delaware, Disaster Research Center. Zurcher, Louis A. 1968. "Social-Psychological Functions of Ephemeral Roles: A Disaster Work Crew." Human Organization. Vol. 27: 281-297.
