Myths and realities of household disaster response

Socially Integrative Responses

Although the influx of people and materials can provide local emergency managers with resources needed to respond effectively to the agent-generated demands of the disaster, convergence can produce unprecedented communication and response difficulties. For example, Kartez and Lindell (1987) described a Louisiana air crash in which fire departments from distant communities appeared at the crash site unrequested. This created a serious strain on the local authorities’ ability to not only deal with the crash but also to handle the logistics associated with additional responders. To complicate matters further, unsolicited materials can also arrive unannounced early in the incident and continue to arrive for days or weeks thereafter. Thus, emergency managers need to be aware of the convergence of volunteers and unsolicited contributions of resources. When convergence processes inundate them with unanticipated people and materials, a potential asset becomes a liability. This makes it essential to develop donations management procedures such as SUMA (Supply Management) which was developed by the Pan American Health Organization (PAHO) to facilitate the integration of volunteers into the response force; the management and care of volunteer labor; and the logistics of receiving, storing, and deploying material and equipment.

The idea spread of taking the refugees into such private homes as had fared less badly. It became the thing to do. The thing to do is social pressure. It may be unwilled and unintended but it is inexorable. It worked effectively upon all who had an unused room.

Since the time Prince conducted his study, further research has been conducted on the extent to which community members not directly impacted by the event support the victims (e.g., Midlarsky, 1968; Vallance & D’Augelli, 1982). Such helping behavior directed at (and among) victims may be seen as a normative response. What is important from the standpoint of emergency management, however, is the positive social climate created by such altruism.

The result of these psychological and social processes is a therapeutic social system in which an unplanned outpouring of personal warmth and direct help provides support to many victims in a time of considerable sorrow and stress (Barton, 1969). This is not to say that these naturally occurring social processes provide complete support for victims or that they entirely mitigate the negative psychological consequences of disaster impact. Both natural and technological disasters are calamitous experiences for many victims. Terrorist events also can be expected to elicit extreme outpourings of help to the perceived blameless victims. The scale of giving following the 9/11 terrorist attacks stands as an extreme example. It is essential for emergency managers to recognize disasters cause indirect positive effects as well as the more direct, and certainly more widely known, negative psychological impacts.

It is also important to appreciate the likely persistence of the therapeutic community response over time. Early researchers saw the therapeutic community as “an outpouring of altruistic feelings and behavior beginning with mass rescue work and carrying on for days, weeks, possibly even months after the impact” (Barton, 1969, p. 206, emphasis added). Regrettably, research on the persistence of the therapeutic community response has been insufficient to permit confident acceptance of Barton's hypothesis of long-term persistence. As Dynes and Quarantelli (1976; see also Quarantelli & Dynes, 1977) suggest, the therapeutic community may not be a long lasting condition. Although their work does not directly test Barton's proposition, they have found that decreases in community conflict and the apparent increase in community consensus following disasters are short-lived phenomena. While the empirical record is too sparse to support specific time estimates, some conflict regarding the distribution of contributed funds and materials began to arise within six months following the 9/11 terrorist attacks. There is agreement regarding the development of a therapeutic community in the short-term aftermath and that it should be seen as promoting positive psychological outcomes for disaster victims. Emergency managers planning for the recovery period, however, are well advised to remember that the duration of situational therapeutic norms is limited.

Other Misconceptions and Erroneous Assumptions

Less prominent, but just as damaging, are some specific misconceptions and erroneous assumptions that about households’ emergency response. Misconceptions are explicit beliefs that are relatively easy to identify and correct. By contrast, erroneous assumptions are implicit and, therefore, more difficult to correct. Whether these erroneous beliefs are explicit or implicit, they produce a distorted image of the processes by which risk area residents receive warnings, evaluate the information, make decisions, and take action to protect themselves, their loved ones, and their property. The most serious misconceptions and erroneous assumptions about household response are discussed below.

One of the most common misconceptions is that there is an undifferentiated “public” when, in fact, there are many population segments that differ in their hazard knowledge, family roles, and household resources. In particular, emergency mangers must distinguish among residents, transients, and special facility populations because these population segments differ in their willingness and ability to evacuate (Drabek, 1996; Urbanik, 2000). Residents are those who live or work in the risk area, whereas transients consist primarily of those who stay in hotels/motels— although day visitors can be a significant concern in rapid onset disasters. However, day visitors are unlikely to be a problem for hurricanes and other hazards with forewarning because they are likely to leave very early. Special facilities include schools, hospitals, nursing homes, and jails (for a more complete list, see Table 6-2). Special facility populations typically must be analyzed separately because their patterns of warning and evacuation are very different from those of residents and transients (Urbanik, 2000).

It is commonly assumed that authorities are the first to warn risk area residents and also that authorities provide almost all the information about a disaster. However, as indicated by the Protective Action Decision Model (recall the discussion in Chapter 4), people also rely substantially on the news media, peers, and environmental cues to determine if disaster impact is likely to affect them. People also observe the behavior of others (peers evacuating and businesses closing) to assess the need for protective actions such as evacuation. They also seek to confirm warnings, regardless of the initial warning source, to verify the information they have received, and to work out the logistics of response (mode of transportation, route of travel, shelter destination). Indeed, in a major disaster, the volume of calls into and within the impact area can overload the available telephone circuits.

The process of warning dissemination generates a distribution of times at which households first receive a warning. Figure 8-1 shows the cumulative distribution of warning receipt over time of the type reported by Rogers and Sorensen (1988). The curve is nonlinear so the rate of warning receipt first increases and later decreases over time. For example, 39% receive a warning in the first hour, another 47% receive a warning in the second hour, only 12% receive a warning during the third hour, and 1%. receive a warning in the fourth hour. The last few people take a long time to receive a warning.

Whether people receive a warning and when they receive one has been the subject of substantial research over the past five decades (Lindell & Perry, 2004). Warning sources differ in their accessibility and these differences can vary by community. For example, Lindell and Perry reported that the majority of those at risk from the eruption of Mt. St. Helens received their first warning from peers (see Table 8-1). In Toutle (a town very close to the volcano), 58% of the population received their first warning from peers, whereas in Woodland (which was farther away), 47% received their first warning from this source. The greatest differences between communities were associated with the news media; 39% of the Woodland residents but only 6% of the Toutle residents received their first warning from this source.

In part, the extent to which any source provides the first warning—and the rate at which the entire population is warned—depends on the communication channels each source uses. Lindell and Perry (1992) summarized the available warning mechanisms as face-to-face, route alert (loudspeaker broadcast from a moving vehicle), siren, commercial radio and television, tone alert radio (which only broadcasts a warning within the home after receiving a special tone from the National Weather Service), telephones, and newspapers. These warning mechanisms differ with respect to their precision of dissemination, penetration of normal activities, message specificity, susceptibility to message distortion, rate of dissemination over time, receiver requirements, sender requirements, and feedback (verification of receipt). Sources using channels that have more rapid dissemination rates will produce steeper warning curves. However, commercial radio and television cannot provide warnings unless they are turned on, which limits their effectiveness when people are asleep or outdoors. Conversely, sirens have limited penetration into homes when people are engaged in noisy activities or are asleep. Thus, the effectiveness of a given warning mechanism varies with the types of activities in which people are engaged. Across a broad range of situations, tone alert radio has been consistently effective in providing rapid warning dissemination.

Figure 8-1. Cumulative Distribution of Household Warning Receipt over Time.


Time (minutes)

Source: Lindell, et al. 2002)

For hazard agents lacking environmental cues such as sights (smoke), sounds (explosions), or smells (chemical odors), people must rely on social sources of information—authorities, news media, and peers. Thus, the time it takes each household to receive a warning from authorities depends upon the warning mechanisms the authorities use. For example, warnings disseminated by route alert depend on the number of vehicles dispatched and the speed at which they travel. By contrast, warnings disseminated by tone alert radio depend upon who owns these devices.

Table 8-1. Source of First Eruption Warning, by Community.

	Toutle		Woodland
	Number	Percent	Number	Percent
Saw environmental cues	27	30	12	14
Authorities	6	6	0	0
News media	6	6	34	39
Peers	51	58	41	47

Source: Lindell and Perry (1992)

The time it takes each household to receive a warning from the news media depends upon whether people have turned on their televisions or radios and are attending to them. Although access to the news media varies to some extent among households, it varies even more by time of day—with the highest levels being reached during the day and early evening (Lindell & Perry, 1992; Rogers & Sorensen, 1988).

The time it takes each household to receive a warning from peers depends the extent to which people are integrated into social networks (Drabek & Stephenson, 1971), which is correlated with a number of variables. Specifically, social integration with both kinship groups (relatives) and community groups (friends, neighbors, and coworkers) is negatively correlated with age (Perry, 1985; Perry, et al., 1981), socioeconomic status (Alvirez & Bean, 1976; Cohen & Kapsis, 1978), and ethnicity (Bianchi & Farley, 1979; Staples & Mirande, 1980). The latter variable introduces some complexities because minority households are more likely to be extended families, to involve multigenerational depth, and to have more than a single family in the same household. Thus, although ethnic minorities are more integrated into kinship networks, there is a greater probability that some family members may not be present, thereby slowing later stages of the warning response until they can at least be accounted for.

There is substantial variation across communities in the relative importance of different warning sources to different ethnic groups. Perry and Mushkatel (1986) found that Hispanics in three communities were consistently more likely to receive their first warning from peers. However, the role of the news media as a warning source for this ethnic group varied among communities. Differences across communities were even more striking for African-Americans. In one community, African-Americans received their first warning from authorities. However, in another community, almost no African-Americans received their first warning from this source. There were also differences across communities for Whites. In one community, the majority of Whites received their first warning from the news media. However, in another community, peers were the most common source of first warning.

For the most part, demographic variables have limited value in explaining people’s warning responses. After an exhaustive review of research, Quarantelli (1980, p. 43) stated “studies dealing with demographic characteristics and evacuation are simply not conclusive.” Baker’s (1991) review of hurricane evacuation studies reached a similar conclusion. This means that source and message characteristics are the most important determinants of household warning response.

Emergency mangers need to recognize that the resident population is nonuniformly distributed within risk areas, but they can estimate population size and distribution with Geographical Information Systems. For example, they can overlay risk area boundaries onto census block group boundaries and use this information to compute each risk area’s residential population (Lindell, et al., 2002b). In addition, emergency managers should recognize that risk area residents are also nonuniformly distributed in time (Alam & Goulias, 1999; Barrett, Ran & Pilai, 2000; Urbanik, 2000). That is, structures are completely occupied at some times, partially occupied at others, and sometimes completely empty. Over 90% of the population is indoors at home from 10:00 p.m. to 6:00 a.m., but only about one-third is there from 10:00a.m. to 3:00 p.m. (Klepeis, et al., 2001). This implies that evacuations initiated during daytime hours should incorporate a time component for travel from work to home (for adults) or school to home (for children when school is in session). Such diurnal variation is especially important for the evacuation of areas around nuclear and chemical facilities that can have incidents with very short forewarning (Hobeika, Kim & Beckwith, 1994; Urbanik, 2000). However, the ample forewarning of a hurricane leads families to stay home in anticipation of an evacuation (Lindell, Lu & Prater, 2005). In such incidents, emergency managers will generally have little need to adjust for diurnal variation in household activity.

There is a tendency for most people to assume that all households have their own private vehicles, but emergency managers should also recognize the transit dependent segment of the resident population (Urbanik, 2000). The transit dependent segment can be as much as 15% or more of the population in some coastal counties exposed to hurricanes and, contrary to most people’s assumptions, this level of transit dependence can be found in both rural and urban counties. Also contrary to most people’s assumptions, much of the transit dependent population evacuates with peers (Lindell, et al., 2005). Thus, the number of buses needed to evacuate the transit dependent is likely to be smaller than expected. However, some emergency managers fail to recognize that the time required to mobilize buses, pick up evacuees at designated staging areas, and travel out of the risk area might be greater than the time required to evacuate households in cars. This is especially likely if the transit dependent must share a limited number of buses with special facilities.

Emergency managers should also estimate the size and distribution of the transient population from local convention/visitor bureau data on the number of hotel/motel rooms (Hobeika, et al., 1994; Lindell, et al., 2002b). In addition, Hobeika, et al. (1994) considered the number of campsites and their occupancy rate. It is extremely important to account for seasonal variation in the transient population because, for example, coastal areas tend to have high rates of tourist occupancy on holiday weekends during the summer and much lower rates during the week after Labor Day. Hobeika, et al. (1994) and Lindell, et al. (2002b) accounted for variation in the size of this population segment by collecting data on occupancy rates for hotel/motel rooms over time. However, these occupancy rates are often reported by month rather than by week or day, so analysts must still make assumptions about variation within month.

Some analysts have assumed that 70-80% of registered vehicles would be used in a daytime evacuation and 90% would be used in a nighttime evacuation (Southworth & Chin, 1987). These analysts also contended that only 75% of registered vehicles would be used in rapidly developing incidents. Unfortunately, there are no empirical data or theoretical foundations for such assumptions. Instead, it is more logical to base the estimated number of evacuating vehicles on the number of evacuating households because fifty years of disaster research has identified the household as the basic unit of evacuation. Indeed, households that are separated when a warning is disseminated almost always attempt to reunite before leaving the risk area (Drabek, 1986; Lindell & Perry, 1992; Tierney, et al., 2001). Data on the number of evacuating vehicles per residential household (EVHH_R) have been collected in a number of household surveys. The EVHH_R values vary substantially from county to county, with a low of 1.10 and a high of 2.15 (Lindell, et al., 2005) but the most probable range is about 1.2-1.5. Data on the number of evacuating trailers per residential household (ETHH_R) is an important consideration in evacuations because many households in coastal areas load boats onto trailers to take when they evacuate. These trailers fill space on the highway, just as the cars towing them do, so they should be included as vehicle equivalents in estimating traffic demand. Lindell, et al. (2005) reported the number of evacuating trailers taken in Hurricane Lili was ETHH_R = 0.3, but values ranged from 0.0 to 1.08 across the five counties.

Warning compliance refers to the percentage of those warned to evacuate who actually do so. Spontaneous evacuation, also known as evacuation shadow, refers to evacuation by those who are outside the risk area and therefore not warned to evacuate. Many evacuation analysts have assumed all persons advised to evacuate will do so and there would be no spontaneous evacuation from areas not advised to evacuate. By contrast, Lindell, et al. (2002a) used data, originally reported by Lindell, et al. (2001), on expected evacuation rates as a function of residents’ risk areas and the Saffir-Simpson hurricane category of an approaching storm. The evacuation percentages in the original data were subject to sampling error, so the data can be smoothed statistically to yields the percentages in Table 8-2 (Lindell & Prater, 2005).

Table 8-2 is consistent with the findings of previous research on evacuations in showing risk area residents’ evacuation expectations differ significantly from the idealized pattern (100% evacuation in areas advised to evacuate and 0% evacuation in areas not advised to evacuate). These data are also consistent with previous evacuation research, although no study other than Lindell, et al. (2002) broke down compliance by risk area. Dow and Cutter (2002) reported 65% compliance during Hurricane Floyd in South Carolina; Riad, Norris, and Ruback (1999) found 42% compliance in Hurricanes Hugo and Andrew; and Prater, et al. (2000) reported 34% compliance in Hurricane Bret. Lindell, et al. (2005) reported that evacuation rates in Hurricane Lili ranged from 11.7 to 86.8% across five jurisdictions and these rates decayed as a function of distance from the point of landfall. These findings are consistent with Baker’s (1991) report that evacuation compliance rates in 15 studies varied significantly from one storm to another at a given location (47-68% in Galveston for three different storms) and from one location to another in a given storm (33-97% in Hurricane Frederic).

Spontaneous evacuation has received increasing attention since Zeigler, Brunn and Johnson (1981) called attention to this phenomenon, but empirical data are sparse. Gladwin, Gladwin, and Peacock (2002) recently confirmed the occurrence of spontaneous evacuation in hurricanes, which Baker (1991) reported to range from 20–50% of residents in areas of “low risk”. Other than Lindell, et al. (2002a), warning compliance and spontaneous rates have not been defined with enough precision to apply to evacuation models. There appears to be no available data on the percentage of transient households complying with an evacuation warning or spontaneously evacuating, but the behavior of this population segment is also likely to be a function of the hurricane’s intensity and the household’s risk area. Based on information collected in interviews of local Emergency Management Coordinators following Hurricane Bret (Prater, et al., 2000) and in Drabek’s (1996) interviews of hotel managers, Lindell et al. (2002a) assumed tourists would produce 100% compliance with a hurricane warning, regardless of their risk area.

There are few reports on the percentage of early evacuees, but data from Lindell, et al. (2005) show the distribution of times when households decided to evacuate from Hurricane Lili. Because of the storm’s high intensity (Category 4) and steady track, local authorities expected the National Hurricane Center to announce a hurricane warning on Wednesday morning 2 October. Accordingly, they announced on Tuesday evening that they would initiate an evacuation on Wednesday morning. Figure 8-2 shows almost two thirds of the households decided to evacuate before the official warning and, moreover, there was a noticeable tendency for people to make their decisions early in the morning so they would have the maximum number of hours of daylight in which to evacuate.

There is a modest amount of empirical data on household departure time distributions. Lindell and his colleagues (Lindell, et al., 1985; Lindell & Perry, 1987, 1992) reported warning and preparation times from four floods and the eruption of Mt. St. Helens and Sorensen and Rogers (1989) reported warning and preparation time data from two hazardous materials spills. These data can be used directly to estimate departure times for these hazard agents and can also be used to construct estimated departure time distributions for other hazard agents as well. For example, Lindell, et al. (2002a) constructed synthetic departure time distributions for hurricane evacuations by combining a warning time and preparation time distributions from two different situations. To approximate a rapid warning, such as might occur with a last minute change in hurricane track (e.g., 1999 Hurricane Bret and 2004 Hurricane Charley), they used data from the 1980 eruption of Mt. St. Helens (Lindell & Perry, 1992). The use of these data is supported by a survey conducted the month before the eruption, which indicated 34% of all residents checked the news media for information about the volcano two or three times per day and another 56% checked more than four times per day (Greene, Perry & Lindell, 1981). This is a level of hazard monitoring that seems to be quite comparable to the attentiveness of those in hurricane risk areas during the days before landfall.

Figure 8-2. Distribution of Evacuation Decision Times.


Source: Lindell, et al. (2005).

To estimate hurricane preparation times, Lindell, et al. (2002a) analyzed data collected by Lindell, et al. (2001), who asked coastal residents to report the length of time they expected to take in preparing prepare to leave work, travel from work to home, gather household members, pack travel items, install storm shutters, and secure their home before evacuating from a hurricane. Later comparison of expected preparation times with actual preparation times during Hurricane Lili revealed statistically and practically significant correlations (Kang, et al., in press).

Drabek’s (1996) research suggests transients are warned at a faster rate than residents and also prepare to evacuate more rapidly than residents, but there are no quantitative data on the size of this difference. Consequently, Lindell, et al. (2002a) derived warning and preparation time distributions for this population segment from their data on residents.

Destination/route choice. Finally, there are four parameters defining evacuees’ destination/route choice. These are evacuees’ ultimate evacuation destinations, their proximate destination/route choices, and their utilization of the primary evacuation route system. The ultimate destination is the place where an evacuating household intends to stay until it is safe to return to their home. The ultimate evacuation destination constrains the proximate destination/route choice because there are few routes that can be taken out of a risk area to the ultimate evacuation destination without significantly increasing travel distance and time. Proximate destination choice (the point at which the vehicle leaves the risk area) is closely related to route choice (the selected route to the exit from the risk area). Some analysts have assumed evacuees choose their routes dynamically when they encounter traffic queues. For example, Sheffi, Mahmassani & Powell (1981) contended drivers have a “myopic view” of alternative routes to their proximate destination and theorized that drivers select the least congested road they encounter at each intersection. However, reports from recent hurricane evacuations (Dow & Cutter, 2002; Prater, et al., 2000) show evacuees tend to take the most familiar routes inland (especially interstate highways), thus overloading those routes and ignoring unused capacity on alternate routes. It seems likely that some evacuees persist in following predetermined routes whereas others improvise in the face of adverse traffic conditions, but there is insufficient data to determine the relative proportion of each type of driver and the conditions affecting this proportion.

Search and rescue

The societal response to disasters, especially disasters having little forewarning, can be understood in reference to zones defining impact and response (Wallace, 1957). In Figure 8-3, the innermost circle is the total impact zone where casualties and damage are the greatest. Immediately adjacent to the total impact area is the fringe impact zone, in which casualties and damage are significant but not overwhelming. The next ring is the resource filter zone, through which information passes from the inner (total impact and fringe) zones to the outer (community aid and regional aid) zones and material resources pass from the outer zones to the inner zones. The community aid zone is the area from which assistance is drawn for minor disasters whereas the regional aid zone is needed to support response and recovery from major disasters.

Source: Dynes (1970)

It has long been known that survivors in the impact area are usually the first to respond to disaster impact by searching for those who are trapped in building debris, extricating them, providing preliminary treatment, and transporting them to hospitals. As time passes, there is increasing involvement by those in the fringe impact area and other zones (Dynes, 1970; Form & Nosow, 1958). These volunteers’ search and rescue efforts can be highly effective when the local building stock comprises wood frame and unreinforced masonry structures. However, collapsed buildings made of steel reinforced concrete require organized teams with specialized equipment.

Medical Transport

It is commonly assumed that authorities transport injured disaster victims in ambulances to the most appropriate hospitals. However, according to Quarantelli (1983), many injured victims arrive at hospitals in their own vehicles or those of peers or bystanders (46% of casualties) rather than in ambulances (54% of casualties). Moreover, the vast majority (75%) of victims are transported to the nearest hospital, which is usually overloaded at the same time as other competent facilities are sent few or no patients (Auf der Heide, 1994). Indeed, a study of 14 disasters found an average of 67% of casualties were treated in a single hospital even though the affected communities ranged from 3-105 hospitals (Golec & Gurney, 1977). Such findings call attention to the need for emergency managers to work with their local EMS agencies to establish procedures for transporting and allocating disaster casualties among the available hospitals.

People commonly assume the majority of evacuees stay in mass care facilities while they are waiting to return to their homes but, in fact, only a minority do so. Mileti, Sorensen and O’Brien (1992) reported the average level of mass care shelter use in 23 well-documented evacuations was 14.7% of the evacuating population. These researchers also found the percentage of the evacuees using mass care facilities varied in the range 5-20% across disasters (Mileti, et al., 1992). The use of mass care facilities could be as low as 5% if evacuations are initiated early in the morning with good weather and evacuation routes have effective traffic management inside and outside the risk areas. Conversely, the level of mass care demand is likely to be higher when the evacuation zone has a high proportion of households from lower income strata or ethnic minority groups. Other conditions causing mass care demand to reach 20% or higher include darkness, bad weather, traffic congestion, or other conditions impeding evacuees’ ability to reach their intended destinations.

In general, evacuees prefer to avoid mass care facilities if other options are available. For example, the Lindell, et al. (2001) hurricane planning analysis found the majority of the respondents to their survey expect to stay with friends and relatives (46.3%), while the next most popular accommodations are expected to be commercial hotels or motels (32.9%). Another 4.3% of the respondents expect to stay in campers or trailers, 3.2% expect to stay in second homes, and 9.8% indicated that they don’t know where they will stay or did not respond. Only 3.4% expect to stay in public shelters. These findings are generally consistent with the findings of previous evacuation research (e.g., Drabek, 1986; Lindell & Perry, 1992; Mileti, et al., 1975; Tierney, et al., 2001).

Emergency Responder Role Abandonment

Another frequently expressed concern is that role conflicts will lead emergency responders to abandon their professional duties in favor of protecting their families. That is, when confronted with the dilemma of protecting the public at large or their families in particular, they resolve the conflict in favor of their families. Contrary to this myth, Quarantelli’s (1982b) careful examination of the Disaster Research Center’s studies of hundreds of emergencies indicates that there is no evidence that role abandonment occurs. The explanation for the nonoccurrence of role conflict is explained by the fact that emergency responders typically develop family emergency plans and some emergency response agencies develop family protection plans to prevent such conflicts from arising. The performance of the New Orleans police in Hurricane Katrina seem to be a well documented exception to the general rule that emergency response personnel do not abandon their roles. Future research will undoubtedly be conducted to explain why police performance in this situation was so different all the previously studied disasters.

In any event, it is important to recognize that conclusions about emergency personnel role abandonment will not necessarily apply to emergency response auxiliaries who are called upon to perform their normal duties but aren’t trained to perform them under emergency conditions (Lindell, et al., 1985). For example, emergency managers might plan for school bus drivers to evacuate transit dependent households. However, they should not automatically assume that these drivers will perform their duties under emergency conditions. Instead, emergency managers should seek to have any emergency response auxiliaries make an explicit commitment to perform their jobs in an emergency and encourage them to make the same preparations as other emergency responders. Identifying all of the emergency response auxiliaries that are implicitly (rather than explicitly) designated in an EOP can be a challenging task. The absence of staff members led to deaths of nursing home residents in Hurricane Katrina and the absence of baggage inspectors prevented people from flying out of Houston before Hurricane Rita struck.

Volunteers and Emergent Organizations

A common theme in many of the previous sections is that volunteers are extremely active in the emergency response and disaster recovery phases. One of the consequences of the large number of volunteers performing tasks such as search and rescue is a need to organize the efforts of those who have not been trained to serve as part of the emergency response organization. To be effective, volunteers must be organized into groups that can perform tasks that are within the scope of their abilities and can interact effectively with the rest of the emergency response organization. Dynes’s (1970) typology of disaster organizations provides a useful way of understanding how to organize volunteers. As Table 8-3 indicates, tasks can be characterized as normal (organizational members perform them routinely) or novel (organizational members perform them only in disasters). Similarly, organizational structures can also be characterized as normal (organizational members conform to their normal roles and authority relationships) or novel (organizational members develop new roles and authority relationships).

		Tasks
		Normal	Novel
Organizational structures	Normal	Established	Extending
	Novel	Expanding	Emergent

This cross-classification produces four different types of disaster response organizations.

It is crucial for emergency managers to understand that people confronted by a disaster tend to respond in ways that seem logical to them given their limited information about the situation and the constraints that limit their options. People generally act adaptively and sometimes even heroically, as in the case of passengers in United Airlines Flight 93 on September 11, 2001. Unlike the passengers on the earlier hijacked flights, they understood the hijackers’ goal. Thus, they organized and attacked their hijackers, choosing to crash the flight in a Pennsylvania field rather than allow it to be flown into a building in Washington DC.

The first basic principle is that people threatened by disaster generally have multiple sources of information, but none of these sources is completely credible nor is any single source expected to have all the information a household needs to protect itself. This frequently produces confusing and conflicting information that people have difficulty resolving because they don’t know which source is correct. For example, Figure 8-4 depicts ratings that residents of Longview and Kelso, Washington, made in 1985 about the degree of hazard knowledge held by themselves, their peers, the news media, and local, state, and federal government. They were asked to make these judgments about hazard knowledge separately for three hazard agents—a volcanic eruption of Mt. St. Helens (about 40 miles east of their communities), a chlorine release from a truck or train on nearby transportation routes, and a release of radioactive materials from the Trojan nuclear power plant (which was less than 10 miles away). The figure indicates most people think they know more about hazards than their peers, but less than the news media, local government, and state and federal government. Moreover, the news media were judged to be significantly more knowledgeable than self or peers but less knowledgeable than local, state, or federal government. However, the differences in knowledge ratings are smaller for the more familiar volcanic hazard than for either of the two technological hazards. Interestingly, the knowledge ratings were consistently higher for radiation hazard than for chlorine hazard, quite possibly because federal licensing regulations required the nuclear power plant to distribute emergency information brochures annually. It is important to note that none of the sources was rated as extremely knowledgeable about any of the hazards (i.e., ratings of 4 or higher). Nor were any of the sources rated as severely lacking in knowledge (i.e., ratings of 2 or lower). Thus, even though there were significant differences among sources, each was credited with some degree of knowledge. Therefore, conflicts in the information provided by different sources would be difficult for people to resolve.

The second basic principle is that people generally experience fear—not debilitating shock or panic. Fear is a normal human reaction to extreme environmental conditions that arises when people expect that they or their loved ones are personally threatened by forces that are high in certainty (“it definitely will happen”), severity (“it will be very bad”), immediacy (“it will be very soon”), and duration (“the effects will last a long time”). Fear rarely is so overwhelming that it prevents people from responding, but does impair people’s ability to effectively reason through complex, unfamiliar problems. Fear is especially high when people lack information about the personal consequences of hazard exposure. Technological hazards and terrorist events involving chemical, biological, or radiological agents inherently involve unknown consequences. Many of these agents are undetectable by normal human senses, so people cannot tell if they are being exposed. In addition, some of these agents have long latencies—it takes many years for symptoms to develop—and they result in conditions such as cancer and birth defects that people dread (Slovic, et al., 1980).

Source: Lindell & Perry (1992).

Therefore, it is important for emergency managers to address people’s concerns directly. This is done most effectively through a strategy of information dissemination. As Chapter 4 indicated, emergency managers should not try to give people a university education on the topic. Instead, they should provide clear, direct, and relevant information about the hazard agent and its potential personal consequences. In addition, people should also be told what the authorities are doing to provide protection from the threat and what sources they can contact for additional information. Contrary to popular fiction, the path to fear reduction is through providing—not withholding— information (Quarantelli & Dynes, 1985).

The third basic principle is that people take action when they think they are at risk. The initial response to a threatening situation might be to seek additional information, but those who ultimately conclude they are at risk will take protective action. It is therefore important that official warning messages include recommended protective actions. If authorities do not provide recommended actions, people will take action anyway—implementing the most appropriate actions they know (or hear about from other sources such as the news media and peers) with whatever resources are available to them. Figure 8-5 illustrates some of the factors that people might consider if told they are threatened by toxic chemical release (Lindell & Perry, 1992). First of all, it is important to recognize that many people would not even think of sheltering in-place and expedient respiratory protection unless these were specifically mentioned in a warning message. Second, many people would be uncertain how to implement sheltering in-place and expedient respiratory protection even if told to do so. Third, Figure 8-5 makes it clear that evacuation is highly attractive because it is believed to be far more effective than either of the other two protective actions. It is also attractive because it takes little more skill than sheltering in-place or expedient respiratory protection. However, it requires more time, effort, and money and is likely to experience more obstacles than the other protective actions. The presence of advantages (being much higher in positively valued traits such as efficacy) and disadvantages (being much higher in negatively valued traits such as time, effort, money, and obstacles) is likely to make people ambivalent about evacuating. In addition, research has found people have additional concerns about evacuation in connection with physically destructive hazards such as hurricanes. For example, some evacuees from Hurricane Lili were concerned that evacuation would expose their homes to looters and to storm damage that they might be able to prevent if they remained at home (Lindell, et al., 2005).

Source: Lindell & Perry (1992).

One of the logical implications of Figure 8-5 is that emergency managers might consider some of these ratings to reflect misperceptions. For example, they might believe sheltering in-place and expedient respiratory protection are much more similar to evacuation in their efficacy and much more different in terms of their time and effort requirements. Alternatively, emergency managers might seek explanations for why sheltering in-place and expedient respiratory protection received such high ratings on obstacles. If the obstacles involve knowledge about how to implement these actions, emergency managers could increase the overall attractiveness of these protective actions by a risk communication campaign.

In summary, a message not accompanied by guidance for protective action fails to provide opportunities to reduce fear. In providing protective action recommendations, it might also be necessary to briefly explain why the action will be effective. Tell people why quarantine at home will reduce their exposure to smallpox, or why climbing a canyon wall will provide better protection from a flash flood better than driving their cars out of the canyon, or why taking potassium iodide will reduce radiation exposure damage. Such information accomplishes two important functions. First, it gives people a rationale for complying with official instructions; and second, it discourages people from inventing other apparently reasonable alternative actions that might or might not ultimately be protective.

The fourth basic principle is that some of those at risk will comply with authorities’ recommendations, but their compliance is rarely automatic. The level of compliance is contingent upon a variety of other factors; it varies by information source, message content, and receiver characteristics as well as situational characteristics such as threat familiarity and apparent urgency for response (usually lead time until impact). In cases such as seasonal floods, where risk area residents are likely to be very familiar with the threat and its environmental cues, compliance with protective action recommendations from authorities is likely to be lower than with less familiar threats such as toxic chemicals. When those at risk are familiar with threat agents, they are more comfortable in their own ability to understand the danger, when and where it will materialize, and what should be done about it. Consequently their personal assessment of the situation might cause them to reject official recommendations, or at least thoroughly examine the basis for of such recommendations. Where threat familiarity is low, for example with some hazardous materials or at the initial eruptive activities of long quiet volcanoes, there is little or no personal experience or knowledge of the threat. In such cases, people are more likely to accept (in the response period) the assessment of authorities. Also, people tend to comply more readily when the warning claims that impact is imminent or the threat is visible, simply because there is no time for reflection.

This is a research-based conclusion from the disaster literature, not wishful thinking by emergency managers (Lindell & Perry, 1992). In times of extreme stress, people look to government for guidance. When the agent of destruction is unfamiliar or intangible, or when the consequences appear overwhelming, people’s expectations of protection and help are especially pronounced. Thus, compliance tends to higher with technological and terrorist threats. For example, national opinion polling following the September 11^th attacks indicated substantial increases in levels of trust in government. The combination of citizen concern and a tendency to feel that taking action is important sets the stage for attention to messages from emergency authorities and enhances a positive attitude for compliance. As has been the case in other types of disasters, people tend to return to their normal skeptical attitudes toward government over time. Nonetheless, there is a window of opportunity for emergency managers in the height of crisis and for some time thereafter.

Particularly during the response phase, people appear to carefully comply with directions from police and fire personnel and sometimes other uniformed personnel. For example, in Phoenix, Arizona, during March, 1999, women who were believed to have been exposed to anthrax underwent nude decontamination by male hazardous materials technicians in a decontamination shelter without roof covering while news helicopters hovered above. One person mentioned concern with modesty, but none of the victims hesitated to follow instructions. Since that time, the Phoenix Metropolitan Medical Response System has acquired enhanced decontamination shelters and the ability to deploy “all female” decontamination teams. Nonetheless, the incident stands as an example of citizen compliance with emergency instructions when the threat is unfamiliar and time for action is limited.

The expectation of compliance also places a special responsibility upon local authorities to manage emergencies responsibly. Namely, they must have current, ongoing vulnerability assessment, detection, and prediction systems for threats when technologically possible. They must recognize the varying degree to which different population segments trust them and plan accordingly. Finally, the plans they establish must be capable of being executed. In the absence of adequate plans, people will hold them responsible through the political process and the courts.

Expectations Regarding Stress Effects

Expectations for Health Consequences

According to Dynes’s (1983) emergent human resources model, emergency managers should assume that many risk area residents have disaster-relevant competencies, just as local emergency response organizations do. In addition, emergency managers should also rely on households’ existing patterns of social and task behavior rather than expect them to learn new ones. Finally, emergency managers should also utilize existing authority structures and communications channels (Dynes, 1994). In particular, they need to understand what percentage of the population can perform different emergency response actions and what are the resources (communications devices, information, vehicles/equipment) households need to respond effectively. Accordingly, emergency managers need to avoid stereotyping all households as being identical. Instead they must recognize the differences among households. Moreover, emergency managers must avoid the error of assumed self-typicality; they must recognize that many households are not self-reliant nuclear families who have reliable cars, can (and will) evacuate when they are advised to do so, and have credit cards to pay for their evacuation expenses. Finally, emergency managers should recognize the problems that can arise from the large number of people who converge on the disaster scene to help. Such volunteers, together with local residents may form emergent groups that lack formal elements of organization and, thus, are difficult to work with (Stallings & Quarantelli, 1985). Emergency managers can avoid the confusion arising from convergence of massive amounts of human and material resources on the disaster scene by devising special units to organize volunteers, developing special locations and procedures for donations management, and asking people to donate money to organizations such as the Red Cross rather than sending food, clothing, and medicines.