Information Trickles Up
On September 11, 2001, millions of Americans witnessed an airplane hit the second World Trade Center tower live on television. After that, TV and radio, as well as just about every message traversing the Internet, were devoted to broadcasting information about the attacks.
Amidst the confusion, citizens received and relayed information through the mass media, email, and telephone. They heard instructions about what to do and where to go from broadcasters, friends, and strangers. What they did not hear were the buzzing tones of the Emergency Alert System.
Created in 1963, the Emergency Alert System (EAS) allows the president to commandeer the airwaves on a moment’s notice in case of emergency to relay information and instructions to the public. However, no president has ever activated the system in the 50 years it has been around, despite hurricanes, earthquakes, tornadoes, the Cuban Missile Crisis, the Oklahoma City bombing, and 9/11.
As then–FCC Chairman Michael Powell explained, “The explosion of 24-hour-a-day, 7-day-a-week media networks in some ways has proven to supplant those original conceptions of a senior leader’s need to talk to the people.” New York City Mayor Rudolph Giuliani addressed local residents—and the rest of the nation—on every electronic medium within an hour of the second plane’s impact, but it was not until 11 hours after the attacks that the president made his first live broadcast.
While some may criticize the president’s timing for other reasons, for the purposes of disseminating information, there was no need for him to speak any sooner. In fact if he had spoken sooner, his speech may have gotten in the way of the flow of useful, practical information.
The traditional view of disaster communications holds that in an emergency, information and instructions must flow from an informed federal government, down to local communities and individuals. This view is dangerous and potentially harmful because it fails to distinguish sharing information from centrally controlling decision making. Traditional post-disaster communication centers on issuing orders rather than sharing information; it tells individuals what to do rather than empowering them to make their own decisions. This experts-know-best mentality can be especially dangerous during and immediately after disasters. Because disasters are unplanned, attempting to conform them to normative prescriptions can create immense harm. Encouraging widespread dissemination of timely information, by contrast, does not make assumptions about what will happen in the future.
Critique of Traditional View
During and after disasters, the most critical resource is information. During a disaster, people want and need information about the nature and specifics of the disaster, where they can find shelter, whether municipal water is safe to drink, how to get medical assistance, and what comes next. Afterward, they need to determine when they can return home, how to repair damage, and how to get back to normalcy. In the long haul of recovery after particularly destructive events like Hurricane Katrina or the September 11 attacks, households seek to establish the “rules of the game” for rebuilding and gather the information necessary to make long-range plans for recovery.
The traditional view of information sharing in times of crisis, as reflected in a variety of public policies and official responses to previous disasters, suggests that information travels from a central point of collection, usually an office or a person in the government, down toward first responders, entrepreneurs, non-profit groups, and ultimately citizens. Newspapers and radio and television broadcasts act as a conduit, bringing information from governments to citizens.
This model, however, rests on two incorrect assumptions: first, that a central authority has access to the information most needed by people on the ground, and second, that this authority can best disseminate this information in a timely, accurate, and intelligible format.
The first assumption has never been correct. During and after a disaster, the people affected by the event are the same ones who disperse information to others. Moreover, much of the most relevant information that these people share during a disaster may seem extraordinarily picayune to people removed from the details of the event. For instance, knowing where gasoline is available and at what price is critical to achieving a successful long-distance evacuation, but no central authority could access all of this information, especially during the chaotic scramble to evacuate a city. Additionally, the idiosyncratic nature of disasters means that crucial information varies from event to event, depending on factors such as local infrastructure conditions, culture, and specific needs.
When faced with disasters, people must make decisions under much greater uncertainty than they face in day-to-day decision making. Nevertheless, people still must make decisions. While nobody knows more than a fragment of all possible relevant information, by assembling collected fragments, people can acquire enough information to decide how to act under the circumstances. Thus, a supply of accurate and timely information reduces that uncertainty, thereby allowing people to make better decisions about whether to remain or evacuate, how to respond to imminent threats, where to find critical supplies, and the myriad other minute-by-minute decisions that present themselves as situations unfold.
The communal production and dissemination of information reflects the fact that people respond to and rebuild after disasters within a larger community. People make decisions not in isolation, but along with their friends, families, neighbors, business partners, and many others. To return to our evacuation-by-automobile example, people cannot evacuate by car if they cannot buy sufficient gasoline to get to safety, so they are dependent upon gas station operators, who are in turn dependent on oil companies (for the supply of the product) and electric companies (for the power to dispense the product). If gas is unavailable, people may seek other means of evacuation, such as carpooling with friends or neighbors, or they may chose to shelter in place. Sheltering requires being able to stockpile food and water, for which people depend on grocery stores and their supply chains.
The second incorrect assumption is that information flows only in one direction: down. This may have been partially correct in a time when information typically flowed through hierarchical networks like radio and television broadcasts that do not allow interaction between end users, but it no longer applies. Networks can and do arise spontaneously. Information is virtually costless to disseminate and sort; social media and online communications have become woven into the fabric of many American communities, especially in younger and middle-aged demographics. The current dire financial straits of many of America’s newspapers illustrate the challenge that new means of discovering, aggregating, and disseminating news and information offer to traditional models and traditional media.
Moreover, even when government officials give specific orders, many people simply choose to ignore them. For instance, despite being under a mandatory evacuation order, between one- quarter and one-third of Galveston Island’s residents did not evacuate in advance of Hurricane Ike. Short of turning pre- and post-disaster areas into police states, there is little that can be done about this. Furthermore, the decision to evacuate or shelter in place in advance of a hurricane, for instance, only appears wise or foolish in retrospect. While it is easy to say after a disaster that a certain action was prescient or foolhardy, in advance of that crisis there is no reliable way to know what is the “best” thing to do, particularly as what is good for one person or group may be bad for another. People have to make their own decisions. To do so, they need access to good information on which to base those decisions.
This traditional view of information sharing after disasters, then, not only incorrectly assumes that it is possible to centralize the dissemination of information but that it is also possible to first aggregate those data. That is, for a central authority like the government to share the information that people need to make decisions during and after crises, the same authority must first pull that information together.
An Alternative View of Information Production and Dissemination
In order to think about the process of aggregating and disseminating information differently, it is important to begin by thinking about how and where information is produced. As Emily Chamlee-Wright has argued, people on the ground play a vital role in generating information about recovery from disasters that empowers others to make informed choices:
Because the information people use to make decisions before, during, and after disasters is generated largely not from centrally issued commands but through interactions with others, it makes sense that information be disseminated on a local level as well.
It turns out the people who produce information share it—consciously, through social media like blogs and Twitter, and unconsciously, through the information that their behavior and actions provides to others. Both formal studies and anecdotes highlight the importance of information technology and social media in disaster response. One study, for instance, found that a third of respondents to a survey of evacuees from the October 2007 California wildfires reported posting information or discussing events online during the fires. In other words, a third of the population escaped their traditional role as passive receivers of information and instruction and became citizen journalists sharing their local knowledge with a broader audience.
An FCC panel charged with investigating communications during Hurricane Katrina noted there was a “lack of activation” of the top-down Emergency Alert System. What was not lacking was the activation of the Internet: Ad hoc blogs and other Web sites popped up with eyewitness reports, lists of those missing, and offers to match victims with those willing to help.
Craiglist.org, for example, became a central hub for unplanned relief efforts. Persons around the country who wanted to offer shelter to victims listed what they had to offer. According to the Raleigh News & Observer, “One offer asked for Christians—no drinkers. One family requested liberals. Another advertised space for a lesbian. Some invited guests to stay as long as they wanted. Others set time limits.”
“The way our site works is that we’ve given people the ability to do what they need without waiting for us,” Craigslist creator Craig Newmark told the Associated Press. “People as a whole do a lot better job than we could ever do.”
SOCIAL MEDIA TO THE RESCUE
When Hurricane Ike hit the Texas coast in September 2008, residents acquired information in a variety of ways. Some were very traditional, such as following television news broadcasts. Others were more innovative. A number of blogs sprang up to document what was happening in real time (including one by one of the authors), and people used a convention on Twitter known as a hashtag to flag information as relevant to people in the hurricane-affected areas. Individuals also used Craigslist.org to buy and sell equipment and locate lost pets. These critical activities would have been significantly slower and immeasurably more expensive if users had not had access to spontaneous networks and decentralized information sharing technologies.
Power and communication failures, while problematic, did not shut people out from accessing electronic information technologies. Despite widespread electrical outages, people were able to recharge cellular phones and laptop computers using automobile chargers or in shopping malls in areas that did have electricity. Cellular coverage remained spotty, but it was available. (Indeed, one of the authors filed an op-ed column from Houston for a New York–based newspaper using an extended-life laptop battery and a cellular modem just hours after Ike had moved through.)
Similarly, when Fargo, North Dakota, suffered from torrential flooding in March 2009, a resident set up a Facebook group to bring together people willing to volunteer to reinforce dikes. Within hours of being established, it had garnered over 4,500 members. By the next day, there were eight different Fargo flood-related groups and events on Facebook with more than 50 members. In one 24-hour period during the height of the flooding, there were over 1,500 Twitter entries, or “tweets,” related to Fargo; some were messages of support, but others contained information of value to other Twitter users.
These changes even affect conventional media like television and radio. After Ike, news stations in the Houston and Galveston areas carried not just information from their reporters, but also from viewers, who sent in information about what was happening in their areas. For instance, the afternoon before the hurricane struck, one television station (KHOU, the local CBS affiliate) put information on its news ticker called in by pet boarding facilities with remaining vacancies—crucial information for people making the decision to evacuate at the last minute. Similarly during Hurricane Katrina, local radio personality Garland Robinette stayed on the air throughout the worst of the storm, relaying information called in and text messaged by listeners.
After the immediate crisis passes, people continue to disseminate information both through formal and informal means. After major disasters, people make decisions about how, where, and when to rebuild based largely on information generated by other people’s actions—commitments to rebuild, store reopenings, and so on. In all crises, people share vital local information such as the location of blocked streets, power outages, and open gas stations and grocery stores through informal networks. Traditional media can play an important role in spreading this information as well. For example, the Houston Chronicle published lists of grocery store reopenings after Ike.
This information, combined with freedom of movement and policies that allowed rather than hampered local entrepreneurship, got the recovery from Ike started on the right foot. As one of the authors argued in the aforementioned column written just hours after Ike,
Because city and county governments are doing what they should do—enforcing the law, sharing critical information and making honest assessments of the status and future of public services—they have cleared the way for the private sector to respond effectively. By yesterday morning, all local grocery chains had reopened at least some of their locations, and their trucks had made it into town and were busy resupplying. This would have been impossible if the city had been locked down, or if employees had been prohibited from coming to work.
Stressing that people should use their judgment rather than trying to freeze movement, officials have created space for what reports indicate is an incredible—and uncoordinated—response by people clearing streets and storm drains. The official attitude that recovery is a grassroots effort, of which government is just one sector that plays a supporting role, means that recovery is already underway, and people don’t have to wait for officials to draw up (and eventually fumble) a complex, top-down plan.
This locally generated and disseminated information may even serve to correct erroneous information coming from official channels. After Ike, the mayor of Galveston announced,“There is nothing to come here for right now…. Please leave.” But when grocery store employees began preparing and selling food out of their store’s parking lot, their actions offered a powerful rebuttal to the mayor’s proclamation. Similarly, by reopening just two days after Ike passed, a Galveston Home Depot created information that both the traditional media and ad hoc networks dispersed about the timing of home repairs and cleanup on the island. These positive signals and concrete information were made all the more relevant because of the poor performance of the Galveston city government, which the Houston Chronicle characterized, a mere week after Ike, as plagued by “petty bickering” and “acrimonious” council meetings.
Why Spontaneous Networks Are Preferable
There are four key elements that suggest that moving from a model of government issuing orders about how to respond to crises to one where information is made available to citizens so that they can make their own decisions would lead to better outcomes.
First, the decentralized model takes advantage of the fact that everybody knows something, but nobody knows everything. As the economist F. A. Hayek argued in 1945,
Information technologies make it possible for people to combine their knowledge and produce information superior to that held by any single person or formal organization. For instance, after the April 16, 2007 shootings at Virginia Tech, a Wikipedia entry on the massacre correctly identified the victims before the university administration had formally released these details. Even in much less dire circumstances, the feedback mechanisms present in marketplaces like eBay, ratings systems offered on products by Amazon, and service reviews sold or given away through Angie’s List and Yelp aggregate information from far more people—and in a fraction of the time—than similar, pre-digital services. Decentralization of information sharing, combined with a democratic digital commons, yields more information with a higher degree of accuracy at a much lower cost than centralization.
Second, decentralizing information sharing reduces the possibility of a single point of failure. If there is only one central source of information or coordination, then the collapse of that source means that no information would be available, and as the destruction of New York City’s emergency management command center at 7 World Trade Center during the September 11 attacks showed, official institutions are as vulnerable as unofficial ones during emergencies.
Additionally, a central source of information can displace other more useful sources of information. One New Jersey resident told the New York Times that during 1991’s Hurricane Bob, he was listening to very useful weather reports on local radio only to have them preempted by “broad and far less useful alert system warnings.” In fact, if the president were to ever activate the national Emergency Alert System, simple on-screen text and a presidential voice-over would replace all programming currently on the air on all channels, no matter how useful the original broadcasts.
Third, central authorities after disasters have incentives that reward appearing to help rather than actually helping. After disasters, there is a significant gap between what officials are capable of accomplishing and what political pressure requires them to appear to be doing. For instance, President George W. Bush was criticized in the days after Katrina for failing to come to Louisiana and Mississippi to see the destruction wrought by the hurricane. But the reality is that even if he had been on the ground, there was little that he personally could have done. Indeed, his presence would likely have been counterproductive to recovery efforts, because security measures would have transferred resources from the recovery to the physical safety of the president.
Governments do not have a good record of disseminating information, much less processing that information effectively. After Hurricane Katrina had passed, it was widely believed that New Orleans had “dodged a bullet.” The Seventeenth Street Canal breach was known to public officials yet not reported through official channels. Officials had been warned that the city was likely to flood, but they did not relay this information. Similarly (though with far less tragic results), after Hurricane Ike, the City of Houston and Harris County announced curfews that applied except in certain circumstances. But neither the Web sites of the city, the county, the Houston Police Department, nor the Harris County Sheriff posted specifics about the curfew, such as what the exceptions were, what the penalties were for violation, and other details.
Empowering people to share and act on local knowledge, rather than requiring them to follow orders given centrally, reduces the incentives for officials to act in ways that are politically expedient rather than expedient to the recovery. And the benefits are not just seen in more accurate and more available information, but may accrue in other ways. As Sutton, Palen, and Shklovski suggest, “sharing of information via text-based sharing sites can serve a dual purpose of providing much needed information to others through a psychologically beneficial practice of talking about traumatic events.”
Fourth, and most importantly, information dispersed and shared through a variety of redundant networks reflects reality. Governments and central planners have remarkably poor track records of trying to organize and coordinate responses to disasters. With the widespread adoption of information technologies that reduce the costs of sharing information between wide groups of individuals—who need not even know one another personally—the cost of obtaining and sharing information is continually decreasing. Just as consumers are increasingly empowered with information about the characteristics and costs of different goods and services, people preparing for and responding to disasters have more information available to them that they use to inform their decisions.
A common critique of decentralized information sharing during disasters is that rumors and untruths can spread. While this is undoubtedly true to some extent, its magnitude tends to be greatly overstated. As Judge Richard Posner has observed, “the blogosphere as a whole has a better error-correction machinery than the conventional media do.” Information put out through informal networks faces virtually instantaneous correction by that network if it is wrong. In all likelihood, the more correct and useful information will be repeated, while false or less valuable information will be drowned out. As one study points out, the “citizen-generated victim lists” after the Virginia Tech shootings “were never incorrect” (emphasis in original) because of self-policing by collaborators.
Moreover, when information flows only in one direction—such as from public officials through traditional media to citizens—there are no error checks on incorrect information. Such systems are less robust, or prone to self-correction, than are network-based information sharing systems. If a blogger makes a misstatement of fact, or if facts change after the blogger posts his, the readers or other bloggers will likely correct the blogger; this correction can take place in a matter of minutes. By contrast, if a public official makes a misstatement which is picked up and broadcast or printed by the media (such as New Orleans police chief Eddie Compass’ completely untrue statement that there were “babies being raped” in the Superdome immediately after Katrina), it can take much more time and energy to correct.
The model that we suggest—where behaviors before, during, and after disasters are governed by information, largely created and disseminated spontaneously and without central direction and not dictates—takes advantage of what is widely called the “wisdom of crowds,” reduces the spread of incorrect information, avoids incentive incompatibility issues, and comports with the reality of the networked world. Nonprofits and social entrepreneurs play a vital role in encouraging this information creation and dissemination.
Along with blogs, the social media service Twitter has become an indispensible communications tool during and after disasters. Earthquakes in Mexico City and Sichuan were first reported on Twitter, as was the January 2009 US Airways plane landing in the Hudson River. The fact that one can read and post messages not only using a computer, but also cell phones and PDAs, means that one is not dependent on electric power or any particular way of reaching the Internet.
During Hurricane Gustav last year, Gulf Coast residents used Twitter to relay local reports. Users spontaneously settled on a common “hashtag” keyword with which they tagged their posts. All posts with a common tag could then be easily aggregated for a comprehensive view of local information. The hyper-localism that Twitter allows is powerful. For example, the May 2009 coal ash spills in Tennessee and Maryland received very little mainstream media coverage, but affected residents coordinated and shared information using Twitter and the hashtag “#coalash”.
Nonprofits as Coordinating Agents
Nonprofits and social entrepreneurs have a critical role to play in creating and disseminating information before, during, and after disasters. As Chamlee-Wright and Virgil Storr demonstrate elsewhere in this volume, social entrepreneurs play an important role in coordinating community recovery; a significant part of this involves collecting and disseminating information.
As discussed earlier, information is only useful in context; social entrepreneurs and entrepreneurial non-profits help not only to collect and disseminate information but also put it in context. Information about what is happening in communities is what economists would call a public good: it is non-rival, meaning my knowing something doesn’t preclude your knowing something, and it is non-excludable, which means that it is very hard to exclude anyone from knowing something. Traditional examples of public goods include national defense, streetlights, and clean air.
Public goods are traditionally regarded as something that governments provide, since the private market is unlikely to supply them because of their non-excludability. However, as we have demonstrated, governments are ill-equipped to provide all of the information people need when making decisions about how to respond to and rebuild after disasters, and information that comes from governments is usually more about controlling actions than it is providing fair and unbiased information.
Into this milieu step social entrepreneurs and nonprofits to provide this information, and most importantly, provide it in a context that allows people to make decisions; that is, information which helps people make decisions about response and rebuilding in concert with their neighbors. In this way, non-profits and social entrepreneurs act as coordinating agents and provide public informational goods in a manner that governments are neither equipped nor inclined to do.
To take one example, New Orleans’ Lower Ninth Ward Neighborhood Empowerment Network Association appointed block captains shortly after Katrina. Their job was to collect information on the whereabouts and future plans of the Lower Nine’s residents and report it back to the association. As a result, homeowners and residents were able to incorporate their friends’ and neighbors’ plans into their own plans for returning and rebuilding.
The local knowledge that these social entrepreneurs generate, share, and act on keeps communities connected and allows citizens, social and for-profit entrepreneurs, and outside groups to better work together. Tronn Moller from the National Council of Churches explains how churches played a connecting role in communities after Katrina,“[Church is] a very interesting connecting piece. Either the faith leader or the congregants… intertwine the communities. They would go walk the community… they know who [the reverend] is, they know who the deacon is, or someone up [at the church] either lived in the community, or have had some kind of tie to the community. And therefore, they brought information to them.”
There needed to be some coordination around [rebuilding and revitalization efforts], so for the last two years, we have actually been convening… every week with a conversation around what everybody is doing… It is the clearing and the space for those conversations to be had.
…[T]hose people who are doing development work, those that are doing investment work, and those people who are doing planning work, could have a place to coordinate so we don’t trip over each other or come into conflict. There are so many strands going together to reweave the fabric that is New Orleans and Central City, in particular, it’s easy enough for the wrong thing to be done for the right reason and with the right effort. That’s one of the places where we’ve been able to, I think, participate and contribute the most.
We were also an incredible emotional respite for people. We always had a pot of coffee going. We always had snacks and food. We tried to create a very homey atmosphere where people were comfortable and didn’t feel intimidated coming in and overwhelmed. We would literally sit and address their issues, point by point, make a list, and start connecting them to the resources that address their needs….
I truly feel like every neighborhood, if they have had a center like this, they would be further along. Because people needed a hub to go to for answers, for resources, to get immediate actions going. We weren’t bureaucratic. There was not red tape to go through. You needed help, you got it.
Social entrepreneurs also help share information provided by other channels. As several of our case studies throughout this volume demonstrate, social entrepreneurs link with one another, thus expanding the networks of available information. Additionally, social entrepreneurs can identify the kind of information people on the ground need and help them obtain it in a less bureaucratic and costly fashion than other means. Indeed, a Department of Homeland Security report based on lessons learned from Katrina identifies nonprofits and non-governmental organizations (NGOs) as critical to providing access to information and the importance of internet access to Katrina evacuees.
In addition to merely sharing this information, social entrepreneurs and nonprofit groups play a key role in creating this information through the signals they send. By reopening doors after a disaster, nonprofit groups signal that there will be a community to return to; if recovery from disaster really means a return to normalcy, then the resumption of the services that non-profits and social entrepreneurs offer hastens that return. Public policies that squelch these signals can, by contrast, slow rebound efforts. As Chamlee-Wright argues, “The return of church services or a functioning school or a medical clinic can send a clear signal that coordinates behavior in the direction of community rebound, but threats of widespread use of eminent domain can, with a pen stroke, eliminate this clarity.”
Moreover, many social media like Twitter and blogs have a strong ethos of voluntary collaboration and cooperation. These media are unlike any that came before in one essential respect: their content is almost exclusively created by unpaid volunteers whose only reward is the joy of expression and perhaps status within a community. Additionally, these new media encourage the formation of small niche communities, whether by geographic region or special interest. As a result, members jealously police their community, and their shared niche affinity encourages collaboration.
Information is only useful in context; knowing that the Earth revolves around the sun matters for the study of the cosmos, but it does not make much of a difference to our daily lives. During and after disasters, the context—the situation on the ground—is constantly changing, and information adjusts with it. Information is most useful when it is timely, accurate, and intelligible; that is, when it helps people make informed decisions about how to best deal with and recover from disasters. And that information is best produced and disseminated on a local level.
The traditional model of how information is produced and disseminated during and after disasters is based on old and disproven assumptions: that a central planner knows best and can disseminate orders to the masses, who must follow these directives. This model was churlish before the widespread democratization of information and communication technologies; today it is positively antiquated. As we have seen in the aftermath of disasters both natural and man-made, information travels in a variety of directions and through a variety of means. Attempts to control the spread of information will only be counterproductive, as people want information on which they can make informed decisions, not orders about what to do.
Because during and after a disaster everyone knows something—and nobody knows everything —the goal should be to aggregate and disseminate accurate information in a timely way. And the best way to do that is to respect that everyday citizens are producers and disseminators, and not just consumers, of all-critical information.
 By “rules of the game,” we mean the institutions, laws, norms, and customs that dictate the rules of formal and informal interactions between people.
 For more on the importance of information and the effects of “information poverty,” see Tisha Slagle Pipes and Jarrod D. Knudson, “Katrina Bankrupts the Information-Rich: Information Poverty in Slidell, Louisiana,” in Learning from Catastrophe: Quick Response Research in the Wake of Hurricane Katrina, Special Publication 40, Boulder: Natural Hazards Center (2006), pp. 403–28.
 Jennifer Latson, Terri Langford, and Harvey Rice, “In a ‘Downward Spiral,’” Houston Chronicle, September 16, 2008.
 Emily Chamlee-Wright, “The Long Road Back: Signal Noise in the Post-Katrina Context,” Independent Review 12, no. 2 (Fall 2007).
 Jeannette Sutton, Levsia Palen, and Irina Shklovski, “Backchannels on the Front Lines: Emergent Uses of Social Media in the 2007 Southern California Wildfires,” Proceedings of the 5th International ISCRAM Conference, F. Friedrich and B. Van de Walle, eds., Washington, DC, 2008.
 Independent Panel Reviewing the Impact of Hurricane Katrina on Communications Networks, Report and Recommendations to the Federal Communications Commission, June 12, 2006, 28, www.fcc.gov/pshs/docs/advisory/hkip/karrp.pdf
 Josh Shaffer, Triangle residents offer victims place to stay, Raleigh News & Observer, A16, September 2, 2005.
 Phil Davis, “Internet company promises to find hurricane evacuees a shelter,” Associated Press, June 3, 2006.
 Brad Hem, “Online networks help keep people in touch,” Houston Chronicle, September 15, 2008. Surprisingly, this appears to be the only article about the role of social media and informal networks that the Chronicle ran within a week of Ike’s landfall.
 Patrick Condon, “Fargo uses social networks to flight floodwaters,” Associated Press, March 26, 2009.
 Author’s observation, www.facebook.com, 3:15 PM EDT.
 Author’s calculation, based on tweets containing “Fargo” between 0:01 GMT and 23:59 GMT on March 26, 2009.
 Emily Chamlee-Wright and Daniel M. Rothschild. “Disastrous Uncertainty: How Government Disaster Policy Undermines Community Rebound.” Mercatus Policy Series, Policy Comment No. 9. (Arlington, VA: Mercatus Center at George Mason University, January 2007).
 Daniel M. Rothschild, “Ike: Recovery That Works.” New York Sun, September 15, 2008.
 “Ill-Conceived: ‘Look and Leave’ Blunder Makes Galveston Mayor Thomas’ Leadership Challenge Even More Difficult,” Houston Chronicle, September 18, 2008.
 F. A. Hayek, “The Use of Knowledge in Society,” American Economic Review 45, no. 4, 519–30.
 See James Surowiecki, The Wisdom of Crowds (New York: Random House, 2004).
 Leysia Palen, Sarah Vieweg, Jeannette Sutton, Sophia B. Liu, and Amanda Hughes, “Crisis Informatics: Studying Crisis in a Networked World,”Third International Conference on e-Social Science, Ann Arbor, October 2007.
 Joseph Geller, An Alert System That Should Stay Mum, New York Times, December 26, 2001.
 Jed Horne, Breach of Faith: Hurricane Katrina and the Near Death of a Great American City (New York: Random House, 2006), 61–62.
 Author’s observation during Hurricane Ike.
 Sutton, Palen, and Shklovski, “Backchannels on the Front Lines.”
 Richard A. Posner, Essay: Bad News, New York Times, July 31, 2005, Book Review Section,.
 Palen, et al., “Crisis Informatics.”
 For more on this, see the case study on Tim Williamson of the Idea Village from Local Knowledge: Is the Gulf Coast Open for Business? (Arlington, VA: Mercatus Center at George Mason University, July 2008), http://www.mercatus.org/localknowledge/Compendiumsplit/16_casestudy5.pdf.
 Heralding Unheard Voices: The Role of Faith-Based and Non-Governmental Organizations during Disaster, Department of Homeland Security, December 2006, 86, www.homelandsecurity.org/hsireports/Herald_Unheard_Voices.pdf.
 Emily Chamlee-Wright and Virgil Storr, ”The Role of Social Entrepreneurship in Post-Disaster Recovery,” (working paper 08-19, Mercatus Center at George Mason University, 2008).
 Clay Shirky, Here Comes Everybody: The Power of Organizing without Organizations, Penguin (2008), 140–42.