Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 15
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary 2 The Science and Practice This chapter summarizes introductory presentations and discussions of the first session of the workshop. The purpose of the session was to introduce participants to the workshop charge (see Box 1-1) and the terminology to be used throughout discussions. Speakers were invited to provide basic information on the sciences and practices of social network analysis (SNA), fostering community resilience, reaching vulnerable populations through social networks, and the use of social networking tools to improve communication. WORKSHOP VOCABULARY During the first day of workshop discussions, the workshop planning committee observed inconsistent use of key terms related to social networks and SNA by participants. To avoid confusion, the committee provided definitions for these terms (see Box 2-1). A social network is a group of people and organizations that form a web of relationships. Social networks were being confused with the tools used to facilitate them (such as Facebook1) or to analyze them. Social network analysis is the process of analyzing the key actors and connections within a social network. SNA can reveal redundancies and vulnerabilities within a network, and can be used to study the changes in all these variables. A product of SNA may be a graphical representation of a network that shows the interconnectedness of network members. An example is provided as Figure 2-1. Issues were also encountered with the use of the term “resiliency.” As described in Chapter 1, resiliency is the ability of a social unit to withstand external shocks to its infrastructure (Norris presentation to workshop participants). 1 Facebook is a free-access, privately owned social networking website. See www.facebook.com (accessed March 24, 2009).
OCR for page 16
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary BOX 2-1 Definitions of Key Social Network Terms The workshop planning committee developed the following definitions of key terms used in the study of social networks and social network analysis. Social network. The interactions between people and organizations, including who knows, works with, or communicates with whom, that can be mapped. The data and information found, for example, in Facebooka and the Enron Email Corpusb are examples of social networks. Social network tools. A set of computational techniques that enable individuals and groups to engage in social networking by monitoring and interacting within the networks with which they are connected. Facebook, MySpace,c and Twitterd are examples of social networking tools. Social networking. The process of creating, maintaining, or altering one’s network to one’s advantage by using the network to gain resources or influence, or to mobilize activity. Social network analysis. The process of analyzing a social network and identifying key actors, groups, vulnerabilities, and redundancies as well as the changes in these variables. Social network analysis tools. The set of tools, technologies, metrics, models, and visualization techniques used for social network analysis. These may include data extraction tools, link analysis, statistical techniques, and graph theory techniques using programs such as AutoMap,e ORA,f UCINET,g and Pajek.h Social network theory. The set of theories for forecasting, reasoning about, and understanding how social networks form, are maintained, and evolve, and the role of variables such as social networking tools, media, and stress in affecting the emergence, utilization, management, and change in social networks. aSee www.facebook.com (accessed March 2, 2009). bSee, for example, ebiquity.umbc.edu/blogger/2006/02/05/search-the-enron-email-corpus-online/ (accessed March 2, 2009). cSee www.myspace.com (accessed April 5, 2009). dSee twitter.com/ (accessed March 2, 2009). eSee www.casos.cs.cmu.edu/projects/automap/ (accessed May 10, 2009). fSee www.casos.cs.cmu.edu/projects/ora/ (accessed May 10, 2009). gSee www.analytictech.com/downloaduc6.htm (accessed May 10, 2009). hSee vlado.fmf.uni-lj.si/pub/networks/pajek/ (accessed May 10, 2009)
OCR for page 17
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary FIGURE 2-1 Graphical representation of a social network. SOURCE: Kathleen Carley, Carnegie Mellon University, Institute for Software Research International (2009). THE STATE OF THE ART IN SOCIAL NETWORK ANALYSIS Kathleen M. Carley of the Center for Computational Analysis of Social and Organization Systems (CASOS) of the Carnegie Mellon Institute was invited to give a presentation on the state of the art in SNA. She provided an overview of the main tools of SNA that focus on defining the characteristics, composition, and structure of existing social networks. Her presentation and subsequent discussion are summarized here. Other topics and issues may be relevant to SNA, such as the study of ego networks, but were not discussed at the workshop and not included in this summary. Unless otherwise noted, the ideas expressed in the following sections are attributable to Dr. Carley. People, units of action, partners, departments, resources, ideas, skills, events, and countries can be graphically represented as nodes in a network (for example, the dots in Figure 2-1). The links—or ties—between the nodes are the interrelations and may represent physical ties such as roads or rivers, or less tangible ties such as alliances, associations, authority lines, transfer of resources, precedence, or who likes or respects whom. SNA can be used to identify and understand the relationships and strengths of the ties within a network, and understand how these ties are vulnerable under certain circumstances. SNA can also be used to conduct sentiment analyses to understand the flow of ideas or feelings. Classic social network and link analysis modeling and basic pattern detection capabilities are readily available but not commonly applied in policy making. Nonetheless research in SNA is rapidly expanding and cutting-edge technologies yield exciting results with sociopolitical ramifications. SNA is mostly unused at the local level with the exception of epidemiological studies, such as the tracking of disease sources,
OCR for page 18
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary and in counterterrorism and counternarcotic investigations. Most of what is known about social networks, however, is not integral to disaster management practice. SNA for any purpose is often thwarted by the discrepancy between the amount of data needed to yield meaningful results and the amount of data available. The lack of technological capacity and the lack of social science skills needed to correctly apply SNA-derived models are also factors. Researchers addressing the data question do not expect to close this gap within the next decade. Cutting-Edge Approaches: Dynamic Network Analysis Traditional SNA focuses on nodes within a network and considers the attributes that make an individual node stand out. The state of the art in the field of SNA is beyond determining who communicates with whom. Newer approaches to SNA consider networks as a whole, and powerful techniques exist that allow the analyses of the what, where, how, why, and when of situations. These techniques enable a user to identify the need for interventions, plan for them, and provide input for policy management. Most agencies, including disaster management agencies, currently collect “trail data,” such as who entered a health department on what day for what information, or who crossed the border at what time. State-of-the-art data collection includes a dynamic network analysis suite of three types of tools to track and analyze trail and other network data. The suite includes (1) data mining tools, such as AutoMap, that collect network data from open sources such as newspapers to identify network components; (2) statistical analysis-type packages, such as Organizational Risk Analyzer (ORA), that take into account social and dynamic network metrics, conduct broader data mining and link analyses, and apply machine learning techniques for clustering; and (3) simulation tools, such as DyNet, Construct, and BioWar, that allow scenario analysis for the consideration of various options. Box 2-2 provides descriptions of some of these tools. Although dynamic social network analysis is used in some applications, many barriers exist to their widespread use for analyzing complex networks. It is difficult to extract text and links from the wide range of required data sources. Many applicable analytical techniques for city-scale networks require extensive computational resources. Additionally, many simulation models are built for a single purpose and cannot be reused, quickly making them obsolete. Finally, interpreting and moving this level of information into the policy realm and scale is not yet a functional reality.
OCR for page 19
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary BOX 2-2 Dynamic Network Analysis Tools AutoMap is a product of CASOS at the Carnegie Mellon University and is a text mining tool that enables the extraction of network data from texts. The tool can extract content analytic data (words and frequencies), semantic networks, and metanetworks. The main functions of AutoMap are to extract, analyze, and compare mental models of individuals and groups, and to reveal the structure of social and organizational systems from texts. SOURCE: www.casos.cs.cmu.edu/projects/automap/ (accessed March 21, 2009). BioWar is a CASOS package that enables community leaders to prepare for biological attacks using computational models of social networks, communication media, disease models, demographically accurate agent modes, wind dispersion models, and a diagnostic error model combined into a single model of the impact of an attack on a city. SOURCE: www.casos.cs.cmu.edu/projects/biowar/ (accessed March 21, 2009). Construct, developed by CASOS, is a multiagent model of group interactions where agents communicate, learn, and make decisions in a continuous cycle. The program takes into account how agents learn through interaction and change their perception of the environment. SOURCE: www.casos.cs.cmu.edu/projects/construct/info.html (accessed March 21, 2009). DyNet is a reasoning support tool developed by CASOS intended to simulate reasoning about dynamic networked organizations under varying levels of uncertainty using computer science, social network, and organization theory. SOURCE: www.casos.cs.cmu.edu/projects/DyNet/dynet_info.html (accessed March 21, 2009). i2 Analyst’s Notebook is a commercial visual investigative analysis tool that allows investigators to organize large volumes of disparate data and conduct link and timeline analyses. SOURCE: www.i2inc.com/products/analysts_notebook/ (accessed May 12, 2009). Organizational Risk Analyzer (ORA) is a risk assessment tool developed by CASOS that examines network information and identifies individuals or groups that are potential risks to a network given social, knowledge, and task network information. SOURCE: www.casos.cs.cmu.edu/projects/ora/ (accessed May 12, 2009). Palantir is a commercially available information analysis platform for integrating, visualizing, and analyzing structured, unstructured, relational, temporal, and geospatial data for security, intelligence, defense, and financial applications. SOURCE: www.palantirtech.com/ (accessed May 12, 2009). R is a computer language and environment for statistical computing and graphics developed by Bell Laboratories. SOURCE: www.r-project.org/ (accessed May 12, 2009). Starlight Information Visualization System is a visualization-oriented user interface for temporal and spatial information analysis and network modeling developed by the Pacific Northwest National Laboratory. SOURCE: starlight.pnl.gov/ (accessed May 12, 2009). UCINET is a commercially available comprehensive package for the analysis of social network data using a variety of network and statistical analysis methods. SOURCE: www.analytictech.com/ucinet6/ucinet.htm (accessed May 12, 2009).
OCR for page 20
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary Functional Applications in Dynamic Social Network Analysis Available tools make it possible to conduct network analysis with open-source, raw-text data input, such as information from newspapers, and then conduct scenario analysis (e.g., what would happen given a certain set of circumstances), and finally conduct analysis that can identify emergent leaders. Analysis is possible, for example, that can connect all of the potential emergency responses in a community to specific emergency responders in order to see where vulnerabilities in a response network exist. Location analysis can be done to see how things are done differently in different areas. Geospatial network analysis is possible, as is information gain and loss tracking. It is also possible to detect changes in organizations and behavior over time within a network using available analytical techniques. If the resources are available, the mapping of belief structures and trends over time is possible and may allow policy makers to identify where people hold certain beliefs, where beliefs are likely to change, who the critical actors are that enable change, and to predict who will be central to the network in the future. Belief forecasting analysis can be conducted for given types of network structures and sociodemographics. The results of the analysis provide policy makers and leaders with powerful information to help them determine how best to communicate with and enlist the assistance of their communities. The resulting data and models from dynamic social network analyses can allow managers to identify critical network features, identify opportunities for intervention analysis or action, and conduct limited types of event forecasting. Issues and Knowledge Gaps in the Application of SNA Traditional SNA technologies that are able to reveal weaknesses in response networks, identify vulnerable populations, target opinion leaders in communities, or conduct text mining to support hot-topic analyses are not regularly utilized in policymaking settings. This is also true in the disaster management community where, in general, networks do not exist that link emergency responders with one another or with networks elsewhere in the community. Some workshop participants expressed the view that SNA could be applied in analysis of the emergency management community and emergency response plans that are in place at the national, state, and local levels. Proper Use of Tools Framework modeling and network statistical analysis tools are readily available to community and disaster managers, but those using them often are not familiar with community social science models. Under such circumstances, statistical analyses may be overapplied, good interpretation of network situations may be missing, and resulting models may be in error. Even scientifically sound network models may be used incorrectly, or metrics for change may be misinterpreted. According to some workshop participants, increased communication between social scientists in the research community and modelers within emergency management communities would be beneficial. A barrier to collaboration, however, is that researchers and practitioners do not use the same
OCR for page 21
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary analysis tools. Different tools are used, in part because of the cost and accreditation of software, and because of the scalability and visualization capabilities of various software packages. Need for Translational Research To realize the benefits of SNA, it is essential that information move from the research realm into practice, also known as translation of research. Research conducted on the best means to translate information is defined as translation research. Translation research can be helpful in developing an educational process that demonstrates how the adoption of new ideas and tools will yield actionable results for practitioners. Methodologies, language, and examples that would be most meaningful to a target audience can be identified. Louise Comfort of the Graduate School of Public and International Affairs, University of Pittsburgh, described experiences promoting the use of networking technologies in different municipalities. She found that new technologies could be quickly accepted if individuals were willing to think beyond their traditional routines (Comfort and Wukich, 2009). She described the interaction and process of learning that occurs on at least three levels to create a learning network: Individuals learn the technologies; Efficient and easier communication occurs among individuals as a result; and Collaboration and validation of information occurs. Generally, younger personnel with access to better equipment were more willing to accept new technologies. Acceptance of technologies into practical use has occurred within the drug-traffic enforcement and healthcare communities. Acceptance more often occurs when champions of the technologies are identified within the communities to collaborate in the development of educational materials. Similar relationships would be useful to the disaster management community. Translational activities validate basic findings for practitioners in language that can be understood, and can result in the decrease in the time to move a concept from the research realm into practice. Dr. Carley indicated that the engineering field engages in translational research that results in a relatively small lag of six years between the inception of an idea and its practical application. She stated that the translation of complex SNA techniques is estimated to occur only after approximately 24 years because the SNA community is not actively engaged in translational research and activities. Without translation research, widespread benefits resulting from the application of SNA will be negligible. In spite of this, no agency is charged with funding such research. Cost There are different but related costs to be factored in by a community when considering SNA technologies for building community disaster resilience: the cost of the necessary analytical tools, the cost of creating the network of individuals to conduct the analyses, and the cost of creating the community networks necessary to develop community
OCR for page 22
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary resilience. It was discussed during the workshop that a complex network analysis for a system at the city level could require between half a million and several million dollars. The cost can vary significantly depending on the data already available and the level and condition of available hardware. The cost of SNA tools may be controlled by taking advantage of free and currently available state-of-the-art tools. Agencies typically use commercially available software at a cost of thousands of dollars. Validation of Models According to Dr. Carley, confidence in models developed using SNA tools is necessary before policy makers will make model-based decisions. However, global datasets essential to validating models do not exist. This is particularly true in the area of disaster preparedness for which large-scale baseline or control data for comparison to projected models are not available. Detailed data may be available regarding specific investigations, for example, arrest records for research documenting specific crimes following a disaster, or health care records documenting a specific disease outbreak. These data, however, are often incongruent, not comprehensive, and not global in scope. When combined they can often lead to baseline models that are inadequate. Legal barriers and unwillingness of agencies or jurisdictions to share data factor into the unavailability of data. Privacy and security issues are a primary reason for this unwillingness to share. A workshop participant indicated that this issue was recognized and discussed in a recent NRC study Successful Response Starts with a Map (2007). The only standards available with which to validate complex social system models are engineering standards. These are not adequate for the task. New technologies for social model validation could result in reduced error and better models. Behavioral Factors Not enough is currently understood about how trust in and reliance on information sources change as a result of stress. A better understanding of the nature of these changes in a technology environment could allow these concepts to be usefully incorporated into network models and decision making. Additionally, researchers may understand how data are collected off the Web and how individuals use their networks, but it might not be understood how the flow of information changes if the status of individuals’ connectivity to the Web changes. Researchers do not know, for example, how Internet penetration in a network changes who the opinion leaders of a network are. This gap is acknowledged and being addressed by the Office of the Secretary of Defense. STUDYING, ASSESSING, AND CREATING RESILIENT COMMUNITIES Fran H. Norris of the National Center for Disaster Mental Health Research of the Dartmouth Medical Center was invited to define community resiliency in a presentation to workshop participants. This section summarizes her presentation in combination with
OCR for page 23
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary the discussion it generated. Unless otherwise noted, conclusions may be attributed to Dr. Norris. What is Resilience? Resilience can be understood as a response to stress and can be considered as (1) a theory that guides the understanding of stress response dynamics; (2) a set of adaptive capacities that call attention to the resources that promote successful adaptation in the face of adversity; and (3) a strategy for disaster readiness against unpredictable and difficult to prepare for dangers. Response to stress can occur at the individual, institutional, or societal levels. The overarching concept is one that can guide research, policy, and the design and sequencing of interventions. Building community disaster resilience is more than a focus on disaster preparedness. It represents a paradigm shift that relies on building economically strong communities whose members can work together and use information to make and act on decisions. The concept of resilience is familiar in many disciplines. In physics and mathematics, resilience refers to the speed with which a material or system returns to equilibrium after displacement. In ecology, resilience refers to the persistence of relationships within a system and the ability to absorb change. In psychological terms, resilience refers to the process of successful adaptation despite challenging or threatening circumstances. In sociology, resilience is the ability of social units, such as communities or cities, to withstand external shocks to their infrastructure. In all these definitions, resilience involves a process and reflects adaptability rather than stability. Resilience is not an immutable condition but rather a set of adaptive capacities to be continuously attended to and modified to prevent their loss. A community that functions well, consists of members that are behaviorally and mentally healthy, and offers a high quality of life is more likely to be able to adapt after a disturbance or adversity. Resilience as a Trajectory of Adaptation Resilience is one of multiple possible stress responses to a disaster that also include resistance, recovery, and chronic dysfunction. Communities resistant to a specific event are barely affected by it. A resilient community may display transient dysfunction that is quickly resolved following an event. For some communities, dysfunction is more slowly resolved, but the community ultimately recovers. Chronic dysfunction represents the failure to adapt to the new circumstances. Factors controlling response trajectories are severity of exposure, the existence (or the perception of the existence) of social supports, and social class. Different levels of distress can be observed within a single community following a disaster, as demonstrated in a case study of mudslide victims in Mexico in 1999. A high degree of property damage, bereavement, and complete displacement was observed following mudslides that destroyed a large portion of a community. Levels of distress, such as symptoms of post-traumatic stress disorder, were monitored among the population over time. Approximately one-third of the community was resistant to stress and was observed to have few symptoms. Another third of the community showed very
OCR for page 24
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary high initial distress that improved at different rates. This group displayed different levels of resiliency but eventually returned to normal function. The remainder of the population showed moderate and high levels of distress that did not improve over a 24-month period. This group was considered chronically dysfunctional. Building Adaptive Capacities to Increase Community Resilience Community resilience emerges from the ability to withstand stress without degradation. It is largely dependent on access to vital community resources. The rapidity with which resources can be accessed and used during and following disruption contributes to resiliency, as does redundancy of vital community elements in the case of failure of an individual element. Community resilience is more likely to result in the building and balancing of different qualities related to: The social capital available to the community through its networks, including organizational linkages within a community, the amount of social embeddedness, the attachment to place, the sense of community, citizen participation, and the real and perceived support in the face of adversity; Community competence—a measure of how ordinary people make decisions. Community competence is dependent on actions of the community, community problem-solving skills, flexibility, creativity, mutual trust in the effectiveness of people working together, and the belief that the community is empowered to control resources; and Information and communication, including competent communication skills and infrastructure, and trusted, responsible, and relevant sources of information. Balanced levels of social capital, community competence, and ability to access and communicate information build resilience by enabling those qualities that help a community function as a community in the face of adversity. Economic development—including the level and diversity of economic resources, the equity of resource distribution, and the fairness of distribution of risk and vulnerability to hazards—is also important in building resilience. Significant improvements in community resilience could result using the public health approach of encouraging small shifts in population response to disaster events. Figure 2-2a shows the distribution of the common response trajectories of victims of the Mexican landslide case study described earlier. Figure 2-2b shows what the responses in the same population could be if only 5 percent of the total population could be made more resistant, 5 percent could be made more resilient, and 5 percent of those who would not have recovered could be assisted into recovery. Such improvements can be achieved by intervening at multiple points, and addressing multiple adaptive capacities, before and during a disaster, and again later in time. Such improvements may include Developing economic resources, reducing of resource inequities, and giving attention to the areas of greatest social vulnerability;
OCR for page 25
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary Meaningfully engaging local people in all steps of mitigation processes; Fostering organizational relationships to rapidly mobilize needed services; Boosting and protecting naturally occurring social supports; Planning for the unexpected by exercising flexibility; and Building trusted information sources that can continue to function in the face of unknowns. The means to measure adaptive capacities, especially in the area of communication, have not been developed. SNA may provide a reasonable means of measuring the adaptive capacities associated with community resilience and in determining how best to intervene to achieve the desired improvements. FIGURE 2-2 Distribution of response over a 24-month period by a population impacted by Mexican mudslides in 1999. 2-2a represents observed responses. 2-2b represents the potential response given a positive shift of 5 percent in both the resistant and resilient groups, and in those able to recover in the longer term. Note the percentage of the population that remained chronically dysfunctional would be dramatically reduced with the application of this public health approach. SOURCE: Fran Norris, Dartmouth, workshop PowerPoint presentation. REACHING VULNERABLE POPULATIONS THROUGH SOCIAL NETWORKS: CASE STUDIES OF EFFORTS TO PREVENT THE SPREAD OF HIV The workshop planning committee sought to explore how social networks and SNA could be used to reach vulnerable populations that may become disenfranchised from the larger community following a disaster. Because of the limited number of case studies on this topic, the committee looked to a case study from the public health community. Carl Latkin of the Johns Hopkins Bloomberg School of Public Health was asked to give a presentation on his efforts to recruit inner-city residents at high risk for HIV infection and transmission to promote positive behaviors within their communities. His presentation was related to how social and personal network approaches were used to both create
OCR for page 26
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary networks among vulnerable populations and potentially influence behaviors within the networks. In the social network of impoverished inner-city populations, a high degree of network linkages exist among impoverished individuals and individuals with chronic physical illnesses, mental illnesses, and drug addiction. Consequently, social network analyses afford a viable approach to reaching these vulnerable populations. A summary of Dr. Latkin’s presentation and ensuing discussion are provided here. Conclusions presented are by Dr. Latkin unless otherwise noted. The Urban Social Service Network Setting Insufficient service capacity exists to deal with the demands of daily medical emergencies in many major northeast urban cities. Emergency Medical Services (EMS) may be overwhelmed by nonemergency uses by substance abusers, the mentally ill, those in nursing homes, and the homeless. The design of social service networks discourages use of public resources by forcing people to wait in long lines, by treating them poorly, and encouraging distrust between service providers and impoverished populations. Building successful networks and services is dependent on building trust, an important adaptive capacity. Interacting with EMS or the fire department is a good way to see the conditions, plights, and social isolation before developing interventions. Understanding the Network The Johns Hopkins Bloomberg School of Public Health used network approaches to reach highly impoverished individuals and change unhealthy behaviors. Inner-city residents at high risk for HIV infection and transmission were recruited in Philadelphia, Pennsylvania, and in Thailand to promote risk reducing behaviors in their communities (Latkin et al., 2009). Systematic study of and establishment of rapport with the communities were necessary to gain the trust vital to the success of the programs. The nature and stability of needle-sharing networks over time were studied. Networks consisted of up to 10 individuals who shared needles and sex. Some network stability existed, but a fair amount of fluidity and turnover of individuals were observed. Some linkages of these smaller individual networks to a larger social network were noted, but not all members of the community were linked. Introducing Interventions Individuals with a desirable combination of skills and natural leadership abilities were enlisted from over 400 of the needle-sharing networks to educate their peers. They were 18 years or older, had weekly contact with active drug users, and were willing to be trained to conduct outreach education and bring network members into the program. The identified leaders were paid to receive training, but received only symbolic rewards for talking to their friends, modeling risk reduction behaviors, and socially influencing the critical behaviors that prevent HIV infection. Given the high level of stressors and traumas within these populations, redundancies were incorporated into the networks
OCR for page 27
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary because of the likelihood a leader would become unavailable. Control groups were established to measure the effectiveness of information diffusion and the potential of behavioral changes. Over a thousand community members participated in the intervention that consisted of six small-group peer-educator training sessions. After 24 months, participating network members were up to twice as likely to have reported not engaging in high-risk injection behaviors. These individuals were also more likely to have engaged in conversations regarding HIV risk behaviors following the training. The efficacy in reducing risk was not established in the studies. Issues Related to Designing Interventions Recruitment methods to identify program participants may substantially influence interpretation of network structures and functions. People identified as central to networks may not be interested in participating in the interventions or may not be the best opinion leaders. Negative reactions may result if a message is received from a source lacking credibility. There may be role conflict for a member of the community representing the program. Individuals need to be trained in how to maintain credibility. Lack of resources within networks is a barrier to effectiveness, but changes in behavior can be associated with a small amount of resources. Workshop participants discussed how interventions have to be carefully planned and translation efforts targeted to be robust and sustainable, especially given the lack of control over the message once it moves into the network. Information has to be disseminated in an appealing and memorable way to withstand mutations as it spreads.2 Using rumors (described as the “grapevine” by workshop participants) to translate information can be an effective way to make messages more durable. Narratives that give experiences shared meaning and purpose are important. Inherent in attempting to change behaviors of individuals in a network is the possibility of changing the network itself. In the case studies presented, individuals whose risk behaviors changed the most were also more likely to drop their ties with the network. This is something to consider when designing intervention networks. How the role of the trained leader is maintained or transitioned at the end of the intervention program so that the message is sustainable is another consideration. Issues Related to Disaster Preparedness Concentrated in many impoverished inner-city neighborhoods is a phenomenally high level of drug and alcohol addiction, chronic disease requiring medication, and mental illness. Individuals afflicted by these ailments often have only the care of others similarly afflicted on which to depend. Community resources are unavailable or not trusted by the population during times of normal operation, but are even scarcer when disaster strikes. 2 For example a more durable approach to informing drug users of the effectiveness of cold water on needles may be to compare the use of cold water on needles with using cold water on clothing to wash out blood.
OCR for page 28
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary The case studies presented here describe intervention efforts targeting small networks whose members may not be attached to larger community networks. Workshop participants discussed that the stability of these networks is fragile at best, and disintegration of the networks is likely following a disaster. Individuals within these communities are at severe risk of being further disenfranchised from the larger community, and may lack any knowledge of, access to, or trust in aid offered in response to a disaster. Workshop participants expressed a need to consider how communication with all members of a community, including individuals within disenfranchised or potentially disenfranchised populations, should occur. The means to communicate and provide services to all community members during each phase of the disaster cycle is essential. Special study of the means to build disaster resilience among fragile communities, such as those described by Dr. Latkin, is desirable. It is important to go into communities to understand the access, feasibility, and reliability of resources, and to understand how many people are reliant on the same resources. This is especially true in disaster management settings when many people may depend on the same resources, or the availability of the resources may change. Social network analyses afford a viable approach to reaching these vulnerable populations. USING SOCIAL NETWORKING TOOLS TO ENHANCE COMMUNICATION In the last decade, numerous digital networking tools have been developed that are changing the way many in American culture communicate. The workshop planning committee invited Michael Byrne of ICF International Inc. to provide background on these tools and relate how emergency managers could use them during all phases of a disaster. This section summarizes his presentation, entitled “Impact of Technology on Collaborative Homeland Security: Web 2.0, 3.0, 4.0 and Beyond,” and the discussion that followed. Many of the experiences relayed were anecdotal but are descriptive of the issues and solutions at hand. A new vision for the Internet began to take shape during the Web 2.0 Conference in 2004.3 Web 2.0 represents a culture shift, with the Internet being controlled by users from the bottom and providing an interactive environment that fosters innovation. Users become active participants rather than observers. The Internet now offers rich user experiences including Web video, interactive maps, timely content, and virtual worlds4 which can be used not only for online entertainment but also for practical purposes such as conferencing and training. The evolution of the Internet will continue beyond Web 2.0 with the development of tools such as autonomous intelligent agents that are programmed to recognize user interests and filter and manipulate information the user sees. This is already being applied to some extent to target advertisements to Internet users. As defined in Box 2-1, social networking tools enable individuals and groups to engage in social networking by monitoring and interacting within their networks. Text 3 A conference held by O’Reilly Media Inc. and MediaLive International (www.web2con.com/web2con [accessed April 5, 2009]). 4 Virtual worlds are computer-based simulated environments in which single or multiple users can communicate and manipulate events within the environment.
OCR for page 29
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary and multimedia information can be easily shared using relatively inexpensive and accessible technologies and distribution networks available for free on the Internet. According to Mr. Byrne, 3.75 billion people in the world have mobile communication devices. Sixty-two percent of all Americans have experience accessing wireless digital data and tools (Horrigan, 2008). Emergency managers who do not use these tools to reach their communities, build networks, and improve communication risk the possible detriment of their communities. Twitter allows members to distribute text messages, called “tweets,” of up to 140-character with their cell phones to geographically-, group-, or friend-based networks. According to the website TechCrunch,5 information regarding the 2008 terrorist attacks in Mumbai was shared worldwide in quasi-real-time using Twitter—faster than news agencies such as CNN reported the events. In London, 62,000 cameras record much that occurs in public spaces. Communities in England are networked, and video information of interest can be shared in real-time. Individuals anywhere in the world can share visual information over networks using tools such as Flickr6 and YouTube.7 The emergency management community has largely missed the networking revolution. Emergency management practitioners would benefit from a new communication paradigm and from studying how others are using social networking tools. For example, the Department of Health and Human Services has used the virtual world Second Life to run training drills;8 law enforcement agencies have used gaming technologies for “shoot, don’t shoot” drills; and America’s Army9 uses gaming technologies to teach basic first-aid skills. This type of training could prove to be an inexpensive alternative to traditional training approaches. Workshop participants heard anecdotal evidence that evacuated New Orleans city staff used social networking tools to organize and get things done following Hurricane Katrina. This was possible in spite of individuals being spread geographically, and far from New Orleans. The electronic convergence of people into widespread information networks can enhance the concept of resiliency in a global sense. Digital Divides Multiple digital divides exists among users of social networking technologies. Currently, social network tools are used most actively by people younger than 25 and older than 50 years (Li and Bernoff, 2008). Fifty-five percent of online teens have created profiles on social networking websites but far fewer online adults have done so (Madden et al., 2007). Other divides exists because of lack of financial or technical resources. Whatever the reason for the divides, penetration of networking technologies in society is far from complete and the incompleteness can be socially stratifying. The most vulnerable populations during and following Hurricane Katrina were those least likely to 5 See www.techcrunch.com/2008/11/26/first-hand-accounts-of-terrorist-attacks-in-india-on-twitter (accessed April 3, 2009). 6 Flickr is a Web-based photo management and sharing tool (see www.flickr.com [accessed April 3, 2009]). 7 YouTube is a video sharing website where users can upload, share, and view video clips (see www.youtube.com [accessed April 4, 2009]). 8 See secondlife.com (accessed April 5, 2009). 9 See www.americasarmy.com (accessed April 5, 2009).
OCR for page 30
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary use or have access to social networking tools. It is still essential to conduct door-to-door searches following an emergency in spite of great advances in networking technologies. A digital divide also exists between many government entities and the public. Many organizations prohibit access to Internet sites such as Facebook and MySpace.10 The organizations have not kept up with networking technologies or have not used them advantageously to communicate with constituents. Government agencies will be unable to ignore networking tools because of the growing reliance on networking (versus traditional communication approaches) by the public. Many municipalities resist the adoption of networking tools because of valid or perceived security concerns. Agencies at all levels often opt to take control of security issues by creating their own networking tools—an expensive approach that could result in quickly outdated technologies. However, some government agencies are now exploring and even embracing the active use of social networking tools. Double-Edged Sword Even though networking tools can be successfully used to spread information, Mr. Byrne also sees them as a double-edged sword. Information may not be well managed, systems may be overwhelmed by a large number of messages, it is difficult to prevent the spread of misinformation, and infrastructure can fail during catastrophic events. High-tech solutions need to be balanced with lower-tech solutions to ensure that redundancies and backups exist. Network tools are easily used by those with both honest and dishonest intent. Workshop participants learned how some terrorist organizations are known to use tools such as Twitter and Google Earth11 to update their networks. The challenge for the emergency management practitioner is to synthesize and analyze the large volume of information available and determine whether the information is correct, actionable, or requires response. Some believe that a large group of people sharing information can arrive at more accurate conclusions than a small group of experts discussing a given topic (Surowiecki, 2004). Many believe this is the strength of the Internet. Interactive connectivity implies constant feedback that makes a system self-correcting. As of 2007, there were over 60 million blogs12 on the Internet (Wyld, 2007). Sites such as YouTube and Facebook create a value beyond what a top-down control model provides. However, bottom-up organizations largely shaped by its members can be disorganized, loosely controlled, unmethodical, and sometimes inconsistent, in part because they are in constant states of flux. Conflicting information may make it difficult to determine which information sources are accurate, and bad information may promote unwanted behaviors. Charles Mackay wrote a book in 1841 entitled Extraordinary Popular Delusions and the Madness of Crowds in which he discusses the perils of the spread of misinformation. The high 10 See www.myspace.com (accessed July 16, 2009). 11 Google Earth is a virtual on-line globe with free downloadable drivers that displays satellite images of the earth’s surface at different resolutions. Users can add their own data and overlay their own images (see earth.google.com/index.html [accessed April 3, 2009]). 12 A blog is a user-generated and regularly updated online journal. An example blog mentioned by Michael Byrne is Disaster Zone: EmergencyManagement in the Blogosphere, maintained by Eric Holdeman (see www.disaster-zone.com [accessed April 3, 2009]).
OCR for page 31
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary level of connectivity in society today creates the potential for major disasters or magnification of disasters through the unintentional or intentional misuse of networking tools. Bad information can be long-lived on the Internet and can persist even on successful social institutions such as Wikipedia,13 where entries are subject to constant review by members. Even considering content error, these institutions remain successful because within their bottom-up organizational structures, ways are available to manage data, look for and fix problems, and recognize and resolve attacks on the system. For application in the disaster management community, it is essential that systems and networks are functioning before a disaster in order for them to be effective during and following a disaster. The physical infrastructure required to operate the networks may be resilient, but infrastructure failure is possible during a catastrophic event. In this situation it is essential to assess the needs of the impacted community and to communicate to people outside the affected area about the contributions they can make. It is also essential that the physical infrastructure be restored. Workshop participants pointed out that emergency response plans generally call only for restoration of communication infrastructure among response agencies, and no authority exists among emergency managers to restore communication and networks used by the public. However, current networking technologies allow the quick localized reemergence of networks wherever a transmitting device is functioning. Some management of network re-emergence would benefit emergency managers. Partnerships among public and private entities could provide solutions. Shifting Paradigms Workshop participants discussed how emergency management organizations within the United States do not employ a command and control structure for disaster management. Instead, a kind of diplomatic structure is in place. This structure can be enhanced and made more effective using networking tools to build community networks and improve disaster resilience. Technology now allows emergency managers to shift from a focus on what needs to be done for the community, to what can be done with the community. There are an increasing number of examples of how emergency managers are using networking tools to change the mindset of community members—changing them from victims to survivors and eventually to community resources. Inexpensive networking tools and devices used by a large proportion of the population. Emergency managers can engage an existing audience, and there is a great potential for positive outcomes and for lives saved. Positive outcomes will only be achieved, however, with an understanding of how a population will behave in response to new information, how a message may change once it is broadcast, and how durable messages are developed to have the desired effect. Many workshop participants expressed the importance of understanding the effective use of social networks and social networking tools during all phases of the disaster cycle while remaining cognizant that the infrastructure for these tools may fail. Understanding how to apply networking 13 Wikipedia is a free, Web-based encyclopedia with entries created, collaborated on, and maintained by volunteer users (see en.wikipedia.org/wiki/Wikipedia:About [accessed April 3, 2009]).
OCR for page 32
Applications of Social Network Analysis for Building Community Disaster Resilience: Workshop Summary innovations at different social and management levels to identify and close disconnects between those that need resources and those that have access to them would be beneficial. Networking tools can be intelligently used to engage meaningfully with the community to positively influence behaviors, influence mitigation and preparedness strategies, build more resilient systems, and improve response and interventions.