By REID A. WODICKA AND S. LORÉN TRULL
When we speak of today’s social networks, most people think of online tools like Facebook or Twitter, which allow us to keep up with friends and coworkers while ensuring that we are constantly informed about the trials and tribulations of our favorite celebrities. Although these are important pursuits, tools like Facebook and Twitter are really a by-product of the development of an analytical method called social network analysis, which was formed in sociology research and measures interactions among a network of actors. It may not be as exciting as keeping track of the musings of Kim Kardashian, but we can use social network analysis to discover the interrelationships among actors within an emergency response network. This article explains social network analysis and how it can be used to analyze the structure of firefighting and emergency medical services (EMS) networks while providing an example of the empirical information that can be gleaned from the use of this method. Also, we will discuss the policy implications of the information that we seek from social network analysis.
A traditional “standing army” station-based firefighting and EMS response system can be theoretically viewed as a network system that exists to provide services beyond the first-due response area of each neighborhood fire station. The advent of mutual- and automatic-aid agreements has created a circumstance in which no one station is ever operating alone, particularly on large incidents. Although in some places fire companies still operate independently of one another, we argue that throughout most of the country, most firefighting and EMS organizations have transitioned to a more cooperative model of service delivery. This demands that those responsible for overseeing the larger response network understand how stations interact with one another without focusing the analysis on one or two individual entities within the network.
A social network is defined as the aggregation of the mutual relationships between and among actors and institutions functioning within a given society or, in this case, the response network. Within the context of a firefighting and EMS response network, when a structure fire occurs in a fire company’s fire-due area, others are automatically assigned to the incident. The responsibility to provide firefighting services now rests not only on the first-due agency but also on the other responding stations. Aid provided by one agency to another on a box assignment can, therefore, be viewed as a network interaction between the two agencies. When we consider all of the interactions within the response network over a given period of time, we can then begin to understand how each station interacts with the rest of the network and where important interactions occur.
To demonstrate how social network analytics can be used to characterize the nature of the relationships within a firefighting and EMS response network, we used data from a rural county on the East Coast that operates as a combination volunteer and career department. Some stations are staffed by career personnel who supplement volunteers; some are staffed 12 hours, others 24 hours, and one is staffed for 10 hours. Other stations in the county have no paid personnel and rely solely on volunteers to provide firefighting services. Twenty-five stations provide services to the county, although eight of them are not actually geographically located within the county. The data on interactions among stations was taken from 18 months of multiple-station responses, using the National Fire Incident Reporting System data provided directly from the county.
In our analysis, we were interested in learning about agencies that are important to the maintenance of the structure of the response network—that is, we looked for fire stations that are so important to the network structure that if the organization failed, the response structure would fall apart. The policy implications for this research suggest that those entities deemed so important that their absence would destroy a network should be monitored closely and supported sufficiently to ensure that they do not fail on any given incident. The way to support the station is a matter of management strategy (adding career personnel or providing financial resources for volunteer recruitment and retention). Allowing an agency to fail endangers its ability to provide the service and those attempting to provide it.
Using a free social network analysis software (there are several on the Internet), we calculated a metric called the “betweeness centrality” for each fire station operating within the county. The betweeness centrality measures the importance of an individual entity that operates within a network. If that entity (called a “node” in this analysis) is eliminated from the network and the county wants to maintain the same level of service, that node will need to be replaced, or new interactions will have to be secured (perhaps changes in the box assignments). This obviously becomes increasingly more problematic as the betweeness centrality of an agency increases because the demand on other agencies will increase dramatically or the replacement will be expensive. Because betweeness centrality mathematically measures a proportion, the scores we calculate for each node range from zero to one. Figure 1 shows the results of our analysis, in which we have randomly assigned a number to each fire station and reported the betweeness centrality.
Figure 1. Betweeness Centrality of Fire Stations in the County |
As you can see from the chart, this is a directional network, which means that we account for the direction of aid provided. For instance, Fire Station 3 gives and receives assistance from Fire Station 16, but Fire Station 13 provides assistance to Fire Station 2—not the other way around. The one-way relationship in this and many other cases is based on the locations of those fire stations. Fire Stations 5, 6, 12, 13, 14, 15, 17, and 20 are outside of the county. Responses of their first-dues are of no concern for fire system leaders in this county. From the data provided in this network analysis, we can easily identify the most important fire stations in the network: Fire Stations 3 (with a score of 0.13) and 10 (0.10) are highly important to the security of the network. Other stations are also important, but less so, such as Fire Stations 2, 11, and 16.
The policy implications from this study relate to its ability to provide a basis for allocating resources. Given the constant scarcity of resources, county officials have to make decisions related to the allocation of financial and personnel resources. If the decision relates to increasing funding, this analysis can determine that directing those resources to Fire Stations 3 and 10 will provide the most benefit to the greatest number of citizens in the county. In addition, if Fire Stations 3 and 10 are independent volunteer organizations that are not necessarily directly accountable to county officials, those organizations should be monitored closely to ensure they don’t demonstrate any signs of future failure because they are so highly centralized in the response network.
Proactive recognition and response to problems help to maintain the network structure for future incidents. Although many cities and counties rely on descriptive statistics like the total number of responses an agency makes, social network analysis can provide a relatively simple tool for fire department administrators to analyze the structure of their response network.
We have provided one example of how social network analysis can be applied to the firefighting and EMS response network structures in existence in just about every jurisdiction in the United States. There are, however, many other metrics that can be used to explore the structure of the response network. Many free online resources can provide a basic understanding of networks, which can be found with a simple Internet search. There are several free downloadable programs on the Internet that perform analytics in a user-friendly manner.
Although it may be more exciting to analyze the social interactions of Kim Kardashian, Justin Bieber, or some other really “important” celebrity, tools such as Facebook and Twitter really only scratch the surface of the potential uses for social networks. By altering the way we think about the interactions that occur within our fire and EMS response systems, we can change the way in which we gather and interpret information. As we have demonstrated, this can alter policy decisions and the way we think about resource allocations.
In an era in which all budget line items are under attack because of declining revenues or changing political priorities, fire service leaders must be prepared to react to funding and functional changes with actions that are based on well-reasoned and methodologically valid research. Social network analysis may be a tool leaders can use when considering the future of the fire service in their jurisdictions.
REID A. WODICKA has a BS degree in public policy and administration and a master’s of public administration degree from James Madison University. He is completing his doctoral dissertation in public policy program at the University of North Carolina at Charlotte. Wodicka has more than 10 years of fire service experience in volunteer and career departments. He serves as town manager and is an active firefighter with Woodstock (VA) Fire Department.
S. LORÉN TRULL has a BA degree in Hispanic studies and a BS degree in sociology from East Carolina University and a law degree from the University of North Carolina—School of Law. She is a PhD student in public policy at the University of North Carolina at Charlotte. Trull has been admitted to the State Bar of New Jersey.
Fire Engineering Archives