Towards Coordination Preparedness of Soft-Target Organisation

Towards Coordination Preparedness of Soft-Target Organisation Mohammed Shahadat Uddin and Liaquat Hossain Project Management Graduate Programme, University of Sydney, Australia Abstract. In this paper, we introduce a network enabled coordination model to examine the coordination preparedness of soft-target organisations such as common public access areas including transit hubs, schools, parks, and sports areas. It is apparent that little attention is given in recent research focusing on the use of network analysis as a way to explore coordination preparedness for this type of organisation. In this study, we emphasise this type of soft-target organisation and propose a model to examine the coordination preparedness to any disasters by testing hypothesis related to network relationship and coordination preparedness. We analyse the dataset entitled Preparedness of Large Retail Malls to Prevent and Respond to Terrorist Attack, 2004, which contains a total of 120 completed surveys of security directors of retail malls. The following questions form the basis of this study: What do soft-target organisations need to be better prepared to respond to disaster? How does network relationship between soft-target organisation and emergency agencies affect the coordination preparedness of soft-target organisation for disaster recovery? Which degree of centrality measure needs to be followed to measure network variables in order to analyse the coordination preparedness? Result shows that soft-target organisation with high level of network relationship with other emergency agencies are better prepared to disaster response. Using this result, the preparedness of a soft-target organisation might be judged for successfully participation in an actual emergency. Keywords: Soft-Target Organisation, coordination preparedness, network relation. 1 Introduction We define soft-target organisation (STO) as an organisation where a large number of people gather regularly, which has multiple entrances and exits making them vulnerable to terrorist attack, and where there is more possibility of higher number of casualties in case of any terrorist attack. This type of organisation does not participate in any emergency response; instead they have to prepare themselves to respond and recover from any kind of disaster. Therefore, there is a need to maintain strong relationship with public emergency agencies, which puts them in a star or wheel (Figure 1) network communication structure with emergency agencies where they are in the center of the network. To measure the network position of this type of actor, degree centrality is found useful, which can be defined as the number of ties incident upon a node [7]. M.A. Wimmer et al. (Eds.): EGOV 2009, LNCS 5693, pp. 54–64, 2009. © Springer-Verlag Berlin Heidelberg 2009 Towards Coordination Preparedness of Soft-Target Organisation 55 Coordination is evident in a network when multiple actors pursue goals together. Coordination can be defined as the additional information processing performed when two or more connected actors pursue goals that a single actor pursuing the same goals would not perform [14]. More precisely, coordination can be referred to as managing dependencies between activities [15]. The following components of coordination form basis for our proposed model: (a) set of actors, public emergency agencies with which STO is connected and STO itself; (b) who perform tasks, all involved security staff in the network from emergency agencies and STO; and, (c) to achieve goals, better coordination preparedness [1] [13]. Studies on disaster coordination suggest that lack of coordination preparedness causes more loss in respect of casualties and finance in extreme events regardless of whether it is naturally occurring such as The New Orleans hurricane or man-made like the World Trade Centre (WTC) bombing [10]. Coordination preparedness encompassing an emergency plan, which describes required services by allowing all actors to provide support and coordination among actors, and instructional guidelines to follow in case of emergency, is considered to be the necessary precondition for handling an emergency effectively. In this paper, we analyse and hypothesise an actor’s network relationship and its perceived coordination preparedness. To do so, we use variables to represent both network relationship and coordination preparedness for an actor, and use dataset to test our proposed hypothesis. We argue that an actor who maintains a strong network relationship is able to show better coordination preparedness in a disaster response. EA2 EA1 EA3 EA5 a EA4 STO EA6 b Fig. 1. (a) a star (b) Similar network structure of STO with emergency agencies (EA) The rest of the paper is organized as follows: a background description of coordination is described in section 2. Then in section 3 proposed coordination preparedness model for STO is presented. We also develop a set of hypotheses for the study in this section. An overview of data collection as well as the methods used for data preparation and data analysis is described in section 4. This is followed by a discussion of the findings and conclusions. 56 M.S. Uddin and L. Hossain 2 Coordination Coordination facilitates social interactions where multiple actors work towards a common goal. Coordination is something that occurs within an organisation or discipline and therefore, it is internal to the organisation. Inter-organisational coordination, on the other hand, is defined as managing capability of inter-organisational collaborations and relationships, which can be in many forms including inter-organisational teams, partnerships, alliances, and networks [8]. It is very difficult to measure coordination due to difficulties to detect the type and nature of group interactions, which define its dependent variables. For example, in an emergency response network, coordination can be measured from how often all organisations in the network share information, and rehearse its emergency protocols. Therefore, we first need to find out key coordination processes, and then measure effectiveness of those processes in order to quantify coordination. In extreme events such as disaster, organisations frequently develop formal and informal network relationships in order to save lives and return a society to a state of “business as usual" [4], and hence, a proper coordination among those organisations make it easy to achieve these goals. Approaches to coordination can be diverse, which mainly depends on the nature of network under consideration. Traditionally, coordination has been justified as hierarchical control mechanism based on top-to-bottom approach. The idea is to maintain coordination by control from the top authority, a ‘coordination by command’ approach [6]. Selecting a suitable governing body make this traditional approach to coordination a contentious issue for a long time. Organisation approaches to coordination seek how to minimize cost by efficient use of available resources [9]. Depending on the way of functioning, coordination might be either centralized, or distributed, or a combination of both at organisational level. A central coordinator decomposes and distributes tasks among others in centralized coordination. To maintain proper synchronization among sub-tasks and a single point of failure are the disadvantages of this approach of coordination [2]. Tasks are often distributed among actors who can work independently and in parallel in a distributed coordination system. Existing studies of emergency response have given emphasis on coordination and management of emergency agencies in the events of crisis. Kruke and Olsen [11] identified coordination challenges during complex emergencies. Comfort and Kapucu [4] developed a guideline for inter-organisational coordination in extreme situations. Also, a study by Riley and Hoffman [18] implies the key properties of emergency coordination in terms of having an awareness of others within the network, defining the accessibility of information flow through the network, and the quality of information that a particular organisation is able to obtain. However, research on measuring coordination preparedness from the perspective of network structure of STO is lacking to date. Though these types of organisations do not participate in case of crisis of others, they need to maintain strong coordination preparedness as they are possible target for terrorist attack [5]. 3 Coordination Preparedness through Social Network (SN) SN theory plays an important role in identifying and quantifying informal network, which functions at level beyond the formal and traditional organisational structure of Towards Coordination Preparedness of Soft-Target Organisation 57 relationships. Prior research suggest that investigations of informal networks are very useful in identifying network properties such as which actor is the most influential, what kind of relationship exist among the actors [16]. Network centrality such as degree centrality of SN theory is very efficient in assessing network behavior, which can unfold existing informal network patterns that are not invented before [3]. The selected approach to studying a social network may be determined by the type of network under investigation and its associate level of data collection. Social network analysis (SNA) is the mapping and measuring of relationships between actors. It has been successfully applied to understand networks and their participants by evaluating locations of actors in the network. Measuring a network location of an actor is finding the centrality of that actor. There are three primary measures of network centrality: (a) degree centrality, (b) closeness, and (c) betweenness. The number of direct links connecting a node determines degree centrality of that node. It highlights the node with the most ties to other actors in the network; indicating having more direct contrast and adjacency than all other actors in the network [19]. Degree measurements mainly relevant in the studies of popularity and activity of actors as it primarily concerns with local point centrality. In the study of coordination, degree centrality is useful depending on the nature of network for measuring local authority, and might be used to compare network conditions and the state of coordination. As STO have a star type of network, degree centrality might be useful in making such a comparison. 3.1 Our Proposed Model The framework of our proposed model, as illustrated in the Figure 2, is intended for the assessment of coordination preparedness during a non-crisis period in order to optimize network performance. The model is constructed with a view to assess the current state of coordination preparedness as a product of attributes of network relation. There is a single moderating variable defined as training score, which is used to cluster our dataset, and also for completing a micro-level analysis. Soft-target Organisation Training Score Network Relation Coordination Preparedness Connectedness Frequency of Information Sharing Tie Strength Degree of Rehearsal Fig. 2. A model for assessing Coordination Preparedness 58 M.S. Uddin and L. Hossain We use connectedness and tie strength as independent variables in our model to measure network relation. Network connectedness or simply connectedness defines the number as well as nature of inter-organisational relationships such as mutual participation to achieve a common goal. It has been defined as interdependency among organisations and their network roles [17]. The second variable for network relationship is tie strength which is taken from the research of Kuti [12]. It defines the strength of network relationship as a source of coordinating activities such as security staff training with public emergencies agencies. In order to assess network relationship against coordination preparedness, we define two dependent variables which form the basis of our coordination preparedness measure. Frequency of information sharing depicts how often actor exchange key information regarding only security intelligence. It does not include normal behavioral organisational information sharing with other organisations. The variable degree of rehearsal refers to all neighbors in the network, an actor rehearse its emergency preparedness plan. As a STO is always in the centre of a star network, this variable depicts the number of emergency agencies that a STO rehearses its emergency preparedness plan with them. 3.2 Research Hypothesis Based on the review of literature and in alignment with our model, we propose the following hypothesis: H1: There is a significant relationship between network involvement and coordination preparedness of an actor, where higher level of network relation produces an increase in coordination preparedness. To assess this hypothesis, we present another four sub-hypotheses (sH) to evaluate and test our principle theory. They are: • • • • (sH1) Connectedness correlates to Frequency of Information Sharing (sH2) Connectedness correlates to Degree of Rehearsal (sH3) Tie Strength correlates to Frequency of Information Sharing (sH4) Tie Strength correlates to Degree of Rehearsal H2: The relation in the H1 is mediated by the quality of security staff training. This means quality of staff training can be used to predict the strength of relation between network relation and coordination preparedness. 4 Coordination Preparedness Dataset The dataset entitled Preparedness of Large Retail Malls to Prevent and Respond to Terrorist Attack, 2004 is located at the Inter-University Consortium for Political and Social Research (ICPSR) website. The study funded by the United States Department of Justice and the National Institute of Justice was developed with the original purpose to assess the level of security in large indoor shopping malls as well as the associated issues of training and legislation of private security forces (Davis et al., 2006). Research investigation was carried out by Robert C. Davis of Police Foundation, Christopher Ortiz of Vera Institute of Justice, Robert Rowe of American Society for Towards Coordination Preparedness of Soft-Target Organisation 59 Industrial Security, Joseph Broz of Midwest Institute of Research, George Rigakos of Carlton University, and Pam Collins of University of Eastern Kentucky. The research agenda was to address the degree to which malls have become better prepared to respond to terrorist attacks in the aftermath 9/11 events. The framework for the collection of data involved sending letters with surveys attached to 1372 security directors across the country in 2004. The researchers also collected survey data from State Homeland Security Advisors. Furthermore, for their research purpose the researchers visited 8 shopping malls in United States and 2 malls in Israel (Table 1). Table 1. Response rate of research sample No. Invited No. Participated Response Rate Shopping Mall Security Advisor 1372 120 8.75% State Homeland Security Advisor 50 33 66% Site Visit (USA) 8 8 100% Site Visit (Israel) 2 2 100% The sampling technique used to invite the participants was the size of enclosed retail malls across the country. The security directors of those retail malls which were at least 250,000 square feet in size were invited to participate in the survey. 4.1 Exploring the Data As noted on the ICPSR site, access to the research is provided by means of SPSS data files and supplementary machine-readable documents and data collection instruments. The first stage of collecting the data required a thorough exploration of the survey instrument to identify possible questions that provide relational data to assess the respondents’ social network, or questions relevant to an analysis of the current perceptions of their emergency preparedness abilities. As the proposed model is for soft-target organisation, we were looking for all degree-based relational data of the respondent in the survey instruments. In searching this degree-based data, three questions (1, 2, and 3 in Table 2) were found which provided information of the perceived interaction of the respondent with other agencies. The response range for question 1 is 0 to 3 organisations. For question 2 and 3, the response range is 0 to 2; where 0 indicates not at all involved and 2 indicate the strongest involvement. The scores of these three questions are combined to form the respondent’s degree of connectedness with other organisations. A further investment of the survey instrument found questions to measure the respondent degree of rehearsal of emergency preparedness protocols (question 4), tie strength (question 5), and frequency of information sharing (question 6) with other agencies. Question 4 has similar response pattern as like question 1. Question 5 had response range of 0 to 2; where 0 indicates less close relation since 9/11, 1 indicates unchanged relation since 9/11 and 2 indicates closer relation since 9/11. For the question 6, the response range is 0 to 2; where 0 means never share key information, 1 means rarely share key information, and 2 indicates regularly share key information. 60 M.S. Uddin and L. Hossain Table 2. Survey questions used to measure model variables Q. No. Degree-based Questions Have emergency response plans developed to coordinate and communicate activities of security staff with local law enforcement, fire, and medical first re ponders in case of attack? If yes, with which agencies? How involved has your state homeland security advisor been in planning, reviewing, or approving your security measures? How involved have local or state law enforcement agencies been in planning, reviewing, or approving your security measures? Have exercise been carried out to rehearse protocols with first responders? If yes, with which agencies? How is your working relationship with local law enforcement since 9/11? To what extent does local law enforcement informs you about key security intelligence? 1 2 3 4 5 6 Histogram for Connectedness Histogram for Tie Strength 30 Frequency Frequency 40 20 10 0 1.25 1.75 2.25 2.75 3.25 Conne ctedne ss 3.75 70 60 50 40 30 20 10 0 0 0.33 0.66 0.99 1.32 Ti e Strength Fig. 3. Histogram for Connectedness and Tie Strength An examination of the connectedness, and tie strength measures in Microsoft Excel reveals common distributions of both of them that follow a non-normal curve. Each graph consists of centralized score having a tapered skew to the left for connectedness, and a small tapered skew to the right for tie strength (Figure 3). These distributions are against a line indicating a non-normal and non-parametric statistical test is needed to be carried out in order to test their correlation with other variables of our model. The Spearman test is a non-parametric alternative to the Pearson test which investigates the relationships between two continuous scores. 4.2 Data Preparation A 2-phase method for data preparation and analysis is used to assess our emergency preparedness model against the survey data. Figure 4 depicts an overview of the software used and the purpose of each phase. The first phase included importing the Towards Coordination Preparedness of Soft-Target Organisation 61 data files into Microsoft Excel by placing the data into columns of Microsoft Excel representing questionnaire responses. Once the data is set up correctly, variables were cleaned and invalid responses such as refusals were removed in order to prevent inaccurate statistical testing. In the second and final phase, all the variables are placed into SPSS to perform some statistical analyses for hypothesis testing as defined in our proposed model. Data Preparation Methods Purpose of Software Microsoft Excel Phase 1 Cleaning raw data file, measure variables, Out-degree centrality. SPSS Phase 2 Perform statistical analysis Fig. 4. Overview of Software and phases of data preparation 5 Result and Discussion We design both macro-level and micro-level test to validate our proposed hypotheses. At macro-level, we considered 117 datasets; exclude 3 datasets with missing values. A Spearman test is carried out to determine if there is a relationship between the continuous independent network relationship variables of connectedness, and tie strength with the continuous dependent coordination preparedness variables of frequency of information sharing, and degree of rehearsal. The result (Table 3) of this test indicates that there is a positive correlation coefficient between each independent variable with all dependent variables where an increase in the measure of any independent variable produces an increase in both dependent variables. Table 3. Results of Spearman Correlation test (macro-level) 1 X Tie Strength X 1 Freq. of Info. Sharing X X Deg. of Rehearsal X X 0.267** 0.345** 1 X ** ** X 1 Connectedness Connectedness Tie Strength Freq. of Info. Sharing Degree of Rehearsal 0.308 0.275 Note: X denotes unnecessary or out of scope testing **. Correlation is significant at the 0.01 level (2-tailed). From the results of the Spearman test, we find that there is a positive correlation coefficient between connectedness and each of the two coordination preparedness variables, where increased connectedness correlated to: • (sH1) increased frequency of information sharing We find from testing the data that the number of maintained relationships displayed by an actor have an effect on its information sharing frequency with other actors in the network. 62 M.S. Uddin and L. Hossain • (sH2) increased degree of rehearsal Our testing of this sub-hypothesis also provides evidence that the number of well maintained relationships displayed by an actor is a contributing factor in determining with how many neighbors an actor rehearse its emergency plan. The relationship between the two dependent coordination preparedness variables against tie strength also produces a positive correlation where an increase in tie strength correlates to: • (sH3) increased frequency of information sharing The Spearman correlation also shows a positive result for this sub-hypothesis that by increasing the strength of relation an actor has over other actors in the network, the more frequently it can share information with others. • (sH4) increased degree of rehearsal Results indicate a positive correlation for tie strength that by increasing the strength of relationship, an actor can rehearse it coordination preparedness plans with more neighbor actors in the network. A subsequent micro-level test for correlation is carried out to analyze our second hypothesis and also to provide support evidence of the relationship as in the first hypothesis. Well-trained security staff is one of the most important requirements for the prevention and respond to any kind of crisis for STO. According to the survey on state homeland security advisors, most of them endorsed improved training for security staff and emergency responders as the most important measure STO could take in order to better prepare against terrorist [5]. We cluster our datasets in three groups on the basis of training score. The questions from survey instrument like “How many hours of training does new staff receive?”, “Do employee receive special training on preventing and responding to terrorism?” are taken to calculate training score. The range for training score is between 0 and 2.40 inclusive. Table 4 shows the training score range for each cluster. A very small difference (0.03) of average training scores between two consecutive clusters (0.54 and 0.51) implies a proper distribution of training score among all clusters. In our datasets, there are few actors having low training score; however, showing strong relation between network relationship and emergency preparedness in our proposed model. This is due to the lower number of full time security staff they have, and more dependency on contract or casual fully trained security staff. To avoid this, we consider 0.25 as our minimum training score for cluster 1. After eliminate all missing values we find 89 training scores out of 120. Table 4. Training score range for each cluster Cluster Number Training Score Range Avg. Training Score (base on all cluster members) Difference Cluster 1 0.25 - 0.75 0.58 Cluster 2 0.76 – 1.25 1.12 0.54 Cluster 3 1.26 – 2.40 1.63 0.51 Towards Coordination Preparedness of Soft-Target Organisation 63 The micro-level results also support our second hypothesis and findings in the macro level. In the micro-level results (Table 5) we find that with the increase in cluster number i.e. increase in training score, there is an increase in the correlation values between the all the combination of independent and dependent variables. For example, the correlation values between connectedness and frequency of information sharing are 0.221, 0.289, and 0.396 for Cluster 1, Cluster 2, and cluster 3 respectively, which provide strong support for our second hypothesis. We see that there a strong positive relation (correlation coefficient score 0.623) between tie strength and degree of rehearsal for cluster 3. This is due to the way we define training score and cluster. We consider ‘special training’ on emergency with other emergency agencies in measuring actor training score. Cluster 2 N = 31 Cluster 1 N = 46 Table 5. Results of Spearman Correlation test for all clusters (micro-level) Degree of Rehearsal Tie Strength Connectedness 1 X X X Tie Strength X 1 X X Freq. Info. Sharing 0.221** 0.302** 1 X Degree of Rehearsal 0.267** 0.208** X 1 Connectedness 1 X X X Tie Strength X 1 X X 0.289** 0.330** 1 X ** ** X 1 X X X Freq. Info. Sharing Degree of Rehearsal Cluster 3 N = 12 Freq. Info. Sharing Connectedness Connectedness Tie Strength Freq. Info. Sharing Degree of Rehearsal 0.314 1 0.342 X 1 X X 0.396** 0.623** 1 X ** ** X 1 0.349 0.354 Note: X denotes unnecessary or out of scope testing **. Correlation is significant at the 0.01 level (2-tailed). N indicates number of training score on that cluster 6 Conclusion In this paper, we measure coordination preparedness using degree centrality of SNA technique for the central actor (STO) of an emergency network, which maintain a star or wheel network structure. We present a model for coordination preparedness based on network relationship. Using this model, an organisation can be reviewed in order to find its current state of coordination readiness and therefore, be judged for its potential to response to an actual emergency. Our research suggests that there is indeed a positive correlation between network relation and coordination preparedness such that 64 M.S. Uddin and L. Hossain by increasing an actor’s involvement within the network, it is likely the ability of that actor to coordination in emergency will also increase. The results from both at microlevel and macro-level seem to complement each other and provide a stronger support for our proposed hypothesis. References 1. Baligh, H.H.: Decision rules and transactions, organisations, and markets. Management Science 32, 1480–1491 (1986) 2. Beaumont, P., Chaib-draa, B.: Multi-Platform Coordination in Command and Control. National Science and Engineering Research Council of Canada (2005) 3. Brandes, U., Fleischer, D.: Centrality Measures Based on Current Flow. In: Diekert, V., Durand, B. (eds.) STACS 2005. LNCS, vol. 3404, pp. 533–544. Springer, Heidelberg (2005) 4. Comfort, L., Kapucu, N.: Inter-organisational coordination in extreme events: The World Trade Center attacks, September 11, 2001. Natural Hazards, 309–327 (2003) 5. Davis, R.C., Ortiz, C., Rowe, R., Broz, J., Rigakos, G., Collins, P.: An Assessment of the Preparedness of Large Retail Malls to Prevent and Respond to Terrorist Attack, A report submitted to U.S. Department of Justice (2006) 6. Donini, A., Niland, N.R.: Lessons Learned, A Report on the Coordination of Humanitarian Activities. United Nations Department of Humanitarian Affairs, New York (1994) 7. Freeman, L.C.: Centrality in Social Networks: Conceptual Classification. Social Networks, 215–239 (1978) 8. Kapucu, N.: Inter-organisational Coordination in Dynamic Context: Network in Emergency Response Managements. Connections, 33–48 (2005) 9. Kirn, S., Gasser, L.: Organisational Approaches to Coordination in Multi-Agent Systems. National Science Foundation, Arlington (1998) 10. Krugman, P.: A Can’t-Do Government, The New York Times, published on September 2 (2005) 11. Kruke, B.I., Olsen, O.E.: Reliability-seeking network in complex emergencies. Int. J. Emergency Management 2(4), 275–291 (2005) 12. Kuti, M.: Cordnet – Towards a distributed behavioral model for emergency response coordination, A thesis submitted to the Faculty of the University of Sydney (2007) 13. Malone, T.W.: Modeling coordination in organizations and markets. Management Science 33, 1317–1332 (1987) 14. Malone, T.W.: What is Coordination Theory? National Science Foundation Coordination Theory Workshop, Massachusetts Institute of Technology, Cambridge, USA (1988) 15. Malone, T.W., Crowston, K.: The Interdisciplinary Study of Coordination. ACM Computing Surveys 26(1) (1994) 16. Mullen, B., Johnson, C., Salas, E.: Effects of communication network structures: Components of positional centrality. Social Networks 13, 169–186 (1991) 17. Rathnam, S., Mahajan, V., Whinston, A.B.: Facilitating Coordination in Customer Support Teams: A Framework and Its Implications for the Design of Information Technology. Management Science 41(12), 1900–1921 (1995) 18. Riley, K.J., Hoffman, B.: Domestic Terrorism, A National Assessment of State and Local Preparedness. RAND (1995) 19. Wasserman, S., Faust, K.: Social Network Analysis: Methods and Applications. Cambridge University Press, New York (1994); Appendix: Springer-Author Discount