Battling uncertainty through designed practices, a comparison of routine and complex operations in Police Dispatch

(1)

Master Thesis Public Administration

Track: Crisis and Security Management

Battling uncertainty through designed practices, a comparison of routine and complex operations in Police Dispatch

An explorative study on coordination practices in a fast-response organization

Author: Joel Vink, s1286226

Supervisor: Dr. J.J. Wolbers

Second reader: Dr. Sanneke Kuipers

(2)

(3)

1. Introduction

1.1 Problem Outline

Crisis management is traditionally concerned with an organization managing a crisis that challenges the routines, structures and survival of that organization (Faulkner, 2001). What might be considered a crisis to some is a mere regularity to others. Some organizations must deal with crises on a continuous basis. Such organizations are referred to in the literature as fast-response organizations. A fast-response organization is an ad-hoc and temporal formation of actors who are capable of reacting rapidly to sudden onset events, in which decisions must be made rapidly and where errors can potentially be fatal (Faraj & Xiao, 2006).

Crisis managing then refers to the actions employed by an organization to return to a state of normality and is often concerned with four separate facets, the so-called four C’s: communication, coordination, cognition and control (Comfort, 2007). To adequately control and coordinate any crisis one needs knowledge on the situation at hand and communicate that knowledge. Without such knowledge, or cognition, the task becomes increasingly difficult (Comfort, 2007). This lack of complete knowledge is particular troublesome in crises, as the information is often incomplete, inaccurate or abundant (Comfort, 2007). The environment of fast-response organizations is especially challenging because they operate in environments in which errors are non-acceptable. These environments are also unique in their complexity and time-constrained. Fast-response organizations are thus forced to rely on unique and different sets of coordination practices to deal with the uncertainty stemming from incomplete information.

Uncertainty is a concept much used in the literature on fast-response organizations and important to this thesis. Uncertainty is defined as the ‘inability to assign probabilities to the likelihood of future events’ (Milliken, 1987, p. 134) and is a function of a highly complex and dynamic environment (Weick, 1993). Uncertainty is thus the inability to predict how one’s practice might influence the outcome of a larger event (Milliken, 1987). Consequently, if the environment is more dynamic and complex, it becomes harder to determine the effects of one’s action on that environment. Uncertainty is not dichotomous, it is not simply existent or non-existent, it has varying degrees. As such, the level of uncertainty is often matched with different practices to overcome such uncertainties (Faraj & Xiao, 2006; Argote 1982, Wolbers, 2016). For example, Argote (1982) found that high uncertainty calls for less-formalized practices and

(6)

2 low uncertainty calls for highly formalized practices in hospital crews. The use of the words high and low is important, as uncertainty in itself is a constant in fast response organizations, one always deal with a minimum level of uncertainty. It is also a dynamic concept as the context in which fast-response organizations operate continuously changes, as such, uncertainty, which is rooted in the context, also changes.

To function in such troubling and uncertain environments, fast-response organizations use different sets of coordination practices. Some are pre-designed, such as protocols and standard operating procedures, and some are based on improvisation and adaptation. Most recent research on coordination in fast-response organizations is based on actual work-practices. Be it trauma teams (Faraj & Xiao, 2006), SWAT teams (Bechky & Okhuysen, 2011), fire departments (Bigley & Roberts, 2001), and emergency commanders (Wolbers, 2018). These observed practices however are all highly work-related; they fit to the specific environments in which the organizations operate. As such, they may show similarities to the work practices of other fast-response organizations, but must differ to fit to the specifics of the context in which they operate.

The Rotterdam Police Dispatch is also a fast-response organization. It must deal with a stressful and uncertain environment in which outcomes are potentially fatal and time is constrained. Moreover, the dispatch room is a very important link in the chain of emergency responders. Police dispatch rooms must collect and distribute information on the emergency and command and coordinate the initial phase of the response (Terrell, McNeese & Jefferson, 2004). This research aims to find out what coordination practices are important to the work of the Police Dispatcher. It seeks to find whether these practices differ in regards to the complexity of the operation. The research question is thus: How do Police Centralists coordinate the dispatch in both routine and complex operations?

The research serves two goals. First, I hope to contribute to the knowledge of the Rotterdam Dispatch Room. By studying its coordination practices, I hope to gain an understanding of what the important coordination practices are and share this understanding with the dispatch room. Second, this research contributes to a broader body of knowledge on coordination practices in fast-response organizations. To achieve said goals, this research is based on an explorative research design. I employ two sub questions. First, how do Police Centralists coordinate the dispatch in routine operations? Secondly, how do Police Centralists coordinate the dispatch in complex operations?

(7)

3

1.2 Academic Relevance

The research is both relevant to academics and society. In terms of scholarly work on the larger subject, coordination in crisis management is widely studied but far from complete. Bundy et al. (2017, p. 1662) stress that: ‘…despite sustained interest across multiple disciplines, recent

commentary on the field suggest that we have only just begun to scratch the surface of understanding crises and crisis management’. Within the study of crisis management falls the

subject of crisis coordination. The work done by Faraj and Xiao (2006) especially showed how much we do not know about the facet of crisis coordination. They employed a practice lens to study what exactly happens in organizations that have to function under pressure. Through direct observation, they learned what mechanisms were in play. The methodology first introduced by Faraj and Xiao (2006) differed from previous works which focused purely on the formalized portion of coordination, on information processing and resource management (Galbraith, 1982). By shifting this focus towards a more practical, or ethnographic approach, a lot of valuable and new information was gathered on the non-formalized portion of coordination practices in fast-response organization. This research uses a similar approach and seeks to contribute further to the growing understanding of coordination practices in fast-response organizations and as such to a larger understanding of crisis management in general. More specifically, it also answers the call from Bechky and Okhuysen (2011) in which they ask for further research on uncertainty in coordination processes.

Additionally, any gained knowledge on the practical functioning of Dutch police dispatch rooms is beneficial and adds to a small but growing body of literature. Some important works includes Van den Eeden (2018), who found that the information provided by dispatchers influences the way in which responders react to a crime-scene. Visser (2014) described how centralists respond to the new concept of a shared dispatch room between the fire department and the police department in Amsterdam. Moreover, Bresser (2013) described the preferred communicative medium Police Dispatch Centralists use. There is however no academic knowledge on how coordination works in Dutch police dispatch rooms and what practices are important. This research on coordination practices seeks to add to that gap in the literature.

1.3 Societal Relevance

Any contributing research into the subject of fast-response organizations is often of societal relevance due to the importance of these organizations to the general well-being. Good research

(8)

4 is then able to improve such organizations and help them in the crucial parts they play in keeping society safe and secure. This especially holds true for the dispatch operations, ‘The ability to

shave of seconds off an emergency response time can make the difference between life and death’ (Terrell, McNeese & Jefferson, 2004, p.706). This importance was showed once more

when The Dutch Red Cross tried to help a young-trained marathon runner who felt ill just meters from the finish line. Faulty preparation to the marathon, ineffective resource allocation and severely lacking communication with the medical dispatch center potentially led to the death of the woman (Red Cross admits mistakes surrounding death of runner, 2018). It is one of many examples that show why research into the practices of fast-response organization is so important. Hillyard (2000, p.9) captures it perfectly: ‘learning together from the event in order

to prevent, lessen the severity of, or improve upon responses to future crises. Research can show

good-practices as well as indicate faults and caveats in such organizations.

Furthermore, scientific research into fast-response organizations is especially helpful as organizational learning is difficult to achieve in the environments fast-response organizations tend to operate. This stems from the fact that such organizations cannot learn through the traditional method of trial and error, as the costs of error are often too high (Moynihan, 2008). Moreover, uncertain environments cause that the scope of learning is much greater than in certain environments and the ambiguity of previous experiences gives rise to faulty lesson

drawing (Moynihan, 2008, p.351). In other words, it is difficult to draw on previous experiences

as they might seem similar, but are in fact different from a current situation. For these reasons, it is beneficial to society to research the practices of fast-response organizations.

1.4 Reading Guide

This thesis starts of by providing a theoretical framework. The framework serves two purposes. It describes what is already known on coordination practices by showing the influential work on which this research is built on. Second, it serves as a mirror to which the coordination practices in the Rotterdam Dispatch Room are compared in chapter five. Chapter three describes the methodology of the research. It contains information on the research method applied, the collection of data and discusses the internal and external validity of the research design. Chapter four contains the actual research results. Finally, chapter five and six contain a discussion of the data and the conclusion to the research. These chapters include possibilities for further research and describe limitations to the research itself.

(9)

5

2. Theoretical Framework

The theoretical framework serves two purposes. First, it shows what is already established on coordination practices. Second, it offers a conceptual exploration of three different categories of coordination practices in fast-response organizations and exemplary coordination practices. These serve as a mirror to which we can compare the coordination practices in the dispatch room.

2.1 Preparing for uncertainty

Studies on coordination in fast-response organization are based on a larger body of research into coordination in organizations. Coordination in organizations is often stymied by uncertainty, ‘the inability to assign probabilities as to the likelihood of future events’ (Milliken, 1987, p. 134). As this was the predominant view in much organizational research, most scholars focused on either an information management perspective or a dependency management perspective to battle said uncertainty. They employed such perspectives to discover how the risks stemming from uncertainty could be mitigated. The information-processing perspective originally provided the first insights into coordination under uncertainty.

Galbraith (1982) for instance argued that any organization could overcome information processing problems by formalizing the roles, processes and people practices (Galbraith, 1982). It is based on the idea that whatever the problem is pertaining to information management, a structure could fix that problem. Thus, even if information is highly incomplete, diverse and contradictory, a structure based in ‘formalized roles, processes and people practices’ could mitigate the problems stemming from the low-quality of information (Galbraith, 1982). Argote (1982) however, in her study of hospital emergency units, concluded that uncertain information is best matched with a less-structured mode of coordination, and vice-versa. A conclusion most interesting to the study of crisis management as it indicates that structures are not always the best way of dealing with coordination problems. Especially interesting as most authors at that time believed it was best to coordinate and program as much as possible (Daft & Lengel: 1986, Huber, O’Connell & Cummings: 1975). Recent literature on information processing is based largely on the concept of sense making introduced by Weick (1993). Sense making is the process in which one tries to make sense of a situation, to understand its complexities and to know what to do next (Weick, 1993). Recent work on information processing include Barton et al. (2015) study on what practices influence successful information processing, Carrol’s

(10)

6 (2015) research into false clarities provided by confident leaders, and Wolbers and Boersma’ (2013) research into using a common operational picture to overcome information management problems.

This focus on information processing was followed up by a focus on managing interdependencies among recourses and activities. These interdependencies range for instance from using the same data resources or work relations with colleagues (Malone & Crowstone, 1990. The work by Malone and Crowstone lay the foundation for this particular method of studying coordination. They aimed to answer the question how dependencies work together, how does x influence y, and what are the results of that in regard to coordination. As such, they drift away from the focus on information management and employ a principal question as they firmly believe that, at a bare minimum, some coordination problems are universal and can thus be solved by the same solutions (Malone & Crowstone, 1990). However, and perhaps symbolic for that phase in research on crisis coordination, is that is this approach still sought to solve uncertainty through formalization (Chen et al., 2004; Carafano, 2002; Jungert, Hallberg & Hunstad, 2006). By looking at interdependencies, scholars arrived at conclusions in which they advised what steps to take to overcome the universal coordination problems stemming from those dependencies (Shen & Shaw, 2004). As such, research was still based on the notion that uncertainty can still be controlled for. Recent work on managing dependencies is mostly based on cross-organizational dependence. Bodin and Nohrstedt’ (2016) study into disaster management is exemplary of this line of research. In their study on collaboration in emergency response to a wildfire, they find that ‘patterns of actor-task interdependencies influence task

response effectiveness’ (Bodin & Nohrstedt, 2016, p.183). They move beyond the assumption

that increased collaboration, and thus a better management of interdependencies, is the universal solution to crises. Instead, different tasks require different sorts of collaboration. Successful solutions are then derived from a good fit of the structure and performance of collaboration, rather than just more collaboration (Bodin & Nohrstedt, 2016).

2.2 Dealing with uncertainty

Information processing and the management of interdependencies help organizations deal with uncertainty. Through standardization of tasks, one could prepare an organization for complex problems.

(11)

7 Rasmussen’s’ division is perhaps the best way to describe the difference in preparing for and dealing with uncertainty: ‘During unfamiliar situations, faced with an environment for which

no rules for control are available from previous encounters, the control of performance must move to a higher conceptual level, in which performance is goal-controlled and knowledge-based (Rasmussen, 1983, p.259). This is similar to what Faraj and Xiao (2006) described, how

in ninety percent of the times, standardized routines, or what they call expertise practices, led to a successful treatment. In the case of their studied trauma center, only ten percent of the patients had to be solved through other coordination mechanisms. They divided such trajectories, in which standard practices were not satisfactory, by the definition of non-habitual task trajectories, vs. the habitual task trajectories (Faraj and Xiao, 2006).

The study by Faraj and Xiao (2006) served as the basis for more research into the matter of non-habitual practices as it was rather contradictory to a large part of the literature on organizational theories. Most literature was indeed based on what is described above, the preparation for uncertainty, and not how organizations deal with uncertainty when it cannot be prepared for. They defined those coordination practices which can be prepared for as expertise coordination practices and those that cannot be prepared for, as the problem is simply to complex or too unique in its nature, as dialogic coordination practices (Faraj and Xiao, 2006). Kroll-Smith and colleagues define what dialogic coordination practices are accurately: ‘The

first-responder resembles in action and character the bricoleur, that person who alters and transforms ideas and materials to create new and innovative approaches to the world (Kroll-Smith, Jenkins, & Baxter, 2007, p.5). Dealing with uncertainty is thus the creation of new and

innovative approaches to solve a task.

Besides Faraj and Xiao’s work on the subject of these new and innovative approaches Beckhy and Okhuysen found through observation that organizational bricolage was important for both SWAT teams and film crews in dealing with uncertainty (2008; 2011). The term bricolage is pointing to the adjustment of that what is designed to such an extent that it fits the new complexity that was not prepared for. Klein and colleagues employed a similar research set up and found that extreme actions teams relied on a highly flexible leadership structure to enable them to successfully complete tasks (2006). Furthermore, the work done by Wolbers is especially helpful to understand how fast-response organizations deal with uncertainty (2014, 2016). By allowing the idea that an organizational cannot control everything, Wolbers,

(12)

8 Groenewegen and Boersma (2018) showed how fragmentation and isolation is sometimes needed to solve the most complex of environments.

2.3 Coordination practices

This theoretical framework is based on three different bodies of literature on coordination practices; designed, emergent coordination practices (Wolbers, 2016) and dialogic coordination practices (Wolbers et al., 2017; Faraj and Xiao, 2016). The term designed points towards a stadium prior to an emergency response, emergent points towards a more continuous process, which requires on the spot adaption, and dialogic coordination practices are only loosely based on pre-designed structures. Designed coordination practices function‘…by structuring the

response operation in advance, and allow professionals to fall back upon well thought out plans of action, known to everybody in the organization’ (Wolbers, 2016, p.17). Emergent

coordination practices require ‘ongoing adaption as fast-paced environments are often too

unstable for aligning coordination mechanisms with predefined contingencies’ (Wolbers, 2016,

p.18). Dialogic coordination practices were first coined by Faraj and Xiao as; ‘…contextually

and temporally situated responses to occasional trajectory deviation, errors, and general threats…’ (2006, p.1159). Such practices are employed irrespective of formalized routines and

structure, and, as such, often cross professional boundaries (Faraj & Xiao, 2007; Wolbers, 2016). These three sets of different coordination practices help organizations deal and adapt with increasingly complex and dynamic contexts.

The following paragraphs explain in detail the main coordination processes discovered in previous empirical research. Furthermore, the framework will show how such coordination practices interact with each other in various fast-response organizations. Such interactions are often necessary because as stated in above, the context in which fast-response organizations operate is highly dynamic. Thus, the interactions of the fast-response organization in that context must also be dynamic. Organizations must adapt to these changing contexts by adapting suitable practices (Argote, 1982). The theoretical framework is thus based on three sets of practices that are not exclusive; they interact and often cross conceptual boundaries. The following paragraphs follow a constant structure. The theoretical concepts important to each category of practice (designed, emergent and dialogic) are explained before coordination practices, exemplary to that category, are described. The resulting framework then serves as a mirror to which one can compare the coordination practices in the dispatch room studied.

(13)

9

2.4 Designed Coordination Practices

Formalized rules form the basis for designed coordination practices. They help organizations prepare for operations in advance. As such, they create certainty in uncertainty. To prepare for its normal use an organization must employ structure that is, it must include ‘standardization,

specialization, formalization, and hierarchy’ (Bigley & Roberts, 2001, p.1281). Even in the

highly diverse environment in which a dispatcher operates, there are often tasks that are more frequent and less complex than others. Formalized practices function well when such tasks are indeed equal in nature. The functioning is thus dependent on the similarity of conditions (Moynihan, 2008). If such conditions are roughly the same, standard operating procedures serve as a quality guideline to solve any new similar case. As such, for a large majority of tasks for any organization formalized rules and practices are an important part of their task-solving capabilities.

Standard operation procedures are thus important to any organization dealing with a regular task set. However, to deal most effectively with all tasks at hand, and as such also with irregular tasks, an organization also needs flexibility (Bigley & Roberts, 1991). This paradox, this reliance on both structure and flexibility is what Adler, Goldoftas and Levine (1999) call one of the starkest challenges for organizations. According to Bigley and Roberts (2001) however, their case study of a fire-departments practices shows that Incident Command Systems (ICS) have the ability to solve this paradox. Even when organizations are faced with ‘extreme uncertainty and instability’ (Bigley & Roberts, 2001, p.1281). ICS’s are not something that is universally equal. ICS’s differ depending on the organization that employs it. ICS’s profit from the bureaucratic formalization in certain environments whilst simultaneously providing an organization the flexibility that is needed in uncertain environments. It is however still based on pre-planned structures and practices and thus matching with the definition of designed coordination. ‘It extends the domain of rationality and bureaucratic organizing to the uncertain

and often chaotic environment of disaster responses. It provides a set of rules and practices to guide the actions of the various organizations responding to disaster, and creates the necessary division of labor and coordination mechanisms among them’ (Buck, Trainor & Aguirre, 2006,

p.1).

ICS’s are rather similar to what Faraj and Xiao call expertise coordination practices. Faraj and Xiao study a trauma team, which in essence rather similar to a fire department as it operates under the same uncertain circumstances. They argue that proper coordination is based on

(14)

well-10 developed expertise coordination practices (Faraj and Xiao, 2006). Expertise coordination practices refers to ‘processes that manage knowledge and skill interdependencies’ (Faraj & Xiao, 2006, p.1160). Managing knowledg, and interdependencies is much alike to what I described as preparing for uncertainty in the body of literature chapter. It is based on pre-planning to make acting under uncertainty easier.

Bigley and Roberts (2001) identify three different elements important to any ICS: Structuring Mechanisms, Constrained improvisation and Cognition Management Methods. The structuring mechanisms allow fast-response organizations to adapt themselves to any specific task without loss of reliability and at the same time improve flexibility and focusses on a top-down hierarchical structure. Through structure elaborating, the process of the construction of a temporary organization, an organization can adept itself to any given situation. Pre-formalized rules give an organization the opportunity to expand and collapse within a pre-determined manner. This helps them adapt to different sort of emergencies. Bigley and Roberts find that this often happens in the same way (2001). Faraj and Xiao call this a reliance on protocols. Their trauma team functions according to standard operating procedures in which tasks, rolls and sequences are formalized beforehand. They are built on ‘consensus among experts based

on reliable experiences’ (Faraj & Xiao, 2006, p.1160). Standard operating procedures, or

structuring mechanisms, help fast-response organizations in the most demanding of

circumstances, they are proven ways to structure interactions and manage the basic resources and expertise interdependencies (Faraj & Xiao, 2006, p.1160).

These processes used to alter the organization in other formalized forms is not enough to deal with uncertainty. As such, fast-response organizations give space for constrained improvisation for the people in the field (Bigley & Roberts, 2001). Important to stress however is that this improvisation is indeed constrained, as such, it is still formalized and nestled in rules and pre-planned practices. It is improvised in the sense that normal procedures might not give a satisfactory result and are thus diverted from. The one in command may give an order without issuing specific information on how to perform the task. He or she relies on the notion that the sub-ordinates know how to perform specific tasks and as such is able to adapt to novel environments on the front-line (Bigley & Roberts, 2001). Constrained Improvisation is necessary to overcome coordination problems associated with explicit and centralized command and thus provides any ICS the ability to overcome tensions nestled in over-formalizing. This is because it gives FRO’s the ability to shift away from centralized to

(15)

11 decentralized- and on-site decision-making processes. Such flexibility is however not without its boundaries. ‘Responding to situational changes is a collective process that occurs within

constraints of organizational values, collaboration, and previously established decision-making routines’ (Bellamy, 2011, p.28).

Designed coordination practices thus help fast-response organizations deal with matters of low uncertainty, as the tasks are often regular and recognizable. Through pre-formalized structures and rules, organizations are capable of adapting to a wide array of regular tasks and problems. Observed coordination practices such as standard operating procedures, overlapping situational representations and elaborating structures are important to the trauma team and the fire department and might show similarities to the coordination practices observed in the dispatch room.

2.5 Emergent Coordination Practices

There are however limits to the use of designed coordination practices to deal with uncertainty. As Lutz and Lindell rightly note, and in line with the limitation mentioned by Bigley and Roberts, such practices are best suitable to ‘existing emergency response organizations

performing their normal tasks with normal staff’ (2008, p.124). They are less effective in

emergency response organizations performing new tasks (Lutz & Lindell, 2008). In such situations, fast-response organizations change their modus operandi and often switch to less formalized practices to deal with higher levels of uncertainty (Faraj & Xiao, 2006).

Less formalized coordination practices are labeled as organizational bricolage (Bechky & Okhuysen, 2011), dialogic coordination practices (Faraj & Xiao, 2006), or emergent coordination practices (Wolbers, 2016). Dialogic and emergent both point to the characteristics of a mechanism which is in flux and not constant and thus contrary to designed coordination practices. They are based on the designed coordination practices but evolve from that point into procedures that are harder to formalize and difficult to capture in protocols.

Bechky and Okhuysen describe this process as restructuring. Organizations, in their research of swat teams and film-crews, restructure their routines to such an extent that they fit the new complexity (Bechky & Okhuysen, 2011). This complexity arises from uncertainty and surprises, which is ‘a break in expectation that comes from situations that are not anticipated or do not

advance as planned’ (Bechky and Okhuysen, 2011, p.239). The contradiction between

(16)

12 difference in environment, which in the case of emergent coordination practices cannot be controlled for. Therefore, organizations must restructure their patterns in order to fit that break in expectation. The study by Bechky and Okhuysen (20100) described how a swat team and a film-crew deal with such a break in expectation by bricolage. Bricolage is the ‘skill of inventing

resources from available materials’ (Cunha, 2005, p.2). It is thus based on the designed

coordination practices stated earlier, but evolves from that into coordination practices that are flexible and more suitable to uncertain environments.

Bechky and Okhuysen then divide the process of organizational bricolage into three main coordination practices and two of these practices are of specific interest to this research. First, the practice of role shifting. It is the process of adjusting one’s own tasks by ‘substituting for

someone else or performing some of the tasks in someone else’s role’ (Bechky & Okhuysen,

2011, p.246). It is conceptually different from role-based coordination or plug-and-play teaming in the sense that such practices are designed for. Role shifting is different as the pre-defined roles themselves change. As such, an operational commander might still be in command, yet at the same time perform activities that are not previously constituted in the formalized task-description of a commander. For such actions to work, one must know, and this is where socio-cognitive processes are important, ‘what everyone is supposed to do’ (Bechky & Okhuysen, 2011, p.246).

The second coordination practice is called reorganizing routines and is defined as ‘reorienting

themselves to a new set of goals and restructuring their activity by switching the routines they were using’ (Bechky & Okhuysen, 2011, p.247). It stems from a disruption of the expected

flow of activities. As the flow is disrupted, the actor might change swap routines to fit the new situation. This touches on the specification described by Faraj and Xiao on habitual and non-habitual trajectories (2006). The use of the word trajectories signals to a continuing dynamic in which the workflow changes. Habitual then describes a trajectory that follows expectations and is thus solved through designed coordination practices (Faraj and Xiao, 2006). Non-habitual describes the situations in which these expectations are thus disrupted and requires the reorganization of routines to fit the uncertainty of the task. Again, this process is based on knowledge acquired beforehand. However, Schakel et al. (2016) study shows just how difficult this is in an escalating environment as it the context of the work is so complex. The expected workflow of surveillance procedures was interrupted because the surveillance target was assassinated. As such, the surveillance operation had to switch practices towards an

(17)

13 apprehension operation, and, whilst a major police force was present at the crime scene, failed to do so (Schakel, Fenema & Faraj, 2016).

Emergent coordination practices are thus largely rooted in integration but are more flexible than designed coordination practices (Wolbers & Boersma 2017). This integration is thus based on collective socio-cognitive resources: the collectively held knowledge about how a task is

performed and how activities advance. In such a sense, they are only viable in somewhat

predictable environments, as one knows what to do in such situations based on this collective knowledge. The given characteristic of uncertainty as a constant in the environment of fast-response organizations does not mean one cannot prepare for it. However, it does entail that such preparations can only reach so far. If the environment is too dynamic, or too unexpected, even emergent practices such as described above are sometimes incapable to offer viable solutions.

2.6 Dialogic Coordination Practices

In such situations, in which both designed coordination practices and emergent coordination practice fail to provide a viable solution to a complex task, organizations engage in even less formalized activities to deal with increasing levels of uncertainty. It is what Faraj and Xiao call dialogic coordination practices (2006). Such practices are needed because of the ‘rapidly

unfolding tempo… and the stochastic nature of the treatment trajectory, dialogic coordination practices are used as contextually and temporally situated responses to occasional trajectory deviation, errors and general threats to the patient’ (Faraj & Xiao, 2006, p.1159). To describe

coordination practices that are used to solve the extra-ordinary Wolbers (2017) uses the term fragmentation. He does so because integration, the use of designed coordination practices, is not always the answer to ‘ambiguity and discontinuity’. This follows from the fact that one cannot prepare for every altercation and uncertainty is thus a constant. As such, it is more helpful to see things in the way they are, as fragmented and non-centralized. In a sense it is much in line with what Argote (1982) described almost 30 years earlier: ‘Programmed means

of coordination made a greater contribution to organizational effectiveness under conditions of low uncertainty than under conditions of high uncertainty. Conversely, non-programmed means of coordination made a greater contribution to organizational effectiveness when uncertainty was high than when it was low’ (p.420).

(18)

14 By focusing on this fragmentation perspective instead of an integration perspective, Wolbers and colleagues (2018) discovered three main coordination practices: working around procedures, delegating tasks and demarcating expertise. These practices all work through the reinstatement of boundaries, be it normative boundaries, functional boundaries or knowledge boundaries. These boundaries reinforce fragmentation within and between organizations (Wolbers, Boersma, & Groenewegen, 2018). The first practice is called working around procedures and is quite similar to the process described by Bechky and Okhuysen as reorganizing routines (2011). It too is concerned with altercations of normal work procedures but differs from that in the sense that they are abandoned all together. In the case of Bechky and Okhuysens’s (2011) definition of reorganizing routines the procedures are switched, or, an pre-formalized procedure is switched for another designed routine (2011). Wolbers et al. (2018) definition entails a more radical practice in which organizations divert away from all procedures as they hinder successful progression. This is what Faraj and Xiao (2016) call protocol breaking and their included example shows how this works: ‘You see that, at least the experienced person

sees that the issue is not ABC but is FGH. You can, in the hands of the experienced person, skip the ABC, go to FGH, and then cover the ABC just as a double-check after you dealt with what you think is the critical problem that the patient has’ (p.1165). Such practices are triggered by

unexpected events and require creative solutions on the spot such as ad-hoc collaborations and a change of tactics (Wolbers, Boersma & Groenewegen, 2018).

According to Wolbers et al. (2018), emergency responders also employ functional boundaries to regain control over a situation. It is triggered by the struggle to create a complete sense of the situation and a common problem to organizations dealing with uncertainty (Weick, 1995). The functional boundaries stem from a process called task-delegation (Wolbers, Boersma, & Groenewegen, 2018). For example, a commanding officer ideally would have full control over the process in an emergency response. However, due to the complex nature of the work of emergency response, it is difficult to retain the overview needed to have full control. As such, they delegate some of their responsibilities to sub-ordinates to such a degree that they have the possibility to regain the overview, or to make sense of the situation at hand. The downside of that is that the one in command loses control over a situation, as the tasks delegated are dynamic in nature. To regain control and sense over their own work, responders reinstate functional boundaries to take responsibility of what needs to happen.

(19)

15 Furthermore, both in the studies by Faraj and Xiao (2006) and Wolbers, Boersma and Groenewegen (2018), expertise is an important part in dialogic coordination practices. Wolbers and colleagues (2018) for example use the term demarcating expertise to describe the process in which someone with specific expertise will assume responsibility for handling a situation that is uncertain or potentially dangerous. This again creates boundaries because if one claims responsibility over what to do, as it is ones responsibility to do so, others are not in the loop anymore. They do not know why certain actions are employed because of this demarcating effect. As such, the coordination is no longer integrated but fragmented. It is comparable to what Faraj and Xiao (2006) call epistemic contestation, which is also based on different fields of expertise. In the process of epistemic contestation, different actors argue over what steps to do based on their previous experience and fields of expertise. This shows that sometimes fragmentation is needed when uncertainty leads to a stalemate. In such situations designed and emergent coordination practices do not offer the emergency responder a viable solution. As such, one might divert from them and use dialogic coordination practices to combat uncertainty and solve the situation.

2.7 Categorization

This leaves us with three different categories of coordination practices in fast-response organizations; designed-, emergent, and dialogic coordination practices. That does not mean however that we will observe all above coordination practices. However, it does provide us with a means of comparing the practices in the Rotterdam Police Dispatch Room to other fast-response organizations. As such, it provides a frame to mirror the actions. The below table combines the three categories and their underlying coordination practices.

(20)

16

Table 1 Coordination Practices in Fast-Response Organizations

Category Coordination Practices Authors 1.Designed Coordination Practices Standard Operating Procedures

Bigley and Roberts; 2001, Faraj and Xiao, 2006,

Constrained improvisation

Bigley and Roberts; 2001

Elaborating Structures Bigley and Roberts; 2001

2.Emergent

Coordination Practices

Role Shifting Bechky and Okhuysen; 2011, Cunha; 2005

Reorganizing Routines

Beckhy and Okhuysen; 2011, Baker and Nelson; 2005

3.Dialogic Coordination Practices

Working around Procedures

Faraj and Xiao; 2006, Wolbers and Boersma; 2018, Argote; 1982

Delegating Tasks Faraj and Xiao; 2006, Wolbers and Boersma; 2018

Demarcating Expertise

Faraj and Xiao; 2006, Wolbers and Boersma; 2018

(21)

17

3. Methodology

3.1 Explorative research design

This research aimed to find out what the important coordination practices are in the work of police dispatchers. To do so, I employed an explorative research design. The research was of explorative nature because there is no knowledge on coordination practices in dispatch organizations. The framework served as a guideline to this explorative research as I expected to encounter similar practices in the Rotterdam Dispatch Room as in other fast-response organizations. This follows from the fact that the environment in which the dispatchers operate shows similarities to the environments of other fast-response organizations. By employing a qualitative approach, the research is based on interpretations of empirical data (Toshkov, 2006). I used a single case research design as I only focused on one dispatch room. Such an approach provides the opportunity to delve deeper into this specific case and zoom in further on the coordination practices at hand (Toshkov, 2006).

The research relies primarily on data acquired through observation and interviews. The non-structured interviews were used to gain a richer understanding of the practices observed. I acquired the data through observing the unit of analysis, the dispatch centralist, in its natural environment. As such, I did not participate in the work. The observational research is thus of non-participant nature in which I, to the best of my ability, did not alter the environment (LeCompte & Goetz, 1982). As I do not wish to infer any causality between variables in this research, and my research goal is exploration and not explanation, an observational or ethnographic approach was suiting.

As I use two sub-questions to answer the main question, it is important to indicate how I theoretically split the two questions. The first question is concerned with police dispatch coordination in routine operations, the second with complex operations. Routine operations are the operations and environments to which the dispatcher is familiar. These operations are plentiful and happen on a regular basis. They are the majority of the events a dispatcher handles. Logically then, an complex operation is an event that is not regular. They happen less often and are more complicated than routine operations due to their nature. They for instance require more police officers on the scene of the emergency. Moreover, complex operations are often characterized by their dynamic nature which constantly changes. The categorization is much

(22)

18 alike to Faraj and Xiao’s (2006) split of habitual and non-habitual cases, as described in chapter two.

3.2 Rotterdam Dispatch Room

Fast-response organizations are characterized by their ability to perform in the most demanding of environments and do so without error. Dispatch rooms are exactly that and could even be seen as the arch-typical fast response organization (Terrell, 2006). The Rotterdam Dispatch Room itself is a multi-disciplined dispatch center; ambulance, fire department and police centralists all reside in the same room. As such, when needed, the communication lines are short and direct. The divisions do reside in different sectors of the room. Furthermore, it is a rather new organization and the product of a merger between the two dispatch rooms of Dordrecht and Rotterdam. It is also the first dispatch room to operate with a new national IT-system, this system enables other dispatch rooms in The Netherlands to cooperate and even substitute for the Rotterdam Dispatch Room in times of need. This researched focused only on the police dispatch centralists.

The Police dispatchers work in two separate teams: intake and dispatch. Intake centralists, often three or more at the time, take in all the calls that the national dispatch room redirects to the Rotterdam Dispatch Room. Dispatch centralists control and command the police officers in Rotterdam and the surrounding area of Rotterdam. Through an Integrated Dispatch Room System (GMS), intake- and dispatch-centralists communicate. Intake centralists retrieve information from the caller and log these bits of information into the GMS, the dispatcher then reads the information, and decides on what to do with it. Besides communication through the GMS, the dispatchers are continuously in contact with each other and communicate face to face about the work.

(23)

19 Both teams are situated closely to each other. Figure 1 illustrates the plan of the Rotterdam Dispatch Room. The dispatch centralists also work in two different teams, Dispatch North and Dispatch South. The Maas divides the operating area of the dispatch teams. The two teams thus operate according to the geographical boundary created by this river. Within the two different dispatch teams, there is continuous communication through a headset connection. If one of the dispatchers communicates with an officer on the street, the other dispatcher automatically listens in. The communication between the two dispatch teams is mostly through face-to-face communication. Through a large screen in the center of the police cluster, the dispatchers know who are available and who is currently on the phone. The Police Dispatch Center works round the clock, 24/7, and a day is typically split into a morning-, afternoon-, and night shift.

At the beginning of a shift, the team leader gives a short briefing on the specifics of that shift. Important information from the previous shift is discussed as well as expectations on the upcoming shift. Besides working in three different shifts, centralists also frequently swap between intake and dispatch and are thus familiar with both sets of tasks.

3.3 Data Collection

The data collection was based on observations and unstructured. The combination of the two data sources created a rich understanding. Spencer, Logan and Coiera (2002) wrote a valuable

(24)

20 manual to observational research and this formed the basis of the research. The authors mentioned four critical standards for non-participant observational research. First, it is important to be familiar with the surroundings of the environment that is being studied. As such, I had a preliminary meeting with my contact in the dispatch room; we talked about the details of the research and the work processes in itself. A small introduction to the office followed and we agreed on the data of the observational period. Second, those observed must know and accept that they will be observed (Spencer, Logan & Coiera, 2002). Again, the officer of service was invaluable and agreed to forward an introductory email of myself to the dispatchers that were going to be observed. Moreover, before each observational period, I introduced the research and myself once more to the dispatcher I was to observe and asked if it was all right to conduct the research. Happily, they never denied me access.

Third, the observer must be as invisible as possible (Spencer, Logan & Coiera, 2002). As the nature of the work is quite complex, the communication is full of difficult terms and there is a large reliance on codes, it was quite hard to remain invisible. Especially in the first two sessions I had to ask a large number of work-related questions on certain procedures and actions taken by the dispatchers. This was necessary to properly understand the work and to derive at valuable conclusions. In the remaining sessions I was able to more properly focus myself on the coordination practices at hand, and ask more questions on the actual practices, rather than the processes. Still, by doing so, I did not remain as invisible as possible but rather made the deliberate choice to not do so and in turn gain a better understanding of the practices. Fourth and final, it must be clear what the specific tasks are of those observed (Spencer, Logan & Coiera, 2002). As such, it was important to familiarize myself first with the dispatch room and the tasks performed in each role. The initial meeting with the officer served as a solid starting point towards such an understanding and further talks with dispatchers filled the remaining gaps. Before I could start my observation period, I was supposed to sign a non-disclosure agreement. We formally agreed that none of the personal information, from either the Police personnel or the civilians was to be included in the final report. No personal information is therefore included in the thesis. Nor is there any detailed information included in the provided vignettes.

The observations itself led to a large collection of field notes. The field notes were recorded with a note-pad and pen, as per requested by the police department. When time allowed, I immediately digitalized these notes and added extra information that I remembered but did not manage to write down at the time of observation. The observational periods were often around

(25)

21 six hours and split between two or three hours per dispatcher, depending on the roster for that day. For every observation I recorded the date, time, actors and the subject. Dependent on the actions undertaken I then recorded extra information on practices, processes and other important notions.

The Rotterdam Dispatch Room allowed me to listen in to the communications through a wireless headset. By logging in to the central system I was able to listen in on the communications between dispatchers and between the dispatchers and secondary sources, such as police officers on the street or other dispatch rooms. Furthermore, as I seated next to the dispatcher I observed, I could also watch the four different screens the dispatchers used. These screens often showed the GMS, any live-camera feeds, opened protocols, live mapping and an extensive communication system. To my understanding and as a result of my position, I had full access to the work done by the dispatchers, as everything was either openly communicated or visible in the screens.

Aside from the data acquired through direct observation, I also conducted interviews with the dispatchers to acquire extra data and test the primary data. There are three sorts of interview methods in qualitative research, structured interviews, semi-structured interviews and unstructured interview (Doody & Noonan, 2013). This research used the last method as structured- or semi structured interviews rely on a set number of questions and, or, exact wording of those questions (Doody & Noonan, 2013). I was not able to conduct such interviews, as I preferred to interview the dispatchers during their work. The unstructured interviews allowed me to ask questions on the practices observed at hand. As I did not know what I specific coordination practices I was to expect, this form of interviewing greatly helped with the collection of rich and deep data as I could improvise on the spot (Doody & Noonan, 2013). A typical interview followed a pause in actual dispatch work. Whereas the work practically never stopped, there were some moments in which the dispatchers did not had to communicate with another dispatcher or a police officer. Such moments were often used to ask questions on the processes and practices just observed. Often the dispatchers themselves started a conversation on the work at hand and chose voluntarily to explain work practices. However, as mentioned above, the work is continuous. As such, the talks were often interrupted and had to be continued later on.

(26)

22

3.4 Conceptualization

Before any valid conclusion could be drawn on the main research question, the raw data had to be translated into theoretical notions. ‘Operationalization is the translation of abstract concepts

and their attributes into less abstract ones than can be detected, classified and measured in the empirical world’ (Toshkov, 2016, p.101). Operationalization thus requires a translation from

text to code to theoretical categories (Saldama, 2008). As such, the data from the textbook – direct usage of a laptop was impractical – had to be typed out. The data was typed out as fast as possible after each session. By doing so, the memory was still fresh and extra notes and information could still be added. This was necessary as one can only write so fast.

The raw digital data then was entered into an excel file, leaving extra space to code the data. This translation can be conducted in two ways. By either using indicators or detectors. Indicators are often used to directly translate a concept, or a part of a concept, into a single measurable object. Detectors however are less specified and primarily used for detecting the presence or absence of concepts in the empirical world (Toshkov, 2016). As such, the use of detectors was suited better to this type of research as observational research is largely based on interpretation. Detectors help shape the collection of data whilst at the same time provide flexibility in research.

After the coding, one can continue to construct conceptual categories (Saldana, 2008). When grouped, the coded transcripts led to these conceptual categories. For example, an observation of the use of certain features in the GMS by one of the dispatchers was coded up as a reliance on standard operating procedures and placed in the category of protocol-usage. After complete codification of the raw transcripts, we thus arrived at a body of categories (Saldana, 2008). The coding progress was far from linear. An initial codification and categorization of an occurrence might seem like a stand-alone coordination practice, but after considerate review of the remaining events, this view might change as it shows similarities to another category (Charmaz & Belgrave, 2007). As such, the progress was back and forwards until I was certain I arrived at the most complete body of coordination practices. As such, I continuously had to review the work done and re-consider the coding scheme.

3.6 Internal and external validity

Employing an observational design or ethnographic approach offers many benefits. The researcher is able to study its subject in its natural environment and is thus easier able to

(27)

23 understand and describe social phenomena (Baker, 2006). However, as with any research approach, it does have its drawbacks. Besides practical obstructions, a number of questions could be asked in relation to the reliability and validity of the research.

A research is valid when it is ‘plausible, credible, trustworthy, and, therefore, defensible’ (Johnson, p. 282). Moreover, scholars use a twofold categorization of validity, internal- and external validity. A research is internally valid if the measurements provide an answer to the that what is asked (Baker, 2006). By wrong conceptualization of theoretic notions, one could arrive at a dataset that unusable in answering the actual research question. There are multiple ‘strategies’ to arrive at an internally valid research design (Johnson, 1997). So-called low-inference descriptors help in achieving internal validity by limiting the influence of the researcher. Low inference descriptors are observations phrased as closely as possible to the participants’ accounts (Johnson, 1997). By using a verbatim coding scheme, I tried to do so, as well as by including an extensive field note roster. Furthermore, Johnson notes that participant feedback is an important way of validating the research. How, for instance, can the observer know for sure a certain set of actions is a dialogic coordination practice, if is also possible that it was actually a formalized routine, quite regular to the work of the dispatch centralist. This is especially troublesome due to the fact that this research is indeed explorative in nature, stemming from the fact that there is not a lot written about fast-response coordination in dispatch rooms. To take away much of these concerns it is important to not only use on data source but also triangulate that data by the use of a secondary data source (Johnson, 1997). By asking the centralist about the performed practices, I determined if the original observation was indeed correct.

External validity is concerned with the possible generalization of the conclusions to the broader population of fast-response organizations. Whilst this is troublesome for any kind of research design in the social sciences, observational research is under extra scrutiny as the results are traditionally only based on observations in a specific period (Baker, 2006). As such, the matter of randomness comes into play. I might have observed a collection of events that were in no way representable of the work practices of the centralists. As such, one might argue that what I have observed is not representable for the practices of the Rotterdam dispatch centralists, let alone the entire population of dispatch centralists in the Netherlands. Ideally, one combats such remarks by entertaining a longer observational period. An observational period of several months, for instance, gives the researcher extra confidence in knowing that his or hers results

(28)

24 are indeed representative for the population. As this was not possible due to the timescale of a master thesis research, I had to employ a second method of data collection.

Triangulation gave me more confidence in regards to generalization of the results. By asking questions prior, after a shift and in between events, I was able to confirm that observed coordination practices actually occurred in the way I saw them, and, that those practices occurred more frequently. Moreover, they often provided much more information on the work processes than one could have observed. Concluding, in regards to the generalization of the results I am confident that the observed practices are indeed valid for the larger population of dispatch centralists.

3.5 Data Presentation

The data was collected through observing the dispatch centralists in their natural environment, the Rotterdam Dispatch Room. The research is built upon five separate shifts of six hours. In these thirty hours, nine different dispatchers were observed, all with different levels of experience and local knowledge. Some dispatchers just finished their training or were still being trained. Other dispatchers were very experienced and served different bodies of the police force for the past forty years. Due to the merger of two different regions into the Dispatch Room Rotterdam, three dispatchers were not as familiar with the specifics of Rotterdam as they wanted to be. As such, some dispatchers were very familiar with the processes of dispatch, but not familiar with Rotterdam, and vice versa. Furthermore, not all dispatchers had both experience on the street and as a dispatcher. Some were specifically trained as dispatchers, and others were first police officer before becoming a centralist.

Thirty-six events were recorded in the five shifts. The first two shifts were dedicated on understanding the processes. This understanding was vital in recording and understanding the events in the last three shifts. These shifts were thus richer in recorded events and insightful interviews. The interviews served as a secondary source of data. According to the dispatchers, the shifts observed were quite calm. As a result, there was plenty of time to ask questions and conduct talks with the personnel. This was partially planned as the first two sessions were scheduled in the morning. As such, it would be easier to understand the processes of the dispatch room as the dispatchers would have more time to talk during their work. Also, the third, fourth and fifth shifts were scheduled later on during the day, with the goal to encounter more serious and complex incidents. However, these shifts were also relatively calm in comparison to other

(29)

25 Thursday and Friday nights. Whilst this might have negatively affected the number of events recorded, there was more time to conduct insightful interviews and ask plenty of questions to the dispatchers.

(30)

26

4. Results

4.1 Findings

This brings us to the results of the research. The analysis presented in this chapter allows me to answer the main research question; how do Police centralists coordinate the dispatch in both routine and complex operations. As mentioned before, I employed two sub-questions to answer the main research question. The first question is as follows; how do Police Centralist coordinate the dispatch in routine operations? The second question is; how do Police Centralist coordinate the dispatch in complex operations. After each sub question, I present a table showing further examples of the practices described. I conclude the chapter with a process model showing how Police centralists coordinate the dispatch.

An event always starts with the intake centralist receiving a call, often redirected from the National Dispatch Room in Driebergen, from a caller who indicated that they required police assistance. To understand the complete process, my research started by observing a shift of an intake centralist. After the observation of an intake-shift, the remaining observational period only focused on the dispatch side. A typical intake event consists of the centralists asking the location and the reason of the call. Vignette 1: The intaker answers the phone and introduces

him or herself. At the same time, the intaker opens up a new event in the GMS. Immediately after introducing itself, the dispatchers asks the location of the emergency. The caller tells the dispatcher that there is an stationary vehicle in the left lane; he does not however mentions the location of the stalling car. The dispatchers thus reminds him of the necessity of knowing the location of the emergency. The caller then tells the dispatchers that it was on the A16, nearing the exit Kralingen. Not being specific enough, the dispatchers asks for the hectometer [a road sign indicating the exact location on Dutch motorways, placed every hundred meters]. With the use of the hectometer, the location was established and the dispatcher could enter this information in the GMS and ask for the specifics of the emergency. The caller tells that the car broke down, the passenger is next to it and that it probably needs towing. The intaker thanks him and enters the final information in the GMS.

When the call comes in, the intake specialist immediately opens up the GMS and enters a new event. Such an event contains a large body of information, including but exclusive: the location of the caller, the number, the time, the reason for the call, and the level of priority. This all happens whilst the caller is possibly still on the phone. In such a way, the dispatch centralist

(31)

27 can start with the actual process of coordinating the response. It is thus essential to gain information on the location first; the centralist can immediately send a police officer to the scene as fast as possible if that is necessary. In such a typical situation as described above, there is often no face-to-face communication between the dispatcher and the intake centralist. The dispatcher is simply notified on one of the four screens they use and then acts on that information.

Vignette 2: The dispatcher sees the notification of a new event. All info is easily accessible through the GMS. A motorist fell from his scooters, sustained physical injuries. The time and location are also visible to the dispatcher. Dispatch calls in a police officer using their personnel four-numbered code. The officer responds and tells the dispatcher to go ahead. The dispatcher then tells the police officer that there was a one-sided accident on the motorway and that a motorist fell from his bike, injuring himself in the process. The police officer tells that he understood the message and that he is on his way and will give an update on the situation. Dispatchers then ends the conversation but before doing so tells the officer that he has permission, [permission is a term used by the police to indicate that the dispatchers gives the police officer(s) permission to ride with sirens]. Again, the police officer indicates that got the message. Event 2 describes the typical manner in which lot events are dealt with. The dispatcher

sends the police officer to the location of the accident and at the same time explains the event. From intake to dispatch normally does not take longer than two minutes, of course depending on the priority of the event – lower priority events are often less important and thus take longer – and the amount of events active at that time.

4.2 Routine operations

The first sub-question is; how do Police Centralist coordinate the dispatch in routine operations? They do so in an orderly and fast manner. Most of the time an event was coordinated in less than a minute. An event came in through the GMS, at that point, either one of the dispatchers read the noted info in the GMS, checked whether there was an available officer on the street and, if there was, sought contact with that police officer. The police officer received the information and moved to the location of the event. The dispatchers often coordinated such events without conversing with each other or asking questions to intake-centralists. The most important coordination practices were standard operating procedures and protocol adjustments.