We Will Always Be Better Than a Spreadsheet

• 1 Introduction

  Crime prevention policing strategies in a welfare state context have hitherto focused on social crime prevention and community policing (Gundhus, 2017; Wathne, 2019). In Norway, dedicated crime prevention units traditionally follow up young people by engaging in continuous trust-building efforts and face-to-face dialogue. Recent shifts in policing have increased the use of intelligence to guide managers’ use of resources. This is described in the intelligence doctrine that was implemented in 2014 (Police Directorate, 2014). These changes have overarching implications for the strategies and organization of the Norwegian police (Gundhus et al., 2022a). Previous studies point to an increasingly entrenched risk-based logic of performance management, where crime fighting is central, which is accompanied by the specialization of sections of the police that collaborate with such agencies as the Tax Administration, Labour Inspection Authority, Custom and Labour and Welfare Administration (Dahl et al., 2022).
  However, as in other fields, the institutional logics in policing are changing in ways that are diverse and not coherent (Terpstra & Salet, 2019; Wathne, 2019). This article analyses how police crime preventers respond to these shifts and develop their practice. We explore how tensions between the long-term strategy of relational crime prevention and the more action-oriented approach intrinsic to intelligence-led policing (ILP) play out. Both are future-oriented, although they have different time frames and requirements for what is defined as valid knowledge. We explore empirically how intelligence is employed in police crime prevention efforts in a vulnerable area and the dilemmas that arise, drawing on institutional logic theory, and examine the rejigging of the pressures and complexities of institutional logic in question (Thornton et al., 2012). We are particularly interested in how a risk-based logic affects the practices of the police crime preventers: the type of data that is seen as important and how it is used. The term “practice” is used to refer to established forms or constellations of socially meaningful activities that can be related to cultural beliefs and materiality (Thornton et al., 2012, pp. 128, 130).
  We will first address previous research and argue why the institutional logic perspective is useful for understanding ILP in practice. We will then present ILP as an institutional logic, which seeks more top-down control of frontline policing and relies on technological solutions and a systematic risk-based approach. After presenting our methodology and the cases, we analyse how an intelligence logic is regarded by the intelligence and crime prevention staff. We will look at how, to a certain degree, interest from below creates autonomy in intelligence practice. We will then identify issues crime preventers continue to have with the logic of ILP. We start by exploring issues connected with the collection and recording of data. Then we examine the processing of information and different perceptions of what valid knowledge is. Thirdly, we analyse conflicting views on the sharing of information about risk. We conclude with a discussion of shifts in relations and the power balance within the police organization arising from the demands of data-driven police practice.

• 2 Context and Previous Research

  Preventive policing has become increasingly intertwined with reactive police work, and since 2000, the pre-emptive approach that targets emerging threats and potential criminals has been essential to understanding risk-based police strategies (Ericson & Haggerty, 1997; Johnston, 2000; Maguire, 2000; McCulloch & Wilson, 2016). A defining characteristic of ILP is its orientation towards the future (Police Directorate, 2014; Ratcliffe, 2016). This is also how pre-emptive policing is framed in the literature, where its aim is seen as being to “enable police forces to foreclose criminal events before they emerge” (Egbert & Krasmann, 2020, p. 906). The risk-based and future-oriented aspects of ILP are also pointed out by Sanders et al. (2015, p. 712) in their empirical research on ILP in the Canadian police:

  Efforts to engage in ILP are described as substantially changing the practice of policing, away from an exclusive focus on reactive crime control, toward proactive security, surveillance and risk management.

  Technologies used for data gathering and analytics are seen as crucial in this context (Kaufmann et al., 2019; Ratcliffe, 2016). ILP is therefore often connected to how the datafication of the police increases possibilities for surveillance (Brayne, 2017). The implementation of ILP in Norway is particularly illustrative of the switch of attention to risk-based logic and a more abstract approach to systemic policing (Gundhus et al., 2022a; Terpstra et al., 2019; Zedner, 2007). ILP depends largely on police officers gathering information about persons, places and phenomena in combination with the adoption of information technologies and data analytics. All data used as the knowledge base has to be registered in information technology systems by police officers; in the Norwegian police, Indicia is the main system used for crime intelligence. The main aim of ILP is to reduce the level of crime-surveillance, and collecting data is a means to this end. However, this has consequences for the less standardized and more relational approaches to crime prevention that are associated with a logic of community officers (Terpstra & Salet, 2019) and less concerned with accountability and the recording of data in police databases. The crime prevention logic puts great emphasis on personal relations, trust-building, direct information, craftsmanship, tacit knowledge, involvement, discretion and experience, in line with the logic of community police officers described in the study by Terpstra and Salet (2019, p. 244).
  Research has shown that the ILP approach leads to more data-driven police (Kaufmann, 2018; Marciniak, 2022), and this raises the question of its effect on the Norwegian police. Drawing on empirical data, this article therefore analyses how the shift towards ILP and digital surveillance affects crime prevention work both theoretically and in practice. We conceptualize this shift as the adoption of an intelligence logic. The article is therefore a contribution to the qualitative scholarly tradition that examines in depth how such changes affect police practice (Brayne, 2017; Burcher & Whelan, 2019; Egbert & Leese, 2021; Fyfe et al., 2018; Innes & Sheptycki, 2004; Innes et al., 2005; Sanders & Condon, 2017; Sanders & Hannem, 2012; Sanders & Henderson, 2013; Sanders et al., 2015; Weston et al., 2019). In this research strand, possibilistic thinking emerges as a prominent police institutional logic, and we will investigate how this is expressed in the organization, with a particular focus on how the intelligence logic is regarded by crime preventers.

  2.1 Institutional Logics in Action

  To analyse ILP in practice, we will draw on the institutional logic perspective (Lounsbury & Boxenbaum, 2013), which originated from the critique by Friedland and Alford (1991) of macrostructural approaches in institutional theory that emphasize the primacy of structure over action. As Scott (2008) notes, institutional theory privileges continuity and constraint in social structures and has been criticized for its inability to explain agency (see for instance DiMaggio and Powell’s (1983) theory of structural isomorphism). Friedland and Alford (1991) offered a metatheoretical framework for analysing the interrelationships between institutions, individuals and organizations in social systems. The institutional logic perspective originally focused on bringing macro aspects back into institutional analysis and was renewed by Thornton et al. (2012) to guide empirical research analysis on both the micro and the macro levels and integrate the two. Institutional logics are defined by them as:

  the socially constructed, historical patterns of cultural symbols and material practices, including assumptions, values, and beliefs, by which individuals and organizations provide meaning to their daily activity, organize time and reproduce their lives and experiences (Thornton et al., 2012, p. 2).

  The institutional logics perspective therefore represents a frame of reference that conditions actors’ choices for sense making, the vocabulary they use to motivate action and their sense of self and identity and their socio-material surroundings. The principles, practices, material objects and symbols of multiple institutional orders give different forms to how reasoning takes place and how rationality is perceived and experienced. The growing body of literature on competing institutional logics suggests that, under conditions of institutional complexity, individuals play an important role in shaping organizational outcomes – which may be described as institutional logic in action (Lounsbury & Boxenbaum, 2013). This shift has led to more fine-grained studies of how logics can be sources for resistance and conflicts, how they affect behaviour and both individuals and groups in and across organizations. It has also inspired research on the partial autonomy of actors from logics (Thornton et al., 2012).
  Importantly for this study, it has also encouraged a renewed interest in material dimensions of logics. Recent research has identified a gap in how materiality has been conceived within the institutional logic perceptive, although “the logics must be understood as simultaneously material and symbolic” (Jones et al., 2013). Since we are interested in what the different use made of technologies by different actors can tell us about institutional logics in action, our research will help fill that gap. What will be of importance in this study is, therefore, how the police crime preventers experience, respond to and react to the intelligence logic. By exploring their practice, we will analyse how the two logics in question may clash, exist side by side in competing dynamics or interact in a collaboration based on territorial division (Pollit, 2013, p. 347). People may ignore, comply or resist situations involving competing logics (Pache & Santos, 2013). According to Lounsbury and Boxenbaum (2013, p. 4), there is a “growing recognition that conflicting and overlapping pressures stemming from multiple institutional logics create interpretive and strategic ambiguity for organizational leaders and participants…”. How sense making and identity construction are processes related to the integration of multiple institutional logics in organizational practice in the police has been theorized by Terpstra and Salet (2019), Salet and Terpstra (2022) and Wathne (2019). Interpretive and strategic ambiguity may be a source of stress and conflict, but different logics may also create room for new action and change (Wathne & Solberg, 2021). Research also indicates that plural logics have implications for how organizations respond to institutional complexity. Since we are studying conflicting institutional logics stemming from socio-technical and material objects, how police staff experience the symbolic, material and cognitive aspects of logics will be of importance.

  2.2 The Intelligence Logic in the Norwegian Context

  We will now describe what the intelligence logic means in the Norwegian context, as against the crime prevention logic. As previously described, a number of police studies have shown that ILP entails a shift from traditional prevention to taking precautions against incidents occurring. Digitalizing police practice is important for achieving this goal. This shift was not made known to police employees or citizens but was an unannounced part of the 2015 reform, in which the Police Intelligence Doctrine is central (Police Directorate, 2014). Intelligence was presented as a (institutional) practice that provides situational awareness that is relevant to the entire police task portfolio: the primary tasks of prevention, investigation and preparedness, as well as administrative duties and civil justice. Nonetheless, an important catalyst for the new intelligence logic and risk-based approach was the desire for improved crisis management and emergency policing, following criticism of the police after the 2011 terrorist attack (Christensen et al., 2018). Concentrating resources on such core police tasks as crime fighting was a huge step in that direction.
  Recent research on the Norwegian police system confirms that there have been changes in strategy, which are associated with the centralization of police forces and a political vision of a police role restricted to reducing crime and catching criminals (Dahl et al., 2022; Fyfe et al., 2013, Larsson, 2017). In our analysis of the reform’s planning documents, we have identified four strategic moves that can be related to changes in institutional logic (Gundhus et al., 2022a). The first is the redefinition of the police’s core duty as being crime control; the second is the adoption of the “Police Intelligence Doctrine” as a new knowledge base to support decisions; the third is stronger management through the operation centres; the fourth is firmer management of the police through centralization and standardization. The intelligence doctrine was presented as a management concept for designing digitalization and work processes, making its logic central to the new practices (Police Directorate, 2018). ILP predominates in the managerial and systemic approach to policing, and it is its institutional logic that we will focus on in this article.
  The aim of ILP is to transfer discretionary decisions from individual police officers to management (Gundhus et al., 2022a). Intelligence-based police work is portrayed as a business model offering new functions and roles and a new organization of knowledge work (Police Directorate, 2018). Intelligence is to be produced to improve managers’ decision-making. The intelligence process is expected to include all police employees: those working on investigation, intelligence and prevention, those in functional or geographical units, and those in frontline or management positions. It aims to make the knowledge process more data driven. What is used as its knowledge base will emerge from the data recorded in police systems. Knowledge-based intelligence will therefore change organizational power structures and make decisions more accountable in police data systems. It will also give frontline police less autonomy by moving decisions further up the police hierarchy. Our objective is to examine how these organizational processes and strategies are reacted to by officers. We will look closely at how the intelligence doctrine as an institutional logic operates in practice and is reacted to by the crime prevention officers who rely on a more community-oriented logic emphasizing personal relations, trust-building, direct information, craftsmanship, tacit knowledge, involvement and discretion and experience (Terpstra & Salet, 2019, p. 244). In this article we will conceptualize this as the logic of crime prevention.

• 3 Methodology

  The study is part of a broader project investigating the implementation of the intelligence doctrine and the use of technology by Norwegian police during the reform that took place between 2016 and 2021. The empirical data includes analysis of documents relating to the implementation of ILP, interviews and observations of officers on patrol.
  The interviews were conducted in two phases. The first set of data, including interviews and document analysis, provides insights into the new strategies, the reform and experiences of ILP. Between 2017 and 2018, we conducted six focus group interviews with 24 patrol officers and investigators, along with interviews with 16 key management-level informants working in the Oslo Police District and Police Directorate. The interviews consisted of open-ended questions about digitalization and organizational reform. In the quotations, respondents from the key informant interviews are referred to as K and those from the focus groups as FG.
  The second set of data provides in-depth knowledge of a case. This is the main data used for our analysis in this article and will therefore be given a more thorough description. This case study was motivated by a particular interest in how intelligence is made use of in police crime prevention efforts in vulnerable areas. “Area-based initiatives” are municipal efforts that focus on living conditions in specific deprived housing and urban areas (Atkinson & Zimmermann, 2018). For several years, the Norwegian government has allocated funds for area-based initiatives in Oslo. One of these is the Oslo South initiative, which started in 2007, and is a collaboration between the Municipality of Oslo and the state. Since 2018, the strengthened police effort has also been felt in Oslo South, funded by the Norwegian Ministry of Justice. There had already been various preventive initiatives – between local prevention teams, the police and schools. However, owing to media attention and the leak of a police intelligence report identifying gang-related crime linked to areas in Oslo South, additional police efforts were made. These include an emphasis on building trust and increasing the citizens’ perceived safety by focused patrolling and police presence in designated areas and holding roundtable meetings and interacting at the library with citizens from the community. The thrust of the effort is directed at weakening criminal groups through investigation, incapacitation and control (i.e. knife and weapon inspections in public spaces). Greater collaboration with other actors is central to preventive work with adolescents, which means the police are one of many stakeholders in efforts aimed at young people who are “at risk”.
  Organizationally, the enhanced police effort is managed by a coordination group with leaders from local intelligence, investigation, patrolling and preventive units, together with a central unit in the Oslo Police District that sets the overarching goals for the effort, and makes final decisions regarding what is prioritized and what projects are carried out as part of the increased effort. The task group, consisting of lower-level leaders in local intelligence, investigation, patrol and preventive units, have the main responsibility for operationalizing the goals set for their units and for acting on them through the frontline police officers who work in investigation, intelligence, patrolling and youth crime prevention. The task group also suggests measures to the coordination group and what should be prioritized. If there are criminal prosecutions or noteworthy incidents, they will have a significant impact on the direction of the effort.
  The empirical data consists of seven semi-structured interviews with leaders with different responsibilities in the strengthened police effort and approximately 32 hours of participant observation of four police patrols. The interviews consisted of open-ended questions about their views on intelligence, crime prevention; experiences of governance and autonomy; and thoughts about the way forward for preventive work. The data was gathered between June 2019 and March 2020, and thus relates to a longish period, making it possible to gain knowledge of changes and development work that took place during that time. Many of the questions included in the interviews were also asked while we were out on observation, in the police car or inside the station. The detailed field notes were written during or shortly after the observation. In total, the empirical data consists of 170 pages of transcribed interviews and field notes. In quotations, interviewees identified by the letters TG were part of a task group, while those marked CG were part of a coordination group.
  Codes and themes across the material were identified with the help of Tjora’s (2018) stepwise deductive-inductive method. This consists of data generation, analysis and theory development and is well fitted to small projects (Tjora, 2018). SDI’s inductive approach works from data to theory, gaining deductive feedback, by looking back at the empirical data from the theory (Tjora, 2018). The SDI method aims at conceptual generalization, whereby concepts, typologies or terms are developed whose importance is confirmed by previous research and theory. In this study, it is demonstrated that several concepts and terms that appear in the data material are well documented in previous police research, especially that on ILP and police cultures. The analysis is developed from empirical data through coding and code grouping. The abductive process addresses theory, and concepts are tested against the analysis.
  By analysing underlying assumptions, values and beliefs regarding reform and ILP, this research design provides a promising starting point for capturing changes in institutional logic and responses to it. It makes it possible to look at how changes in policing strategies, methods and tasks unfold and interprets forms or constellations of socially meaningful beliefs and values that can be related to cultural performances, together with shifts in the use of material objects (Thornton et al., 2012, p. 128).
  All data collection was conducted after approval had been obtained from the Norwegian Police Directorate and the Chief of Police in Oslo Police District. The project has also been assessed by the Norwegian Centre for Research Data (NSD), which is responsible for enforcing ethical guidelines in research. All informants received written information prior to the interview and were given the opportunity to ask questions or withdraw from the study: thus informed consent to participate was gained from all informants. The management of personal data was also reported to and approved by NSD.

• 4 Results

  The practice of ILP is based on a doctrine developed by the Police Directorate, which lays down procedural guidelines for how to conduct intelligence work correctly (Police Directorate, 2014). The intelligence process is based on two cycles, an intelligence analysis cycle topped by a management cycle. The Police Directorate (2014) divides the intelligence process into four subphases. The intelligence cycle starts by identifying management’s need for information on an intelligence problem, followed by phase two – “acquisition”, which is carried out by searching in police databases or collecting information from human sources. In phase three, information obtained is processed, analysed and assessed. Finally, intelligence products are delivered to managers, who decide the priorities for future projects (Vestby, 2018).
  However, our interviews indicate that there is a lack of clear guidance on how the process described in the intelligence doctrine should be put into practice. Despite the objective of tighter control, the police officers say they have considerable scope for discretion in the use of intelligence in their daily work. Employees find their own way of using it, following their own judgments. This interviewee makes this clear:

  Yes, I can tell from my own experience … I just had this doctrine put on my desk, and I was told something like, “You can read this if you have time.” And that is all the training I had on this doctrine (TG3, Crime Prevention, 2019).

  Similarly, a frontline prevention officer shakes his head when asked whether training has been provided and whether there is a commitment to following the doctrine. Lack of commitment to the intelligence cycle is also acknowledged by a unit leader, and the intelligence model is perceived as too time-consuming to follow in detail. Variation in practice in adopting the intelligence doctrine seems to prevail, both when registering information in the Indicia criminal intelligence system and when responding to requests to collect information from the information coordinator in the intelligence unit. This suggests that interest and motivation are of great importance in shaping intelligence practice. Conversations during the observations also suggested there are variations in interest in meeting requests to record information the intelligence analysts have said they need.
  Observations of frontline crime preventers suggest they find their own room for action through the digital tools used by police officers to update each other and interact:

  We bring pads and telephones, and you can go straight into the PO [police operative system] and keep up with what’s going on … We have several such systems, like Signal … you can go in such systems when there is an incident and keep up with developments. We have a lot … Lots of tools are available, that actually facilitate very good interaction (CG3, Crime Prevention, 2019).

  These systems also provide more current, new and “fresh” information. Real-time information via smartphones is important for police practice. This is in line with “the good old” actionable information that conforms more closely to the traditional practices of those on patrol out on the streets, but crime preventers also value up-to-date knowledge about relationships from colleagues. The observations also showed how officers communicate through the encrypted Signal app, which enables those who are not on duty to participate in discussions. When asked why Signal is preferred to other ways of sharing information, one ordinary patrol officer said that he believes it improves interaction, because everyone uses it at work and “all sorts of things are shared there” (Observation 4, March 2020). In one example of such use, pictures of some young people on a surveillance video in a shop where there had been a disturbance were shared in a Signal group for a local police crime prevention unit, to see if anyone knew them (Observation 2, November 2019). This immediate sharing of information is therefore also part of building up more long-term knowledge of context, relations and young people.
  The findings indicate huge possibilities for discretion in everyday practice. Moreover, other digital tools than the more time-consuming ILP process are available. We will therefore look further into why the crime prevention logic is at odds with the intelligence logic. What issues do officers have with the intelligence logic?

  4.1 Criminal Data and the Logic of Prevention

  The intelligence logic emphasizing crime fighting is embedded in the digital tools used by the intelligence analysts. All data registered in the Indicia system, must be connected to crime fighting. This also symbolically expresses the objectives of intelligence, because it is directed at identified crime problems, whereas other actors (i.e. the childcare services) have responsibility for what might arise. In this case study, the strategy was to identify which individuals should be labelled as low, medium or high risk, in relation to gang violence. The aim of the intelligence was to draw up lists of high-risk (potential) offenders and to incapacitate the most prolific ones. However, this strategy creates tension between the intelligence staff and crime prevention officers about the police methods or practices used to identify, record and respond to emerging trends in youth crime.
  Intelligence officers gather data and are involved in analysing it and disseminating it throughout the police organization, in the way described in the Police Intelligence Doctrine (Police Directorate, 2014). These practices are systematic and determine how information is gathered and processed. When asked what intelligence ideally is, an intelligence manager described it as follows:

  Yes, it’s simply about being able to sort of create a good enough basis for decision-making to see which measures and preventive strategies are right, and give a picture of the world, and define problem areas that are comprehensive. We must identify where the police should take preventive measures and show what the knowledge base for this is. (CG2, Intelligence, 2019).

  Prevention officers described their work as “a strategy choice, a mind-set, a thought process” and insisted that prevention should not consist solely of measures or initiatives deriving from intelligence products but also include relational non-recorded information. This insistence breaks with the intelligence doctrine in fundamental ways. The crime prevention officer expresses the professional community policing logic that values human beings and argues that police performance also requires professional discretion and contextual judgments (Wathne, 2020).
  From a community police perspective, crime prevention officers need frequent contact and meetings with the public and other collaborators and a detached command centre. They also need to employ soft methods of policing in preventive ways, especially when dealing with children and juveniles (Larsson, 2010). One crime prevention officer expressed concern about the emphasis on crime fighting and having a contingency mindset:

  That’s why I’m probably a bit worried too, when I see there’s a danger of us engaging more in crime fighting than in crime prevention. Because it is very much based on criminal intelligence. And it’s through these glasses that we look at the young people who have already been identified on the basis of some indicators … How does this affect our presence and availability and relationships with the whole local community? We’re going to miss a lot. Including information that could help increase our understanding of these analyses … (TG3, Crime Prevention, 2019).

  During our observations, frontline crime prevention officers expressed the feeling that they are “feeding” a lot of information into the intelligence system, while receiving little in return. In contrast to the intelligence analysts, these officers are primarily expected to provide input to the intelligence system. In the preceding quote we see that the officer is concerned that the crime fighting approach may affect relationships with the local community. His concern is connected to his belief in contextual information, and he values this more than information in the police register. The crime prevention officer sets a high value on relationships and is reluctant to record the information, in contrast to the intelligence staff, who are most concerned with data gathering and information processing.
  At an institutional level, the assumption that preventive practices involve relational and informal non-recorded information breaks with the intelligence doctrine in key ways. The officer quoted above invokes the logic of crime prevention, where the performance of police tasks also requires professional discretion and contextual judgments (Terpstra & Salet, 2019). Central principles are that the police should control society but that, equally, society should control the police, which presupposes that the police are connected to society through relationships. The contrasting values of “objective” and “clean” information provided by the intelligence system and contextual information provided by relationships indicate the competing logics that inform different police practices (Thornton et al., 2012.

  4.2 Valuable Knowledge for Crime Prevention Officers

  In the preceding quotation, the officer warns of the danger of wearing the “risk oriented” glasses that intelligence products can produce. His concerns are connected to his belief in contextual information, where relationships are more important than hits in a police database. The different glasses can be interpreted as different logics based on opposing values, assumptions and beliefs. The officer is aware of the rise of the new approach in the police organization – the new logic he must deal with – and he is concerned. The quote also illustrates the reactive aspect of proactivity: those who are already considered suspects may become even more noticeable to the police because of previous incidents. This was a matter of great concern to the crime prevention officers.
  Intelligence products depend mostly on data recorded in the Indicia register. However, despite managers’ requests for information, what is collected is selective, since only a minority of crime prevention officers register what they know in the system (see also Handegård & Berg, 2020). Officers say they are reluctant to register data, because only information linked to crime and criminal activities is supposed to be registered. This requirement and the format itself restrict what is considered as knowledge:

  Because we are constantly told that what is not in Indicia is not information. And for those of us who are out a lot on the streets with a focus that differs from just the reactive one, there is a lot of information that does not go into Indicia (TG3, Crime Prevention, 2019).

  Crime prevention officers, together with investigators and prosecutors specializing in youth crime, routinely follow up young people reported for criminal incidents four times or more. Searches are regularly carried out in police registers to monitor those with several registered convictions. However, the crime prevention officers also want to identify young people who are not yet registered. As one officer points out:

  Being targeted means taking measures against those on whom there is a lot of information in the registers. If we only go after those young people, we already have a lot of information on them. And then I think it is also difficult to work on recruitment [to criminal groups and gangs], and on general early intervention as well. Because, hopefully, they are probably not registered yet (CG3, Crime Prevention, 2019).

  Identifying individuals early in their criminal career is central to the logic of crime prevention, where relationships, professional discretion and contextual judgments are of importance. The crime prevention officers we interviewed believe that relationships, trust-building, discretion and contextual judgment are of the greatest importance. One officer was not in favour of selecting young people directly from an intelligence system that labels them low-, medium- or high-risk. He explains that only about half of the information they normally use when taking decisions comes from police registers. The other half comes from partners such as schools, outreach groups and the childcare service. The prevention staff make a distinction between knowledge from the police systems and information from collaborative partners. They also take part in a lot of “relational conversations” with young people, which are not entered into the database, to avoid the risk they will be labelled as “criminal”. They often talk with them about things that have nothing to do with criminality. Such interactions show how the practice of the crime prevention officers includes emotional and relational work (Hochschild, 1983).
  Since there is no clear guidance on when data should be entered into the system, a lot of discretion is exercised on this. Officers’ values and the logics they follow when interacting with young people are important, as this officer describes, when reflecting about when to record data in the police register:

  It’s very individual. I think that if there are conversations that address something that has either happened or is going to happen, then it will probably be registered. If it’s more like this person only needed half an hour’s attention from an adult … or something like that, then you don’t do it. I’d estimate, of all the relational conversations we have with young people, we probably register a little less than half (TG3, Crime Prevention, 2019).

  For the officers, personal relations and trust-building efforts are the most important concern, and registering data is also about maintaining trust. This is in line with the institutional logic of community officers, which puts a great deal of emphasis on personal relations, trust-building, direct information, craftsmanship, tacit knowledge, involvement and discretion and experience (Terpstra & Salet, 2019, p. 244). These things are of great importance to crime prevention officers and are perceived as being threatened by the new demands set out in the intelligence doctrine. When analysing the interviews with the prevention staff, we find a repeated view that there is a need for more documentation and follow-up on what is done by the prevention unit. A central finding is that the prevention staff clearly have a concern about the unintended consequences of the logic underpinning these new practices and the datafication processes. It is perceived that registering information about young people in the criminal intelligence database has a net-widening effect, since it results in more of them being labelled as “risky”.
  The prevention officers portray Indicia as a reactive system. The feedback loops reinforce the selection of those that are already registered. The decision on what to register in Indicia is experienced as a dilemma. Although officers describe themselves as respecting the intelligence system requirements for information gathering, there is also a desire for a different ICT tool that can be used for prevention within a logic of crime prevention. TG3 also talks about how crime prevention officers have been criticized for not submitting all the knowledge they have to the criminal intelligence system but insists they are trying to get better at submitting the most necessary information. Officers provide personalized and local information that results from a more contextualized view of knowledge. If a person has red trousers, for example, they do not think it is important to enter this into the system, but they might be criticized for not doing so:

  But I think our contribution is personal knowledge, local knowledge … and I think this will help fill in the picture as well, because we sit in interagency meetings. So, when the police suddenly think something is a problem, we can step in and say, “None of the other agencies think this is a problem,” because I don’t think we’re good enough at using other people’s and other agencies’ analyses (TG3, Crime Prevention, 2019).

  The frontline police officers on patrol, who are often the ones who entered the information, say in interviews that intelligence is perceived as consisting of known conditions and events. It offers little of what they would like to know more about, such as what is expected to happen in the immediate and more distant future. They miss additional information about upcoming risks that they are not aware of. On the other hand, crime prevention officers talk about the need for relational knowledge:

  I really believe in knowing people … we will always be better than an Excel document. But what intelligence is good at is using research, and that’s where we have to learn! After all, [the operational analyst programme] is for gathering information … to learn from research what are the risk factors for a child. That it’s not just older brothers, that it’s not just dropping out of school: it’s all the other things that [the system] picks up on through parenting situations and … yes. The stuff they feed the matrices with…. Both parts can learn from each other, and we can either use those matrices, double check if they’re relevant to the young people we work with, or if there is something we have missed…. But I want, when I ring a doorbell, to be ringing it because I know that someone who knows that child is worried about them … not because an Excel document I printed out at work says so (TG4, Crime Prevention, 2019).

  In the preceding quote, the officer emphasizes how intelligence may enrich the crime preventers’ work and at the same time relates this to the crime prevention logic that focuses on human being, informal relations and the value of face-to-face interaction and trust. However, we sense an element of ambiguity in that he also acknowledges the intelligence system’s ability to use research, but with one big difference: it is the child that is of central concern, rather than the possibility that the young person may pose a risk to society. The substantial difference between the two concerns is linked to different institutional logics: concern for the young person is connected to a logic of crime prevention, while the concern about risk to society connects to that of the intelligence system.
  The worst-case scenario for crime prevention officers is to be made into cogs in an intelligence process. There have been signs of this happening, which have met with resistance from officers. The belief that intelligence should be seen as something that enriches their knowledge, rather than being the solution, is often expressed in the interviews. As was found in the study of Terpstra and Salet (2019), a logic of community policing is prevalent. In this case, there is a strong determination not to be governed too much by the “intelligence cycles”. Prevention is portrayed as a way of thinking, not an instrument to be made use of.
  We will now turn to the last theme of the analysis: the different views about how intelligence should be shared in the organization.

  4.3 More High Policing Than Sharing Knowledge

  The aim of producing intelligence is to increase police officers’ understanding of situations and to support managers in the decision-making process. The sharing of knowledge starts long before the data becomes visible in the data system, since it first has to be observed and coded (Flyverbom & Schade, 2022). Inevitably, the intelligence staff have an important intermediary role in the coding of the data. They contribute what they themselves call a “cool gaze”, as CG2 describes:

  It’s a more systematic approach than just what might be experienced [by police officers themselves], or basic assumptions in the organization. So, it’s taking a “cooler look” at something that many people attach a lot of emotion to. But it can contribute to doing the opposite if it is not done well. It reflects the skill the organization has (CG2, Intelligence, 2019).

  The term “a cooler gaze” refers to the more systematic processing of data. Interviews with an intelligence manager (K2) reveal there is a reluctance to share intelligence information in the police organization, both internally and externally. The intelligence staff is cautious, owing to risks of leaks to the media and a less mature way of handling intelligence reports elsewhere in the organization:

  What we have had are leaks. We make our living by producing reports and the information must not be made public, because that also helps to escalate situations. For example, the intelligence report on criminal gangs … which was leaked, without any context, without any description of what it actually says about things. And perhaps without senior management involvement, without anyone used to commenting on such information and dealing with reports of that sort. And without knowledge that this report only looks at a small part of the big picture of the reality in South Oslo (CG2, Intelligence, 2019).

  All reports are stored in Indicia, so one way of providing information is by making one’s colleagues aware of reports. To reach more people, individual reports can be posted on the police network, and joint information meetings (paroles) can be held both across police departments and units and in geographical operating units, such as police stations, and in the individual sections. Verbal briefings, known as “briefs”, are also given.
  What intelligence should provide to different police departments is also discussed. The systematic “cool gaze” is also made use of to assess who needs to know what and how much they need to know. One intelligence manager says that he has concerns about sharing too much information or “going all out”, and stresses, among other things, how intelligence is communicated in meetings:

  and before we throw the information out to everyone, we need to know that the organization has the ability to manage it. Information sharing must be linked to governance. Because not everyone can chase this person or the other (CG2, Intelligence, 2019).

  This is demonstrated by the example of how information initially intended to increase situational awareness can have the opposite effect: if it is disseminated at meetings, it may single out certain people for particular attention. If guidance is not provided along with the information, it may lead to disproportionate resources being deployed against a target group:

  Now we have two brothers who have been mentioned in all the meetings and it’s like … where is Osama bin Laden? … It’s getting like the hunting Osama bin Laden feeling…. Every patrol drives by where they live and hang out. Yes, it must somehow be connected to the management (CG2, Intelligence, 2019).

  The interviewee admits that intelligence should be disseminated more widely in the long term, but for the moment, he thinks it is too soon. Managers have the huge responsibility of directing frontline staff, he says. It is the responsibility of management to distinguish between intelligence and the measures that should be taken. On one side are those who produce and disseminate intelligence and on the other those who receive and consume what has been assessed as needing to be disseminated, and the two sides might not agree.

  This can be illustrated by opinions expressed by the patrols. They do not consider the information they receive to be particularly useful: they want more concrete information about persons and places. This emerges clearly both from interviews and from observations of those working on incidents and in preventive patrols:

  We don’t get knowledge we can use on patrol. Who are we going to catch? Who are we supposed to talk to? Who are we going to do this with, and who are we not going to do that with, and so on. The intelligence is not there, the information that we need is not available…. That is the biggest problem (CG1, Operational Manager, 2019).

  They require more of what they call “useful intelligence” to be shared – for instance, operational intelligence about the risk level of organized crime (Observation, November 2019). By contrast, the crime prevention officers want more strategic intelligence about long-term trends. TG4, for instance, thinks too much information is shared at daily meetings about individuals with a capacity for violence, who are going to be controlled or apprehended (TG4, Crime Prevention Officer, December 2019). The same interviewee stresses the importance of giving short overviews of the intelligence reports to his subordinates, since it seems that it is very much up to the individual whether they read full reports:

  I always forward them, but I also always write, “If you don’t want to read it, please read my short summary, and take this with you at least” (TG4, Crime Prevention, 2019).

  When asked what could improve the integration of intelligence processes within the organization, one informant mentioned more information sharing on status, goals and priorities, and, not least, on why something has been prioritized (TG5, Investigator, 2019). If this is not provided, motivation suffers.
  In focus groups, police officers also expressed concern about the secrecy surrounding the origin of information, which makes it difficult to explain to citizens why they are being stopped for ID checks. The secrecy is connected to the intelligence logic, with its “need to know” criterion for the police on the street. By contrast with the crime prevention logic, the intelligence logic emphasizes information and activities behind the scenes, rather than direct personal relations; risk minimization rather than trust-building; standardization rather than craftsmanship, involvement and discretion. ILP may therefore be regarded as rooted in Taylor’s management principles: greater division of labour, with some actors in the organization having the whole picture and others doing specialized tasks (Gundhus et al., 2022a). Police practice rooted in the intelligence doctrine means that some officers feel like retrievers of “bits and pieces” of data – knowledge becomes an abstraction, and they are constantly trying to sift and retrieve relevant info. Fears are also voiced about patrols being turned into confirmation machines that just collect data ordered by the specialist units and are unable to observe the environment more inductively:

  You have to collect as much data as possible. In a way, it becomes part of a generalist’s task to report into the intelligence registers (K, 2017, Oslo Police District).

  Unlike intelligence produced by visibility and the registration of possible pre-crime, a lot of knowledge about prevention is therefore supposed to be hidden from the patrols and not something they can use in interactions with citizens. Policing that is more intelligence and data driven is therefore at odds with prevention officers’ reliance on less tangible data, tacit knowledge, trust and relations, as described here:

  We want a system that is different from the reactive system … because there is a lot of important information in prevention officers’ heads. We are told that what is not in the police system is not information. We need some kind of preventive system, which can add to the intelligence products. Of course, as prevention officers, we must be good at deciding what should go into such a system and in that way contribute to good analyses. And I think the most important thing is that we must understand the importance of presence, relationships, trust … and make room for them – that it is also part of this. Understand that it doesn’t come at the expense of … the crime fighting approach when we’re just out on the street, hunting and findings things out and … then fighting (A3, Crime Prevention, 2019).

  The prevention officer would also like to share information but not information about “young people at risk of offending”. These different and conflicting responses to the intelligence logic therefore provide an area for understanding how digitalization and datafication affect police knowledge production. The crime prevention logic clearly emphasizes other ways of getting to know the young people than by taking information from computers:

  I want a local police station that knows the population well enough not to need to use computer systems to find who to work with. [The operational data analysis system] probably provides a more neutral entry into working with young people. And then I have a dream … we will be able to work together to treat everyone equally, because we know them … and not just because an Excel document says something (TG4, Crime Prevention, 2019).

  The preceding quotation summarizes our findings on the contested views of the intelligence logic. It also reveals how material objects such as data entry in the computer gives power to new logics that inevitable shape practice. The need for data to be visible in the police registers for it to be accepted as valid knowledge opens the way for the datafication of police work. This is in contrast to crime prevention officers’ approaches, which emphasize less tangible data, trust and relationships.

• 5 Intelligence Logic in Action

  By drawing theoretically on the institutional logic perspective, our study shows that intelligence logic diverges from the logic of crime prevention as regards what is considered valid data. The intelligence doctrine aims to make police officers’ data gathering more visible in the police systems. However, officers are required to collect data for superiors or information coordinators. This means frontline officers collect it for reasons that are unknown to them – it is data that is required for decision-making at the senior levels of the police organization. For crime prevention officers, these demands for recording data are particularly unwelcome, since the logic behind them conflicts with their own, where personal relations and creation of trust with target groups are paramount.
  In general, intelligence is seen as a support for managers, who make the decisions, and the intention is that the level where discretion is exercised should move up the organization, thus reducing frontline officers’ autonomy. The shifts in the power structure created by the intelligence logic, together with digitalization and the new management approach, narrow and constrain the type of knowledge that is seen as valid. The demand for data to be made more visible in the system is perceived as problematic and exemplifies the conflict between the traditional logic of crime prevention and the intelligence logic.
  As is described in international studies, the primary task of policing is increasingly seen as doing whatever is necessary to prevent crime from occurring, which may include stopping or disrupting it (Hebenton & Seddon, 2009, p. 346). This goes along with an emphasis on threats, the anticipation of potential criminals and crisis management: crime is seen as risks to be pre-empted. However, this goes against the crime preventers’ professional vision and values, although there is a move towards more documentation and greater standardization. The way work is now organized clearly indicates an adoption of more control-oriented intelligence, which is criticized by the crime prevention officers we have interviewed. As Brodeur (2007) points out, new police methods remove the need for proximity and face-to-face interaction and make for greater abstractness in policing (Terpstra et al., 2019). This applies even to crime prevention. The police become less visible and focus their attention on “things the public do not know, but which affect them”, as is described in the interviews. Formerly separate police areas, such as intelligence and prevention, are merged when intelligence provides the basis for decisions. Patrol officers in a focus group told us the intelligence products resulting from this are perceived as being manufactured for senior management. The intelligence reports they receive from specialist units are seen as abstract and as lacking both the timeliness and the long-term perspective they need. They do not see the value of either decontextualized knowledge or data put into standardized templates and forms, and crime prevention officers also find them of little use.
  Analysis of the production of knowledge offers contradictory findings about the types of information that are of importance for preventive police practice. On one side, intelligence is framed as being based on the visibility of pre-crime, while, on the other, much of the knowledge required for prevention remains invisible, owing to the logic of community collaboration and relations. Possibilistic thinking is contested and taken on board in different ways by intelligence, prevention and patrol officers’ primary institutional logic. There are tensions between the long-term strategy of the preventive approach and that of ILP that sets out to be action-oriented. Patrols do need concrete data on persons, locations, addresses and possible violent incidents, and real-time data is particularly valuable.
  The intelligence logic is not all-encompassing. We find a lot of discretionary assessment, and less quantified risk assessment (Brayne, 2017)¹x See Sarah Brayne (2017, p. 985) for inspiration on the framework. She identifies five shifts in the degree of displacement of traditional police practices towards big data surveillance.. Although data is used to predict, there is only moderate systematic use of the automatic alert information system. The merging of previously separate data systems only involves police data, which limits the spread of surveillance into a wider range of institutions. The main instance of data surveillance found in this study is the lower threshold for inclusion in law enforcement databases, which now include individuals who have had no direct contact with the police. This is what the crime prevention officers are most critical of: they do not think data from collaborators should be put into the police system. Our findings show that the demand to register more and more data and put it into the system is an issue discussed within the police organization, where different institutional logics collide. Crime prevention officers are particularly against the demands that increasing amounts of data are entered into the crime intelligence system. They are worried that intelligence will eclipse prevention, rather than enhancing and supporting relevant data. Intelligence staff too are careful about what type of risk information is communicated. The policing of “dangerous” crimes and the “criminal other” pulls in the direction of militarization and combating crime, but there is resistance to this within the police organization.

• 6 Concluding Remarks

  ILP as an institutional logic aims at more top-down control of frontline policing, relying on technological solutions and a systematic risk-based approach. However, our analysis shows that despite this, the intelligence logic is not all-encompassing. The logic is viewed differently by intelligence and by community police officers. We have shown how, to a certain extent, interest from below creates autonomy in intelligence practice. We have also identified what issues crime prevention officers have regarding the logic of intelligence. The first is that crime preventers have reservations about the type of information that should be entered into the system. There are also indications of differences of view about how data should be processed by the criminal intelligence system and different views about what is perceived as valid knowledge. The last topic analysed is the disagreement about sharing risk information. The intelligence managers have decided not to share risk information widely in the organization, and this has consequences for the contextual information needed for frontline patrols’ information gathering and their direct interaction with citizens. To conclude, our findings indicate there have been shifts in relations and the power balance within the police organization, owing to the demands of data-driven police practice.
  The power structure created by digitalization and by the intelligence logic and new management approach changes what type of knowledge is seen as valid. Thus, crime prevention is increasingly embedded in a risk-based logic, but the institutional logics in policing are changing in ways that are diverse and incoherent, so there still remains room for discretion and resistance (Terpstra & Salet, 2019; Wathne, 2019). In our analysis we show how this is manifested through the gaps in ILP and how this may direct the focus onto people or areas, based on the information officers have decided to obtain. This means there is more information about the same people or areas. ILP may thus equate in practice to an amplification of police interests.
  Risk leads to an increase in security and makes the question of the legitimate limits of security pertinent. What matters is to pre-empt crime and threats, and the desired outcome is to avoid, disrupt or stop threats by being anticipatory and precautionary. These maximal tendencies of risk-based proactivity are not smoothly integrated into the logics of the police, reducing the potential for ever more invasive policing. However, further research is needed to know more about how the digital logics shape the way the police practise ILP.

References

• Atkinson, R., & Zimmermann, K. (2018). Area-based initiatives – a facilitator for participatory governance? In I. H. Heinelt & S. Münch (Eds.), Handbook on participatory governance. Edward Elgar.

• Brayne, S. (2017). Big data surveillance: The case of policing. American Sociological Review, 82(5), 977-1008. https://doi.org/10.1177/000312241772586

• Brodeur, J.-P. (2007). High and low policing in post-9/11 times. Policing: A Journal of Policy and Practice, 1(1), 25-37. https://doi.org/10.1093/police/pam002

• Burcher, M., & Whelan, C. (2019). Intelligence-led policing in practice: Reflections from intelligence analysts. Police Quarterly, 22(2), 139-160. https://doi.org/10.1177/1098611118796890

• Christensen T., Lægreid, P., & Rykkja, L. (2018). Reforming the Norwegian police between structure and culture: Community police or emergency police. Public Policy and Administration, 33(3), 241-259. https://doi.org/10.1177/0952076717709523

• Dahl, J. Y., Fyfe, N. R., Gundhus, H. O. I., Larsson, P., Skjevrak, p. E., Runhovde, S. R., & Vestby, A. (2022). Old, new, borrowed and blue – Shifts in modern policing. British Journal of Criminology, 62(4), 931-947. https://doi.org/10.1093/bjc/azab085

• DiMaggio, p. J., & Powell, W. W. (1983). The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields. American Sociological Review, 48(2), 147–160. https://doi.org/10.2307/2095101

• Egbert, S., & Leese, M. (2021). Criminal futures: Predictive policing and everyday police work. Taylor & Francis.

• Egbert, S., & Krasmann, S. (2020). Predictive policing: Not yet, but soon preemptive?, Policing and Society, 30(8), 905-919. https://doi.org/10.1080/10439463.2019.1611821

• Ericson, R. V., & Haggerty, K. D. (1997). Policing the risk society. University of Toronto Press.

• Flyverbom, M., & Schade, F. (2022). Coded visions: Datafied visibilities and the production of political futures. In A. M. Brighenti (Ed.), The new politics of visibility: Spaces, actors, practices and technologies in the visible (pp. 32-43). Intellect Books.

• Friedland, R., & Alford R. R. (1991) Bringing society back in: Symbols, practices, and institutional contradictions. In W. W. Powell & p. J. DiMaggio (Eds.), The new institutionalism in organizational analysis (pp. 232-263). University of Chicago Press.

• Fyfe, N. R., Gundhus, H. I., & Vrist Rønn, K. (2018). Introduction. In N. R. Fyfe, H. I. Gundhus, & K. Vrist Rønn (Eds.), Moral issues in intelligence-led policing (pp. 1-22). Routledge.

• Fyfe, N. R., Terpstra, J., & Tops, p. (Eds.) (2013). Centralizing forces: Comparative perspectives on contemporary police reform in Northern and Western Europe. Eleven Publishing.

• Gundhus, H. O. I. (2017). Discretion as an obstacle: Police culture, change, and governance in a Norwegian context. Policing: A Journal of Policy and Practice, 11(3), 258-272. https://doi.org/10.1093/police/pax012

• Gundhus, H. O. I., Talberg, N., & Wathne, C. T. (2022a). From discretion to standardization: Digitalization of the police organization. International Journal of Police Science & Management, 24(1), 27-41. https://doi.org/10.1177/14613557211036554

• Gundhus, H. I. O., Talberg, N., & Wathne, C. T. (2022b). Investigating at a distance: Abstractness in detective work. In J. Terpstra, R. Salet, & N. R. Fyfe (Eds.), The Abstract police: Critical reflections on contemporary change in police organisations. Eleven International Publishing.

• Handegård, T. L. & Berg, C. R. (2020). Kunnskapsbasert politiarbeid – kunnskap til å stole på? Nordisk Politiforskning, 7(1), 39-60. https://doi.org/10.18261/issn.2703-7045-2020-01-04

• Hebenton, B., & Seddon, T. (2009). From dangerousness to precaution: Managing sexual and violent offenders in an insecure and uncertain age. British Journal of Criminology, 49(3), 343-362. https://doi.org/10.1093/bjc/azn085

• Hochschild, A. R. (1983). The managed heart: Commercialization of human feeling. University of California Press.

• Innes, M., Fielding, N., & Cope, N. (2005). The appliance of science? The theory and practice of crime intelligence analysis. British Journal of Criminology, 45(1), 39-57. https://doi.org/10.1093/bjc/azh053

• Innes, M., & Sheptycki, J. W. (2004). From detection to disruption: Intelligence and the changing logic of police crime control in the United Kingdom. International Criminal Justice Review, 14(1), 1-24. https://doi.org/10.1177/105756770401400101

• Johnston, L. (2000). Policing Britain: Risk, security and governance. Longman.

• Jones, C., Boxenbaum, E., & Anthony, C. (2013). The immateriality of material practices in Institutional logics. In Lounsbury, M. & Boxenbaum, E. (Eds.), Institutional logics in action, part A (pp. 51-75). Emerald Group Publishing Limited.

• Kaufmann, M. (2018). The co-construction of crime predictions. Dynamics between digital data, software and human beings. In N. R. Fyfe, H. O. I. Gundhus, & K. V. Rønn (Eds.), Moral issues in intelligence-led policing (pp. 143-160). Routledge.

• Kaufmann, M., Egbert, S., & Leese, M. (2019) Predictive policing and the politics of patterns. British Journal of Criminology, 59(3), 674-692. https://doi.org/10.1093/bjc/azy060

• Larsson, P (2010). Ideology as a cover-up: Community policing in Norway. Policing Europe, Journal of Police Studies, 1(0), 39-48.

• Larsson, P (2017) From integration to contact. A community policing reform? Nordisk Politiforskning, 4(2), 170-186. https://doi.org/10.18261/issn.1894-8693-2017-02-05

• Lounsbury, M., & Boxenbaum, E. (2013). Institutional logics in action. In M. Lounsbury & E. Boxenbaum (Ed.), Institutional logics in action, part A (pp. 3-22). Emerald Group Publishing Limited.

• Maguire, M. (2000). Policing by risks and targets: Some dimensions and implications of intelligence-led crime control. Policing and Society, 9(4), 315-336. https://doi.org/10.1080/10439463.2000.9964821

• Marciniak, D. (2022). Algorithmic policing: An exploratory study of the algorithmically mediated construction of individual risk in a UK police force. Policing and Society, Online first. https://doi.org/10.1080/10439463.2022.2144305

• McCulloch, J., & Wilson, D. (2016). Pre-crime: Pre-emption, precaution and the future. Routledge.

• Pache, A.-C., & Santos, F. (2013), Embedded in hybrid contexts: How individuals in organizations respond to competing institutional logics. In M. Lounsbury & E. Boxenbaum (Eds.), Institutional logics in action, part B (Research in the sociology of organizations, Vol. 39 Part B) (pp. 3-35). Emerald Group Publishing Limited.

• Police Directorate. (2014). Etterretningsdoktrine for politiet. Versjon 1.0. Politidirektoratet.

• Police Directorate. (2018) Plan for implementering av etterretning og etterretningsbasert politiarbeid 2017-2010. Politidirektoratet.

• Pollitt, C. (2013). The logics of performance management. Evaluation, 19(4), 346-363. https://doi.org/10.1177/1356389013505040

• Ratcliffe, J. H. (2016). Intelligence-led policing. Routledge.

• Salet, R. & Terpstra, J. B. (2022). Conflicting institutional logics in the control of crime and disorder. The double strategy of administrative and criminal law. Criminology & Criminal Justice, Online first. https://doi.org/10.1177/17488958221112061

• Sanders, C. B., & Condon, C. (2017). Crime analysis and cognitive effects: the practice of policing through flows of data. Global Crime, 18(3), 237-255. https://doi.org/10.1080/17440572.2017.1323637

• Sanders, C. B., & Hannem, S. (2012). Policing the “Risky”: Technology and surveillance in everyday patrol work. Canadian Review of Sociology, 49(4), 389-410. https://doi.org/10.1111/j.1755-618X.2012.01300.x

• Sanders, C. B., & Henderson, S. (2013). Police ‘empires’ and information technologies: uncovering material and organisational barriers to information sharing in Canadian police services. Policing and Society, 23(2), 243-260. https://doi.org/10.1080/10439463.2012.703196

• Sanders, C. B., Weston, C., & Schott, N. (2015). Police innovations, ‘Secret Squirrels’ & accountability: Empirically examining the integration of intelligence-led policing in Canada. British Journal of Criminology, 55(4), 711-729. https://doi.org/10.1093/bjc/azv008

• Scott, W. R. (2008). Institutions and Organizations: Ideas and Interests. Sage.

• Terpstra, J., Fyfe, N., & Salet, R. (2019). The abstract police: A conceptual exploration of unintended changes in police organisations. The Police Journal: Theory, Practice and Principles, 92(4), 339-359. https://doi.org/10.1177/0032258X18817999

• Terpstra, J. B., & Salet, R. (2019). The contested community police officer. An ongoing conflict between different institutional logics. International Journal of Police Science and Management, 21(4), 244-253. https://doi.org/10.1177/1461355719889465

• Thornton, p. H., Ocasio, W., & Lounsbury, M. (2012) The institutional logics perspective. A new approach to culture, structure and process. Oxford University Press.

• Tjora, A. (2018). Qualitative research as stepwise-deductive induction. Routledge.

• Vestby, A. (2018). Policy-making without politics. Overstating objectivity in intelligence-led policing. In N. R. Fyfe, H. O. I. Gundhus, & K. V. Rønn (Eds.), Moral issues in intelligence-led policing (pp. 255-282). Routledge.

• Wathne, C. T. (2019). Welfare Professionals in transformation: The case of police officers in Norway. In H. Hviid & E. Falkum (Eds), Work and wellbeing in the Nordics – Critical perspectives on the world’s best working life (pp. 219-220). Routledge.

• Wathne, C. T. (2020). New public management and the police profession at play. Criminal Justice Ethics, 39(1), 1-22. https://doi.org/10.1080/0731129X.2020.1746106

• Wathne, C. T., & Solberg, A. G. (2021). To virksomme tiltak for kjønnsbalanse – i spennet mellom institusjonelle logikker. Tidsskrift for Arbejdsliv, 23(4), 43-60. https://doi.org/10.7146/tfa.v23i4.130011

• Weston, C., Bennett-Moses, B., & Sanders, C. (2019). The changing role of the law enforcement analyst: Clarifying core competencies for analysts and supervisors through empirical research. Policing and Society. 30(5), 532-547. https://doi.org/10.1080/10439463.2018.1564751

• Zedner, L. (2007). Pre-crime and post criminology?. Theoretical Criminology, 11(2), 261-281. https://doi.org/10.1177/1362480607075851

Noten

• * Acknowledgements: This work was supported by the Norwegian Research Council grant: 313626 ‘Algorithmic Governance and Cultures of Policing: Comparative Perspectives from Norway, India, Brazil, Russia, and South Africa’ and Nordforsk grant: 106245 ‘Critical perspectives of predictive policing’. We are grateful to the anonymous reviewer who helped us clarify some points and encouraged us to develop others.

• 1 See Sarah Brayne (2017, p. 985) for inspiration on the framework. She identifies five shifts in the degree of displacement of traditional police practices towards big data surveillance.