• 1 Introduction

  In the mid-1960s, the United States was convulsed with urban protests, as African Americans and their supporters expressed outrage towards systemic racial inequality and discriminatory police practices (Walker, 2016). In response, President Lyndon B. Johnson formed the President’s Commission on Law Enforcement and Administration of Justice (1965) and the National Advisory Commission on Civil Disorders (1967). Their respective reports, the Challenge of Crime in a Free Society (1967) and The Report of the National Advisory Commission on Civil Disorders (1968), are powerful reminders of the enduring challenges to democratic policing and accountability (Kelling, 1999, p. 15).
  Both reports acknowledged the tremendous personal discretion of police officers at the street-level in choosing how to respond to any given incident. The problem was how best to control discretion to eliminate (or at least minimize) misuses of police authority, and to ensure quality in decision-making. In response to this challenge, the respective Commissions advocated for the development of clear and enforceable policies. This administrative rulemaking approach remains the dominant model for governing police discretion in the United States today (Friedman, 2017; Walker, 1993), and it is a key feature of police organization and professionalism in Europe and elsewhere (Cockcroft & de Maillard, 2021).
  Just as they were over fifty years ago, the arbitrary, intrusive and racially discriminatory uses of police discretion remain of great concern to the public and police reformers (Camp & Heatherton, 2016; Purnell, 2021). The most reckless and egregious examples of police officers abusing their authority, such as excessive force, deserve to be central to any efforts to reform the police (Friedman, 2017; Thacher, 2016, p. 535), but our focus is on considering discretion in more routine police-public encounters, with consequential but less alarming outcomes. As one of the most powerful and visible systems of state power, people care very much about how they are treated when they engage with the police, no matter how commonplace or mundane the encounter. These contacts often involve low-key problems or “human troubles,” such as officers mediating interpersonal conflicts, managing rebellious youths, responding to noise complaints, or protecting the rights of people to walk or drive in neighbourhoods (Black, 1971, p. 1090; Lum et al., 2021). While these encounters might not make the news, civilians are comfortable judging the quality of a patrol officer’s actions, even when a situation is resolved informally (as is most often the case) (Mastrofski, 1996). In developing our vision for reform, we continue to wrestle with the full extent of patrol officers’ discretionary leeway, and with some of the challenges of formulating an alternative approach to governing police discretion to supplement the administrative rulemaking model.
  In this article, we outline the key dimensions of a craft learning model we have discussed elsewhere (Willis & Toronjo, 2022). This is centred on first-line supervisors using body-worn cameras (BWCs) to apply craft-based performance standards derived from patrol officers’ experiences to a learning process of reflection-in-action. Here we build on this earlier work and structure this article as follows: first, we seek to advance a conceptual understanding of this model by situating its key elements in relation to Egon Bittner’s observations on two distinct mechanisms of police organization and control: legality and workmanship; second, we illustrate the largely overlooked potential of BWCs for advancing craft knowledge and skills by drawing on more recent in-depth, case studies we have conducted on the use of BWCs in three U.S. police departments (Koen et al., 2019, 2021; Willis, 2022);¹x Two of these case studies were conducted in small police agencies (50-100 sworn) located in the mid-Atlantic region of the United States. One had implemented BWCs in 2012, the other in 2016. The third agency was located in the mid-West, and had between 250 and 300 sworn officers. It implemented body-worn cameras in 2014. We promised these agencies that we would not reveal their identity. One case study relied on interviews only, the other two combined interviews with observations. and third, we address three challenges to what we propose and consider how these might be mitigated or overcome: (1) resistance from the police culture; (2) the limited role of first-line supervisors; and (3) the current lack of community participation in guiding decision-making on the front lines. Although our focus is on policing in the United States, the themes we address on discretion and its regulation are applicable to police organization, practice and reform more generally.
  We do not present new research findings here. Rather our overarching purpose is to develop the value, logic and nature of the craft learning model and encourage others to consider alternative models or mechanisms for potentially enhancing patrol officers’ capacity for both practical and moral reasoning in the kinds of situationally complex and indeterminate encounters that characterize everyday patrol work.

• 2 Two Models for Governing Police Discretion – Workmanship and Legality

  For Bittner, high-quality patrol work depends on the capacity of patrol officers to make good decisions relatively quickly in situations that are often complex, dynamic and uncertain, and under circumstances where people might be upset or uncooperative. This is a much more demanding decision-making environment than the public, and even the police themselves, generally acknowledge (Bittner, 1990, p. 8). Trusting the police to respond to “critically serious social and human problems” (Bittner, 1990, p. 10) in ways that were not arbitrary or unprincipled presents police organizations with “two relatively distinct adequacy problems”, which call for “two correspondingly distinct control mechanisms” (Bittner, 1983, p. 2).
  The first adequacy problem is legality, which seeks to restrict police actions through general bureaucratic rules and procedures. Compliance is then enforced through regulatory supervision, often involving the threat, or practice, of sanctioning those who violate these rules. Legality comports with the image of the police officer as a “snappy, low-level, soldier-bureaucrat” mechanically following orders, rather than a skilful and autonomous professional, such as a doctor, social worker or teacher (Bittner, 1990, p. 260). Bittner lamented the “overwhelming preference” for regulatory supervision in policing, which was based on the mistaken assumption that “police work, especially the work of uniformed patrol, is a low-grade occupation and that not much can be expected of the persons engaged in it” (Bittner, 1983, p. 5). This reliance on ensuring that patrol officers complied with a broad “scheme of strict internal regulation” (Bittner, 1990, p. 253) for fear of punishment was not easily applied to addressing the kinds of complex and serious problems common to everyday police work (Bittner, 1983, p. 4). Moreover, the focus of post-action reviews was generally limited to relatively rare critical incidents, and typically narrowed to reconstructing a precise account of what happened to determine rule and procedural violations, and to impose an appropriate punishment (Bittner, 1983, p. 4; Worden & Dole, 2019, p. 47). Such a formal review process was best suited to establishing minimal levels of acceptable performance, or judgments that the “assessed person did nothing wrong” (Bittner, 1983, pp. 5-6; Thacher, 2020).
  Bittner also recognized a second adequacy problem that bedevils street-level performance, or workmanship. By this he meant an officer’s ability to make decisions by calling “upon resources of knowledge, skill, and judgment to meet and master the unexpected within one’s sphere of competence” (Bittner, 1983, pp. 2-3). Workmanship is synonymous with the police craft and is revealed when patrol officers recognize some of their fellow officers as master craftspeople. These are those select officers, who have developed a level of sophistication or artistry in their work, which serves as an example to others (Bayley & Bittner, 1984).
  The control mechanism corresponding to workmanship is accountability in the form of judgments by one’s peers, whose professional experience and knowledge best qualify them to assess their colleagues’ work. In addition, the community or its representatives are entitled to judge patrol officer performance, because police work is being done on the community’s behalf, and police are ultimately accountable to the public (Bittner, 1983, p. 4).
  According to Bittner, police organizations take few deliberate or systematic actions to raise expectations around the police craft (Bittner, 1983, p. 4). And yet, as others observe, at its best, craft involves

  making good arrests, deescalating crises, investigating crimes, using coercion and language effectively, abiding by the law and protecting individual rights, developing knowledge of the community, and imparting a sense of fairness to one’s actions. (Mastrofski, 1988, p. 65)

  Nor do police organizations try to articulate criteria for evaluating officers’ performance as craftspeople as a basis for learning and improvement. In comparison with administrative rulemaking and its emphasis on explicit rules, the domain of workmanship rests on an “inchoate lore” (Bittner, 1990, p. 262).
  Nothing much seems to have changed since Bittner made his keen observations. Police reformers continue to champion the promise of specific rules to “properly constrain and direct decision-making” and to espouse the virtues of stricter accountability mechanisms (Friedman, 2017; Kelling, 1999; Ponomarenko, 2019, p. 4; Worden & Dole, 2019, p. 42). Much less attention has been paid to workmanship, or how to improve officer discretion by pooling the reservoir of craft knowledge and officer experiences and using it in ways that might help advance more skilful and judicious decision-making. In the next section, we propose what such a model might look like.

• 3 Sketching a Craft Learning Model

  3.1 Developing Craft-Based Standards

  In some of our earlier research we explored the possibility of deriving craft-based standards from patrol officers’ experiences that could be used to judge performance. We summarize this research here (including its findings), to provide a context for some of the challenges we anticipate to a craft-based learning model. Readers interested in more specific details about the interview sample, methods and results can consult Willis and Toronjo (2022).
  A major challenge to efforts to improve street-level performance is the assumption that “standards of excellence” cannot be “fully formulated in advance of the occasions of use”, due to the extraordinarily varied situations police can be required to handle (Bittner, 1983, p. 3). Police practitioners push this claim even further by asserting that every situation is different and that their actions cannot be assessed fully or fairly by those not present at the actual scene. Any form of “Monday Morning Quarterbacking” by others (the tendency to criticize with the benefit of hindsight) is generally scorned and seen as intrusive, arbitrary and unhelpful (Bayley & Bittner, 1984).
  However, Bittner recognized that even if street-level encounters did not lend themselves to strict and narrow evaluative criteria, it was still possible to identify broad operational goals worth achieving, which could then provide a valuable basis for assessing decision-making (Bayley & Bittner, 1984). In short, while it might not be possible to distil these goals into simple rules that could be automatically applied to a given situation, that did not mean that performance dimensions could not be identified ex ante for guiding discretion along more “principled routes” (Thacher, 2016, p. 533). After all, any evaluation of police discretion (and a potential framework for reviewing and exploring its use) first requires the possibility of identifying defensible criteria for its assessment (Mastrofski, 2004). Consequently, based on their own research experiences, Bayley and Bittner (1984) suggested that police officers generally sought five operational goals in police-public encounters.²x These goals are meeting departmental norms, containing violence and controlling disorder, preventing crime, avoiding physical injury to themselves and avoiding provoking the public into angry retaliation that threatens their careers (Bayley & Bittner, 1984). These were subsequently developed by Mastrofski (1996, p. 213) into seven performance standards: violence containment and disorder control, problem diagnosis, problem resolution, citizen response to police, people’s safety, lawfulness of police response and economy in police response.
  While police scholars have long recognized the possibility of using police officers’ collective experiences to improve discretion, we noticed that few attempts had been made to advance this insight empirically, or to develop the logic, substance and uses of a craft-based learning model, especially in relation to the dominant rulemaking alternative (Bayley & Bittner, 1984; Bittner, 1983; Mastrofski, 1996; Thacher, 2008; Willis & Mastrofski, 2017).
  Thus, in our earlier research, we were curious whether patrol officers would identify similar performance standards, and we tested this directly (Willis & Toronjo, 2022). Of course, we expected officers to think differently from one another about what comprises skilled police work in any specific situation. At the same time, we wondered if there was a sufficient level of consensus among patrol officers on what constituted skilled performance to act as a useful tool for learning and channelling discretion in beneficial ways. We could then use these standards to develop a model for improving police discretion, particularly in more perplexing or problematic encounters where it was not easy to determine how best to respond. We also wanted to explore how this framework could help advance police officer understanding of the essential normative or moral elements of their decisions. These value judgments, powerfully shaped by situational contexts, tend to be implicit and rarely exposed to careful examination. And yet disagreements over normative considerations regarding what constitutes “doing the right thing” on the front lines often lie at the heart of public controversies over police actions (Mastrofski, 1996, 2018).
  With this in mind, we conducted in-depth interviews with thirty-eight patrol officers in one medium-sized police agency in the United States (around 300 sworn officers). Our questions around police performance were part of a much broader study examining individual officer perspectives on patrol officer performance and police culture (Willis & Mastrofski, 2017). As part of the interview, we asked respondents to react to the actions of two officers in a video-taped incident involving an actual dispute between two neighbours in an apartment building. We chose the clip because it was a good depiction of the kind of fairly common situations that police confront. Moreover, because law and policy seemed to offer little guidance, this scenario was a high-discretion situation. This increased the likelihood that our interviewees would have to rely on their own experiences in sharing their own choices about good decision-making.
  One neighbour was upset and complained to the police that the neighbour in the adjacent apartment was slamming their door too loud. The second neighbour retaliated by pounding on the complainant’s door with a heavy metal weight and making threats. We forewarned respondents we were going to stop the video clip at a key decision point as the encounter was still evolving, and ask them to assess these officers’ decisions and what they would have done differently. This was an attempt to replicate the evolving nature of the police decision-making environment, so that respondents would not have the benefit of knowing how the encounter ended in making their assessment.³x For more details on the methodology, please see Willis and Toronjo (2022).
  Using a combination of deductive and inductive analyses, we derived seven craft-based standards: accountability, lawfulness, problem diagnosis, repair of harm, economy, safety and order, and fairness. Accountability is concerned with officers holding people responsible for their actions, especially when their behaviour harms others (Mastrofski, 2018). In the neighbour dispute, the officers in the clip tried to persuade the woman accused of damaging her neighbour’s door that she was blameworthy for her actions, that she should not make threats and damage property, and that she was responsible for the consequences.
  Another key dimension of police performance to emerge was the lawfulness of an officer’s actions. A fundamental element of democracy is the public’s expectation that the police obey the law. In interpreting whether and how the law applies, officers must know its provisions and assess what the evidence warrants, including whether it supports the serious deprivation of a person’s liberty in the form of an arrest. Some key legal questions raised in the clip were whether the officers did anything not permitted by law, or omitted something required by it (Mastrofski, 2018). Our respondents disagreed whether the neighbour dispute was best characterized as a criminal or civil matter, but the majority did not feel an arrest was appropriate.
  Mitigating or resolving a problem depends in no small measure on an officer’s ability to make an accurate diagnosis of the nature of the problem (Muir, 1977). In the case of the neighbour dispute, we anticipated several questions, many of which were raised by our respondents: What prompted such a hostile reaction from the second neighbour to the door slamming? What exactly had the neighbours done to try to resolve this issue informally, and how were those efforts received? Was this the first time the police had been called to this address? Was the problem escalating? Despite the crucial importance of problem identification for the selection and implementation of a response, police agencies pay little attention to assessing the problem-setting dimension of police performance (Mastrofski, 1996).
  In their assessment, our respondents criticized the officers in the clip for minimizing the seriousness of the dispute, for failing to do more to improve the immediate situation before leaving, and for doing too little to ameliorate the risk that the problem might re-occur in the future. Whether patrol officers strive for shorter or longer term outcomes, the police craft seems to afford considerable importance to their efforts to repair the harm in a given situation. Mastrofski (1999, p. 2) refers to this performance element as competence:

  The public judges police competence primarily in terms of the tangible things they can readily observe. They do not use crime statistics or other so-called outcome measures. They watch the officer at work and make judgments about his or her ability to get the job done.

  The expectation that police provide a “definitive resolution” to almost every encounter must be anchored in the reality of the resource constraints within which patrol officers must operate (Mastrofski, 1996, p. 215). When it comes to a performance standard for economy, officers’ time is one of their most precious resources, but currently police departments offer little guidance on how much time and effort should be spent on a given encounter, such as this dispute. Manager, supervisors and police officers themselves pay much more attention to safety and order at the scene. The potential for danger and injury are well-documented features of the police culture (Paoline, 2003). Failure to maintain safety and order also undermines an officer’s ability to accomplish any of their other objectives.
  The final performance standard to emerge from our interviews was fairness of the process or procedural justice. Even if police officers cannot deliver the outcomes different parties desire, people care very much about how they are treated (Tyler, 2004). It is important that those involved have a chance to tell their side of the story (or the opportunity to have a voice), that the officer listens attentively and that the officer’s decisions are not biased towards one party or the other. People also expect the police to show that they care about whatever problem a person is experiencing, and to make a good faith effort to resolve it.
  These seven craft-based performance standards suggested to us that despite the “unavoidably complex” realities of police practice (Thacher, 2019, p. 280), it was possible to identify criteria applicable to assessing performance in police-public encounters, in this case a neighbour dispute. Although some of these standards might be applicable more generally, how they apply would likely vary across different types of encounters distinguished by common features (e.g. disputes between strangers, ongoing disputes between friends, traffic stops). For example, few would argue that it is inappropriate for police officers to voice their concerns, but in a situation with a high threat of violence, it might well be appropriate for an officer to initially issue commands authoritatively that could be perceived as brusque or dictatorial. It would likely make sense for departments to identify and define which types of encounters to prioritize for the development of their performance dimensions. We imagine a process similar to the one we described earlier carried out by police agencies, but one designed around a small seminar setting. The focus of the discussions would be on providing plenty of opportunity for the sharing of different viewpoints, careful deliberation and for building consensus, even if it is unlikely that everyone would agree on the best course of action (Mastrofski, 1996).
  Importantly, unlike the more constraining properties of “if-then provisions in rules and procedures” (Engel & Worden, 2003, p. 132), standards allow for greater flexibility in making judgments about the goals that practitioners should strive to meet when making decisions (Bittner, 1983). Professionals set standards to evaluate their own performance in specific contexts because they want their “behavior to measure up to the standards as much as possible” (Cohen & Feldberg, 1991, p. 40). Similarly, standards for judging police responses to everyday encounters can help distinguish higher from lesser degrees of ability (Bayley & Bittner, 1984, p. 51). Thus, there is an aspirational quality to craft-based standards grounded in the realities of actual practice that is not easily captured by abstract and general rules. In the same way that craftspeople are distinguished by their virtuoso skills, creativity and desire to “do a job well for its own sake”, craft standards can help distinguish superior performers from those who merely meet their minimal obligations (Sennett, 2009, p. 9).
  While the focus of our interview data in this earlier work was to explore empirically the feasibility of deriving craft-based performance standards, we went one step further by envisioning a model for how these standards might be integrated into a more systematic learning process for potentially improving practical decision-making and moral reasoning. Such a model, designed around meaningful reflection on one’s actions, is consistent with an emerging interest in making policing a more critically reflective practice and in taking advantage of BWC technology for this purpose (Charles, 2000; Christopher, 2015; Phelps et al., 2018).

  3.2 Applying Craft-Based Standards through Reflection-in-Action

  The rapid spread of BWCs across police agencies in the United States and worldwide over the last few years means that, unlike patrol work in the past, many police supervisors now have access to high-quality audio and video recordings (White & Malm, 2020). These recordings can be used to help appraise how officers behave during their more routine encounters with the public (Lum et al., 2020; Nix et al., 2020). Body-camera footage can capture a tremendous amount of detail, including variations in body language and tone of voice, and it can be played slowly, stopped and even replayed to help officers recall what they were thinking in the actual moment and context of deciding.
  Unfortunately, research suggests that BWCs have been harnessed tightly to the features of the legalist or administrative model, namely to try and strengthen compliance with bureaucratic rules, to resolve citizen complaints, and to identify and punish police misconduct (Koen et al., 2019; Koen & Mathna, 2019; Lum et al., 2020; Willis, 2022). In comparison, police have shown less interest in exploring the potential of BWC technology as a more constructive learning tool for capturing and improving craft knowledge and skills (Willis & Mastrofski, 2017).
  First articulated by John Dewey (1933) and later developed by Donald Schön (1983), reflection-in-action is part of an epistemology of practice whose point of departure is “the competence and artistry already embedded in skillful practice” (Schön, 1983, p. xi). Key to this artistry is the ability of some practitioners, especially experts, to “reflect in the midst of action without interrupting it”, particularly when dealing with situations characterized by “uncertainty, disorder, and indeterminacy” (Dewey, 1933; Schön, 1983, p. 16, 1987, p. 26). Since routine responses are less likely to apply to these problematic situations, reflection-in-action may lead to a practitioner changing the initial framing of a problem, reconsidering underlying assumptions, rethinking a particular strategy, and conducting “on-the-spot” experiments, to adapt to changing circumstances, and to pursue an approach likely to be more successful (Schön, 1983, pp. 62-63). Given this background, using video footage to enable practitioners to think actively and carefully on how they reflected-in-action during an encounter with the public, particularly around how they adjusted in those moments which were unexpected or particularly challenging, could play an important role in the development of practical skills and new knowledge, which then could inform future actions (Schön, 1987, pp. 31).
  Because first-line supervisors work closely with individual officers and instruct them on how to perform their craft, they are well situated to periodically review BWC footage of what initially might have appeared to seem a fairly routine encounter, but that an officer then experienced as problematic (Engel & Worden, 2003; Muir, 1977). It is common for BWC policies to grant supervisors permission to access BWC footage to conduct individual performance appraisals, but studies suggest this is rare in practice (White et al., 2019). Since video footage can provide a close approximation to the “surrounding world of actual life-related practices”, it can serve as part of a process of open-ended enquiry designed to help expose and interrogate the implicit decision-making processes essential to professional practice (Yanow & Tsoukas, 2009, p. 1342). This focus on thoughtful reflection and learning in the context of actual incidents as they evolve distinguishes craft-based learning reviews from an administrative rulemaking model. A craft learning model might bring attention to what policy requires, but its assessment of an officer’s choices goes well beyond this. Furthermore, it does not insist that officers justify their responses retroactively in an environment carrying a high risk of censure or punishment.
  Evaluations of what constitutes desirable exercises of police authority can obviously be complicated and controversial, particularly when law and policy offer little guidance. How then would a supervisor help an officer understand whether their performance was substandard, merely adequate or truly exemplary? As David Thacher observes, there is no simple answer to this question “for Dewey’s alternative model of rationality is fundamentally open-ended”. It is worth noting that this challenge is not unique to policing. Professionals in other fields, such as medicine, engineering and law, may struggle to reach a “firm and uncontroversial evaluation of a particular decision” (Thacher, 2020, p. 762).
  From the perspective of the craft learning model, this observation opens up rather than closes down discussions during the review process. The key task of supervisory review is to help patrol officers reflect on their actions, articulate how they tried to meet or exceed key performance standards and consider areas for improvement. This same philosophy applies to potential conflicts between the standards (i.e. the tension between economy and the amount of effort and time needed to try and resolve a problem). Supervisors can help clarify expectations, but keeping the tensions in mind (rather than denying them) is essential for informing decision-making and for helping avoid complacency (Marx, 2006, p. 278).
  Although it cannot offer a definitive guide to action, reflection-in-action structured around performance standards could offer a number of benefits. In narrowing the scope for evaluating what should matter in handling an encounter to a practical set of relevant criteria, standards may help officers consider more intentionally the most relevant factors when making decisions and feel less overwhelmed by available choices. Making these priorities explicit also helps reduce the “vulnerability” of patrol officers to “arbitrary performance evaluations after the fact” (Mastrofski, 1996, p. 223), which can undermine morale and exacerbate tensions between management and street cops (Reuss-Ianni & Ianni, 1983). Moreover, conversations around specific standards could help identify important gaps in an officer’s knowledge and skills that might require additional training, and clarify to what extent an officer’s choices and goals matched with the police department’s own priorities. Should these be misaligned, attempts could be made to understand where major differences lie and to address them. Giving patrol officers’ ownership in defining and developing performance criteria, rather than imposing regulations from above, might also increase acceptance of this model for reviewing and guiding discretion.
  Second, learning about officers’ attempts at spontaneous improvisation and the understandings behind them might help supervisors understand how an officer framed an initial problem, and how this might have shaped subsequent choices, perhaps even unduly limiting their scope. It could also lead to the identification of innovative solutions that could be disseminated to other officers to become part of a broader repertoire of “examples, images, understandings, and actions” for handling encounters in the future (Schön, 1987, p. 66). In doing so, this kind of review process could help collect knowledge about surprising situations, where more routine responses fell short of expectations. This knowledge could then accumulate and become a valuable resource for organizational learning (Schön, 1987, p. 66). For example, in the case of our neighbour dispute, one officer suggested putting padding around the door, so it would no longer slam shut.
  Finally, a non-punitive reflective review process could contribute to patrol officers’ normative understanding of important social values, as they are encouraged to consider how the meaning and application of these values influenced their choices (Thacher, 2008). Albert Reiss once noted that, as a profession, policing requires officers to make decisions affecting the “fate of people” that were not just tactical but involved making “moral judgments” (Reiss, 1971, p. 21). The implications of a patrol officer’s decisions for an individual’s liberty or dignity, for example, may become taken-for-granted over time, and thus lead to complacency about the ideals a police officer should aspire to and the obligations they should obey (Thacher, 2006). Reflection-in-action could benefit officers by helping them identify, clarify and revise their understanding of varied and ambiguous values, and develop ideas about acceptable trade-offs between them (Thacher, 2001). For example, even if it were legally permissible to make an arrest in the neighbour dispute, could it be normatively justified in the face of alternative solutions with fewer costs and more benefits to the second neighbour? Arrests can be invasive, humiliating and worsen a suspect’s employment and financial prospects, and should not be disproportionate to the goals they are intended to serve (Harmon, 2016). More “mindful” and “self-critical” interpretations and justifications (key features of other professions such as medicine) could then support more principled decision-making in future encounters (Epstein, 1999, p. 833). Notably the requirement to explore the normative dimensions of one’s decision-making in the context of an actual encounter is different to simply adhering to a more abstract code of ethics (Kleinig, 1996). The latter bears close resemblance to the logic and limitations of the administrative model, and while it can be difficult to justify ethical judgments based on abstract principles, “we often have more confidence in concrete judgments about what should be done in a particular case” (Thacher, 2008, p. 55).

• 4 Three Challenges to a Craft-Based Learning Model

  We recognize there are significant challenges to the model we propose, especially at a time when public distrust of the police is so high, and there is a strong desire to shrink the police “footprint”, but there are reasons for optimism (Thacher, 2022). A recent study showed that even brief meetings between officers and sergeants for talking about street encounters could lead to “more measured responses to later incidents” (Owens, 2020, p. 1). Those officers who met regularly with their supervisors were much less likely to resolve incidents with an arrest than members of a control group, and to be involved in incidents where force was used. The authors attributed this change to discussions prompting officers to reflect on their thought processes and actions. These meetings helped “slow down” officers’ thinking on the street, so that they paid closer attention to the reasons for, and possible implications of, their actions (Owens, 2020, p. 2). There is also evidence that patrol officers can play a central role as agents of police reform (Toch, 2008). Still, we anticipate at least three major challenges to a craft learning model, including (1) resistance from the police culture; (2) the limited role of first-line supervisors; and (3) the current lack of community participation in guiding police discretion. We identify these challenges based on our interviews, the literature on police reform and discretion, and our own broader experiences conducting research with the police. We briefly examine the nature of these problems before suggesting some possible means for their abatement.

  4.1 Resistance from the Traditional Police Culture

  In response to pressures from the police occupational and organizational environments, including uncertainty, danger and internal scrutiny, police officers can develop similar attitudes and beliefs (Paoline, 2003). These include a preoccupation with critical incidents and the dangerousness of police work, concerns over a punitive system of bureaucratic accountability, suspiciousness towards civilians’ motives and complaints, the importance of officer autonomy in deciding how and when to exercise their authority, and an emphasis on in-group loyalty, particularly in relation to management oversight (Loftus, 2010). While the police culture is not monolithic (Muir, 1977), these elements are often described as resistant to reform efforts (Lum & Koper, 2017; Mastrofski & Willis, 2010).
  When it comes to BWCs, research suggests that police officers have embraced BWCs more fully when they are used as a tool to address civilian complaints, and to review critical incidents (Koen et al., 2019; Lum et al., 2020; Owens & Finn, 2017; Sandhu, 2017). It appears that these specific functions of BWCs align closely with features of the traditional police culture, as they can help protect officers from false or malicious accusations of wrongdoing, and contribute to officer and public safety.
  In comparison, officers appear to be much less enthusiastic about using BWCs in ways that threaten their traditional autonomy, including efforts to review their street-level decisions through more proactive supervision. In a review of BWC policies submitted by 304 U.S. police agencies which received federal BWC awards between FY 2015 and FY 2018, “over 90% of 2016 and 2017 funded departments allow[ed] supervisors to review line officers’ BWC footage” as part of a broader assessment of patrol officer performance (White & Malm, 2020, p. 68). However, based on field studies we have conducted in three separate police agencies of different sizes and location, we concluded that first-line supervisors rarely availed themselves of this opportunity.
  For example, in one small agency in the mid-Atlantic, consistent with the police culture’s commitment to patrol officers’ sense of autonomy, sergeants largely limited their review of BWC footage to addressing civilian complaints or, on occasion, to reviewing critical incidents (Van Maanen, 1983, p. 288). Our other two other agencies demonstrated a similar pattern. Patrol officers viewed BWCs as a tool that would protect them from frivolous and false complaints, and that would help show the legal elements justifying their use of coercive force. Moreover, BWC policies were unambiguous in mandating the conditions for supervisory review of footage. Supervisors eschewed the idea of reviewing BWC footage beyond policy guidelines to understand how their officers navigated routine citizen encounters. Footage would instead be reviewed on an “as-needed” basis in response to citizen complaints, uses-of-force, and/or when a supervisor suspected an officer of some form of wrongdoing.
  This was particularly pronounced at the mid-Western agency, as supervisors and patrol officers were suspicious of the command staff and their intentions for implementing BWCs, viewing the technology as a “gotcha mechanism” with which to conduct “witch hunts” to sanction trivial violations of the department’s policies. To supervisors, reviewing footage beyond policy criteria would be an inefficient use of their limited time and would undermine the trust between them and their patrol officers. The only exception to this rule was when supervisors sought to point out an officer safety issue, most often within the context of a high-intensity situation.
  In comparison, BWCs had not been embraced by officers or their superiors as a tool for proactively learning about the many day-to-day “particularities of police work as it is experienced by serving officers and by analyzing that experience and making it available to future police officers” (Bayley & Bittner, 1984, p. 36). Rather, BWCs were closely associated with the traditional and punitive bureaucratic accountability process, and with officers’ preoccupation with the physical dangers of the job.
  Moreover, the high value a department’s occupational culture may place on patrol officers’ loyalty to one another, and on mutual respect, subverts interest in using footage more frequently to enhance training on the craft of police work. In this environment, BWC review is resisted because it may invite unfair criticism and foster ill will. We observed this at all three of our police agencies. The culture of police departments is very unlike hospitals, which have a learning culture that encourages an honest review of mistakes through peer evaluation and feedback. These qualities are on display at a hospital’s regular Morbidity and Mortality Conferences, where doctors assemble to discuss their struggles and errors with their peers to identify areas for improvement and promote professionalism (Gawande, 2002). We asked some officers across the three agencies if they could envision something similar in their own department, but all baulked at this idea.
  Fostering habits of mind around enquiry, self-criticism and intuitive expertise in an organizational environment characterized by internal control systems that are formalized, strongly hierarchical and “essentially negative” presents a formidable challenge, but a reasonable place to start would be at the top (Weisburd et al., 2003, p. 446). Research suggests that strong leadership can influence organizational change by communicating a compelling vision and by implementing strategies, such as training, that support this vision (Fernandez & Rainey, 2006; Weisburd et al., 2003). Fortunately, the craft learning model’s involvement of patrol officers in the creation of standards might increase the likelihood of success, given that widespread participation in any change process can lower internal resistance (Toch, 2008). Building support among key external stakeholders would also be essential to building support and sustaining interest, as would providing sufficient resources, such as freeing up time for first-line supervisors to conduct regular performance reviews (Boyne, 2003).

  4.2 Limited Role of First-Line Supervisors

  Because they work closely with patrol officers, first-line supervisors are usually the most important organizational members in a patrol officer’s professional life, acting as both a “boss” and a “critical audience” of their performance (Van Maanen, 19783, p. 279). However, how supervisors give direction to their officers and hold them accountable is complicated by the high degree of “mutual dependence and reciprocity” that governs these supervisor-subordinate relationships, and by the varied mechanisms supervisors can use for influencing an officer’s behaviour (Van Maanen, 1983, pp. 279-280). Some supervisors may enforce bureaucratic rules, while others may broker compliance by protecting their officers from department discipline in return for their being productive and staying out of trouble (Engel & Worden, 2003). Not only does this give rise to different styles of supervisory practice that might not be well suited to a craft learning model (Engel, 2001), but supervisors are often reluctant to infringe upon their officers’ decision-making autonomy, which officers may guard jealously (Van Maanen, 1983, p. 288).
  The latter was clearly revealed at two of our research sites, where the departments’ BWC policies restricted reviews of officer performance to civilian complaints or training issues, and where first-line supervisors showed little interest in expanding their use of cameras for broader learning purposes. In fact, we asked patrol officers directly to put themselves in the role of a first-line supervisor and imagine how they might use BWC footage to help improve street-level performance. Half of the respondents said they might review footage under limited circumstances, such as in response to specific complaints (e.g. an officer’s rude demeanour). The other half said they might use them for training, but they provided several conditions. These included obtaining an officer’s consent, avoiding Monday Morning Quarterbacking and being careful not to criticize in a way that undermined morale.
  At a third site in the mid-West, command staff were univocal about not using BWCs as a training tool. The police chief said, “Maybe some time in the future it can be considered, but now body cameras are not used in training situations.” Supervisors had discretionary latitude to restrict their use of BWCs footage for training to high-intensity situations. During roll call, after obtaining the consent from the officer, supervisors would play BWC footage to all the officers present and point out certain “officer survival issues” (e.g. an officer moving into the line of friendly fire).
  Other police scholars have suggested the importance of supervisors behaving as coaches, and similarly the craft learning model encourages police agencies to consider first-line supervisors in this role rather than as simple rule enforcers (Goldstein, 1990; Muir, 1977). As coaches, they would rely less on their formal authority, and more on their experience as skilled professionals to support officers in being reflective, flexible and creative as a basis for solving problems and making sound judgments. Muir (1977) suggested the potential influence of this mode of supervision on officer behaviour by having sergeants take opportunities to teach their officers lessons on the importance of empathy and on understanding the tragedy of the human condition, and on the moral complexities of using coercion. Becoming a good coach is challenging, and it is difficult exposing oneself to “scrutiny and fault-finding” as part of a process of self-improvement, but there is some evidence from other criminal justice contexts suggesting supervisors can be taught how to coach, which could be instructive for policing (Gawande, 2011, p. 13).
  In probation, for example, the nature of officer-supervisor relationships is similar to that of police sergeants and patrol officers. Probation supervisors vary in the degree to which their supervisory style aligns with coaching (Kras et al., 2017). They perform their (mostly administrative) duties in a risk-adverse environment. And like all supervisors, once promoted they no longer have the “sanctioned luxury of acceptable dependence” on a superior, which can foster the belief that they must always be problem solvers (Kadushin & Harkness, 2014, p. 219). Despite this, the current dominant approach to working with probationers requires officers to use a variety of professional skills that can only be developed via experiential learning and continuous coaching (Bonta & Andrews, 2016; Toronjo, 2019a). Research suggests coaching probation officers is a promising practice, which can help improve both officer practice and probationer outcomes (Bonta & Andrews, 2016; Lowenkamp et al., 2012; Robinson et al., 2012; Smith et al., 2012). These coaching models in probation take inspiration from reflective practice models used in other ‘helping professions’ such as nursing, social work and teaching, and aim to train front-line supervisors as coaches (Kadushin & Harkness, 2014; Rex & Hosking, 2013; Toronjo, 2019a).
  Two in-depth qualitative studies on these kinds of models illustrate the promise and peril of supervisor coaching in a risk-adverse organization. Toronjo (2019b) interviewed twenty-one supervisors in four U.S. agencies across three states, and Ainslie et al. (2022) interviewed thirty-three front-line officers and twenty-eight supervisors in the United Kingdom. Both studies found that supervisors who were engaged in a reflective practice model of coaching appreciated the spirit and underlying principles of the model. Supervisors believed there was a need for coaching to help officers improve their professional practice, and to help officers navigate the emotional demands of the work. Supervisors also believed coaching practices were appropriate for their position (Ainslie et al., 2022; Toronjo, 2020).
  Despite the perceived acceptability and appropriateness of coaching practices, barriers to practice included time, supervisor confidence and the nature of some supervisor-officer interpersonal relationships. Taken together, these issues relate to the ‘ecology of supervision’ (Harvey et al., 2016). Coaching with reflective practice requires supervisors to navigate the tension between a risk management milieu and a pro-learning culture that encourages experimentation and treats mistakes as learning opportunities. The continued dominance of bureaucratic processes in probation agencies illustrates the need for a concomitant shift in occupational culture and values with the rise of reflective practice. One in which the field of probation might be poised to make and from which policing might learn (Toronjo, 2019b).

  4.3 Lack of Community Participation

  Finally, a core feature of legality or administrative rulemaking is injecting more democracy into policing processes by soliciting civilian feedback on police policies (Davis, 1975). In addition to the democratic principles of transparency and community participation, a virtue of this approach is that policies (and the police practices they are intended to govern) will be improved “because those policies that cannot withstand public scrutiny will be minimized or eliminated” (Klockars, 1988, p. 108). Similarly, our vision for a craft learning model includes attempts to strengthen democratic policing by using BWC footage as an opportunity to provide community members with the ability to react to video clips of an officer’s performance. This could be part of the process for distinguishing better from worse police work, and for identifying police responses consistent with community values and expectations. Currently, police departments do little to solicit input tailored towards guiding street-level discretion. While “generally, it is a good idea to involve citizens” in police operations, this raises complex issues about who should provide the input, the community’s role and potential risks to the police department (Kelling, 1999, p. 42).
  Public participation in local policing and government tends to be low, and it would be easy for police chiefs to select motivated community stakeholders to participate in the craft learning model, who are merely supporters of the department and its leadership. But this would give an agency a distorted view of what the public desires from their police (Worden & Dole, 2019). The same problem often hinders community policing initiatives, as certain groups (e.g. white middle-class property owners) are more likely to participate in police-community meetings. These groups may not reflect the views of others, or a broader consensus on neighbourhood values (Mastrofski, 1988, 2018). Thus, police leaders should consider soliciting diverse perspectives to comment on police officer performance, including from those who come from disadvantaged neighbourhoods, lack political influence and are more likely to experience the brunt of police power. It might also be worth considering involving other professionals, such as domestic violence advocates, who could provide useful insights that might not have occurred to the police (Mastrofski, 1996). This approach would go far beyond police agencies selectively releasing BWC footage to the public to help alleviate public outcries of police wrongdoing and without any corresponding mechanisms for sharing diverse perspectives and addressing disagreements.
  Given how people’s judgments of police behaviour are influenced by their general impressions of the police, specific past experiences, and a broader set of psychological and cultural dispositions (Kahan et al., 2009; Willis & Toronjo, 2019), it is unlikely that a clear consensus will emerge on the appropriateness of an officer’s actions. Civilians tend to judge police performance differently from police practitioners (Carlson & Sutton, 1988; Worden & McLean, 2017), and they often disagree in their assessments of police behaviour (Waddington et al., 2015). However, the purpose of the community input would not be consensus but having the opportunity to “voice their reactions to what they see and hear” (Mastrofski, 1988, p. 66). In this advisory role, feedback from community members could help shape the creation of standards and how they are applied.
  An obvious concern of the craft learning model is that an over-reliance on practitioner knowledge and experience will give rise to practices that are potentially harmful or biased against certain groups. The kind of community input we envision could help increase police accountability and mitigate this problem. While examples of troubling or ineffective applications of the police craft will undoubtedly remain, creating a system for critically reviewing officer performance to help increase transparency, foster debate, solicit community input and improve external accountability would be a large step forward in monitoring and managing everyday uses of police discretion (Mastrofski, 1996). In addition, the identification and clarification of where major disagreements lie over an officer’s choices could be a useful means for submitting these “different assumptions to empirical test and possibly to provide a starting point for negotiating an agreement over at least some differences” (Mastrofski, 1996, p. 233). The same logic of using scientific evidence to help resolve disagreements between the community and the police could apply to differences between police leaders and patrol officers on the most effective and appropriate tactics for a given situation (Lum & Koper, 2017).
  Publicly acknowledging the highly discretionary nature of patrol officer decision-making is a risky proposition for police departments. It could easily make policing more difficult by undermining the police “image of impartiality and objectivity” and exposing officers to the challenges of civilians who feel that they are being treated unequally or unfairly compared to others (Klockars, 1988, p. 107). At the same time, ignoring the fact that discretion lies at the core of police work perpetuates the current unsatisfactory status quo, and it does little to recognize the degree of professionalism that high-quality decision-making in complex situations demands.

• 5 Conclusion

  The underlying assumption of this article is that street-level policing is a situationally complex task that requires patrol officers to make difficult judgments about what to do and how to do it. Under these conditions, a key question then becomes, how do you help officers make more effective and reasoned decisions? Rules, laws, science and values may all provide useful guidance, but ultimately, patrol officers must identify and interpret how these apply to a given context, before choosing a course of action that seems most appropriate. What choices the officer makes have important implications, especially for those who are subject to those decisions. In this respect, policing is very much like other professions, such as law and medicine, but these higher status professions do much more to recognize honestly and openly the moral and practical uncertainties and struggles confronted by decision-makers. They also do more to recognize how hard-won experience gives rise to professional standards of excellence whose “evaluative components are learned through engaging in and with the practice, not through thinking about them” (Yanow & Tsoukas, 2009, p. 1344).
  Despite these distinguishing features of everyday police work, police reformers continue to focus on developing rules and disciplinary processes to help restrict and standardize police behaviour, rather than considering learning alternatives for distinguishing good from bad choices. In no small measure, the success of these alternatives will depend on their ability to account for the unavoidably adaptive, intuitive and non-routine nature of police discretion. Moreover, police researchers and policymakers often gravitate towards department-level evaluations of police programmes and initiatives and the promise of formal training regimens for police reform (Willis & Toronjo, forthcoming), and less to considering ways to improve street-level practices by “discovering those good qualities of police work that already exist in the skills of individual practitioners” (Bittner, 1990, pp. 146-147, emphasis in original). The prospects of guidance and control for the craft learning model we envision here is consistent with a post-bureaucratic focus on public service organizations learning from their limitations to reform their practices (Sabel & Simon, 2017), and in contrast to the current dominant paradigm of increased administrative rulemaking and bureaucratization. The current policing crisis might be an opportunity for reform advocates to pursue bolder and more innovative responses to the challenges of doing good police work and to foster higher levels of police professionalism that such work demands.

References

• Ainslie, S., Fowler, A., Phillips, J. & Westaby, C. (2022). ‘A nice idea but ….’: Implementing a reflective supervision model in the National Probation Service in England and Wales. Reflective Practice, 23(5), 1-14. https://doi.org/10.1080/14623943.2022.2066075.

• Bayley, D. H. & Bittner, E. (1984). Learning the skills of policing. Law & Contemporary Problems, 47(4), 35-59. https://doi.org/10.2307/1191686.

• Bittner, E. (1983). Legality and workmanship: Introduction to control in the police organization. In M. Punch (Red.), Control in the police organization. MIT Press.

• Bittner, E. (1990). Aspects of police work. Northeastern University Press.

• Black, D. (1971). The social organization of arrest. Stanford Law Review, 23(6), 1090. https://doi.org/10.2307/1227728.

• Bonta, J. & Andrews, D. A. (2016). The psychology of criminal conduct. Routledge.

• Boyne, G. A. (2003). Sources of public service improvement: A critical review and research agenda. Journal of Public Administration Research and Theory, 13(3), 367-394. https://doi.org/10.1093/jopart/mug027.

• Camp, J. T. & Heatherton, C. (2016). Policing the planet: Why the policing crisis led to Black Lives Matter. Verso.

• Carlson, H. M. & Sutton, M. S. (1988). A multimethod approach to community evaluation of police performance. Journal of Criminal Justice, 9, 227-234. https://doi.org/10.1016/0047-2352(81)90072-6.

• Charles, M. T. (2000). Police training: Breaking all the rules: Implementing the adult education model into police training. Charles C. Thomas Publishing.

• Christopher, S. (2015). The police service can be a critical reflective practice … if it wants. Policing: A Journal of Policy and Practice, 9, 326-339. https://doi.org/10.1093/police/pav007.

• Cockcroft, T. & de Maillard, J. (2021). Abstract police and police ‘professionalism’: Contemporary contradictions in British policing. In J. Terpstra, R. Salet & N. R. Fyfe (Eds.), The abstract police: Critical reflections on contemporary changes in police organisations (pp. 35-55). Eleven.

• Cohen, H. S. & Feldberg, M. (1991). Power and restraint: The moral dimension of police work. Praeger.

• Davis, K. C. (1975). Police discretion. West Publishing.

• Dewey, J. (1933). How we think: A restatement of the relation of reflective thinking to the educative process. D.C. Heath & Co Publishers.

• Engel, R. S. (2001). Supervisory styles of patrol sergeants and lieutenants. Journal of Criminal Justice, 29, 341-355. https://doi.org/10.1016/S0047-2352(01)00091-5.

• Engel, R. S. & Worden, R. E. (2003). Police officers’ attitudes, behavior, and supervisory influences: An analysis of problem solving. Criminology, 41, 132. https://doi.org/10.1111/j.1745-9125.2003.tb00984.x.

• Epstein, R. M. (1999). Mindful practice. Journal of American Medical Association, 282(9), 833-839. https://doi.org/10.1001/jama.282.9.833.

• Fernandez, S. & Rainey, H. G. (2006). Managing successful organizational change in the public sector. Public Administration Review, 66(2), 168-176. https://doi.org/10.1111/j.1540-6210.2006.00570.x.

• Friedman, B. (2017). Unwarranted: Policing without permission. Farrar, Straus, and Giroux.

• Gawande, A. (2002). Complications: A surgeon’s notes on an imperfect science. Metropolitan Books.

• Gawande, A. (2011, September 26). Personal best. The New Yorker.

• Goldstein, H. (1990). Problem-oriented policing. Temple University Press.

• Harmon, R. A. (2016). Why arrest. Michigan Law Review, 115(3), 313-320. https://doi.org/10.36644/mlr.115.3.why.

• Harvey, M., Coulson, D. & McMaugh, A. (2016). Towards a theory of the ecology of reflection: Reflective practice for experiential learning in higher education. Journal of University Teaching & Learning Practice, 13(2), 1-20. https://doi.org/10.53761/1.13.2.2

• Kadushin, A. & Harkness, D. (2014). Supervision in social work. Columbia University Press.

• Kahan, D. M., Hoffman, D. A. & Braman, D. (2009). Whose eyes are you going to believe? Scott V. Harris and the perils of cognitive illiberalism. Harvard Law Review, 122(3), 837-906.

• Kelling, G. (1999). Broken windows and police discretion. U.S. Department of Justice.

• Kleinig, J. (1996). The ethics of policing. Cambridge University Press.

• Klockars, C. B. (1988). The idea of police. Sage Publications.

• Koen, M. C. & Mathna, B. E. (2019) Body-worn cameras and internal accountability at a police agency. American Journal of Qualitative Research, 3(2), 1-22. https://doi.org/10.29333/ajqr/6363.

• Koen, M. C., Newell, B. C. & Roberts, M. R. (2021). Body-worn cameras: Technological frames and project abandonment. Journal of Criminal Justice, 72(August 2020), 101773. https://doi.org/10.1016/j.jcrimjus.2020.101773.

• Koen, M. C., Willis, J. J. & Mastrofski, S. D. (2019). The effects of body-worn cameras on police organisation and practice: A theory-based analysis. Policing & Society, 29, 968-994. https://doi.org/10.1080/10439463.2018.1467907.

• Kras, K. R., Portillo, S. & Taxman, F. S. (2017). Managing from the middle: Frontline supervisors and perceptions of their organizational power. Law & Policy, 39(3), 215-236. https://doi.org/10.1111/lapo.12079.

• Loftus, B. (2010). Police occupational culture: Classic themes, altered times. Policing & Society, 20, 1-20. https://doi.org/10.1080/10439460903281547.

• Lowenkamp, M. S., Robinson, C. R., Koutsenok, I. & Lowenkamp, C. T. (2012). The importance of coaching: A brief survey of probation officers. Federal Probation, 76(2), 36-39.

• Lum, C. & Koper, C. S. (2017). Evidence-based policing: Translating research into practice. Oxford University Press.

• Lum, C., Koper, C. S., Wilson, D. B., Stoltz, M., Goodier, M., Eggins, E., Higginson, A. & Mazzerole, L. (2020). Body-worn cameras’ effects on police officers and citizen behavior: A systematic review. Campbell Systematic Reviews. https://doi.org/10.1002/cl2.1112.

• Lum, C., Koper, C. S. & Wu, X. (2021). Can we really defund the police? A nine-agency of police response to calls for service. Police Quarterly, 1-26. https://doi.org/10.1177/10986111211035002.

• Manning, P. K. (1977). Police work: The social organization of policing. MIT Press.

• Marx, G. T. (2006). Forget big brother and big corporation: What about the personal uses of surveillance technology as seen in cases such as Tom I. Voire. Rutgers Journal of Law and Urban Policy, 3(2), 278.

• Mastrofski, S. D. (1988). Community policing as reform: A cautionary tale. In J. R. Greene & S. D. Mastrofski (Eds.), Community policing: Rhetoric or reality (p. 65). Praeger.

• Mastrofski, S. D. (1996). Measuring police performance in public encounters. In L. T. Hoover (Ed.), Quantifying quality in policing (pp. 207-241). Police Executive Research Forum.

• Mastrofski, S. D. (1999). Policing for people. Ideas in American policing. Police Foundation.

• Mastrofski, S. D. (2004). Controlling street-level police discretion. The ANNALS of the American Academy of Political and Social Science, 593, 100-118. https://doi.org/10.1177/0002716203262584.

• Mastrofski, S. D. (2018, March 19). Do the right thing: Evaluating police performance at the street level. [Inaugural Stephen D. Mastrofski Lecture], Fairfax, VA, United States, George Mason University.

• Mastrofski, S. D. & Willis, J. J. (2010). Police organization: Continuity and change. In M. Tonry (Red.), Crime and justice: A review of research (pp. 55-144). University of Chicago Press.

• Muir, W. K. (1977). Police: Streetcorner politicians. University of Chicago Press.

• National Advisory Commission on Civil Disorders. (1968). The Kerner Report. U.S. Government Printing Office.

• Nix, J., Todak, N. & Tregle, B. (2020). Understanding body-worn camera diffusion in U.S. policing. Police Quarterly, 23, 396-422. https://doi.org/10.1177/1098611120917937.

• Owens, C. & Finn, W. (2017). Body-worn video through the lens of a cluster randomized controlled trial in London: Implications for future research. Policing: A Journal of Policy and Practice, 12, 77-82. https://doi.org/10.1093/police/pax014.

• Owens, E. (2020, 2 June). How do you build a better cop? By making them slow down. Washington Post.

• Paoline, E. A. (2003). Taking stock: Toward a richer understanding of police culture. Journal of Criminal Justice, 31, 199-214. https://doi.org/10.1016/S0047-2352(03)00002-3.

• Phelps, J. M., Strype, J., Le Bellu, S., Lahlou, S. & Aandal, J. (2018). Experiential learning and simulation-based training in Norwegian police education: examining body-worn video as a tool to encourage reflection. Policing: A Journal of Policy and Practice, 12(1), 50-65. https://doi.org/10.1093/police/paw014

• Ponomarenko, M. (2019). Rethinking police rulemaking. Northwestern University Law Review, 114(1), 4. https://doi.org/10.2139/ssrn.3333804.

• President’s Commission on Law Enforcement & Administration of Justice. (1967). The challenge of crime in a free society: A Report. U.S. Government Printing Office.

• Purnell, D. (2021). Becoming abolitionists: Police, protests, and the pursuit of freedom. Astra House.

• Reiss, A. J. (1971). The police and the public. Yale University Press.

• Reuss-Ianni, E. & Ianni, F. A. (1983). Street cops and management cops: The two cultures of policing. In M. Punch (Red.), Control in the police organization (pp. 250-274). MIT Press.

• Rex, S. & Hosking, N. (2013). A collaborative approach to developing probation practice: Skills for effective engagement, development and supervision (SEEDS). Probation Journal, 60(3), 332-338. https://doi.org/10.1177/0264550513499002.

• Robinson, C. R., Lowenkamp, C. T., Holsinger, A. M., Van Benschoten, S., Alexander, M. & Oleson, J. C. (2012). A random study of Staff Training Aimed at Reducing Re-arrest (STARR): Using core correctional practices in probation interactions. Journal of Crime and Justice, 35(2), 167-188. https://doi.org/10.1080/0735648X.2012.674823.

• Sabel, C. F. & Simon, W. H. (2017). The duty for responsible administration and the problem of police accountability. Yale Journal on Regulation, 33, 165-211.

• Sandhu, A. (2017). I’m glad that was on camera: A case study of police officers’ perceptions of cameras. Policing & Society, 29, 223-235. https://doi.org/10.1080/10439463.2017.1285917.

• Schön, D. (1983). The reflective practitioner: How professionals think in action. Basic Books.

• Schön, D. (1987). Educating the reflective practitioner: Toward a new design for teaching and learning in the professions. Jossey-Bass Publishers.

• Sennett, R. (2009). The craftsman. Yale University Press.

• Smith, P., Schweitzer, M., Labrecque, R. M. & Latessa, E. J. (2012). Improving probation officers’ supervision skills: An evaluation of the EPICS model. Journal of Crime and Justice, 35(2), 189-199. https://doi.org/10.1080/0735648X.2012.674826.

• Thacher, D. (2001). Policing is not a treatment. Journal of Research in Crime & Delinquency, 38, 387-415. https://doi.org/10.1177/0022427801038004003.

• Thacher, D. (2006). The normative case study. American Journal of Sociology, 111(6), 1631-1676. https://doi.org/10.1086/499913.

• Thacher, D. (2008). Research for the front lines. Policing & Society: An International Journal of Research and Policy, 18, 46-59. https://doi.org/10.1080/10439460701718567.

• Thacher, D. (2016). Channeling police discretion: The hidden potential of focused deterrence. University of Chicago Legal Forum Article, 13, 533-535.

• Thacher, D. (2019). The aspiration of scientific policing. Law & Social Inquiry, 41, 280. https://doi.org/10.1111/lsi.12367.

• Thacher, D. (2020). The learning model of use-of-force reviews. Law & Social Inquiry, 45, 755-786. https://doi.org/10.1017/lsi.2019.80.

• Thacher, D. (2022). Shrinking the police footprint. Criminal Justice Ethics, 41, 62-85. https://doi.org/10.1080/0731129X.2022.2062546.

• Toch, H. (2008). Police officers as change agents in police reform. Policing & Society, 18(1), 60-71. https://doi.org/10.1080/10439460701718575.

• Toronjo, H. (2019a). Gut check: turning experience into knowledge. In Ugwudike, P., Graham, H., McNeill, F., Raynor, P., Taxman, F. S. & Trotter, C. (Eds.), The Routledge companion to rehabilitative work in criminal justice (pp. 352-382). Routledge.

• Toronjo, H. (2019b). A corrections workforce for the 21st century. Federal Probation, 83(1), 3-7.

• Toronjo, H. (2020). Do you really want to help me? Practitioner perspectives on a new coaching model for probation front-line supervisors [Thesis]. George Mason University.

• Tyler, T. R. (2004). Police legitimacy. The Annals of the American Academy of Political and Social Science, 593, 84-99. https://doi.org/10.1177/0002716203262627.

• Van Maanen, J. (1983). The boss: First-line supervision in an American police agency. In M. Punch (Red.), Control in the police organization (pp. 275-317). MIT Press.

• Waddington, P. A. J., Williams, K., Wright, M. & Newburn, T. (2015). Dissension in public evaluations of the police. Policing & Society, 25, 212-235. https://doi.org/10.1080/10439463.2013.833799.

• Walker, S. (1993). Taming the system: The control of discretion in criminal justice, 1950-1990. Oxford University Press.

• Walker, S. (2016). Governing the American police: Wrestling with the problems of democracy. University of Chicago Legal Forum, 615-670.

• Weisburd, D., Mastrofski, S. D., McNally, A. M., Greenspan, R. & Willis, J. J. (2003). Reforming to preserve: COMPSTAT and strategic problem solving in American Policing. Criminology and Public Policy, 2, 446. https://doi.org/10.1111/j.1745-9133.2003.tb00006.x.

• White, M. D., Flippin, M. & Malm, A. (2019). Key trends in body-worn camera policy and practice: A four-year policy analysis of US Department of Justice-funded law enforcement agencies. ASU Center for Violence Prevention and Community Safety.

• White, M. D. & Malm, A. (2020). Cops, cameras, and crisis: The potential and the perils of police body-worn cameras. NYU Press.

• Willis, J. J. (2022). Culture eats strategy for breakfast: An in-depth examination of police officers’ perceptions of body-worn camera implementation and their relationship to policy, supervision, and training. Criminology and Public Policy. https://doi.org/10.1111/1745-9133.12591

• Willis, J. J. & Mastrofski, S. D. (2017). Understanding the culture of craft: Lessons from two police agencies. Journal of Crime and Justice, 40(1), 84-100. https://doi.org/10.1080/0735648X.2016.1174497.

• Willis, J. J. & Toronjo, H. (forth.). A way ahead: Re-envisioning the relationship between evidence-based policing and the police craft. In D. Weisburd, T. Jonathan-Zamir, B. Hasisi & G. Perry (Eds.), The future of evidence-based policing. Cambridge University Press.

• Willis, J. J. & Toronjo, H. (2019). Translating police research into policy: Some implications of the national academies report on proactive policing for policymakers and researchers. Police Practice and Research, 6, 617-631. https://doi.org/10.1080/15614263.2019.1657631.

• Willis, J. J. & Toronjo, H. (2022). Exploring a craft learning model for reviewing patrol officer decision-making in encounters with the public. Law and Social Inquiry, 1-28. https://doi.org/10.1017/lsi.2022.57.

• Worden, R. E. & Dole, C. (2019). The holy grail of democratic policing. Criminal Justice Ethics, 38, 41-54. https://doi.org/10.1080/0731129X.2019.1586217.

• Worden, R. E. & McLean, S. J. (2017). Research on police legitimacy: The state of the art. Policing: An International Journal, 40(3), 480-513. https://doi.org/10.1108/PIJPSM-05-2017-0062.

• Yanow, D. & Tsoukas, H. (2009). What is reflection in action? A phenomenological account. Journal of Management Studies, 46(8), 1339-1364. https://doi.org/10.1111/j.1467-6486.2009.00859.x.

Noten

• 1 Two of these case studies were conducted in small police agencies (50-100 sworn) located in the mid-Atlantic region of the United States. One had implemented BWCs in 2012, the other in 2016. The third agency was located in the mid-West, and had between 250 and 300 sworn officers. It implemented body-worn cameras in 2014. We promised these agencies that we would not reveal their identity. One case study relied on interviews only, the other two combined interviews with observations.

• 2 These goals are meeting departmental norms, containing violence and controlling disorder, preventing crime, avoiding physical injury to themselves and avoiding provoking the public into angry retaliation that threatens their careers (Bayley & Bittner, 1984).

• 3 For more details on the methodology, please see Willis and Toronjo (2022).