Introduction

Police officers, as representatives of their departments, and thereby the state, have more frequent interactions with the public than any other branch of the judicial system. Some estimates suggest that as much as 97% of the daily tasks of police-officers involve citizen interactions (Myers et al. 2008) and that more than 25% of US adults interact with the police on a yearly basis, and about half of those being involuntary interactions (Simckes et al. 2021). Many of the police-public interactions also involve some form of conflict where police officers either are, or become, part of the conflict or where they act as mediators (Rosenbaum and Lawrence 2017; Sun 2006). To handle their duties, police have unique mandates in using coercive measures, including use of force. The right to use of force has been hailed as a defining aspect of police work, and albeit sometimes necessary, something that a democratic and procedurally just police should always strive to avoid (Bittner, In Todak and James 2018). Thus, conflict management and use of force are closely interlinked, and the use of force could be viewed as a result of either a failure in conflict management, as a result of available tools not being sufficient, or a means for conflict management.

In a democratic society, a central tenet of policing has long been to maintain positive relationships to the community and its citizens. To be effective, modern police officers need to cooperate with citizens, and for the police to be effective, citizens need to perceive the police as legitimate and procedurally just (Shjarback and White 2016). Citizens have high demands on officers to act in a respectful and at all times professional manner, where demeanor, competence, and strategies for conflict management are important parts (Rosenbaum and Lawrence 2017). Therefore, given the frequency with police officers interact with the public, the frequent occurrence of conflicts, and the consequences if they are even perceived as unprofessional or unjust, it is essential for officers to have solid interpersonal, communicative, and conflict management skills.

While these skills are necessary for officers in all areas of police work, whether working with primarily investigative tasks or patrolling duties, patrolling officers are perhaps most often subject to rapidly changing situations, in uncontrolled environments, thereby demanding an even greater need for dynamic and professional conflict management skills.

Despite this, research on patrolling officers is scarce (Ahmad et al. 2019), something that also appears to hold true also for research into police involved conflicts. In the preliminary searches conducted for this review, no scoping reviews that provided maps of the evidence base for the interventions in terms of training and theories used in police conflict management were found.

Given this, the objective of the present scoping review is to map the existing research on the topic of police conflict management, with a primary focus on patrolling officers (i.e., not specialized police capacities) in terms of the underlying theories and evidence for interventions and training that are in use in modern policing.

Review Questions

  1. 1.

    What theories provide the foundation for interventions, training, and methods employed in modern policing with regard to police involved conflicts?

  2. 2.

    What evidence base does the current research on the topic provide?

  3. 3.

    What areas of research has been given attention, and what knowledge gaps can be identified in the current research?

Inclusion Criteria

For the current review, the mnemonic suggested in the JBI Manual for evidence synthesis was used, i.e., PCC (population, concept, and context) (Aromataris et al. 2020; Munn et al. 2018). The inclusion criteria are elaborated on below:

Population

Police officers, with primarily patrolling duties, needed to be the primary focus of included articles. The rationale for this choice is that patrolling police officers can be assumed to be exposed more frequently to conflict situations in their daily work. Thus, officers with specialized capacities such as negotiators or traffic police were not included. Articles whose primary focus was the perceptions of individuals that had been in contact with the police were excluded.

Concept

The primary phenomena of interest in the current review are police involved conflicts, and the underlying frameworks in terms of theory and methods. Given that the terms conflict management and conflict resolution are used somewhat interchangeably, both terms were included. As described above, use of force can be seen both as a method for conflict management, a means for, or a result of conflict management, and was therefore included.

Context

The context of interest in the current review is police involved conflicts with no limitations as to geography. The review is limited to research that is applicable to individual officers, albeit at an aggregate level, thus excluding articles that focus on the political or organizational levels of conflict. In addition, the conflicts studied should be related to police and citizen encounters, rather than workplace conflicts (i.e., conflicts among peers).

Further, in terms of study designs and publication types, only articles published in peer-review journals, that were not pure opinion pieces such as letters to the editor, were included. For grey-literature, reports and PhD theses were included.

Methods

In the initial stages of this review, the intention was to conduct a systematic literature review. Thus, a protocol was submitted in PROSPERO (2022 CRD42022278909). During the initial screening of the records found in the searches, our familiarity with the area of research grew, and given the broad research questions posed for the review, it was decided that a scoping review would provide a more feasible methodological framework. For example, the breath of the phenomena studied in the records that were found would not enable a confirmation or refusal of whether current practices in police conflict management were based on relevant evidence, which is typically a central aim in systematic literature reviews (Munn et al. 2018). Rather a systematic mapping of the available evidence would be more relevant for the field of study. To the best of our knowledge, no such broad mapping of the research area in question exists, and conducting one could provide valuable insight into the types of available evidence, and clarification of key concepts, thereby helping to identify and analyze knowledge gaps in the current research (Peters et al. 2015).

The review was conducted guided by recommendations provided by Joanna Briggs Institute (JBI) and the JBI manual for evidence synthesis (Aromataris et al. 2020) as well as the PRISMA-ScR checklist (Tricco et al. 2018).

Search Strategy

The searches were limited to the last 20 years from the start of the review, i.e., from 2001. Full texts to be included in the search included English, Swedish, Norwegian, and Danish.

Electronic searches were conducted in the following databases: Academic Search Elite, ERIC, Socinfo, and Psycinfo. The search term that was used was SU (police or law enforcement or officer or cop) AND SU (conflict management or conflict resolution or problem solving or communication or interpersonal conflict or reconciliation or “use of force”). To broaden the scope of the search, a novel tool for literature searches, Research Rabbit, was also used. Research Rabbit uses algorithms to search for publicly available records of scholarly publications, including dissertation and reports (i.e., grey literature). For the current project, the tool was used to enhance the search results by first mapping the area through the “traditional” databases. Relevant search terms were then defined through theories identified in the initial search. The search term, for example “police AND procedural justice” was used to find relevant articles. In the next step, the “similar research” function was used, and abstracts sifted for inclusion/exclusion. Next, “later work” was similarly explored for studies that met the inclusion criteria.

Search results were downloaded and inserted to the Zotero software, to enable removal of duplicate records, before uploading the results to Rayyan. In Rayyan, two reviewers (first and second authors) conducted the initial screening of abstracts, blinded to each other. Disputes over inclusion and exclusion were settled through discussion. In case that disputes had not been settled, the third reviewer would be involved. Full-text screening of the results was conducted by the first and third authors, again independently with possible disagreements being settled through discussion. The third reviewer randomly selected and cross-checked approximately 10% of the articles included in this stage to ensure the robustness of the inclusion and exclusion of articles.

Data Extraction

A charting table was developed in Excel based with extraction fields based on recommendations by Arksey and O’Malley (2005), adapted to suit the aim and research questions of the current review. Extraction fields included bibliographical data: author, year (publication), and title; country of study and aim and purpose; methods and theories, population (recruits/officers), type of intervention, theory (theoretical underpinning of the study), and key findings. To further enhance the robustness of the mapping process, the open-source software Obsidian was employed. For each article included in the study, a note was created that included extracted annotations as well as notes regarding its content. Also, Obsidian enabled linking articles that for example had contrasting or confirming results or conclusions.

Results

Identification of Studies

As indicated in the PRISMA flow diagram (Appendix), the initial search of databases included a total of 1545 results. A number of duplicates were identified when uploading the results in Zotero, and where subsequently removed. In the next phase of the identification of eligible studies, abstracts (N = 1485) of the included records were uploaded to Rayyan.ai and analyzed: a total of 1340 studies were removed due to not matching the eligibility criteria. In total, 145 studies were retained for the final part of the study selection, with 47 records being included for the final mapping of the data.

Study Characteristics

The vast majority of the included studies were conducted in the North America (n = 32) with a smaller proportion using data from European countries (n = 14) and Australia (n = 3). Among the studies included from the Research Rabbit search, some (n = 4) did not use original data, and therefore, country of study was not specified. The subsequent search in Research Rabbit did yield a larger variety in terms countries studied, although proportions remained largely the same. Further, the majority of the included articles (n = 26) focused on use of force and the study population for the included studies were mainly focused on police officers, with only four studies focusing on police recruits. It is worth noting that in several studies, the samples include a small number of departments per study, mostly one or two, with exceptions being found in studies that examine use-of-force registers, making generalizability of the results more difficult. Overall randomized controlled trials (RCTs) which are often considered the “gold-standard” for building evidence-based practices (Hariton and Locascio 2018) constituted a small proportion of the included studies, with nine out of the total 88 studies employing such a research design. This result is in line with results from other reviews in police research (Engel et al. 2020; Rosenbaum and Lawrence 2017). Also, all the included RCTs found were published between 2018 and 2022, except for a randomized field trial (RFT) published in 2013 (Table 1).

Table 1 Number of studies by continent and number of RCTs
Full size table

Use of Force

As stated above, many of the studies included from the initial search were related to use of force, an area that has been described as one of the most well-researched areas of police research, and that includes meta-analytic studies such as one by Bolger (2015). When mapping the data pertaining to use-of-force, some challenges become apparent. To begin with, use of force incidents are rare events in terms of relative frequency, making it sensitive to random variations or outliers (Bennell et al. 2021). Previous research has pointed to that a small number of officers, the so called “power few,” account for a large number of use of force incidents (Boxer et al. 2021; Manzoni and Eisner 2006). In addition, when looking at data at an aggregate level, the likelihood of data-being affected by outliers is attenuated further, in that several officers could report the same incident.

As for data available for research, lack of data from official sources and the suitability or comprehensiveness of available data is explicitly addressed in a number of the studies (Hickman et al. 2015; Lim and Lee 2015; Swencionis et al. 2021). Further, the lack of national level data has long been a concern for police scholars and calls for such a system have been raised repeatedly (Shjarback and White 2016). In the USA, the FBI have developed a reporting system for national level data, the NUFDC but providing data for the system is voluntary, and levels of”no-response” have been high. The system also only includes reports of fatalities, serious bodily injury, or when a firearm is discharged by officers (FBI 2022). Researchers have instead opted for open-source data provided by for example news outlets or NGOs. One potential effect of the difficulties in acquiring relevant, and up-to-date data from official sources, is the time between data-collection and publication. The mean time between data collection and publication in the 15 articles that use data collected either directly from the departments or similar registers was 6.9 years (min 3, max 16).

Definitions

Further, the definitions of use-of-force and suspect resistance appears to differ greatly between departments that provide data. The National Bureau of Justice has provided a definition, but it is only related to what constitutes excessive force (Motley and Joe 2018). Garner and colleagues have adapted a definition provided by the National Academy of Sciences, and this definition has been argued to be the solution to the questions raised about definitions: “… behaviors by individuals that intentionally threaten, attempt, or inflict physical harm on others” (Bolger 2015). The definitions and criteria for reporting found among the included studies are provided in Table 2, but in summary vary in granularity — ranging from 3 to 8 levels of force. In addition, criteria for when officers are required to report use of force vary greatly, from mere presence of officers to only incidents where a person is injured. Similar discrepancies exist in regard to definitions of suspect or citizen resistance, further adding to methodological challenges (McCarthy et al. 2019).

Table 2 Definitions and reporting criteria
Full size table

In terms of the theoretical underpinnings of the included studies, several articles have employed use-of-force continuums and variants thereof such as the force factor method (Hickman et al. 2015). Other theories identified among included studies include theories related to decision-making and use of force such as naturalistic decision-making (Hine et al. 2018), network exposure (Ouellet et al. 2019), self-efficacy (Torres 2020), and ego-depletion (Staller et al. 2018, 2019).

Many of the studies included use sociodemographic variables such as age, sex, race, and educational levels as predictors, with less focus on individual differences such as personality, which can be assumed to assert a significant influence on use of force and conflict behaviors (Chapman 2012) and where much of the research has been called outdated (Abrahamsen and Strype 2010).

Policing Approaches

Among the included studies, three major models, or approaches to policing related to police conflict management, could be discerned: procedural justice (PJ), de-escalation, and the elaborated social identity model (ESIM). Although they are separate constructs, there also common denominators among the three. One particularly salient commonality between these constructs is that they, albeit from somewhat different points of departure, focus on legitimacy and relational approaches to policing.

Among the three, de-escalation stands out in terms of lacking a definitive or more broadly accepted definition, something that has been raised as a criticism in some of the included articles (Wolfe et al. 2020). Procedural justice (PJ) on the other hand appears to be more well defined in terms of both what it is and its underlying theories. Extant research on PJ has overall shown positive effects in several key aspects, such as affecting willingness to obey and the perceived legitimacy of the police. Moreover, in later years, some important criticisms have been forwarded, and knowledge gaps have been identified related to these models. For example, some researchers argue that despite an extensive body of research, the causal claims made on the effects of PJ should be questioned, stating for example that the lack of longitudinal data and experimental designs does not make it possible to account for third common causes or confounders (Nagin and Telep 2017, 2020; Pina-Sánchez and Brunton-Smith 2020; Schaap and Saarikkomäki 2022).

The literature on ESIM similarly appears to have a more distinct theoretical underpinning than de-escalation. In short, ESIM emanated as a reaction to classical views on the “pathological crowd” and argues that behaviors in a crowd should rather be understood as a dynamic social identity. Similar to research on PJ, many of the included articles point to positive effects of policing models based on the theory. In terms of criticism, it has been argued that the ESIM places too much emphasis on police and crowd relations and is therefore not sufficient to account for other important aspects such as cultural and contextual antecedents (Lydon 2018). Among the included studies, two suggest other theoretical models such as the flashpoint model (Lydon 2018; Waddington 2013) and Lydon’s (2018) contextually driven model of police legitimacy. An important note in relation to these criticisms is that critics do not suggest that the relationally based policing models of PJ or ESIM are to be discouraged. The criticisms are rather related to what conclusions can be reasonably derived from the research conducted thus far. Of note is also that ESIM-based models of policing in fact contain definitions of the de-escalating principles that however do not appear to have been fully adopted in de-escalation research.

Training Interventions

For both the de-escalation and procedural justice, there appears to be a more or less salient consensus in that more research is needed into the efficiency and effects of training interventions geared towards these models. For de-escalation, the criticism is again largely related to the lack of definition and that it is therefore difficult to draw conclusions and compare training interventions (White et al. 2021; Engel et al. 2021). Another important criticism that could at least in part be related to lack of definition are concerns over officer safety in relation to de-escalation. One such concern is that de-escalation could make officers hesitant and that de-escalating tactics might slow officers down in critical situations (Engel et al. 2021). For PJ, the criticism is rather related to the lack of randomized trials that are needed to make conclusions on the causal pathways between training and behavior (Nagin and Telep 2020). Among the included studies, there is however a growing, albeit limited, body of research that have used creative research designs to assess such training interventions. A number of studies have for example used systematic social observation (SSO), employing body-worn camera footage to study effects on behaviors (see for example Dai 2020), as well as variants of randomized field trials (RFTs) (Mazerolle et al. 2013) and in one study a “mystery shopper” design (Canales et al. 2020).

Discussion

The purpose of the current scoping review was to map the research and evidence base of research related to police involved conflicts, in terms of both studies focusing on use of force as well as conflict management. The results indicate that there is a growing body of research in several areas related to the topic, but that there are also some potential gaps of knowledge that needs further research to be filled.

Most of the studies included in the present study have been conducted in North America, and while the search conducted in Research Rabbit did yield a somewhat broader result, proportions remained largely the same. Importantly, as in any review study, using other search terms, databases, or broadening, the search to include other languages such as Spanish, German, or French could have provided other results. However, many journals that are indexed in major search engines publish exclusively in English. Therefore, to disseminate research to a wide circle of stakeholders (academics and practitioners alike), either sufficient language skills or financial resources for translation and language reviews become a necessity. One drawback with such obstacles is that global perspectives, including lessons and experiences from around the globe, are lost. Also, while many aspects of policing likely are at least similar in democratic societies, cultural, socio-economic, and historic aspects (to name but a few) can yield different results and make the generalizability of the results more difficult.

Definitions

In terms of gaps of knowledge, and difficulties in examining the evidence base for the constructs included within the current study, many are related to lack of uniform definitions. De-escalation for example has been criticized for lacking in terms of definition, making it difficult to scientifically assess as well as opening for perhaps in part unwarranted criticism. The research of among others Todak and colleagues have provided valuable insights and steps toward definitions of the construct however (Todak 2017; Todak and James 2018). Similarly, the same lack of unified definitions appears to hamper progress in use of force-research where many departments employ different definitions of what constitutes use of force, and when it is required to be reported. To advance the research on this area, of central importance for police work, departments, not only in the USA but globally, need to have at least some consistency in how use of force is defined and what types of incidents are reported. Such definitions are also important in developing training, methods, and approaches for police work, and these models need to be firmly based on robust theoretical underpinnings, that makes it possible to assess not only observable behaviors but also the pathways leading up to behaviors.

Access to Relevant Data

Some of the included studies point explicitly to lack of relevant and sufficiently comprehensive data, not least related to use of force (Hickman et al. 2015; Lim and Lee 2015; Swencionis et al. 2021). The results in the current review indicate that lack of data is still a salient issue in conducting research on the topic and that differing regulations of when to report use of force can influence findings and the generalizability of the results. One possible effect of difficulties in acquiring relevant data is perhaps evident in that the time between data-collection and publication of the included studies sometimes being several years. Thus, the results drawn from the research conducted are not necessarily a representation of the current state of affairs, but rather of what it was. Expanding what is reported and when reports are filed would not benefit researchers but could influence perceived legitimacy of police departments positively, by potentially obviating concerns of lacking transparency. However, for such endeavors to be successful, it is important to firmly establish it among not only police-executives but also among the rank-and-file of police departments — requiring a careful balancing act between acquiring information and bogging down officers already burdened by administration. Therefore, the fact that such considerations have been taken in some of the included studies is not to be underestimated in terms of its importance (Boivin and Lagacé 2016; Gallo et al. 2008). Also, some of this work could perhaps, in close cooperation with authorities, be at least partially automated through text-mining techniques with already available data such as incident reports and memos.

Research Designs

As highlighted in some of the included studies, in several areas, there have been calls for more longitudinal, as well as experimental (RCT) studies. In the present review, only a few of the included studies used experimental trials with randomized samples and control groups. Interestingly, the studies that where found had all (except one) been published between 2018 and 2022. This could be indicative of researchers answering calls for such research, which would be an important step towards a deeper understanding of police conflict management. Such research designs are inherently difficult to conduct, in terms of both time and resources, and require a close collaboration between researchers and the objects of study. However, not least with regards to examining the effects of training, both RCTs and longitudinal studies will be essential to continuously improve police conflict management practices and training.

Use-of-force Vs Conflict Management

Further, while the research on use of force does provide important insights into police conflict management, and the research conducted thus far provides an important base for improving police work and training, an added focus on conflict management and how to avoid use of force could also provide more badly needed knowledge and evidence on what works. While there certainly can be important work published that was not found in the searches conducted for this study, in the case of police conflict management, when looking solely on situations where force has been used, we risk not learning from all the other (majority) of cases where force was not used. There are good of examples of research that does in fact examine successful police and citizen interactions such as studies by Todak and James (2018), as well as procedural justice related research and studies on ESIM-based models of policing (see, for example, Mazerolle et al. 2013; Stott et al. 2019) thereby providing important evidence for what works, not only what does not.

Conclusions

In conclusion, it appears that despite efforts from the research community, given how central conflict-management is to policing, there is surprisingly little research on the topic and important gaps of knowledge that remain to be filled. Cogent claims can be made both in favor, and against some of the models and theories that have been studied thus far, but overall, more research is needed. Salient among the criticisms that can be directed to police departments are the apparent problems associated with accruing relevant data. But, perhaps the most important criticism should not be directed towards the research that has been done, the researchers that have done it or the departments that make some effort to provide data. Rather, the criticism should be directed towards the research that has not been done or departments that do nothing to facilitate research. Also, the apparent thresholds associated with disseminating research from other countries than those included in this study should be addressed. There are significant both practical and economic hurdles to be overcome for a researcher from a non-English speaking country, with limited resources to publish their findings in international, ranked journals. But there are certainly lessons to be learned and experiences to be drawn from such countries.

Also, there is a significant body of research based on aggregated data and sociodemographic factors related to not least use of force, with less focus on individual differences. While there certainly are inherent hurdles associated with conducting such research, it could provide valuable insights and add to the evidence base for police conflict management.

One way forward for police research, not least in relation to studies on conflict management, could be to in a formalized manner create multidisciplinary networks of scholars and practitioners. Such networks could be a driving force in providing definitions based on scientific evidence as well as disseminating research within the area, to draw from experiences and knowledge from the global research community.