787022 CRJ0010.1177/1748895818787022Criminology & Criminal JusticeSowatey and Tankebe research-article2018 Article Criminology & Criminal Justice Doing research with police 1 – 17© The Author(s) 2018 Reprints and permissions: elites in Ghana sagepub.co.uk/journalsPermissions.nav httpDs:O//dIo: i1.o0r.g1/107.171/177/4187848985985188178788770022 journals.sagepub.com/home/crj Emmanuel Addo Sowatey University of Ghana, Ghana Justice Tankebe University of Cambridge, UK Abstract Much of our methodological insights from researching policing in sub-Saharan Africa comes from studies of frontline officers. Consequently, many important methodological questions about research on senior police officers remain unanswered. This article addresses this gap by drawing on insights from interviewing senior officers in Ghana. It focuses on the challenges and opportunities in negotiating access, establishing trust during interviews and dealing with ethical dilemmas. We highlight the role of informal social networks and cultural practices of surprise visits, what we have termed strategic ambush, in securing formal approval for our research. However, this represented mere or putative access for which deference towards institutional gatekeepers was key to its actualization. Deference towards officers and extensive knowledge of the policing environment helped to put the senior officers at ease, and enhanced the chances of a successful interview. Finally, we offer reflections on our responses to unexpected ethical dilemmas that we faced in the field. Keywords Deference, Ghana police, intruding outsiders, police elites, strategic ambush Introduction Until recently, there had been little research on policing in Africa. Now, however, a size- able literature is accruing. It includes ethnographic accounts of everyday police work Corresponding author: Justice Tankebe, University Lecturer, Institute of Criminology, University of Cambridge, Sidgwick Avenue, Cambridge, CB3 9DA, UK. Email: JT340@cam.ac.uk 2 Criminology & Criminal Justice 00(0) (Beek, 2012; Beek et al., 2017; Faull, 2018; Hornberger, 2011; Sowatey and Atuguba, 2014), quantitative analyses of police–public relations (Boateng et al., 2014; Bradford et al., 2014; Tankebe, 2009), and enquiries into the challenges posed by democratic police reforms (Faull, 2016; Hills, 2007, 2008). Even so, much of what we know about policing in Africa, despite its abundance of methodological insights, derives from studies of frontline officers (see, for example, Beek, 2012; Faull, 2018; Marks, 2003; Tankebe, 2010, 2014). We know almost nothing about research with senior police commanders. As Hills (2007: 407) notes, ‘little is known about the role, background, and political or administrative functions of commissioners. There are no systematic comparisons of their social origins, career paths, rewards, or philosophies of policing.’ The paucity of research on elite police officers in Africa means that we have no equivalent to the methodological insights emanating from Robert Reiner’s (1991) study of chief constables in England and Wales. What special challenges arise in conducting research on senior police officers in an African context, and how can they be overcome? How can access be negotiated? How do we handle the interview in such a way as to maximize data quality? To address these gaps in our knowledge the present article draws on experiences of conducting in-depth interviews with 81 senior police officers in Ghana. The project, titled Chief Police Officers in Ghana, sought to gain an insight into aspects of democratic governance in a post-colonial society by exploring the views of senior officers on police institutional capacity, police governance and accountability, and orientations to human rights. The interviews sought to answer the following questions: how do senior police officers in Ghana perceive their roles and legal powers? What are their perspectives on crime and its control? What are their management and policing philosophies? What is the nature of their relationship with elected officials, traditional authorities and other elites? Who are these officers? From what social backgrounds have they come to occupy their current positions of responsibility in implementing general policy and day-to-day opera- tional matters? Ghana has a national police service headed by an Inspector General of Police (IGP). It is structured into 11 administrative regions, with each region sub-structured into divisions and each division sub-structured into districts, and districts into police stations. This arti- cle is based on the authors’ experience of interviewing 11 regional commanders, 24 divi- sional commanders and 46 district commanders between March 2016 and March 2017. Senior officers in Ghana fall under three broad ranks: Inspector General of Police and deputy; commissioner – comprising commissioners, deputy commissioners and assistant commissioners; and superintendent – which comprises superintendents, deputy superin- tendents and assistant superintendents. The regional commanders interviewed were either assistant or deputy commissioners, while divisional and district commanders fell within the superintendent ranks. The interviews each lasted between one and three hours, and all but one were audio-recorded. Only in seven interviews were both authors present. In them, we took turns to cover topics that we had agreed beforehand. This arrangement seems to have worked well. There was only one exception, when a regional commander appeared terse and cold in his responses to questions by one of the authors. We both rec- ognized the posture. The questioner, therefore, tendered an excuse and left the command- er’s office. The commander then appeared more relaxed, and the rest of the interview Sowatey and Tankebe 3 proceeded smoothly. Both at the time and subsequently, it was unclear why the com- mander had felt uneasy. The article is structured as follows: we begin with a discussion of the process in obtaining access for the study, both at the Ghana Police Service headquarters and at each of the 11 police administrative regions across the country. We then proceed to explore issues of positionality, focusing on the multiple identities that we assumed, their advan- tages and shortcomings. Next, we discuss the practicalities of planning the fieldwork and how we dealt with exigencies in the field. We discuss how we handled the interviews in ways that encouraged respondents to discuss topics they had initially considered too sensitive. We highlight the roles of demeanour and deference (Goffman, 1956). The final two sections focus on ethics and risks in undertaking research with senior police officers. We hope that by sharing our fieldwork experience of the Ghana Police Service we can suggest some practical steps and insights that may simplify the task for those who wish to study senior police officers in Africa. Negotiating Access at the Top Access is a major challenge to research on senior police officers. The senior police offic- ers we interviewed may be viewed as part of Ghana’s elites, in the sense that they are ‘persons who are able, by virtue of their strategic positions in powerful organisations and movements, to affect political outcomes regularly and substantially’ (Higley and Burton, 2006: 7, emphasis in original). As elites, they are ‘difficult to penetrate. [They] establish barriers that set their members apart from the rest of society’ (Hertz and Imber, 1993: 3). Given the centralized structure of the Ghana Police Service, the starting-point for us was to seek official approval from the office of the Inspector General of Police, the head of the police. This we managed to achieve within a month. Two key resources facilitated our access. The first was informal networks with middle- level officers, with whom we had developed long-standing relationships over several years. We held informal meetings at the national police offices in Accra to discuss our research objectives and plans. The middle-level officers were chief superintendents who themselves had extensive networks at the top hierarchy of the police organization and deep insight into administrative procedures. They knew the (invisible) power dynamics within the police service and could offer critically important advice. Our experience therefore chimes with observations by Walker and Lidz (1977: 115) that access to restricted worlds requires an ‘individual who will establish the [researcher’s] credentials [and who is] well thought of by the other participants in the system’. Ostrander (1993: 12) makes a similar point: ‘you get in and get useful data from them (i.e. elites) if you know others that they know and respect’. However, having such friends is not sufficient. Far more important are those friends who understand and appreciate the potential value of academic research for police work. They wish to cultivate a strong relationship with academia but are confronted by a culture that does not necessarily encourage such a collaboration. We explained that there was a paucity of research capturing senior officers’ voices and experiences and that, without a systematic study of those officers, policy initiatives to reform police work would be based on distorted information or anecdotes. We further explained that the research had the potential to pro- vide key information that could enhance policy and strategy. Once persuaded on the 4 Criminology & Criminal Justice 00(0) potential merits of the research, the chief superintendents became key advocates for our study and offered advice on negotiating and maintaining access. During our preparation and engagement with the police we realized that, although telephone calls served an important purpose, a physical presence was particularly advan- tageous. It conveyed a symbolic message about our commitment. Our fieldnotes illus- trate this last point: A staff officer at the national office told us over the phone: ‘My brothers, these calls are good and important, but if you are able to make time off your busy schedule and come over to the national office it will show to my superiors that you are serious about your project.’ (Extract from fieldnotes) Within a month, we made as many as 15 – often fruitless – pre-arranged visits to the national offices of the police. Attempts to book appointments by telephone or text usu- ally yielded the response that ‘We are currently busy due to upcoming special events so why don’t you come next week, or we will call you when it is convenient?’ Marks (2003: 57) found that, in doing research with frontline officers in South Africa, time spent ‘hanging around’ with officers was not entirely fruitless; as she found, it offered oppor- tunities for ‘informal interactions [that] were crucial not only to building trust and famili- arity, but to coming to grips with the everyday and more mundane thought processes of the police, as well as their everyday inter-actions with one another’. While ‘hanging around’ the national police offices, our police friends introduced us to two regional com- manders who were visiting these offices for various reasons. These initial, albeit fleeting, contacts proved useful when we went to the jurisdictions of those commanders. Nonetheless, to avoid those unproductive visits, we adopted what we called strategic ambush access. It involved attempts to gain an audience with senior officers without prior appointment. We simply showed up at the offices of the national police services, hang around – sometimes for several hours – for chance encounters with the relevant officers who could help track our application and possibly introduce us to the key decision mak- ers. Our ambush visits were helped by the receptionists, to whom we became familiar faces, not demanding strict proof of an appointment; we merely had to indicate the name of the officer. It was on such ambush trips that we were informed of the approval of our application made two months earlier, and the approval letter was handed to us. Thus, even at this early stage in the research programme, we learned an important practical lesson about access to the police: reliance solely on bureaucratic requirements of advance appointments and waiting for formal communication could be a futile exercise. Thus, strategic ambush access was the dominant approach at the regional, divisional and district levels. This was not because we doubted their claim of busy schedules, not least because our fieldwork started in a national election year. The election period had both advantages and setbacks for us. One advantage was that all police personnel and commanders were directed by the Inspector General of Police to remain at post. This meant that the chances of meeting commanders were high. On the other hand, we could not be certain about the commanders’ availability and movements. Elections in Ghana are often characterized by low-intensity violence. This means that commanders are often occupied in meeting their junior commanders and/or planning for political campaigns Sowatey and Tankebe 5 and other outside activities. At other times they were in the field, supervising the security detail for politicians in their jurisdictions. Consequently, strategic ambush at police division and districts served a key purpose: whenever an appointment was rescheduled at short notice or interrupted halfway, we hurried, without prior appointment, to another station or division that we had selected for the study. Our recourse to ambush arose from a choice between gentle nudging through phone calls to schedule appointments or causing a temporary nuisance with the risk of incur- ring the displeasure of these officers. Ambush visits are a key feature of traditional Ghanaian culture. It is commonplace for family members, friends, acquaintances or even strangers to pay a person a surprise visit for various reasons; in some cases, they travel long distances, risking the disappointment of unfruitful journeys. It is not unknown to these officers, which might have explained why none of them openly remonstrated. However, we found that the Ghanaian cultural practices of ambush visits had important implications for our interviews. There was a constant risk of an inter- view being postponed, paused for a couple of hours or truncated at any time not only on account of an unforeseen security-related emergency. It was also on account of ambush visits by a senior politician, influential locals, or friends. Herein lies the down- side for the researcher of the Ghanaian culture of ambush visits: there are no assur- ances against disruption of interviews. To respond to the risks of interruptions, we learned to prioritize certain thematic areas such as motivations for joining the police, and views on recruitment, institutional capacity, leadership styles and police integrity. However, these unexpected events always had a cascading effect on our travel plans and interview schedules. Strategic ambush also carried the risk of fruitless visits. Apart from commanders not being at post for the reasons already noted, some commands existed only on paper, as exemplified in this correspondence: Researchers: D COP (deputy commissioner of police), good morning. Please, we are leaving the location for your office. Commander: I am not available now. I am moving with the Vice President. Researchers: Ok we will be in town tomorrow morning at 8 to see you. Commander: It is too early, I will be with Vice President at 9 a.m. Researchers: We will then go to (___) to interview the commander and return. Commander: There is no police station at (____). Negotiating Local Access Once approval had been granted by the national office, we entered the next level of access negotiations. These took place at each of the 11 regional command points across the country. The letter approving our application informed regional commanders that ‘approval has been granted to liaise with respective regional commanders to conduct the research exercise’. However, as documented by other researchers, initial official access to the police for research does not necessarily guarantee a fixed and permanent space (e.g. Marks, 2003; Reiner, 1991). It merely confers putative access that was to be 6 Criminology & Criminal Justice 00(0) consolidated through continuous (re)negotiation with commanders at the regional, dis- trict and divisional levels in a spatially fluid context. The officers who first met us at the station scrutinized our letter from the national office and directed us to the orderly room. At the orderly room, we showed a copy of the letter of approval to the officer in charge (often an inspector or chief inspector), who further scrutinized it and asked us to see the staff officer to the commander. The staff officer also examined the letter and asked the same routine questions: ‘Who are you? Where are you coming from? Do you have an appointment?’ At this point, we could tell from his verbal and non-verbal communication that he regarded us as outsiders. He called an officer from the orderly room to check from their files whether they had received the approval letter (called a ‘signal’) from the national police office. After 30 minutes, the assigned officer returned to say there was no such signal on file. This situa- tion, where there was no trace of official approval on file, meant that technically the staff officer could ask us to leave since they acted only on official directives received directly from the national office. While waiting for that decision, we engaged the officer in friendly chats and tried to find ways and common grounds that could shift us from being regarded as outsiders to partial insiders or, at worst, non-intruders. Having taught at the police college and being familiar with the college system, we tried to guess the year that officer had entered the police college. ‘You are course 000, right?’ The ability of the researcher to put the rank to a year group often caught the attention of officers by con- veying to them an indication of our familiarity with the police service. In this case, we were close to his course number. This helped us; the officer suddenly opened up. Eventually, we were permitted to see the commander even without the letter of approval from the national office. Our ability to identify and understand how we were perceived (i.e. as outsiders), coupled with our capacity to negotiate our way around the contours of this administrative barrier, helped us to obtain an interview. Without this prior knowledge, we would have wasted 10 hours’ travel time from and to Accra because there was no letter of approval on their file. We had moved from being regarded as intruding outsiders to outside-insiders, and were then extended some courtesies because we could understand and identify how we were likely being perceived and what appro- priate steps we had to take immediately. The assistance from the staff officer largely stemmed from our knowledge of the Ghana Police Service. Our experiences in Ghana were, therefore, similar to those reported by Marks (2003: 45) in South Africa: ‘access in the police requires constant negotiation. Researchers have to negotiate access to different levels in the police, and to different police settings.’ At each level we encountered the assistance – or resistance – that staff officers could pro- vide once they perceived us as outsiders of credibility. In fact, there are various barriers and opportunities that researchers studying police bosses in an African country may encounter as they seek to gain, and perhaps exploit, access. During our fieldwork we had encountered ‘vertical’ and ‘horizontal’ barriers and opportunities. The horizontal barriers were those at the same level of command, and the vertical ones were those emanating from a higher or lower level of command. We illustrate these terms from our field diary: The divisional commander is unwilling to speak to us. He is one of the few who is being difficult, although a colleague had recommended us to him and briefed him on the purpose of Sowatey and Tankebe 7 our study. We have been to his office on four occasions, but he had not been helpful. We need him to tell his district commanders we will be interviewing them. This is quintessentially a vertical barrier. (Extract from fieldnotes) The officer we met at the commander’s office is unfriendly. We went to the orderly room and the person in charge was rude. He said we should sit down and wait. Then, after 30 minutes, we went back to remind him he had asked us to stay. He gets infuriated and shouts, ‘Why are you bothering me? When we come to your office we have patience.’ Then the bodyguard of the commander also tells us to wait. We wait for an hour but he says nothing. He sees us as intruding outsiders. These are horizontal barriers! Barriers at the same level. Luckily, the commander comes out and ushers us into his office for two hours. The bodyguard is shocked and afraid that we are unknown notables. (Extract from fieldnotes) It follows that the researcher’s ability to position her- or himself as an appropriate insider or friend of the police becomes an essential identity and an asset for successful field- work. The reason is that, from our experience, it is not enough to be an insider, since not everyone regarded by officers as an insider can be trusted with sensitive information. Thus, whether they are an outsider or an insider, a critical variable is whether a person is perceived as trustworthy. On Being Outside Insiders? Some scholars argue that there are no certain advantages or disadvantages associated with stances taken by researchers. Being an ‘insider’, for example, can facilitate research but it can also mean that a researcher encounters problems that might compromise effective fieldwork (Goffman, 2014; Jipson and Litton, 2000; Koenig et al., 2003; Thurman, 2017). Nevertheless, some researchers have challenged this compartmentalization. Thus Gokah (2006: 63) argues that ‘conceptually, an “insider” may become an “outsider” within the same country, province, region or even social grouping’. The inherent limitations of a binary insider–outsider status became apparent during our fieldwork. We had distinct backgrounds and identities that proved facilitative. The lead author had worked closely with the Ghana police at different levels for over a decade. For instance, he provided expert support on security management at election times, and on youth violence and community policing in low-income urban communities. Additionally, from time to time he taught senior officers at the national police college. This gave him extensive knowledge of the different cohorts of officers and of the training programme, knowledge which was to prove decisive in some of the interviews. The second author had spent several years living at a police barracks, where he became acquainted with police practice; he subsequently developed a professional relationship with the police, and undertook research with frontline officers. In the months before seeking formal approval for the current project, both authors had engagements with the police manage- ment team at the national office on a pilot study of body-worn cameras which subse- quently fell through. We expected that these positions would facilitate access, as indeed they did, since we were sometimes perceived as outside-insiders. However, our positions were not static. They evolved in response to repeated assess- ment by some of the officers. Thus, officers might initially perceive us as mere outsiders, 8 Criminology & Criminal Justice 00(0) but our identity could change as we became outsiders of some standing and, further, came gradually to be viewed as outside-insiders of high repute who could be trusted with sensi- tive information. However, a researcher who is considered as an outsider of high repute can come, within a short time, to be regarded as an outsider because she or he has breached the norms and informal rules and regulations that govern outside-insider status. In other words, one needs to understand the ethics of a trusted outside-insider in order to maintain that status and the privileges that go with it. If a researcher is a vocal anti-corruption advo- cate and consistently criticizes police corruption or leadership in the media, they can be re-assessed and viewed as an intruding outsider whose access has to be limited. Police personnel, just like researchers, have multiple and overlapping identities (Bourke, 2014), and each of these identities contains both perceptions and expectations. For instance, some police personnel gave us information because they perceived us to be trustworthy insiders, even though we had criticized some police actions on corruption and human rights abuses. Such information included details of personnel strength and operational resources such as vehicles and weekly fuel supply, which we believed would be useful for assessing the institutional capacities of these police areas. A few of them refused, but those who acceded to our requests would often tell us: ‘I’m not supposed to give you this information, but because I have known you for years…’, or, simply, ‘I would usually not give out this kind of information’. Our experience further suggests that a researcher needs to develop the ability to assume and maintain different and overlap- ping identities in order to access vital information (Kezar, 2002). This does not breach generally accepted principles of research ethics. Rather, it is a practical way to harness and use the benefits embedded in these identities. During our fieldwork, how we were perceived and how we wanted police officers to regard us evolved from intruding outsid- ers to reasonable outsiders to outsider insiders of high standing and back to intruding outsiders. This simplified categorization obscures a highly contingent and complicated affair, characterized by unexpected difficulties of varying stubbornness. The complicated undercurrent of this enterprise of studying senior police officers can easily deceive a naive researcher, who might be tempted to assume that this is a natural and smooth flow. In the case of the first author of this article, it took over a decade of consistent infor- mal and formal working relationships with officers at various levels of the Ghana Police Service to gain the ‘privileged’ status of a trustworthy outsider–insider. In other cases, access was facilitated by recommendations from respected senior officers. Our experi- ence was consistent with the observation by Reiner and Newburn (2008: 357) that access often depends on a researcher’s ability to ‘forge links with crucially placed individuals’. As they go on to argue, ‘the researcher’s relationship with different individuals being studied requires constant and delicate negotiation, and also the reflexive awareness of how the subjects’ perceptions of the researcher can alter the material’ (2008: 358). During our in-depth interviews, different kinds of spaces were created between ourselves and some of our respondents. There were some who widened a space (for discussion) based on our questions and how they perceived us. Our fieldnotes show this in detail: When we asked him the question about promotion in the service, the respondent got angry at the police as though he had been discriminated against and wanted someone who would listen to his pain and frustration. ‘I have been at one rank for over six years instead of four or five Sowatey and Tankebe 9 years, and without any disciplinary charges. Even my juniors have been promoted over me.’ He went on to show how there are informal structures in a service that favours cronies. He went on and on and on. He had widened the space available to us to explore in more detail issues that were connected to perceived (or real) injustices and discrimination in the service; and how being connected to a ‘godfather’ enhances one’s promotions and career prospects in general. On the other hand, when we asked about corruption among his officers, he said it was a perception and did not want to stay on that subject too long. He had shrunk the space and time available to scrutinize the embarrassing and thorny subject of officer corruption. In being asked this question about officer corruption, he may have perceived us as intruding outsiders, and his personal interest was at play in any space he created for whatever subject/discussion. (Extract from fieldnotes) Thus, the willingness of a respondent to answer and fully explore a question may be indicative of how they perceive a researcher. This issue of common space confirms the assertion by England (1994) that research brings an investigator and his or her informant into a shared space in which identities and mutual perceptions, and how the other should perceive them, all come into play. As Bourke (2014: 1) noted: ‘identities of both researcher and participants have the potential to impact the research process. Identities come into play via our perceptions, not only of others, but of the way in which we expect others to perceive us.’ Handling the Interview A few senior officers were initially sceptical, and hesitant to provide information that they considered ‘sensitive’. This was evident in their tone of voice, facial expressions, brevity of response to initial line of interviewing and other non-verbal cues. The initial indicators typically signified that we were being perceived as intruding outsiders. Whenever we saw these early warning signals, we sought to change our positionality from intruding outsiders to respectable outsiders who could be trusted with sensitive information. We developed a flexible strategy, with elements of adaptability, focus and innovation intended to help us create the right space and environment for both access and interviews. For instance, in what we called the chameleon interviewing technique, we adapted quickly to the peculiarities of particular commanders and situations by changing both the sequence of our interview schedule and how we approached and answered ques- tions from gatekeepers. Indeed, the elements of adaptability enabled us to adopt the right positionality, which created the appropriate shared space and a congenial environment for interviews some- times lasting up to three hours. We did this by finding the common background, staying on respondents’ interests for a while and listening to their frustrations and enthusiasms. It was often time-consuming, but useful as preparation for a lengthy session. For instance, our interview schedule incorporated five interconnected themes: (1) personal and career history; (2) police function; (3) crime and crime control; (4) internal management; and (5) international experience. However, in developing the right rapport for the interview, we tried to identify areas that engaged the respondent and focused on those themes for a little longer before moving to other themes. This created a delicate situation and 10 Criminology & Criminal Justice 00(0) a ‘tight-rope’ to walk. The reasons were complex. The fieldwork covered the whole country, and this meant a great deal of travel, sometimes deep into the night, with impli- cations for safety on account of road accidents and highway robberies. We had to maxi- mize the value of our time by focusing on major issues relating to our broader research aims. In practice, it meant assessing the responses and making a judgement as to when to ask difficult questions and when to ask the friendly ones that interested them most. We realized that it was ill-advised to stay on difficult questions for too long because doing so appeared to create an avoidable anxiety. Moving as quickly as possible to more innocu- ous topics often relaxed our respondents in quite profound ways. For example, when the respondents were more relaxed, we then gradually asked a combination of what we judged to be sensitive questions. These included questions on police corruption, political interference in police work, or suspects’ rights; friendly questions included those on welfare, promotion, recruitment for United Nations peacekeeping missions and the com- mander’s personal legacy. Moreover, the context of the interviews should be borne in mind. We were recording serving police commanders on sensitive issues during elections and there was an under- standable anxiety that, if sensitive responses became public, politicians could victimize officers. These risks are present in almost all countries and to varying degrees. However, the Ghanaian political culture appears – in comparison to the United Kingdom, for exam- ple – particularly vindictive. Ghana is reputed as a ‘beacon of democracy’ in Africa (Gyimah-Boadi, 2010) and is rated highly by international pro-democracy organizations. Yet, there is minimal protection for public office holders, including police officers. Perhaps the most potent vindictive tool against a police officer is police transfers. It is common practice for officers perceived to be aligned to opposition political parties to be transferred to remote police posts or be assigned to roles that keep them from the public space. These features of the political and policing milieu meant that the interview envi- ronment had to be congenial. We, therefore, sought constantly to develop an appropriate balance between a need for a relaxed atmosphere and our intention to ask challenging questions. The essence of having the right balance and sequence of questioning was to ensure we were not perceived as intruders or typical outsiders who were prejudiced against the police and wanted data to confirm such a prejudice. This strategy became the main vehicle that helped us to assume different positionalities, from that of being per- ceived as untrustworthy intruding outsiders to persons whom commanders could trust to maintain confidentiality when handling interview data. A key tool that put officers at ease and encouraged them to discuss what they initially considered ‘too sensitive’ to be discussed and tape-recorded was deference. Jourdan et al. (2017: 234) explain that deference ‘is a strategic behavior aimed at establishing and maintaining relationship with others reluctant to do so’. Deference can be either asym- metric, as between people of unequal status and power (Collins, 2000; Henrich and Gil- White, 2001; Jourdan et al., 2017), or lateral, as between peers (Collins, 2000; Fragale et al., 2012). Deference can be verbal or non-verbal and is generally intended to ‘con- vince the recipient that the message sender is yielding, appeasing, and honouring the recipient’s position in the rank order’ (Fragale et al., 2012: 374). Demeanour is an impor- tant element in deference. As Goffman (1956: 489) argued, ‘good demeanor is what is required of an actor if he is to be transformed into someone who can be relied upon to Sowatey and Tankebe 11 maintain himself as an interactant, poised for communication, and to act so that others do not endanger themselves by presenting themselves as interactant to him’. Jourdan and his colleagues argue that deference is a mechanism for gaining accept- ance to groups that are inclined to view outsiders as ‘impure foreigners’ (Jourdan et al., 2017). Deference was also key in dealing with the commanders themselves. As a rule, we always began our interview by telling respondents that we were there to learn from their rich experience and that we were, as it were, their students. We never immediately indi- cated that we held different opinions or showed disapproval through verbal and/or non- verbal communication. Rather, we nodded approvingly throughout answers to specific questions to show that we were not challenging our informants. However, where any claims were factually inaccurate, or we held contrary interpretations of facts, we would return to those much later. This was especially true in their reactions to certain pro- nouncements by the courts. Ghanaian culture is known for its respect for authority, both in traditional and formal-bureaucratic spheres. In disputes with the elderly or those in authority, the cultural presumption is that such persons are always right, or they should not be openly challenged and contradicted (Assimeng, 1999; Nukunya, 2003). To be sure, Ghanaian culture encourages robust debate even across generations and gender. However, these are governed by norms and principles violation of which can prove con- sequential. For example, we never addressed any of the officers by their first names as that would have been culturally a faux pas. We were, therefore, sensitive to cues that might be misinterpreted as disrespectful and arrogant, and how such misinterpretations might affect the interviews. Consequently, putting contrary views to these commanders required us to be mindful of the tone. Throughout the interviews, there was a certain degree of formality that reinforced the situated power asymmetries between us and the officers. It can be said that we consistently assumed the position of ‘inferiors’ or ‘impure foreigners’, and as Jourdan et al. (2017: 242) note, such deference has ‘self-esteem cost’. Both deference and a congenial environment allowed officers to speak more candidly and extensively on the various topics of interest. Our experience reinforces the claim by Reiner and Newburn (2008: 358) that: Although gaining entry is always a problem, and trust needs to be continuously cultivated, many police officers are only too glad to tell you of their views and experiences once initial barriers have been overcome, and exiting may be as problematic as gaining access in some instances. Ethics Many researchers have discussed various ethical issues in diverse settings (Goffman, 2014; Guillemin and Gillam, 2004; Hammersley and Atkinson, 1995). A key ethical issue in our study concerned confidentiality. Our research participants were serving police officers invited to discuss a wide range of topics, some of which – such as politics and police, corruption within the organization and aspects of internal management – many of them found highly sensitive. The interview schedule included an information sheet detailing the purpose of the study, why we had chosen particular informants for the study, assurances of confidentiality and informants’ rights. It also included permission to 12 Criminology & Criminal Justice 00(0) tape-record the interviews. We either read these out or gave a copy to the commanders for their own perusal. In all cases, consent was verbal. However, once the tape-recording started, we sought confirmation that they had granted consent for the interviews and the recording. We did this in order to obtain a record of the consent that had been sought and granted. We offered assurances and promised, as Reiner and Newburn (2008: 358) advised, to turn off our dictaphone if they worried about specific responses. Only in five of the interviews did an officer make such a request. Other respondents paused at sensi- tive questions to seek reassurance that we were academics as we claimed and not under- cover journalists. Only one commander refused the tape-recording, and so handwritten notes of that interview were made. Overall, we found that assuring officers of confiden- tiality and specifying their rights to refuse participation or withdraw from the interviews without having to give any reasons proved sufficient. Researchers often face ‘ethically important moments’; that is, ‘difficult, often subtle, and usually unpredictable situations that arise in the practice of doing research’ (Guillemin and Gillam, 2004: 262). There is a widespread public perception that the Ghana police is corrupt, and media reports often document instances of corrupt behaviour from street- level to the top hierarchy (see Citifmonline.com, 2017; Foltz and Opoku-Agyemang, 2015). We, therefore, feared we would encounter situations where gatekeepers demanded bribes to facilitate access to the commanders. However, our fears did not materialize. We faced different sorts of ethical moments. The first concerned the declaration of our inten- tions. There were a few occasions where we thought it would have been easier to obtain rapid access to a commander if we had deceived gatekeepers by saying that we were police officers from the national office or officials from the Ministry of the Interior, which had political oversight for the police. This temptation arose in situations in which some officers had assumed we were officials conducting an inquiry for the Ministry. The situation reminded us of what Hammersley and Atkinson (1995: 56) had observed with respect to ‘deceive or not to deceive’. We chose to resist the temptation for three inter- related reasons: first, it would have been unethical to engage in deception; second, mak- ing false representations to a public official with the intent of deceit is a criminal offence in Ghana; third, we would have left a negative legacy where other researchers would suffer on account of our dishonesty. A different kind of ethical dilemma arose when our vehicle developed a mechanical fault. It was about midnight, while we were travelling to northern Ghana for interviews with divisional and district commanders. Residents at a hamlet assisted us with flashlight torches to assess the extent of the fault and advised us to drive to a town that was 20 kilo- metres away where wayside mechanics could help us fix the problem. Given the time of the night, we decided, upon arriving at the town, to approach the local police station for advice on neighbourhoods that would be safer for overnight accommodation. When we introduced ourselves to the officers, they appeared to have formed the impression that we were insiders or associated police officers. We had some pictures we had taken at the national police offices with senior police officers, some of whom the officers at the town knew. Our intention was to re-assure them of who we claimed we were, but it had the unintended consequence of an ethical dilemma: the officer in charge assigned the official patrol vehicle to drive us to a hotel in town and the driver was instructed to pick us up in the morning and take us to a mechanic. Herein laid the dilemma: to acquiesce to this Sowatey and Tankebe 13 directive could entail unwitting abuse of research power or status; yet, without knowl- edge of the security assessments that might have informed the directive, to refuse it could have meant risk to our safety. In the end, we acquiesced. A final ethical dilemma arose from a need to strike a balance between our research needs and those of people who were in queues to see the senior officers. In almost every police station we visited, we encountered these queues. Were these relatives or friends on ambush visits, or were these victims of serious crimes in need of urgent attention? Our dilemma was twofold: first, do we press ahead with our interviews, which could last three hours without sensitivity to the waiting public? In the earlier stages of our field- work, we sought to resolve the dilemma by offering to take breaks to allow the com- manders to attend to the visitors. Most commanders welcomed the offer. However, what was meant as a short break soon became a long wait because some of the visitors appeared to have problems that were too complex for a speedy resolution. Some commanders felt exhausted and would ask for interviews to be rescheduled. This tended to disorganize our fieldwork plans, with important budget implications. As these experiences became fre- quent, we became less sensitive, choosing to complete our interviews without humanitar- ian breaks. The other aspect of our queuing-related dilemma concerned having to ‘jump queues’. On countless occasions, gatekeepers ushered us into the offices of the commanders with- out having to wait our turn in long queues. MacCormick (2007: 14) argues that unless a person has a compelling justification, jumping a queue amounts to ‘unwarranted self- preference’ that violates the ethics of efficiency and fairness governing such a social practice. As a ‘socio-moral institution’ (MacCormick, 2007: 15) queuing in Ghana appears to have its own power dynamics such that whether a person waits in a queue or not, and for how long, may depend on the distribution of power – symbolic or material. However, we had, in many of those instances, travelled several hours, and that might have informed the decision to facilitate our access without having to wait our turn in queues. Yet, in jumping queues, a researcher potentially gains unfair advantage over the powerless who might have been waiting hours for their turn. Not surprisingly, we some- times overhead murmurings of resentment from those who had been waiting in these queues. Conclusion The commanders in our interviews occupied critical roles in policing; they were ‘the prime movers of law and order policy within their areas’ (Reiner, 1991: 4). Globally, police studies have focused disproportionately on frontline officers, or what Beek and his colleagues (2017) called ‘street level view’. That focus has produced rich data, with important insights into the character of governance, the rule of law, human rights and accountability. While these are global problems, they are acute in sub-Saharan Africa where scholarly work on criminal justice agencies, especially the police, remains unde- veloped. This article draws on our qualitative study of senior police officers in Ghana to explore the practicalities of conducting research with elite police in the African context. Two key issues arose from our fieldwork. 14 Criminology & Criminal Justice 00(0) The first concerns access negotiation, especially the centrality of informal networks within the police organization. We use the term strategic ambush to capture a willingness to move beyond the procedural formalities of appointments and gradualism towards mak- ing surprise visits. A fear that these visits risked incurring the commanders’ displeasure proved unfounded; ambush visits are a feature of Ghanaian social life, and therefore not unknown to the officers we interviewed. On several occasions, our interviews were inter- rupted by such from friends, political and traditional elites, and relatives of the officers. Our strategic ambush visits proved fruitful through the symbolic message of commitment that they conveyed. Any impression that the quest for access was perfunctory would only accentuate the delays inherent in an already inefficient bureaucracy. However, it would be a mistake to attribute access difficulties only to the perceived perfunctory nature of a research project. As Blaustein (2015: 89) argues, in his ethnographic study in Bosnia and Herzegovina, the reluctance to grant access can be motivated by the view that: allowing an outsider to access [organizations] for the purpose of interpreting the activities and discourses that influence policy meaning and content is potentially risky because this level of transparency has the potential to undermine the ability of the host institution to legitimate their participation in the policy translation process. The second key lesson from our fieldwork concerns the dual role of deference. We encountered many ‘institutional gatekeepers’ (Blaustein, 2015) as we negotiated access to the commanders at the local level. These were staff officers whom we needed to facili- tate our access at the district, divisional and regional levels. The police sub-culture litera- ture suggests that police officers sometimes maintain a ‘them-versus-us’ mentality when they interact with civilians (see, for example, Herbert, 1998; Skolnick, 2011). Marks (2003: 64) also notes that ‘police organisations are generally closed organisations’. We found that deference towards gatekeepers was critical to obtaining access to the critically important sites of the local police stations where the interviews were conducted. Deference tendencies are common feature of traditional Ghanaian culture, being ‘more pronounced in the domain and presence of state authority’ (Hasty, 2005: 74). New researchers should be aware of the verbal and non-verbal cues of deference, giving them a resource that helps to overcome institutional and procedural barriers. During inter- views, deference towards commanders helped to put them at ease and to create an envi- ronment to enhance the quality of interviews and data. We also found that our knowledge of police history in Ghana and even of apparently minor things such as our knowledge of the cohort to which a senior officer belonged made a difference. Finally, the Ghanaian research landscape has its ethical challenges. Guillemin and Gillam (2004: 264–265) define an ethical dilemma as ‘a situation in which there is a stark choice between different options, each of which seem to have equally compelling ethical advantages and disadvantages’. Should a researcher use deception to overcome bureaucratic hurdles mounted by institutional gatekeepers? Does the researcher take up offers to use public resources or let them pass? How should the researcher resolve the conflict between research needs and the unknown but potentially pressing needs of visi- tors to police stations? The researcher has to decide how to respond to these dilemmas. These were some of the dilemmas we faced. We believe we made appropriate choices. Sowatey and Tankebe 15 Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) received no financial support for the research, authorship, and/or publication of this article. References Assimeng JM (1999) Social Structure of Ghana: A Study in Persistence and Change. Accra: Ghana Publishing Corporation. Beek J (2012) ‘There should be no open doors in the police’: Criminal investigations in northern Ghana as boundary work. Journal of Modern African Studies 50(4): 551–572. Beek J, Göpfert M, Owen O, et al. (2017) The Police in Africa: Street Level View. London: Hurst. Blaustein J (2015) Speaking Truths to Power: Policy Ethnography and Police Reform in Bosnia and Herzegovina. Oxford: Oxford University Press. Boateng FD, Makin DA and Yoo J (2014) Let me speak: Officer perceptions of community mem- bers in Ghana. International Criminal Justice Review 24(1): 22–38. Bourke B (2014) Positionality: Reflecting on the research process. The Qualitative Report 19(33): 1–9. Bradford B, Huq A, Jackson J, et al. (2014) What price fairness when security is at stake? Police legitimacy in South Africa. Regulation & Governance 8(2): 246–268. Citifmonline.com (2017) Timbillah fired over 2015 recruitment scam. Available at: http://citif- monline.com/2017/01/25/timbilla-fired-over-2015-recruitment-scam/ (accessed 19 May 2018). Collins R (2000) Situational stratification: A micro-macro theory of inequality. Sociological Theory 18(1): 17–43. England KV (1994) Getting personal: Reflexivity, positionality, and feminist research. The Professional Geographer 46(1): 80–89. Faull A (2016) Measured governance? Policing and performance management in South Africa. Public Administration and Development 36(2): 157–168. Faull A (2018) Police Work and Identity: A South African Ethnography. London: Routledge. Foltz JD and Opoku-Agyemang KA (2015) Do higher salaries lower petty corruption? A Policy Experiment on West Africa’s Highways. Department for International Development, 1 January, 1–47. Fragale AR, Sumanth JJ, Tiedens LZ, et al. (2012) Appeasing equals: Lateral deference in organi- zational communication. Administrative Science Quarterly 57(3): 373–406. Goffman E (1956) The nature of deference and demeanor. American Anthropologist 58(3): 473– 502. Goffman A (2014) On the Run: Fugitive Life in an American City. Chicago, IL: University of Chicago Press. Gokah T (2006) The naïve researcher: Doing social research in Africa. International Journal of Social Research Methodology 9(1): 61–73. Guillemin M and Gillam L (2004) Ethics, reflexivity, and ‘ethically important moments’ in research. Qualitative Inquiry 10(2): 261–280. 16 Criminology & Criminal Justice 00(0) Gyimah-Boadi E (2010) Another step forward for Ghana. In: Diamond L and Plattner MF (eds) Democratization in Africa: Progress and Retreat. Baltimore, MD: Johns Hopkins University Press, 137–151. Hammersley M and Atkinson P (1995) Ethnography: Practices and Principles. London: Routledge. Hasty J (2005) The Press and Political Culture in Ghana. Bloomington, IN: Indiana University Press. Henrich J and Gil-White FJ (2001) The evolution of prestige: Freely conferred deference as a mechanism for enhancing the benefits of cultural transmission. Evolution and Human Behavior 22(3): 165–196. Herbert S (1998) Police subculture reconsidered. Criminology 36(2): 343–370. Hertz R and Imber JB (1993) Fieldwork in elite settings: Introduction. Journal of Contemporary Ethnography 22(1): 3–6. Higley J and Burton M (2006) Elite Foundations of Liberal Democracy. Boulder, CO: Rowman & Littlefield. Hills A (2007) Police commissioners, presidents and the governance of security. Journal of Modern African Studies 45(3): 403–423. Hills A (2008) The dialectic of police reform in Nigeria. Journal of Modern African Studies 46(2): 215–234. Hornberger J (2011) Policing and Human Rights: The Meaning of Violence and Justice in the Everyday Policing of Johannesburg. Oxford: Routledge. Jipson JA and Litton CE (2000) Body, career and community: The implications of researching dangerous groups. In: Treweek GL and Linkogle S (eds) Danger in the Field: Risk and Ethics in Social Research. London: Routledge, 147–167. Jourdan J, Durand R and Thornton PH (2017) The price of admission: Organizational deference as strategic behavior. American Journal of Sociology 123(1): 232–275. Kezar A (2002) Reconstructing static images of leadership: An application of positionality theory. Journal of Leadership Studies 8(3): 94–109. Koenig BA, Back AL and Crawley LM (2003) Qualitative methods in end-of-life research: Recommendations to enhance the protection of human subjects. Journal of Pain and Symptom Management 25(4): S43–S52. MacCormick N (2007) Institutions of Law: An Essay in Legal Theory. Oxford: Oxford University Press. Marks M (2003) Policing ethnography. Society in Transition 34(1): 38–69. Nukunya GK (2003) Tradition and Change in Ghana: An Introduction to Sociology. Ghana: Ghana University Press. Ostrander SA (1993) ‘Surely, you’re not in this just to be helpful’: Access, rapport, and interviews in three studies of elites. Journal of Contemporary Ethnography 22(1): 7–27. Reiner R (1991) Chief Constables: Bobbies, Bosses or Bureaucrats? Oxford: Oxford University Press. Reiner R and Newburn T (2008) Police research. In: King R and Wincup E (eds) Doing Research on Crime and Justice. Oxford: Oxford University Press. Skolnick J (2011) Justice without Trial. New Orleans: Quid Pro. Sowatey EA and Atuguba A (2014) The complexities of local notions of (in)security and (in) justice. In: Albrecht P and Kyed HM (eds) Policing and Politics of Order Making. London: Routledge. Tankebe J (2009) Public cooperation with the police in Ghana: Does procedural fairness matter? Criminology 47(4): 1265–1293. Tankebe J (2010) Identifying the correlates of police organizational commitment in Ghana. Police Quarterly 13(1): 73–91. Sowatey and Tankebe 17 Tankebe J (2014) The making of ‘democracy’s champions’: Understanding police support for democracy in Ghana. Criminology & Criminal Justice 14(1): 25–43. Thurman L (2017) Somewhere between green and blue: A special police unit in the Democratic Republic of Congo. In: Beek J, Göpfert M, Owen O, et al. (eds) The Police in Africa: Street Level View. London: Hurst. Walker A and Lidz C (1977) Methodological notes on the employment of indigenous observers. In: Weppner R (ed.) Street Ethnography. Los Angeles, CA: SAGE. Author biographies Emmanuel Addo Sowatey is a Research Associate at the Centre for Gender Studies and Advocacy, University of Ghana. He is also a part-time Lecturer at the Ghana Police Academy, and the Ghana Police Command and Staff College. Emmanuel provides expert commentary on a range of secu- rity-related issues for numerous national and international media outlets. His research interest includes policing, elections security management, youth and violence/security, and illicit prolif- eration of illicit small arms. His most current research projects are: (1) police elites in Ghana; (2) understanding why micro disarmament fails in relatively stable African states; (3) plural legal systems in Africa and women in the informal economy of Africa. Justice Tankebe is a University Lecturer in Criminology and a Fellow at St Edmund’s College at the University of Cambridge. Justice’s research interests are in policing, legitimacy and criminal justice, vigilantism, corruption and counter-terrorism. His most recent projects include police elites in Ghana, corruption among prospective elites in Ghana and an EU-funded project on coun- ter-terrorism legitimacy and recruitment into terrorism in the UK.