Conducting Prison Research in a Foreign Setting

This paper discusses the process of conducting prison research in France. Drawing on a study conducted with a sample of prisoners in a maximum‐security facility in Paris, this article outlines the major challenges relating to access, data collection, and dissemination of results in correctional research. It also addresses some of the barriers that are inherent to prison research conducted in a setting foreign to the researcher. The value and place of prison research in the field of criminology are also discussed.


Introduction: France as a setting for prison research
In 2006, the Council of Europe revised the European Prison Rules, originally developed in 1973. 2 These rules provide guidelines to the member states of the Council of Europe on the humane and just treatment of prisoners (47 member states at the time of this writing; 46 in 2006). Following these recommendations, a correctional law was enacted in France in 2009, which guaranteed certain rights to incarcerated individuals. These include the right to obtain identity papers, to vote, to have access to social aid and employment opportunities, to participate in training programs, to maintain family ties, to have reasonable access to telephone services and, for some categories of individuals (for example, elderly prisoners), to benefit from reduced prison time. These recent changes created a timely opportunity for a research study investigating the prisoners' perceptions of the quality of life in prison, and how these perceptions relate to their assessment of prospects for desistance from crime and reintegration into the community after release.
In addition to the opportune climate for prison research resulting from the revision of the European Prison Rules, the socialist government in France has made efforts to reform sentencing, prison and reentry policies. In his 2012 electoral campaign, President François Hollande promised to reverse the punitive policies implemented by the Sarkozy administration (2007)(2008)(2009)(2010)(2011)(2012), and more specifically repeal a law which had increased the recourse to mandatory minimum sentencing (Law #2007-1198, enacted on 10 August 2007). Consistent with these priorities, the Justice Ministry organized a consensus conference on the topic of recidivism prevention in February 2013. I was invited to present at this conference, along with other IJCJ&SD 114 Online version via www.crimejusticejournal.com © 2015 4(1) academics, practitioners, and victims of crime. This initiative highlighted the Justice Ministry's interest in gaining a better understanding of issues affecting the prisoner population and its willingness to invest in initiatives that promote reintegration after prison. The conference led to a series of penal reforms, spearheaded by the Justice Minister and enacted on 15 August 2014, which aimed to reduce recidivism and minimize the burden on the prison system. These reforms were widely criticized by right-wing opponents for being 'soft on crime'. The two most significant features of this law include a call for increased recourse to probation as an alternative to incarceration, and an effort to minimize 'dry releases' (that is, returns to the community that do not entail any form of pre-release preparation or post-release support) by conducting an assessment of the prisoner's circumstances at two thirds of the sentence. In short, the proposed research was timely given the current French government's commitment to prison reform.
Notwithstanding the increasing prevalence of international research, the large majority of studies on prisons, desistance from crime, prisoner reentry and offender reintegration have originated from Anglo-Saxon countries. Non-English-speaking countries are more susceptible to being excluded from dialogues among academics, practitioners and policy-makers about the most effective correctional practices that may promote desistance from crime and prisoner reintegration (see Tubex 2013 for similar observations). France is a prime example of this exclusion; while it has produced some of the most influential scholars in the fields of criminology and sociology (namely, Durkheim, Foucault, and Bourdieu), we know very little about contemporary French criminology. Part of this situation stems from the controversial discussions about the creation of the discipline of criminology in France. Ideological differences across disciplines have impeded these efforts, contributing to the further exclusion of France from dialogue with the international community of criminologists (for a detailed discussion of this issue, see Colson 2013;Herzog-Evans 2012). In addition, the fact that English is the dominant language of most prominent journals in the field as well as the major criminological associations (for example, the European Society of Criminology and the American Society of Criminology) further contributes to the exclusion of France and other non-Anglophone countries from conversations about correctional and other criminological issues. 3

The current project
In the spring of 2013, I conducted in-depth interviews with a sample of 59 male inmates housed in a maximum-security facility (maison centrale) in Paris. 4 The study sought to assess the perceptions of quality of life in prison as reported by the inmates themselves, and to document potential barriers to desistance from crime and a successful eventual return to the community. The research employed a mixed-methods approach and included both structured quantitative scales and in-depth narratives. In order to better understand the major impediments to the reentry process of inmates after their release from prison, particularly among individuals serving long sentences, the interview assessed how individuals adapt to prison life. Interviews documented detailed information about the physical and mental health of prisoners, substance use habits, experiences of victimization, and support network. The research also investigated the various programs and services implemented to prepare individuals for their return to the community and to enhance efforts to desist from crime, and questioned prisoners about their perceptions of the quality of these services. Study participants were asked about the requirements for a successful social reintegration after prison. The project ultimately aimed to inform the correctional administration about effective practices with regards to preparation for reentry, desistance from crime and from other problematic behaviors, and thus the prevention of recidivism.
This paper describes the processes involved in conducting this research, some of which are specific to the foreign context in which I was operating and others that are applicable to prison research more generally. As a researcher with a French-Canadian background working in the IJCJ&SD 115 Online version via www.crimejusticejournal.com © 2015 4(1) United States, I discuss some of the issues that may arise when conducting prison research in a foreign setting. This paper also underlines the crucial importance of research linking up the worlds of prisons, desistance and prisoner reentry.

Background: A brief overview of the French prison system 5
Because there is limited information published in English on the French correctional system, I offer a brief overview of the structure of the prison system. In France, correctional services fall under the authority of the Justice Ministry. Correctional statistics are updated monthly on the Justice Ministry's website (Direction de l'Administration Pénitentiaire 2013 There are five general types of adult correctional facilities in France: maison d'arrêt, maison centrale, centre de détention, centre de semi-liberté and centre pour peines aménagées. A maison d'arrêt houses three categories of individuals: those awaiting trial, those serving sentences of less than two years, and those who have been sentenced to prison but are waiting to be transferred to a correctional facility. All individuals with a prison conviction begin their sentence in a maison d'arrêt. A maison centrale is the equivalent of the maximum security prison. It houses individuals serving long sentences, recidivists, and offenders who are considered to pose a threat to the community. A centre de détention houses individuals who are deemed to have high potential for social reintegration. It may house individuals serving very long sentences; the nature of the offense is not necessarily a guiding principle in the assessment of dangerousness. Individuals usually come to a centre de détention after having spent some time in a maison centrale. Finally, a centre de semi-liberté or a centre pour peines aménagées enables individuals to maintain employment or to participate in training or treatment programs in the community, but require them to be present at the facility on nights and weekends.

Gaining access to prisons
At the time that I submitted the proposal for this research, the political climate in France was ripe for this type of project. The team that worked closely with the Justice Minister was highly supportive of my project. I was also fortunate to have met, on a previous trip to visit French correctional facilities, a high-ranking correctional staff member who was sympathetic to research, prisoner reintegration efforts, and restorative justice initiatives. He later became the director of a major maison centrale in a suburb of Paris. I contacted him before I submitted my proposal to the correctional authorities in order to discuss the objectives of the study, the resources required, and the value of the project for the prisoners, staff, and the facility. This dialogue was essential in order to ensure the director's willingness to cooperate. He was very enthusiastic about the project, and his cooperation was crucial to the success of the research study.
Despite the prison director's support, many other obstacles had to be overcome before gaining access to the facilities. All research involving human subject participants requires approval by the university's research ethics committee, otherwise known as the Institutional Review Board (IRB). The City University of New York's (CUNY) IRB required a letter of authorization from the French correctional administration in order to even begin reviewing the application. I was required to be present and meet with the French correctional administration staff in person to negotiate the conditions of the study. I arrived in Paris on 1 January 2013 for a six-month sabbatical leave. I had a series of meetings with the French equivalent of the research ethics committee, which was comprised of members of the correctional administration's research team. These meetings were challenging for several reasons. First, although my study employed a mixed-methods approach, there was some skepticism about the use of structured instruments. The quantitative approach is far from being the dominant methodology in French criminology. Most criminological work in France is either theoretical or draws on qualitative methods. I was challenged for using a number of validated instruments that have been widely employed in prior research (for example, Alison Liebling's Measuring Prison Quality of Life instrument: Liebling Hulley and Crewe 2012; Cathy Widom's Lifetime Trauma and Victimization History scale: ; personality scales; self-control measures, and so on). Second, there was a limited understanding among the research team of why I was conducting this work. A high-ranking correctional official later explained to me that academics in France seldom engage in applied research, and rarely conduct the fieldwork themselves; academics tend to produce theoretical pieces, and fieldwork is generally conducted by the correctional administration's research staff. Overall, with some notable exceptions since 2004 (Chantraine 2004;Chauvenet 2006;Mouquet 2005;Rambourg 2009;Rouillon et al. 2004), there are few prison-based studies in France, or they are poorly disseminated to the international research community. The absence of a 'culture of research' also explains the skepticism about the project that was widespread inside the facilities, which I describe below.
Once I was granted approval by the French research ethics committee, the CUNY IRB reviewed my proposal. This was an extremely stressful and lengthy process that stretched over several months. The seemingly over-cautious attitude of the IRB most likely stemmed from fears of legal ramifications resulting from a research study involving a vulnerable population in an international setting. These concerns have a historical precedent. In the late 1940s, federally funded American researchers deliberately infected nearly 700 Guatemalan inmates, mental patients and soldiers in order to test the effectiveness of penicillin. While several decades have passed since these experiments, they have re-emerged in the media in recent years (see McNeil's (2010)  In addition to the excessively restrictive requirements imposed by the IRB, I was also challenged about the questions that were to be included in the interview, which was highly unusual and beyond the IRB's prescribed mandate. For instance, I was told that I should not ask questions about expectations for reentry of prisoners who were not within two years of release, on the basis that interventions are only effective during this time period. This argument was inconsistent with the discourse of many prisoners who participated in the current study. A prevalent theme in the narratives of prisoners related to the idea that preparation for release should begin from the point of entry into the system, regardless of the length of the sentence. The IRB's argument was also inconsistent with a large body of research in criminology which has argued that desistance is best perceived as a gradual process rather than an event that occurs abruptly (Bottoms et al. 2004;Kazemian 2007;Maruna 2001). In short, the research ethics committee's comments suggested that we should overlook the needs of individuals serving life or otherwise long sentences, a view with which I strongly disagreed.
After months of negotiation, I was granted permission to begin my research. Given the variability in conditions from one type of facility to another (Liebling assisted by Arnold 2004; National Research Council 2014), the original plan was to conduct interviews with prisoners housed in three types of correctional facilities. Although the higher correctional authorities had granted me authorization to conduct my research in three facilities, the local authorities (that is, the prison directors) could, in principle, deny me access and prevent prisoners from participating in the study. As a result, the implementation of the research and the opportunity for prisoners to share their viewpoints largely depended on the support of the prison director. These power dynamics at the local level constituted an important obstacle to the implementation of the research.
With the exception of one enthusiastic maison centrale director, the prison directors of the other two facilities which I was authorized to access were not responsive to my requests. As a result, all interviews were conducted in just one maison centrale. Although the prison directors who were unresponsive never provided reasons for declining to cooperate, it appears reasonable to assume that they did not perceive this research as a priority. Again, this may have resulted from the limited presence of academics and researchers inside French prisons, and from the scarcity of correctional research more generally; this is particularly true in French jails (maisons d'arrêt), where overcrowding is a significant problem, staff are overwhelmed by the large number of prisoners under their supervision, and the presence of researchers is likely to be particularly burdensome.

Recruitment into the study
Once all authorizations were in order, letters were distributed to all prisoners in the selected facility to inform them of the research study. Those who were interested in learning more about the project responded to the letter in a sealed envelope dropped in a locked box designated for this purpose. I met with these potential participants individually to explain the purpose of the study; these steps were undertaken in addition to the informed consent process, which occurred at the beginning of each interview, once individuals had agreed to participate. Every individual who met with me agreed to participate in the research (though two individuals later changed their minds).
Efforts to recruit participants were somewhat hampered by the lack of a culture of correctional research, and a limited understanding of what I was trying to accomplish. This was resolved when I met with the potential participants and explained the objectives of the study. However, mistrust is ubiquitous in the prison environment, among inmates and staff alike (Liebling assisted by Arnold 2004). Several participants reported that members of staff discouraged prisoners from participating in the study. Some prison offficers allegedly told prisoners that the content of the interviews would not be confidential. I encouraged participants to reassure individuals with those concerns that this was completely false and to meet with me, with no obligation of participating in the study, in order to allow me to rectify this misinformation.
Being female and a foreigner most likely influenced a higher participation rate than if I was male and/or of the same nationality as those being interviewed. Jewkes (2012) drew attention to the fact that the special dynamics of female researchers conducting interviews with male inmates (or of male researchers conducting research with female prisoners) required particular consideration (for a counter-argument, see Crewe, 2014). Similarly, Claes et al. (2013) argued that the gender of the researcher inevitably involves specific difficulties and challenges in the prison environment, and that male prisoners and staff alike may modify their behaviors in the presence of a woman. Prisoners welcomed the opportunity to interact with a woman who was not in a position of authority, who addressed them in a polite manner, and who treated them with respect and courtesy. Several expressed that they were delighted to speak to 'a friendly young woman'. Claes et al. (2013) rightly suggested that it is difficult to determine if and how gender impacts the research process and findings. Specifically, the authors asked whether it is gender that matters or rather personality, friendliness, open-mindedness or other character traits.
Some study participants revealed that they had been motivated to participate in the study on the basis of my physical appearance (for similar observations, see field notes from Hanne Tournel's prison study, cited in Claes et al. 2013: 65). Ultimately, while these comments felt highly inappropriate and made me uneasy, they did not change the outcome of the interviews: most individuals became highly engaged in our exchange as the interview progressed, irrespective of the reason that motivated them to meet with me in the first place.
My status as a foreigner was also advantageous. Because of my French-Canadian background, participants did not question me when I reassured them that I did not work for the correctional administration. Some playfully mocked my accent. I found humor to be the best tool to lighten the mood and dissipate suspicion and doubt. In addition, some participants stated that they were more amenable to meeting with me because of my ethnic background. Some of the participants felt that because we had similar cultural backgrounds, I would understand their customs and code of values and thus be less prone to judging them.

Conducting the interviews
Entering a maximum-security prison in France was probably not unusual from a European perspective, but it was very different from the American experience. Like prisons in most countries, metal detectors are employed and occasional physical searches are conducted upon entry. While staff members wear distinctive uniforms, prisoners are not required to wear uniforms, nor are they bound by handcuffs or any other restrictive devices; this is fairly standard in European facilities. With the exception of meal times, when inmates remain in their cells, individuals are generally free to walk around in the facility during most of the day (though this is not the case in all French facilities). Interviews were conducted in the area of the prison where prisoners meet with visitors. Most interviews took place in a private room that was devoid of cameras. No officers were present in this room, and we were generally left alone for the duration of the interview (approximately two and a half hours, on average). This was appreciated for the purpose of the research, but surprising; almost all individuals who participated in the study were convicted of serious offenses (homicide or murder: 77.7 per cent, n=45; rape: 8.6 per cent, n=5; armed robbery: 8.6 per cent, n=5).
In addition to the obstructionist attitudes of a few staff members, which may have discouraged some inmates from participating in the study, there were other challenges. Inherently, prison research cannot be anonymous, at least during the data collection phase. While the answers are confidential, an individual's participation in the research is inevitably known to staff and inmates, as the participant needs to be called to meet with the researcher. This may have deterred some individuals from participating. Another challenge related to the difficulty in maintaining a structured schedule. While a detailed appointment list for the interviews was created, schedules were often disrupted and I had to return to the prison on separate occasions to complete the interviews. Lockdowns occurred periodically, which caused delays. In a few cases, I made my way to the prison early in the morning for an appointment only to be told that a prisoner who had previously agreed to meet with me 'was not in the mood' and refused to come down from his cell. 7 In this high-tension environment where moods change frequently and depression, sadness and despair are not uncommon, the unreliability of participants is inconvenient but not surprising. These issues are all inherent to prison research, regardless of the cultural setting.
In order to maximize the depth of the data, the study drew on the mixed-methods approach, and included both structured (quantitative) and open-ended (qualitative) questions. Various cognitive factors, such as the decision to change, or the development of a positive self-image and identity, have been found to be predictive of successful desistance efforts (see Kazemian 2015). These indicators are difficult to measure through structured questions and are typically excluded from quantitative analyses. Like Liebling (1999a), I have difficulty envisioning any prison-based data collection endeavor that would not include both structured scales and openended questions that allow participants to take the conversation in any direction that they deem relevant. Both sources of data were extremely informative, but the combination of the two was particularly valuable. Participants sometimes had trouble answering structured questions with limited response options, as they often felt the need to elaborate further. In these cases, I made note of the comment and we revisited the topic during the open-ended portion of the interview. For instance, the Measuring Prison Quality of Life survey inquires about whether the prison provides adequate facilities and means to maintain a presentable appearance. The possible answers range from strongly agree to strongly disagree. In some instances, while the respondents generally agreed with this statement, the open-ended portion of the interview revealed a more complex situation. The 'hygiene kits', which include basic personal hygiene products such as a toothbrush, toothpaste and soap, were not distributed to the prisoners. They had to go fetch them on their own in a specific area of the prison. This was a highly stigmatizing and humiliating process; requesting a hygiene kit suggests that one cannot afford to purchase basic products. Many expressed that, while the kits were available, they preferred not to collect them in order to avoid being subjected to humiliation. Without these detailed narratives, I would have wrongfully concluded that since the kits are generally available, prisoners in need are receiving basic hygiene products.

Cross-sectional data: A snapshot in time
On a related note, because moods are often transient in the prison environment and can fluctuate with situational factors, there can be some degree of variation in the responses provided by the participants on a 'good' versus a 'bad' day. This became evident in cases where we were unable to complete the interview in one sitting and had to meet again. In one case, at the time of the first meeting, the participant was unhappy with his life and his treatment in the facility. His discourse was characterized by anger and despair. We met a second time after he had been granted a day furlough. His mood was markedly different. He was noticeably less critical of life inside the facility and more optimistic about his life after prison. In another case, only a few days prior to our second meeting, the individual's request for early release had been denied. While he had been relatively calm during our first meeting, he was visibly upset, angry, and discouraged during the second meeting: ' mind in order to capitalize on periods when individuals feel optimistic and, conversely, to assist them when feelings of frustration and despair are particularly significant.
To what extent do situational circumstances taint perceptions of the quality of prison life and other dimensions as reported by study participants? As researchers, how do we address the issue that perceptions may differ significantly at different points in time? This issue is not specific to prison research; in all contexts, cross-sectional data only provide a snapshot in time. We know, for instance, that criminal careers are characterized by a high level of intermittency (or zig zag patterns; Piquero 2004), and that attitudes and behaviors may differ widely at different times or periods of the life course. Decisions to desist from crime may involve several relapses and reversals of decisions before reaching the final point of termination from crime (Burnett 2004). As such, attitudes and behaviors are also likely to fluctuate throughout this process, but we know very little about the extent of this variation. This issue requires further attention. The most effective way to assess the degree of change in responses across situational settings and prison life circumstances is to collect longitudinal data on prisoners and to systematically document changes that occur during periods of incarceration.

Is researcher objectivity the golden standard in prison research?
Liebling (1999a) raised the question of whether objectivity can be achieved in prison research, and whether it is even a desirable outcome. Her description of the emotional experience resulting from prison research remains highly accurate; researchers may feel distress, exhaustion, or satisfaction after conducting interviews with prisoners. Liebling (2001) also discussed the role of sympathy in prison research, and asked whether it enhances or undermines 'professional integrity'. She argues that empathy and the ability to relate to the individual in the context of an interview are key components of the research process. To balance this argument, Crewe (2014: 393) offered a word of caution against giving too much credence to 'field emotions, especially compared with other forms of knowledge', as these states represent, to some extent, social constructs. Crewe (2014: 401) (citing Les Back) also reminds us that the writer is the 'least important person there'.
I have been involved in various data collection endeavors with vulnerable populations (that is, formerly incarcerated individuals, adjudicated populations, and so on). As a researcher who has predominantly engaged in quantitative research, I have always felt the need to remain as objective as possible, and not to show feelings or reactions to the study participants' accounts. This was nearly impossible in the context of the current study. Interviews included questions about experiences of victimization, both in prison and on the outside. Some inmates described traumatic childhood experiences, which made it difficult to remain stoic and to avoid showing any emotion. In those instances, it became impossible to simply move on to the next question without expressing sympathy, asking about the recourse to available resources to address the trauma, and how they were coping with these traumatic experiences. While I occasionally felt as though I was introducing researcher bias and that I was not being true to my role as an objective researcher, not reacting to such stories felt unnatural. Ultimately, the participants expressed that they appreciated the attentive listening and compassionate response. Qualitative researchers remind us of the importance of reflexive thought, and raise important reflections about the role of the researcher in the research process (Crewe 2009;Jewkes 2012;Liebling 2014;Phillips and Earle 2011). These discussions have certainly changed my perception of the role of the researcher. While it may appear to be easier to remain 'objective' and to leave emotions out of the equation when working with structured questions (which are usually employed in quantitative analyses), the methodological approach is less relevant than the nature of the questions. For instance, Cathy Widom's Lifetime Trauma and Victimization History scale is a quantitative instrument, but it elicited some emotional responses that required a compassionate response from the researcher and warranted additional time for reflection and discussion.

The benefits of participating in prison research for inmates
If research ethics committees tend to over-emphasize the potential risks associated with prison research, they also underestimate the benefits for the participants. Most prisoners expressed their appreciation for the opportunity to talk to someone who was willing to listen (see also Crewe 2009;Jewkes 2002;Liebling 1999a, for similar feedback from prisoners). One study participant who reported physical and sexual abuse in childhood stated that he felt comfortable talking to me because he sensed that I genuinely listened to him, and that he would 'talk much more about these things' if he had a better rapport with the prison psychologist. Another participant stated: 'They need to listen to people. You are here today, you are listening to me, that's already a lot'. Once the researcher has established a rapport of trust with the participants and that the word spreads among potential participants that it is 'safe' to work with the researcher, there can be a transition from a status of 'outsider' to that of a (semi) 'insider' (Jewkes 2012). As stated by one of the participants: 'We feel that you are on our side, that you are here to help us'. This shift in perceived status led to more meaningful exchanges with the participants; once mistrust and skepticism were shed, there was more scope for in-depth dialogue about life in prison and beyond.
Interestingly, many of those who expressed gratitude were individuals who reported incidents of victimization and who answered the sensitive questions that could potentially cause distress. This suggests that the distress caused by the interviews may have less to do with the nature of the questions and more to do with the quality of the rapport with the interviewer, which underlines the crucial importance of interviewer training on one hand, and the researcher's genuine interest in the individual and his/her ability to empathize with the participant on the other. This hypothesis needs to be further explored in future studies by better documenting the quality of interactions between interviewers and research participants, and the respondents' feedback about their participation (see . While some prisoners felt indifferent at the end of the interview and were skeptical about the impact that the study would actually have on prison practices, none of the participants reported any negative feelings, despite the sensitive topics addressed. This is an important point for research ethics boards, which often focus on the risks of data collection involving prisoners or other vulnerable populations and neglect to consider the benefits of their participation. Bosworth et al. (2005: 259) were right in stating that '… institutional review boards have created a culture in which the protection of the scholar's institution from lawsuits has become more important than the emotional security of the research participants'. In a study inquiring about lifetime trauma and victimization experiences,  found that, while vulnerable individuals were more likely to have emotional responses in interviews, they were also more willing to continue their participation, to regard their participation as meaningful, and to agree that they were treated with respect and dignity by the interviewers. The authors concluded that: … researchers and IRBs should not be wary of conducting research on sensitive topics with potentially vulnerable populations, particularly research that has the potential for further understanding the characteristics or needs of these kinds of vulnerable populations. (Widom and Czaja 2005: 136) That no study participant expressed feelings of distress may be the result of the fact that interviews were not merely a mechanical data collection endeavor but rather a proactive dialogue in which a productive exchange about the prison experience was encouraged. When prisoners expressed discontent about a particular dimension of prison life or claimed that they would most likely reoffend when they returned to the community, the follow-up questions encouraged them to think of solutions to the problem or to better explain what could be done for them not to engage in behaviors that may get them into trouble again. A dynamic dialogue with the researcher enables the participant not only to identify potential problems but also to reflect on realistic solutions to current or future challenges. While we tend to assume that the researcher's role is merely to 'collect data', relevant and targeted questions can stimulate productive thinking that can compel the participant to reassess his role in the process of change, during the interview and beyond. While this resulted in lengthier interviews, it served as a reminder that, with an appropriate framing of questions, data collection endeavors can be beneficial to the development of both the research community and the study participants.

Disseminating results
Swift dissemination of results is important in order to demonstrate to the correctional authorities that the project was productive and would provide useful knowledge for the management of correctional facilities and preparation for eventual release. Separate reports were produced for the prison director (containing information relevant to the management of the facility and realistic recommendations on short-term changes that can be implemented) and the higher correctional authorities. Results were also presented to the prison staff, prisoners, and the correctional administration. It is equally important for correctional authorities and prisoners to feel that the research was not a waste of time and resources. It is the responsibility of researchers to demonstrate that they are trustworthy in order to ensure that correctional authorities and staff, as well as prisoners, will not be reluctant to participate in other research projects in the future.
One of the challenges in communicating findings to the correctional authorities is striking a balance between maintaining the integrity of the data and avoiding the impression that the researcher is overly critical of the system, which may lead to future reluctance to grant access to prisons for research purposes. Liebling (2001) discussed the notion of 'taking sides' in research endeavors and the delicate balance between social and political values, empathy, and social science research. Sparks (2002) also offered a word of caution about the prison researcher's role in generating some potentially unintended consequences of prison research, which may occur with some selective interpretations of research findings by authorities and the media.
This challenge was somewhat addressed in two ways. First, because the interviews provide a balanced view of life in prison, findings highlight both positive and negative dimensions (even though prisoners tend to focus on the latter). Second, instead of simply underlining the limitations of prison life, the final report offers concrete short-and long-term suggestions to improve quality of life in prison and efforts to prepare prisoners for release.

Prison research in criminology
Some important prison research has emerged over the last two decades, as evidenced by some of the studies cited in this paper. Nonetheless, there are still substantial gaps in our knowledge regarding the long-term impact of incarceration on prisoners. While '… psychological research has often found that the effects of imprisonment are largely minimal … or that prisoners cope surprisingly well despite an initial period of disorientation and serious anxieties about family and friends…' (Liebling 1999b: 284), sociologists and criminal justice scholars have repeatedly highlighted the damaging effects of imprisonment. Liebling (1999b: 287) argues that the former claim is partly biased by issues of operationalization of harm, and '…the failure of research on the effects of prison life to ask the right questions or to ask in an appropriate kind of way how imprisonment is experienced'. In addition, some prisoners may feel that they are well-adjusted to life in prison and not realize that these adjustments may affect them in a negative way upon release (Jamieson and Grounds 2005).
Much of the research that has suggested the absence of deterioration in the well-being and adjustment of prisoners over extended periods of incarceration (for example, MacKenzie and Goodstein 1985) is dated and may no longer reflect the reality of contemporary prison life. Similarly, while we know that there is a significant prevalence of traumatic experiences and mental health disorders among the prison population (see National Research Council 2014 for a review of this research), it is unclear whether prison time leads to the development of these problems, or whether it merely exacerbates a preexisting condition (Schnittker, Massoglia and Uggen 2012).
We also know very little about psychological and other changes that occur throughout a prison sentence. While it has been established that prison is a highly stressful environment (Clemmer 1958;Hassine 2004;Johnson and Toch 1982;National Research Council 2014;Sykes 1958), much remains unknown about how individuals progress and change over the course of a prison sentence. Studies documenting the perspectives of prisoners, especially over long periods of confinement, are not abundant. We need contemporary, prison-based longitudinal studies to revisit the question of the effects of incarceration over prolonged periods of time. Such studies would enable us to better understand how the desistance process operates during periods of incarceration. In addition, as stated by Jamieson and Grounds (2005), attitudes and behaviors that may be adaptive to the prison environment may be maladaptive outside prison, and longer follow-up periods extending beyond the period of incarceration are needed to better understand this issue.
Although the fields of imprisonment, desistance, and prisoner reentry are closely linked, these bodies of research have generally been poorly integrated (for a well-integrated piece on these topics, see Maruna and Toch 2005). While desistance research has been primarily driven by theoretical advancements, research on prisoner reentry has focused on the practical implications of the desistance process of formerly incarcerated individuals as they return to the community. Few researchers engage in both prison and community research, and thus few have a thorough understanding of both contexts. Findings drawn from desistance research have obvious implications for reentry practices, but these two areas of study often appear to be disjointed. The obstacles faced by formerly incarcerated individuals upon release from prison share many similarities with the impediments to desistance identified in the literature, namely strains on family relationships, physical and mental health issues, substance abuse, housing issues, lack of marketable skills, restrictive laws and policies, and unemployment (see National Research Council 2014 for a comprehensive review). Prison-based research is an appropriate bridge for these often disconnected fields of study.
Reconciling the academic and prison worlds is no easy task. These systems pursue different missions and have distinct priorities (Quina et al. 2007). The reluctance of correctional officials to collaborate with researchers may stem from the perception that academics can be overly critical of the prison system and unsympathetic to correctional staff and authorities. To establish productive partnerships between academics and correctional authorities, it is important to have an open dialogue about how research in a prison setting can benefit not only the advancement of knowledge, but also inform programs and policies that target inmate needs, improve correctional practices, reduce prison maladjustment and misbehaviors, and ultimately better prepare for release. This dialogue requires compromise on the part of researchers. For instance, in the research conducted in France, the prison director expressed interest in collecting data about restorative justice initiatives. These questions were ultimately included in the interview. 8 In my view, this was an ideal researcher-practitioner collaboration. I did my utmost to minimize disruptions to the normal routine of the prison and to avoid creating additional tasks for staff as a result of my presence, and the prison director reciprocated by not obstructing data collection efforts.
Because of the increasingly heavy burden on correctional systems and relatively limited resources, access to prisons for research purposes is inevitably selective, and it is incumbent on the researcher to demonstrate that a given study is of practical value to the correctional authorities and prisoners.

Conclusion
This paper described the processes, challenges and benefits involved in conducting prison research, specifically in a setting that is foreign to the researcher. While many of these issues are universal and have been addressed by veteran prison researchers, conditions of access and the need to adapt to the distinct cultural context of the host country are added challenges to research involving a vulnerable population in a foreign setting.
While there is still progress to be made in French prisons in order to attain the standards laid out in the European Prison Rules, it remains that efforts are being undertaken, at least in the formal discourse and public policies, to improve the situation. In my view, research that can guide these efforts is more useful than research emphasizing that French correctional facilities are not up to par with most other European countries. The former approach is likely to lead to a more productive dialogue with French correctional officials, whereas the latter approach is likely to elicit a defensive response and result in increased reluctance to collaborations with researchers.
We know that conducting research in prisons is a difficult endeavor (Bosworth et al. 2005;Liebling 1999a). Liebling (1999a: 163) argued that it is 'an enterprise made complex by the human nature of the researchers and the researched', in 'an intense, risk-laden, emotionally fraught environment', and that it 'makes demands on fieldworkers which are at times barely tolerable'. There is no question that prison is a highly stressful environment. However, in my experience, the part of the research that entailed sitting in a confined space with an individual who has committed violent acts, without supervision or cameras, was the least stressful part of the process. The administrative process required to gain access to prisons is extremely arduous. This is particularly true for research conducted in a foreign setting, as the project needs to comply with standards of both the host country and the researcher's country of origin. Given the stringent conditions imposed by the IRB, it is not at all surprising that academics are increasingly reluctant to engage in original data collection involving vulnerable populations. This is unfortunate because the potential benefits of prison research are substantial, for prisoners, researchers, administrators, policy-makers, and the scientific community. Analyses suggested that the study participants were representative of the population of prisoners housed at the facility; no significant differences emerged with regards to various individual and criminal record characteristics (nationality, age, sentence length, conviction offense, criminal history, history of violence, etc.). However, the facility included more individuals convicted of homicide and murder (66.5%, n=135) when compared to the national prevalence rate in maisons centrales (50.3%, n=802); this difference was statistically significant (χ 2 = 26.06, df=3, p < .001). 5 For a more detailed overview of the French correctional system, see Kazemian and Andersson (2012). 6 While most countries calculate the number of prisoners divided by all inhabitants, in Australian statistics only adult inhabitants are counted. 7 In some of these instances, the interviews were rescheduled. Over the course of the project, two prisoners had originally agreed to participate but changed their mind on the day of the interview and were not interested in rescheduling. 8 This decision was contested by a member of the IRB committee, who stated that it was unnecessary for a practitioner to provide input about my interview questions.