Abstract
Cross-cultural prison research is a growing field, and it is increasingly common for researchers to adapt existing instruments to new contexts to quantitatively evaluate prison climate. Yet, comparative penology remains an underdeveloped field, and there have been limited efforts to contextualize adaptations of prison-based instruments. Various obstacles arise when adapting validated prison-based surveys to diverse cultural settings. Drawing on the Measuring the Quality of Prison Life (MQPL) survey, this study examines these challenges and provides a framework for prison researchers seeking to adapt survey instruments to varied sociocultural contexts. We carried out a systematic search of studies that have used the MQPL survey and conducted interviews with nine researchers and practitioners who adapted the instrument in eight different countries. We identified four adaptation challenges related to linguistic, measurement, contextual, and worldview differences. Findings highlight the importance of regarding translation as cultural mediation, improving measurement through qualitative methods, adapting the scales to fit their new cultural setting, and thoroughly testing for survey reliability. The paper discusses the implications of these observations for future cross-cultural prison research and for the field of comparative penology more broadly defined.
Introduction
In 2013, a French-speaking, US-based criminologist initiated a project aiming to study the quality of prison life and the process of desistance from crime in the French carceral system. The initial sample included individuals who were incarcerated in a correctional facility in the outskirts of Paris; many were transferred to different facilities across France over the course of the follow-up period. The project's findings were published in a book, which examined the circumstances under which individuals may, against all odds, develop strategies to find meaning to their time in prison (Kazemian, 2020). In addition to detailed interviews, study participants completed a series of surveys over the course of the study, including the Measuring the Quality of Prison Life (MQPL) survey. This instrument was developed by Liebling (2004) and seeks to assess the “moral quality,” culture, values, and social environment of individual prisons.
Several challenges arose in the process of adapting the MQPL survey to the French context, leading to many questions about the application of the survey to different cultural settings. “Culture” here is defined “in the broader anthropological sense,” referring to “all socially conditioned aspect of human life” (Snell-Hornby, 1988: 39). What aspects of the survey were not quite transferrable to cultural settings that are different from the original English context? How can we better adapt survey instruments to the distinctive cultural setting of other countries? Lastly and most importantly, what are the implications of these lessons for future prison research employing the MQPL or other validated instruments in different cultural settings?
The fundamentally harmful and repressive nature of the incarceration experience has been well documented in prior research, but there is also evidence to suggest that some prisons may be less painful than others (Liebling et al., 2019). High incarceration rates, prison overpopulation, and the over-use of solitary confinement are unequivocal crises that greatly contribute to the pains of imprisonment. Regulatory agencies and international governing bodies have established set rules, standards, and oversight structures aiming to reform prisons, improve their conditions and safeguard human rights (United Nations, 2021). However, these international standards are based on a “reductive understanding of carceral harms” and frame the prison as “fixable” (Kemp and Tomczak, 2024: 1686). Scholars have noted that existing research tends to “focus more on the quantity of punishment (i.e., imprisonment rates) than its form or experience” (Crewe et al., 2023: 425). To guide large-scale efforts towards meaningful punishment reform, there is a crucial need for comparative prison research that examines the experience of incarceration from the perspective of those who live it.
Despite the inherent value in examining and comparing different penal practices, comparative penology remains an underdeveloped field (Brangan, 2020). Different countries may hold vastly different punishment philosophies; these are rooted in varying cultural sensibilities and political dispositions, or “political cultures” (Brangan, 2023). Distinctive approaches to penalty reflect different perspectives on “how society should be organised, and how should it be governed” (Brangan, 2023: 950). The comparison of different prison systems enables researchers to identify the social, political, historical, and cultural factors that may explain imprisonment trends (Garland, 2018). Seminal works comparing criminal justice and penal systems across the world have guided the development of comparative penology (Garland, 2001; Wacquant, 2009). Yet, comparative prison research remains scarce, partly because of the wide variations in the conceptualization and operationalization of basic metrics such as recidivism (Fazel and Wolf, 2015). While comparative penology is crucial to understanding the cultural factors that may drive reform efforts in various parts of the world, many important areas of inquiry remain unexplored.
This paper aims to provide a framework for prison researchers seeking to adapt survey instruments to different socio-cultural contexts. Cross-cultural research bears many complexities, and the methodological research on survey adaptation across cultural contexts generally focuses on linguistic translation, cultural dissimilarity, and different categories of biases (van de Vijver and Leung, 2021). In the context of prison research, there have been limited efforts to contextualize adaptations of prison-based instruments. The MQPL survey is a reputable instrument in prison research, and it has been adopted in different countries. It seems appropriate to examine the challenges associated with the adaptation of this instrument to different cultural settings. We conducted a systematic search of studies that have used the MQPL survey outside of the UK and interviewed nine researchers and practitioners who adapted the instrument. Drawing on these interviews, we identified four categories of challenges that emerge in the process of adapting the instrument to different cultural settings. The analysis outlines variations in applications of the MQPL across different languages and contexts. These adaptation obstacles were linked to linguistic, measurement, contextual and worldview differences. By identifying the types of challenges that may arise in the process of adapting an instrument to different cultural settings, this analysis provides valuable insights for future prison research.
Quantifying the quality of prison life
Academics and practitioners in the field of criminal justice have increasingly prioritized quantitative and, when possible, experimental designs in research that aims to evaluate programs and policies (Sampson, 2010). With the expansion of punitive carceral systems across the globe, a growing body of research has emerged on the observation and measurement of prison environments. Some early efforts include the Correctional Institutions Environment Scale (CIES; Moos, 1975), which sought to measure correctional climates through an evaluation of the effectiveness of treatment programs in correctional settings. Toch's (1977) Prison Preference Inventory also aimed to capture the nature of interactions between the prison environment and the incarcerated individual, by identifying the features of the prison that may be conducive to positive or negative experiences of incarceration.
In the decades following these seminal works, the heyday of both prison expansion and empirical fervor saw the emergence of several new prison-based surveys aiming to quantify the subjective experience of incarceration. The Prison Social Climate Survey, spearheaded by the United States Federal Bureau of Prisons, collected large-scale quantitative data measuring the organizational climate of prisons (Saylor et al., 1996). The Essen Climate Evaluation Schema (EssenCES) was developed in Germany to assess the social climate of secure psychiatric wards, though it has also been used in non-psychiatric carceral facilities (Schalast et al., 2008). In the Netherlands, the Dutch Prison Climate Questionnaire (PCQ) aimed to overcome some of the psychometric shortcomings of previous prison-based climate surveys (Bosma et al., 2020). In the United States, drawing on a community engagement approach, the Urban Institute's Prison Research and Innovation Initiative (PRII) developed a survey to assess prison environments. Researchers worked with incarcerated individuals and prison staff to produce and validate a survey that documented stakeholders’ knowledge and expertise of the prison environment (Crocker and Fox, 2024). The survey and its findings have not yet been published but the PRII represents a novel, boundary-pushing approach to quantifying perceptions of prison quality of life using participatory-action research. Overall, prison climate surveys vary in their approaches and operationalizations of “quality of life,” but some of the most consistent themes that emerge across different studies include dimensions related to safety, personal well-being and mental health, and perceptions of prison staff and operations. However, these instruments are not without methodological and conceptual limitations. Assessments of the prison environment relying solely on quantitative data may lack depth, scope, and context to allow for different interpretations.
Measuring the quality of prison life (MQPL)
The MQPL instrument, developed by Alison Liebling (2004) in England, set out to address some of the limitations of prior surveys. The instrument uses a mixed-methods approach and draws on two fundamental tools. First, it involves interviews with incarcerated individuals and prison staff using the appreciative inquiry framework, a method that involves posing generative questions with a focus on strengths and positive experiences (Liebling et al., 1999). Second, the MQPL includes detailed survey questions about the quality of life in the facility for incarcerated individuals and prison staff. At its core, the MQPL attempts to empirically assess a prison's moral performance, defined as “those aspects of a prisoner's mainly interpersonal and material treatment that render a term of imprisonment more or less dehumanizing and/or painful” (Liebling, 2004: 473).
The instrument includes five core dimensions (Liebling et al., 2011). Harmony taps into the interpersonal and relational aspects of prison life; it measures the extent to which interactions between staff and incarcerated individuals are trusting, fair, and supportive, and whether the environment is characterized by kind regard and concern for the person. Professionalism examines how prison work is carried out and authority is wielded, the transparency and responsiveness of the prison system, and the perceived impartiality of its procedures. Security reflects the aspects of prison life concerned with the rule of law, the use of authority, and the provision of safety. Conditions and family contact measure the ability to maintain relationships with loved ones during the period of incarceration. Finally, the well-being and development dimension explores incarcerated individuals’ perceptions of their own sense of progression. Each dimension contributes to a comprehensive understanding of the incarceration experience from the perspective of the detained population.
Methods
This research involves two components: (1) a systematic search of studies that adapted the MQPL to a different cultural context, which yielded ten studies and (2) in-depth interviews with nine researchers who adapted the MQPL to different cultural settings.
We conducted a systematic search to identify all studies that implemented the MQPL survey across different countries and cultural settings. We included all peer-reviewed studies with abstracts in English, French or Spanish that used the MQPL survey in a carceral setting outside of England and Wales. We searched four databases in December 2023 and again in October 2024: SCOPUS (six results), Criminal Justice abstracts (six results), Sociological Abstracts (13 results) and Google Scholar (296 results). Following guidance from prior work (Papaioannou et al., 2010), we also consulted with various experts who helped to identify studies that were overlooked in database searches. The systematic search yielded ten studies across ten different countries.
Following the systematic search, we conducted interviews with researchers and practitioners who had adapted the MQPL to their respective cultural settings. The selection of interview participants was purposive. We contacted the investigators of the ten studies that were identified through the systematic search, as well as five other researchers that were recommended by leading scholars in the field. We reached out to fifteen individuals and conducted interviews with nine; six individuals declined to participate, did not wish to be identified, or failed to respond. Nearly all interviews, except for one, were conducted via the Zoom platform. The interviewees were from diverse geographic locations and provided insight from a wide range of cultural contexts. The seven MQPL studies that were the object of the interviews are summarized in Table 1, along with the key features of each study's adaptation of the MQPL. 1
Key characteristics of the MQPL studies discussed in the interviews.
Dr Guillermo Sanhueza, a social worker and prison researcher, sought to investigate the experience of incarceration in Chile. Using the MQPL as a guiding framework and in consultation with Alison Liebling and her team, Dr Sanhueza created a version of the survey that was adapted to the Chilean context. He aimed to create a survey that was practical, accessible, and maintained “the original spirit of the MQPL.” His version is one-third of the length of the original MQPL and includes a new dimension on infrastructure and food. The resulting questionnaire is a nine-dimension measure of the moral performance of Chilean prisons (Sanhueza and Pérez, 2019).
Dr Nazirah Hassan used the MQPL in the context of her PhD research project, which investigated the impact of perceived institutional quality of life on well-being and development among adolescents housed in juvenile detention centers in Malaysia (Hassan et al., 2019). She used the survey component of the MQPL and translated an unaltered version of the original survey.
Dr Lila Kazemian initiated a longitudinal mixed-methods study with individuals serving long-term sentences in French prisons. The study was conducted in a maximum-security facility located in a suburb of Paris. The survey portion of the project included various validated instruments, including the MQPL. The original MQPL survey was translated and supplemented with additional questions relating to the French judicial process. Drawing on Liebling's (1999) appreciative inquiry methodology, the research also included in-depth interviews to supplement the survey data.
Also opting for a mixed-methods approach, Dr Jennifer Peirce investigated the impact of reforms in the prison system in the Dominican Republic. She used an adaptation of the MQPL, which she supplemented with interviews. The Chilean and Dominican contexts have many common features: similar histories of colonization, underdeveloped infrastructures and prison overcrowding (Peirce, 2021: 197). Dr Peirce adopted the infrastructure questions included in the Chilean adaptation of the MQPL and added questions on the role of governance organizations led by incarcerated individuals. Her final version included six dimensions, largely following the same structure as the Chilean adaptation.
Two separate interviews were conducted with two US-based researchers. In 2021, a coalition was formed between the Pennsylvania Prison Society, the Correctional Association of New York (CANY), and the John Howard Association. Aidan King, a director at CANY, worked with Alison Liebling's team to use a shortened 46-question version of the survey. They piloted the survey in New York and Pennsylvania, and the John Howard Association distributed the MQPL remotely across all 28 state prisons in Illinois. Gwyn Troyer, an associate director with the organization, provided valuable feedback on the ambitious project's successes and shortcomings. The effort culminated in the creation of a public data dashboard (John Howard Association, 2023).
Kenneth Owusu Ansah is a Ghanaian researcher in the field of psychology. He used the MQPL (among other survey instruments) in a mixed-methods study of peer victimization among juveniles in the Senior Correctional Center of Ghana (Owusu Ansah et al., 2022). The study aimed to assess the impact of victimization on psychological distress experienced by adjudicated adolescents. He chose to use only four of the MQPL's subscales—caring for the vulnerable, fairness, policing, and security and prisoner safety. To our knowledge, this is the only published study that has used the MQPL in the continent of Africa.
Dr Milena Milićević is the principal investigator of an ambitious three-year project examining the perception of quality of life in Serbian prisons (Međedović et al., 2024). It is the first of its kind in Serbia, aiming to collect national data with over 500 incarcerated individuals. The Serbian version of the MQPL closely resembles the original survey. Dr Milićević's research was ongoing at the time of the interview, and she provided valuable insights about the process of implementing the MQPL in a large-scale project.
Dr John Rynne, an Australian forensic/clinical psychologist and prison researcher, examined the impact of custodial environments on Aboriginal and Torres Strait Islanders. In the process of piloting the survey, he noted that the original version of the MQPL did not accurately reflect the realities and experiences of the Aboriginal population. The researcher opted to bypass the MQPL survey entirely and adapt the qualitative questions to better reflect the experiences of the Aboriginal communities of Western Australia (Leeson et al., 2016).
Reliability versus validity
In seeking to operationalize theoretical constructs, researchers strive for measures that achieve both reliability and validity. Reliability refers to the degree of consistency in the measurement of a construct (i.e., whether the construct is stable across time, subgroups or indicators), while validity refers to the more abstract notion of “truthfulness” (i.e., whether the construct can accurately measure the social reality that it purports to assess; Carmines and Zeller, 1979). There are three main types of reliability. First, stability across time is assessed through the test−retest method, wherein an instrument is distributed to the same population at different points in time. Second, internal consistency reflects the interrelatedness of the items included in a given construct, often assessed through indicators such as Cronbach's alpha or Kuder-Richardson. Third, interrater reliability, which is primarily applied to qualitative data, assesses whether two or more individuals agree in their assessment, observation or rating of a given phenomenon.
While the assessment of reliability is relatively straightforward, it is more challenging to test the validity of theoretical constructs. An instrument achieves greater validity when it is perceived to be credible by the sources and stakeholders involved in the study, and when it is context-bound (Guba, 1981). Validity can also be maximized by triangulating data sources and collecting “thick” descriptive data (i.e., detailed, thorough contextual information). Van de Vijver and Leung (2021: 14) regard “bias” as a form of validity specific to cross-cultural research, which is regarded as “all nuisance factors threatening the validity of cross-cultural comparisons.” The authors highlight three types of bias. Construct bias occurs when the measured concept is not synonymous across cultural groups. Method bias refers to variations in specific characteristics of the instrument (e.g., mode of administration, scale type, etc.). Item bias denotes measurement issues at the item level, such as inadequate translation. Our analysis for the current study predominantly focuses on the issue of validity/bias in adaptations of the MQPL, but we also discuss the relevance of testing for reliability in future research.
Results: The four impediments to cross-cultural adaptation
The researchers who worked to adapt the MQPL to their respective sociocultural contexts each faced vastly different research environments, priorities, methodologies, timelines, and objectives. Yet, interviews consistently highlighted similar challenges in efforts to adapt the survey to different cultural settings. Many of the difficulties faced by researchers touch upon the validity (or bias) of their survey adaptations, though they do not always directly correspond to the four categories of validity that are typically highlighted in quantitative research methods (i.e., face, content, criterion, and construct). Guided by van de Vijver and Leung's (2021) three types of bias (item, method, and construct), we identified four broad themes that reflect the challenges of cross-cultural adaptation: language, measurement, context, and worldview. These challenges may bear relevance for other prison-based research that seeks to adapt instruments to different cultural settings.
Language
Translating a survey into a different language is an arduous task. While translation may appear relatively straightforward, literal (or “close”) translations are often insufficient because they fail to negotiate the gaps that are distinctive to each culture. When a construct's features vary across different cultural groups, it must be “adapted” rather than simply “adopted” (He and van de Vijver, 2012). To develop a culturally appropriate translation, feedback from individuals who are familiar with the local culture is essential (McGorry, 2000). Many of the interviews with researchers who adapted the MQPL highlighted issues with the linguistic translation of the British survey.
Most researchers opted to use back-translation in their adaptation of the survey. Brislin's (1970) model of back-translation is often cited as an effective method in cross-cultural research. The method consists of re-translating content from the target language back to its language of origin, using independent translators to ensure accuracy. The Ghanaian version required the services of a translator who spoke Twi, the most common local dialect in Accra; the Twi version was then translated back into English. In the Malaysian version of the MQPL, Dr Hassan went beyond the process of back-translation. Before engaging in the translation, she enlisted the assistance of a criminological expert to thoroughly examine each item in the English version to ensure a strong understanding of the meaning of each question. After this iterative exchange, the researcher proceeded with the process of back-translation.
The original MQPL includes idioms, the passive voice, and expressions that are specific to the prison setting in England and Wales. Expressions such as “doing time,” “the pecking order,” and “if your face fits” required some modifications. Other linguistic challenges arose. For instance, in both French and Spanish, the same word is used to convey “fairness” and “justice,” and these concepts are often conflated. Even when the survey was adapted to an English-speaking country, some modifications were necessary. In the USA, Mr King emphasized that the MQPL had to be “Americanized,” a process which involved replacing British expressions with their American counterparts. The Serbian translation of the MQPL required extensive clarification. Following the first iteration of the survey's translation, the research team put together two focus groups. The focus group participants deemed that the wording in many of the questions was unclear. In response to the questions about “prison staff” and “prison employees,” they did not understand to whom the questions referred: “The treatment group? The guards? Health care professionals?” Drawing on cultural knowledge and feedback from survey pretests, researchers made careful choices to develop surveys that were adapted to their respective cultural settings.
Measurement
The format of the MQPL led to some adaptation challenges. Van de Vijver and Leung (2021: 46) discuss “measurement” adaptations of cross-cultural studies, which refer to changes applied to the format of the original instrument, including the phrasing of questions and the scales used to measure responses. The original version of the MQPL was intended to be distributed to random samples of incarcerated individuals, who were expected to complete the survey on their own. However, in countries that grapple with prison overcrowding and that struggle to keep a reliable count of the incarcerated population, random sampling is inconceivable. Moreover, because of literacy challenges and attention deficit issues among the incarcerated population, researchers reported that many participants were unable or unwilling to complete a survey of this length on their own. In France, Dr Kazemian opted to administer the survey to each of her participants verbally. She read out each question to the participants and documented their responses. She viewed this method as the most efficient way to gather survey responses and to address the fatigue that inevitably occurs when completing a 136-question questionnaire. In New York, the survey was also administered verbally because the oversight teams were not permitted to distribute paper surveys to incarcerated individuals. This was a tremendously time-consuming process, which would be very costly to implement on a larger scale.
The Likert scale format of the survey led to some confusion in the Chilean context. While very common in English-language surveys, the scales were counterintuitive to many of Dr Sanhueza's study participants: Let's say, ‘Here in this prison, guards beat me too much.’ And in the pilot questionnaire, inmates were responding, in that violent prison, that we knew that we had corruption, we had mistreatment by guards… You know, we already knew that. In the pilot, most inmates were saying: ‘completely disagree, completely disagree, disagree, completely disagree.’ And we said, ‘what's going on here, guys? We know that this is not…’ And they reply to me, ‘hey, professor, would you like me to agree with this thing happening to me?’ They were interpreting the question in a sense that they agreed with the fact that they were beaten very often. (Dr Sanhueza, Chile)
Context
Another key challenge to cross-cultural adaptation pertains to the diverse economic, political, and social contexts that characterize different countries and prison systems. The original MQPL survey does not include a measure of the sociopolitical context of the prison. It was conceived for the prison system of England and Wales, which is relatively homogeneous and centrally governed. Many of the countries in which the MQPL was implemented have fundamentally different social, political, and economic landscapes. It stands to reason that the standards for acceptable living conditions may be quite different in emerging economies than in the United Kingdom. Living conditions that may be regarded as acceptable or comfortable in one cultural setting may be deemed intolerable in another.
All researchers who adapted the MQPL to countries with higher rates of inequality and poverty stressed the need for an additional survey dimension to document prison infrastructure. Dr Hassan noted that study participants often referred to the displeasing physical structure of the prison and its living conditions. The Chilean and Dominican adaptations of the MQPL included an additional “infrastructure” dimension, which included questions documenting access to food, water, and bathrooms, as well as the sanitary and sleeping conditions in the facility.
The threshold for satisfactory (or even bearable) conditions can vary even in a wealthy nation. Ms. Troyer (the USA) noted that the survey lacked questions about basic services, specifically about healthcare in American prisons. She questioned the relevance of using the MQPL's more abstract questions in a context where “people are concerned about basic necessities.” Dr Rynne (Australia) described an encounter with an incarcerated individual in Alice Springs, one of the hottest and driest parts of the world. When the researcher asked the participant how he was doing, he responded, “Doing great, thanks!,” which shocked the researcher. I said, “You're good? It's 50 degrees!” And they used to lie on the concrete because the concrete was so cool. It's part of the center of the prison. And he said, “well, I get a meal, I get fed, I'm sort of safe. I can get whatever drugs I want to use and I know I'm not going to get bashed.” And that man, to him, that was good because life in the community was about drinking, violence, abuse, destruction. In the prison, he felt safe. Now, if we give the MQPL, you're going to get a guy saying, “hang on, this is great.” Which is totally wrong. (Dr Rynne, Australia)
In this context, respondents may provide high scores on a scale measuring the perceived quality of life despite subpar living conditions that violate international standards. Similarly, many of Dr Peirce's (DR) participants indicated they had sufficient personal space in their cells, which seemed inconsistent with the reality of their daily lives in prison: “And then you ask how many people are in this completely cramped area. And it's like 80 people in an area that was built for ten. So, like, it just doesn’t add up, right?” Peirce's (2021) study specifically contrasted the Dominican Republic's “new model” of prisons (featuring new buildings, programs, and improved staff training) with the “traditional” prisons that continue to operate in tandem; the latter are characterized by overcrowded barracks, deteriorated infrastructure, and limited police or military officers as guards. Counterintuitively, many respondents preferred the traditional prisons to the new model of facilities because the traditional model was more lax on regulations, which facilitated access to certain freedoms and privileges that would otherwise not be accessible: Basically, what I found in my study is that people, even when they did say that their conditions were bad, and they didn't have access to clean water, the food is terrible, they're always hungry. They still said, I would rather live here where I don't have COs [correctional officers] beating me up and putting me in solitary every day, and where I have access to cash and a cell phone, and my family can come visit basically with no regulations because the system is so corrupt and porous. (Dr Peirce, DR)
The benchmark or frame of reference for “normal” or “acceptable” living conditions fundamentally influenced the participants’ responses, posing a significant challenge to adapting the concept of quality of life to different cultural settings. It would be wise for future adaptations of the survey to consider the broader social, economic, and cultural characteristics of the countries in which the research is conducted, as these elements indubitably have an impact on subjective and objective prison conditions and standards.
Worldview
The final obstacle to cross-cultural adaptation is the most abstract in nature. Some of the interviews highlighted inconsistencies in participants’ collective understanding of the world. Hofstede (2001: 3) argues that human beings’ behavior and subjective understanding are guided by “mental programs,” or intangible constructs that operate at the individual, collective, and universal level. Mental programs at the collective level are passed on, learned and shared by communities and cultures; we refer here to such “collective programming of the mind” as differences in worldview (Hofstede, 2001: 4). In the Dominican Republic, when asked if they agreed with a statement about their rights (“my rights are respected here”), Dr Peirce reported that many participants pushed back: “‘What are rights?’, ‘what are my rights?’, or ‘of course I have no rights here in the prison.’ So they just kind of rejected the premise of the question.” If individuals do not have knowledge of their rights, or do not believe that they are entitled to rights, it does not seem inappropriate to inquire about whether their rights are respected. This is a fundamental challenge when adapting surveys to vastly different cultural and prison settings.
In his reflections on the profound differences between Western and Aboriginal conceptions of justice, Dr Rynne highlighted the significant hurdles in translating certain concepts from traditional prison research to the Aboriginal context. The concept of the prison is foreign to Aboriginal culture. The Aboriginal conception of justice emphasizes the reestablishment of balance through proportional actions, or “payback.” Dr Rynne argues that the language, philosophy, and logic of Western prison systems are fundamentally colonial; thus, many of the concepts of the MQPL cannot be translated to this context without acquiring a profound understanding of the culture. Dr Rynne also noted the ways in which his interactions with incarcerated Aboriginal individuals were tainted by his own identity as a White male researcher: They'll tell you what you want to hear. And if you come from an impoverished racialised background where you don't really understand English, you'll say the thing that'll get you into least trouble. And soon we started to very quickly understand that the Aboriginal people will answer white people's questions in a way that the white person wants to hear. (Dr Rynne, Australia)
Dr Rynne described the vast differences between “White fellow's justice” and conceptions of justice in the Aboriginal community, emphasizing the inability of standardized surveys to capture unique experiences that are specific to a culture and shared history. He opted to omit the quantitative portion of the MQPL altogether, choosing instead to adopt the survey's concepts and dimensions as guiding principles for qualitative research. In his view, “if we took the basics of the MQPL, translated it into an Aboriginal language and gave it to them, all we would find is a White version of… We wouldn’t find what humanity is for them.” Dr Rynne valued the MQPL as a guiding tool but stressed that the survey in its original form is inherently at odds with Aboriginal culture.
Discussion
Most methodological reflections on the intricacies of conducting prison research have drawn on the experiences of qualitative and ethnographic researchers (Jewkes, 2014). While criminology remains a predominantly quantitative and positivist discipline (Jacques, 2014), qualitative research seems to be the “favored” approach of prison researchers, with fewer large-scale, quantitative studies conducted in carceral settings. This paper has outlined the challenges faced by researchers in their efforts to adapt the MQPL survey, one of the most recognized quantitative instruments in prison research, to different cultural contexts. Many of the obstacles raised by the study participants are, at their core, issues of validity; researchers were often uncertain of whether the MQPL's constructs were applicable to their participants’ experiences. The observations emerging from the interviews with international researchers bear implications for future prison research conducted in different countries and cultural settings. Four broad recommendations emerge from our research.
The importance of framing the translation process as a form of cultural mediation
Researchers adapting the MQPL often faced linguistic issues, ranging from idiomatic impediments to lack of clarity over the object or meaning of a survey question. Katan and Taibi (2021) explained that multicultural communication is akin to the work of a mapmaker, who must make choices about which information to include to create a “meaningful and useful” product (p. 142). These choices require a deep understanding of both the cultural context of the original text, and of the object of transposition. Scholars have pointed to the potential role of translators as “cultural mediators,” whose knowledge allows them to mediate by “interpreting the expressions, intentions, perceptions, and expectations of each cultural group to the other” (Taft, 1981: 53). Language should not be viewed as an isolated phenomenon, but as “an integral part of culture;” translation, then, is a “cross-cultural event” (Snell-Hornby, 1988: 39). The task of cultural mediation can be particularly valuable in crafting research questions in comparative studies; international researchers collaborating on a project may view the translation process as the first step in identifying potential contrasts and coherences between their distinctive cultures.
Ideally, researchers seeking to adapt any prison-based survey to a different cultural setting would have the ability to access and participate in both the culture of origin and the culture in which the survey is administered. Linguistic skills are important but should be complemented with a knowledge of cultural history, customs, values, and taboos. In a German translation of the MQPL, researchers found it challenging to convey the concept of “decency” due to the word's inextricable ties to the Nazi regime: “in German ‘decent’ is a term that is both morally charged and historically corrupted” (Neubacher et al., 2023: 1455). Prisons are also reflections of a state's criminal justice philosophy and policy. Prisons in Nordic countries are widely regarded as integral parts of the state's leanings towards social democracy; to follow the Nordic example is to “buy into a whole package of welfare state solutions” (Shammas, 2015: 5). In contrast, prisons in Latin America are influenced by security concerns and the militarized practices in the region, which are rooted in enduring political and historical forces (Darke and Karam, 2016: 461). History matters. To grasp the nuances of the original instrument and to produce an output that conveys the intended meaning, it is important for researchers to thoroughly understand the political and historical context of different cultural settings.
Improving measurement through qualitative methods
The MQPL survey was developed as part of an innovative project aiming to identify the issue that matter to incarcerated individuals and staff through dialogue, interviews, and observations. While it is a mixed-methods instrument, some researchers have opted to only carry out the survey or the interview portion of the MQPL. In the context of cross-cultural prison research, we do not recommend bypassing the qualitative component of the survey. The detailed narratives that emerge from the interviews provide valuable contextual perspective, which serves to improve validity and reduce bias. In the Ghanaian adaptation, Mr Owusu noticed a validity issue when engaging in the qualitative portion of his study. The quantitative measures of “prisoner safety” were mostly neutral, indicating that respondents did not express significant safety concerns in the prison. However, interviews painted a different story and raised safety issues that were not uncovered in the survey. The stark difference between survey responses and interview narratives suggests some response bias. In some cases, respondents may conceal the truth or provide false responses (Furnham, 1986).
Some of the measurement issues discussed in this paper could be addressed by spending more time at the research sites, and by engaging in dialogue with incarcerated individuals and staff through interviews or unstructured observation. Researchers can implement Guba's (1981) criteria for assessing and improving trustworthiness, or validity. While often associated with qualitative research, techniques such as prolonged engagement at a site, the collection of detailed descriptive data and triangulation can reinforce a survey's validity during the process of cross-cultural adaptation.
Modifying survey scales to fit the cultural setting in which they are applied
We use surveys in the social sciences to quantitatively assess theoretical concepts that may be challenging to measure and operationalize. Researchers prioritize the use of standardized and validated instruments with established reliability and often operate on the assumption that constructs are measured similarly across different settings, but the reliability and validity of standardized instruments may be compromised in cross-cultural comparisons (Gjersing et al., 2010). While some disciplines have strict guidelines and are more skeptical about modifying items in standardized scales, the social sciences generally acknowledge that validated questionnaires may require some modifications when applied to a new language, time, culture, and context (Finn and Kayande, 2004).
Researchers who conduct cross-cultural prison research should not be reluctant to make changes to survey instruments. In our interviews, the two researchers who implemented the MQPL in Latin American prisons noted that it was essential to add a new dimension to the survey to measure elements of the prison infrastructure. The political culture may also compel researchers to modify their surveys, as in France, where Dr Kazemian was unable to include questions about racial and ethnic background because the country prohibits the gathering of data on race and ethnicity (Simon, 2015). Researchers can take steps to identify the required modifications to improve validity. These include engaging in dialogue with research participants, stakeholders, and experts, and conducting pilot testing and focus groups prior to the implementation of the final instrument. Community-engaged research is a crucial component of cross-cultural adaptations of research instruments. Participatory prison research acknowledges the unique perspectives of incarcerated individuals, which help to generate a more comprehensive understanding of the incarceration experience and its effects (Farrell et al., 2021). Integrating community-engaged research principles into adaptations of the MQPL, or any other prison-based instrument, can strengthen and validate findings while also promoting more ethical and innovative research practices.
Testing reliability in future research
Insights derived from the interviews primarily touched upon issues of validity. While this was not directly addressed by researchers who participated in the current study, the assessment of reliability is also crucial to cross-cultural adaptation. If a survey does not produce consistent results, this will inevitably affect the validity of the instrument. As a preliminary analysis, we examined the internal consistency (measured through Cronbach's alpha coefficients) reported in prior studies that used the MQPL survey (see Appendix A). This analysis compares Cronbach's alpha coefficients for each MQPL dimension across ten studies that used the survey in their research (in addition to the original UK survey). In Appendix A, the studies are presented in ascending order, from least (Serbia, Malaysia) to most (Chile, Dominican Republic) extensive changes made to the original instrument. The reliability coefficients are highlighted in different colors according to their strength: strong (green), average (yellow), weak (orange), and very weak (red). We used Cronbach's alpha scores because these were the only consistently reported measure included in the international MQPL studies that emerged from our search. While alphas are an imperfect comparison metric due to the differences in survey adaptation and research design across studies, they provide a preliminary standardized comparison of the MQPL's reliability across cultural contexts. Overall, the harmony dimension had the highest reliability coefficients across all sites, while the security dimension displayed the lowest reliability coefficients. Some research has underlined the cultural differences in the perception of risk and danger (Bontempo et al., 1997), and feelings of safety may be conceptualized differently across cultures. In contrast, the notions of “humanity,” “decency” and “respect” may be transferrable because these are fundamentally human notions that may transcend intricate social and cultural constructions.
We recommend that researchers who seek to adapt the MQPL to different cultural settings take the analysis beyond the threshold of acceptable alpha scores to achieve reliability. Constructs should be clearly conceptualized to ensure that survey questions are interpreted in the same manner across all participants. The survey can be modified to add clarifying information, as Dr Milićević's team opted to do in Serbia. One MQPL question inquired about the privileges gained and taken away in prison, but the type and nature of these privileges were unclear to participants in the focus groups. As a result, the Serbian team added brackets to clarify the question: “[extended rights to receive packages, number of visits going out in the city…]” Most interviewed researchers highlighted the value in conducting a pilot study and/or focus group prior to distributing the surveys to all research participants. Pilot studies allow researchers to preemptively identify potential reliability issues and implement changes to the instrument to maximize its reliability.
The assessment of internal consistency is crucial, but cross-cultural comparisons should also prioritize examining reliability across time as well as validity issues, which are insufficiently addressed in quantitative research. To maximize reliability and validity, we recommend that researchers consider all four recommendations included above when adapting existing surveys to different cultural settings.
Conclusion
The limited body of research devoted to comparative criminology/penology and cross-cultural adaptations of quantitative instruments has generally focused on the linguistic and methodological challenges associated with this process. This paper has outlined four key challenges inherent to cross-cultural prison research, which include linguistic, measurement, contextual, and worldview differences, to provide a guiding framework for criminological researchers seeking to adapt the MQPL or similar instruments to different cultural settings. We may also need to recognize that some concepts are socially unique, and that they may not be easily transferrable to a different cultural context.
Given the significant challenges that are inherent to the adaptation of a standardized instrument to distinctive prison systems, an obvious question arises: should we even attempt to adapt surveys intended for a specific site to other cultural settings? If phenomena that are the object of research are intrinsically tied to the time and context in which they unfold (Guba, 1981), should studies construct entirely new surveys to meet the realities of their own cultural context? Researchers have drawn on the principles of community-engaged research to develop new surveys as a collaborative and iterative process with target populations (Crocker and Fox, 2024). Some researchers were inspired by Liebling's original methods of instrument development but have ultimately created their own prison climate surveys using the “spirit” of the MQPL (Crewe et al., 2023). While certainly more “trustworthy,” constructing a robust survey is costly and time-consuming. A considerate, culturally responsive adaptation of the MQPL can serve to elevate research from the Global South and build the groundwork for other researchers to pursue prison research in neighboring regions (Carrington et al., 2016). For researchers who lack access to large-scale funding to support such an undertaking, particularly in geographic areas where prison research is lacking, a judicious adaptation of the MQPL is likely the best option.
If we accept that surveys can and should be adapted across different cultural contexts, we might also question whether, given the challenges outlined in this paper, comparative criminology in its purest form is in fact possible. In an increasingly globalized world, in which public services are shaped by global market forces (Walsh, 1995), it is crucial to study the unique shifts and patterns that arise in different prison systems. To simply implement the recommendations provided in this paper and deploy the MQPL across different countries would be inadequate; researchers would benefit from focusing on trends identified in existing research as well as specific questions and issues that emerge from their own cultural contexts. For instance, Dr Peirce's (2021) findings on the differences in quality of life between “old” and “new” prison paradigms in the Dominican Republic could be used as a point of comparison across countries implementing penal reforms to assess whether similar patterns emerge. Despite notable measurement differences across cultural settings, the MQPL provides a strong baseline instrument for comparative prison research.
Footnotes
Acknowledgments
We are grateful to the researchers who agreed to participate in the interviews conducted in this study. We would also like to offer special thanks to Professor Alison Liebling, who has always been gracious with her time when we had queries about the MQPL, and who facilitated our initial connection. Lastly, we are thankful to the anonymous reviewers who provided valuable feedback on an earlier draft of this manuscript.
Notes
Appendix A. Average reliability coefficients (Cronbach's alpha) for MQPL dimensions measured in 10 countries and in the original UK site.
| UK a (original) | Serbia | Malaysia | Norway | France b | US a | Canada | Belgium | Ghana | D.R. | Chile | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Harmony | 0.80 | 0.8 | 0.75 | 0.82 | 0.92 | 0.71 | 0.84 | 0.74 | 0.8 | 0.84 | 0.77 |
| Professionalism | 0.84 | 0.85 | 0.76 | 0.62 | 0.89 | 0.62 | 0.7 | − | 0.82 | 0.86 | 0.83 |
| Security | 0.74 | 0.76 | 0.7 | 0.68 | − | 0.45 | 0.57 | 0.66 | 0.74 | 0.8 | 0.65 |
| Conditions and family contact | 0.68 | 0.82 | 0.83 | − | − | − | 0.58 | 0.67 | − | 0.79 | 0.7 |
| Wellbeing and development | 0.76 | 0.74 | 0.69 | 0.8 | 0.79 | 0.72 | 0.64 | 0.49 | − | 0.74 | 0.74 |
Cronbach's alpha scores were only available for each sub-dimension in the UK and US studies and not for the overall dimensions. Using weighted averages, we created estimates of Cronbach's alpha for each of the five main dimensions based on the number of questions included in each sub-dimension.
Adjustments were necessary in the French analysis due to the smaller sample size. Reliability coefficients are not presented for two of the five dimensions because these failed to meet the minimum factor loadings and eigenvalues required to produce reliability coefficients on a sample size < 100 (Guadagnoli and Velicer, 1988).