Insider and Outsider Fieldwork Challenges in Medellín, Colombia

High levels of violence and conflict in Latin America have attracted the interests of local and international researchers to further understand how to reduce this violence and prevent current and future outbreaks. Conducting research in any environment is challenging. However, the obstacles facing not only researcher safety but also data collection methods are particularly complex in settings with high levels of violence and conflict. This article contributes to the methodological literature, as it provides reflections from two researchers, an insider from the Global South and an outsider from the Global North, each collecting data in Medellín, Colombia. It also argues the importance of collaboration between Global North and Global South researchers investigating violence and insecurity.


Introduction
In 2018, 42 out of the 50 most violent cities in the world were located in Latin America (Seguridad, Justicia y Paz 2017). These high levels of violence have attracted interest from local and international researchers in various disciplines to further understand how to reduce this violence and prevent future outbreaks. To ensure the reality of this violence is accurately represented, researchers are encouraged to take a qualitative approach to their work, such as conducting fieldwork. Conducting fieldwork in any environment is challenging. However, the obstacles to not only researcher safety but also data collection are particularly complex in locations with high levels of violence and conflict. This article makes an important contribution to the methodological and urban violence literature, as it presents the experiences of an outsider researcher from the Global North and an insider from the Global South, collecting qualitative data in Medellín, Colombia. While this city has recently experienced a reduction in homicides, different forms of violence continue to affect the daily lives of residents (Dávila 2018;Doyle 2019aDoyle , 2019b. The experiences of both researchers provide important insights for both insider and outsider researchers investigating violence in other Latin American cities, notably by highlighting the necessity and complexity of conducting fieldwork in volatile environments. First, this paper reviews the methodological literature on the range of challenges qualitative researchers can face in locations with high levels of violence and conflict. The review acknowledges how a researcher's individual identity can affect the data collection methods used and their ability to maintain and obtain participant trust. Also discussed is the importance of collaboration between Global North and Global South researchers when investigating violence and conflict in Latin America. We define the Global South as African, Asian, Latin American and Middle Eastern countries and the Global North as the Group of Eight and the five permanent members of the United Nations Security Council (Medie and Kang 2018). 1 Second, the article moves to the experiences of an insider and an outsider researcher collecting qualitative data in Medellín, Colombia, from 2014 to 2016. The narratives are presented in first-person format, as this is an effective mechanism through which to communicate the challenges faced in a volatile research environment (Behar 1994;Doyle and McCarthy-Jones 2017;Shenaz Hossein 2016). The researchers' experiences focus on their physicality, obtaining and maintaining trust with participants, obtaining reliable data and the challenges faced while collecting qualitative data amid violence and crime. The article concludes that although conducting research in a volatile environment does come with particular challenges, the findings from the qualitative data make an important contribution to academic debates. Further, Global North researchers should collaborate with Global South researchers in their efforts to not only understand the realities of urban violence, but also to ensure that structural imbalances in the literature can be addressed. Finally, the article provides recommendations to the global academic community as to how some of these imbalances can be addressed.

Literature Review
In an effort to gain further understanding of the realities of violence in Latin America, researchers are encouraged to take a qualitative approach to their work, such as by conducting fieldwork (Abello-Colak and Pearce 2015; Auyero, Burbano de Lara and Berti 2013;Baird 2012;Berents and ten Have 2017;Moser 2004;Moser and McIlwaine 2004). Although fieldwork in any form is challenging, the multitude of risks researchers face in environments with elevated levels of violence and conflict have been widely acknowledged in the literature (Belousov et al. 2007;Gill 2004;Goldsmith 2002;Nordstorm and Robben 1995). In these environments, researchers must compromise between ambient risks and situational risks. Ambient risks can derive from a hazardous environment, such as a war zone, and situational risks can be evoked by the researchers' own presence or actions, such as asking questions about sensitive topics (Lee 1995). Not all of these risks can be fully comprehended before exposure to the field. As Gill (2004: 1-10) commented on her experiences collecting qualitative data in the Dominican Republic, 'being street smart is not something one can necessarily teach in a classroom … making it impossible to have known the insecurity of the field site ahead of time'. Researchers can also struggle with local crime data before arrival due to lack of quantifiable data, distrust of the police or poor record keeping by authorities (Goldstein 2014 shown that in Latin America, there are discrepancies between perceived levels of crime and reported levels of crime (Corbacho, Philipp and Ruiz-Vega 2012;Doyle 2019b;Restrepo and Moreno 2007;Sanguinetti et al. 2014).
There are also challenges with the data collection methods used in these environments. Researchers may find it difficult or impossible to apply rigorous methods used in other research to obtain data reliability and validity (Barakat et al. 2002: 991). In qualitative research, data validity refers to the degree that a description is credible with one's interpretation of a reality (Smith and Heshusius 1986). While all qualitative research is predicated on establishing rapport and trust, in environments with high levels of violence and conflict, obtaining trust and establishing rapport can seem 'unending' (Goldstein 2014: 3). In these environments, there can be high levels of trauma which can affect social relationships and networks (World Bank 2011: 34). For example, in Guatemala, Moser and McIlwaine (1999) discussed how in some communities the history of political violence meant there was reluctance among residents to share their experiences and trust had been severely eroded. These levels of distrust can affect not only the researchers' access to participants but also the types of techniques they can use (Belousov et al. 2007;Nordstrom and Robben 1995). Residents may also be conditioned to 'dissemble' when outsiders are asking questions, as individuals with these characteristics are often tax collectors or police authorities (Goldstein 2014: 3). Therefore, researchers in these environments are encouraged to use purposive or snowballing sampling to assist them in identifying potential participants (Goldstein 2014).
Researchers can also face the challenge of fragmentary data and cannot assume the existence of a body of verifiable data (Barakat et al. 2002: 992). Barakat et al. (2002) discussed how in conflict zones, sources tend to be incomplete and unreliable because interruptions caused by conflicts mean that information is not systematically collected and is commonly lost or destroyed. Crosschecking and controlling for accuracy can also become far more difficult, especially in the absence or one-sidedness of written evidence (Helbardt, Hellmann-Rajanayagam and Korff 2010: 358). For example, Caracas, Venezuela was recently excluded from the 2017 Igarapé Institute's list of the most dangerous cities in the world due to local data discrepancies on homicide rates (Aguirre and Muggah 2017).
Researchers have acknowledged how in a wide range of fieldwork environments, a researcher's identity can have implications on their experiences in the field and on their findings (Coffey 1999;Hume 2007;La Pastina 2006;Roguski and Tauri 2013). For example, Hume (2007) argues that key to the epistemological challenge of researching violence is a researcher's identity, which shifts and is highly relational, as it is shaped by the subject of the research and its participants. Therefore, the 'reality' of the research topic is based on the subjective experiences and circumstances of the researcher (Fawcett and Hearn 2004). One area of researcher identity concerns if a researcher is an insider or an outsider. Ferrell and Hamm (1998) discussed how individuals involved in illegal activities may talk to 'outsider' researchers about some aspects of their lives, but hesitate to discuss these illicit activities with 'insiders'. From conducting research on second-generation Muslim immigrants involved in informal economies, Bucerius (2013) found how her outsider status encouraged young men to trust her with insider information that they would have not otherwise shared with 'insiders'. Bucerius discovered that her identity markers (being an outsider) were actually key to garnering 'insider' information and facilitating effective research. Further, Brownlie (2009) discusses how it is not possible to predict how participants will position a researcher and how their commonality or difference will be constructed or perceived. While insiders may start with more in-depth knowledge and perspectives about the research area, an insider may fail to notice pertinent issues because of their inability to step back from the situation (Kitchin and Tate 2000). Further, while an insider may come from a particular community, individuals do belong to several communities simultaneously and can, therefore, be both an insider and an outsider (Narayan 1993). Previous research has even acknowledged how both insider and outsider researchers investigating violence can experience very real fear and anxiety (Hearn 2004;Hume 2007). From conducting research in El Salvador, Hume (2007) documented how despite having lived in the country, she experienced very real fear. However, she, along with other scholars (see Edwards and Alexander 2011;Pickering 2001), acknowledged the ability to disconnect from the research is a freedom only awarded to outsiders.
www.crimejusticejournal.com IJCJ&SD 90 2020 9(3) Another theme acknowledged in the methodological and urban violence literature is the importance of researchers from the Global North collaborating with those from the Global South to ensure the realities of violence are accurately communicated. Generally, the Global South refers to African, Asian, Latin American and Middle Eastern countries that are also members of the Group of 77, 2 and the Global North generally includes the Group of Eight and the five permanent members of the United Nations Security Council (Medie and Kang 2018). Scholars from various fields have acknowledged how the global production of knowledge is skewed to the Global North (Connell 2007;Hogg, Scott and Sozzo 2017) and that researchers from the Global South are under-represented in top, international, peer-reviewed social science journals (Medie and Kang 2018). Researchers in the Global North generally have a wider global reach and are considered to be at the forefront of knowledge production and dissemination, whereas South-based scholars are seldom part of major debates and conversations (Medie and Kang 2018). The latter also have less access to the same funding as those from the Global North, not only limiting contact with scholarly journals but also preventing travel to international conferences, which are often held in the Global North (Minai 2018). Recently, scholars have challenged this North-South divide in areas of violence and conflict by introducing southern criminology. Their aim is to be more 'inclusive of and responsive to the global problems of justice and security in the 21st century' (Carrington, Hogg and Sozzo 2016: 2), taking into account the histories and patterns of crime, justice and security outside the Global North.

Medellín, Colombia
In As such, this article does not intend to explain the significant reduction in homicides in Medellín but rather the methodological challenges of conducting qualitative research in environments with high levels of violence and conflict. The following narrative presents the experiences of an outsider from the Global North and an insider from the Global South collecting qualitative data in Medellín. Given the limitations of relying on homicides rates alone to measure levels of violence ( interviews with six government officials, five former government officials, three members from NGOs, five community leaders, seven academics, one business leader, one social commentator and one journalist. I also conducted a focus group with local urban violence researchers.

Researcher Identity
I knew from previously living in Latin America that being a curly, blonde-haired female would distinguish me from the locals. I also knew that people would be 'surprised' when they heard my Spanish accent, which is a mix of slang I had acquired from different Latin American countries where I had studied and worked.
My housemate in Medellín added to this vocabulary by teaching me what he called 'paisa avanzado' (advanced Medellín slang). Most interviewees were also surprised that a woman from Australia was researching the levels of violence in Medellín. However, their astonishment often provided a type of 'icebreaker' for the interviews, as I would pre-emptively explain to participants that I was 'muy mona' (local term for a blonde female) so they would recognise me and, upon meeting, they would laugh in agreement: 'you are not wrong, you are muy mona'. This provided a good opening to start the interviews and establish rapport with the participants. My Spanish accent also helped to ease discussions, as many would comment on how 'you look muy mona but sound like you have lived in Latin America for years'.
I considered it important to take public transport and taxis and even walk to venues to gain more understanding about daily life in Medellín. As I often had meetings in different neighbourhoods throughout the city, I considered these were safe and reliable ways to commute. I found that because of my identitybeing a mona with an odd Spanish accent-people were interested in knowing what I was doing in Medellín. In particular, chatting to taxi drivers provided me with perspectives of day-to-day life in the city, especially since we were often stuck in traffic and, thus, able to have long conversations. These taxi drivers would tell me stories about how they used to work for Pablo Escobar (former leader of the Medellín cartel), that the day before they had seen a shootout on the street or that they could identify various gang members. Sometimes, I wanted a day off from my research, but my identity meant that inquisitive taxi drivers would always want to chat with me.
My interviews with NGO members and community leaders often required me to travel to more marginalised neighbourhoods. Despite organising a meeting time with participants beforehand, they often arrived late or did not arrive at all. This meant that I was required to wait for long periods in public places in which I could not blend in, given my appearance. Often, locals would approach me or stare at me, as it was uncommon to find a mona in these neighbourhoods. Forced to wait in areas such as public parks, I often felt unsafe checking my phone to see if participants had confirmed their arrival time with me. For meetings conducted in public venues, such as coffee shops, I noticed some participants 'cautiously' looking over their shoulders or lowering their voice when discussing certain topics, such as details of local gangs. In one interview conducted in the San Javier Library Park, one participant and I heard the sound of a camera clicking, as if someone were taking our picture, and even noticed another eavesdropping on our conversation. The person taking the picture could have been doing so because they had never seen a mona in these neighbourhoods or because they knew I was a researcher. Situations like this made me realise that while my particular identity helped in terms of establishing rapport, it was also a disadvantage, as I was not able to easily 'blend in' throughout the city and because participants might have been taking a risk talking with me.

Obtaining Reliable Data
My acquired understanding of the local context combined with my various networks enabled me to crosscheck data and ensure that I was obtaining reliable information about violence in the city. However, given the complexity of the explanations I uncovered, it was not always possible to obtain a complete understanding. For example, one evening I took a taxi with a friend from a middle-class neighbourhood and the driver told us he had just been robbed in one of the marginalised neighbourhoods. He told us that this was the first time he had driven to this neighbourhood because he had previously been afraid to do so. My friend and I had given him an extra COP$50,000 (USD$12) after hearing his story. When I told my more privileged friends about this, they laughed and said 'I shouldn't be so stupid to believe his story. It is a scam and you fell for it.' Another day sitting in traffic, a young, male taxi driver told me how gangs control the city and this is why the homicide rates are so low. He has knowledge of this because he was in a gang but left to work as a taxi driver. While he knew he would not earn as much money working as a taxi driver at least he could leave his house without fear of being assassinated. Indeed, his friend had been killed the previous month over a disagreement about a phone, prompting the realisation that life was worth more than a 'silly' device. When I finally reached my destination, I wished the taxi driver good luck and paid him the exact fare. I told an upper class friend about the conversation and he replied, 'you shouldn't believe these ex-paras [local term for paramilitaries]. They are all criminals. I am surprised that he did not rob you.' Moments like these made me question who was telling me the truth-the taxi drivers or my elite friends? These are just some examples of the different 'realities' of violence in Medellín and the importance of crosschecking data to ensure they are reliable.

Challenges of Researching Violence
Homicide rates were the lowest 'in 35 years' when I was conducting research in Medellín. Despite this, I would encounter different forms of violence on a daily basis. On at least three occasions, I saw on a morning run the police pulling dismembered bodies from the Medellín River. On a weekly basis when leaving the University of Antioquia, I would encounter riots by encapuchados (hooded vigilantes). One afternoon when arriving at EAFIT University, I narrowly missed a targeted assassination by a sicario (trained assassin). Given these levels of violence, I was required to take certain precautions to ensure I was safe. I only conducted interviews during the daytime and researched the particular neighbourhood I was visiting beforehand. I also advised a local contact of my whereabouts for the day. Often, my housemate would remind me 'ten cuidado' (be careful) when I departed for interviews. However, I soon came to realise that I had more knowledge of the marginalised neighbourhoods than the residents from the more elite neighbourhoods.
The issue of researcher safety was evident during my time in Medellín. In 2015, three human rights leaders were assassinated in less than a month and a local journalist was threatened shortly before my arrival (Garcia 2015;Laverde Palma 2015). One of the potential participants for my research was shot 25 times in a marginalised neighbourhood in broad daylight in May 2015. I wanted to interview this community leader about the knowledge he had about the local gangs co-opting the participatory budgeting process. Some media reported that he had been threatened and assassinated by the gangs because of this knowledge. Following this, I was advised by a local researcher to avoid any 'unnecessary risks'. I was not sure how to interpret this, as I was taking a risk simply by being in the city conducting research. What, then, constituted 'an unnecessary risk'? When I questioned another researcher, he told me not to worry: 'you are at a lower risk given that you are foreign female-no one would threaten you because of your identity. However, you should be more careful now.' I interpreted this advice as limiting my visits to the marginalised neighbourhoods and would only return with company. I also had to acknowledge that this limitation would likely affect the data I could collect for the remaining fieldwork period.
www.crimejusticejournal.com IJCJ&SD 93 2020 9(3) As a researcher from the Global North, I struggled to comprehend the 'reality' of the violence. Everyone I encountered told me about their experiences, from having a family member assassinated or kidnapped to watching someone be murdered in broad daylight. I had to constantly remind myself that my role as a researcher was to observe and I was there to try to understand the realities of the violence from the perspective of residents. Some weekends I would leave the city to nearby towns to try and 'escape' my research. Yet, it would follow me. One weekend, I was staying in a hotel in a small town outside of Medellín and a doctor confided in me how she did not care for the perpetrators of violence, as in the last month a gang member had interrupted a surgery she was conducting on a gunshot victim. The gang member put a gun to her head and screamed, 'if you save this man, I will shoot you, your work colleagues and your family'. While I had wanted a small 'escape' from my research, I had to remind myself that insiders did not have the privilege to simply leave. As an outsider, I could choose to leave and I could choose never to come back.

The Insider
I am a Colombian male researcher who has lived most of his life in Medellín. My research is qualitative with an interdisciplinary focus between sociology, political science and criminology. Violence has been present my entire life, and I have lost count of the different forms of violence I have encountered and observed. I have been the victim of multiple robberies, my relatives have been assassinated and a classmate made death threats against me over poor grades. As an example of my experiences with violence, in the early 1990s, my father, who worked as a taxi driver, received a threat against his life. One afternoon, the neighbourhood guard approached me and my brother and told us that a couple of days ago there had been a shootout between the local gangs, and a gang member had been injured. My father had seen the injured gang member and, similar to other drivers, was too scared to stop and help. However, the gang wanted revenge because my father did not stop, and offered a stolen motorcycle as 'payment' for his assassination. Following this, my father rarely left our home, and my brother and I were too scared to tell anyone about the threat. We were lucky that the gang members did not find my father, but it was a strong lesson of how much, or little, life was worth in Medellín. Someone was prepared to kill my father for a stolen motorcycle, but another victim could have been worth less than this. This was my reality growing up in Medellín. You could not escape the brutality. These lived experiences mean that I am not an outsider when it comes to researching violence. Other Colombian researchers such as Sanchez (2006) and Guzmán Campos, Fals Borda and Umaña Luna (2010) have reflected on their experiences as insiders, studying the violence in Colombia. In Snyder and Duran-Martinez (2009), Duran-Martinez provides her reflections as an insider researching these trends in Colombia, but also as an outsider investigating patterns of violence in Mexico.

Researcher Identity
Being a young (18-to 35-year-old) male Colombian researcher conducting interviews and focus groups in the marginalised neighbourhoods of Medellín, I have had to mitigate risks. For example, I need to think about how I look when I visit these communities, where residents may initially think that a 'smartly dressed' young man is a police officer or a government employee-which can be, and has been, grounds for assassination. Further, in most, if not all, of the marginalised neighbourhoods, there are 'spies' who work for the local gang and report to gang leaders any 'strangers' (such as young men) entering the area. These 'strangers' are also at risk of being assassinated. Therefore, if residents are seen talking to unfamiliar people, it can carry both ambient and situational risks for the researcher and the research participants.
While there is no denying my status as an insider, there are moments when I feel like an outsider. I reside and work close to the elite neighbourhoods where the homicide rate is similar to those in some North American cities (nine per 100,000 people), whereas in marginalised neighbourhoods the rates are equal to or worse than some cities in Central America (150 per 100,000). Therefore, the overall homicide rate for the city (27 per 100,000) does not reflect the daily reality of violence for many residents. Inequalities are also noted in the quality of public services and the provision of security services. In the elite neighbourhoods, similar to the gated communities in privileged parts of other Latin American cities, residents can afford to pay for private security guards. In the marginalised neighbourhoods, residents www.crimejusticejournal.com IJCJ&SD 94 2020 9(3) instead rely on local gangs to provide them this service (Dávila 2018). These residents would tell me how they did not feel like they 'live' in Medellín, as when they visit the city centre they say they are 'going down to Medellín'. There are many different realities in the city and, as a researcher, I can feel like both an insider and an outsider on a daily basis.
As a resident of Medellín, I feel responsible for ensuring that my participants do not feel 'used' by researchers. Some participants have told me how they invested their time and integrity participating in research but never received any benefits. There was one case in which an international organisation took pictures of local children without their permission and used these photos to obtain monetary benefits, while the community did not receive anything. As a minimum, researchers should provide participants with the research findings, such as a final report or publication. As an insider, I am in a better position to do this than an outsider.
I also noticed how some outsiders collect their data, take the knowledge, publish it and receive international recognition. Meanwhile, the participants do not receive any recognition and, if anything, are viewed by outsiders with 'pity'. Outsiders are paid to return to the city to deliver their findings in conference rooms at five-star hotels, with audiences anxious to hear the 'secrets' of those in marginalised neighbourhoods, but the participants were denied an invitation themselves. There have also been situations in which researchers from the Global North work with local scholars in collecting data, only to be published under the name of the Global North researcher. Local researchers are not even recognised in Q1-ranking publications or in bestselling books.

Obtaining and Maintaining Trust with Participants
I find research participants through snowball sampling. To contact potential participants, I use my mobile phone and the 'WhatsApp' application. There are some risks with using this app, as participants can see my picture and my name, but I have found that revealing this information is an important way to gain their trust.
Before organising focus groups with community organisations in marginalised neighbourhoods, I will arrange to meet with their leader. This is an important part of gaining their support as well as participants' trust, and provides a much better position for researchers to enter a neighbourhood safely, as the leader has 'recommended' you. The leader can also explain the 'informal rules' of a neighbourhood. For example, if you arrive in a car, you must have the windows down and, at night, you must have the interior car lights on. Similarly, when arriving on a motorcycle, you cannot wear a helmet. These rules ensure that the locals can identify any 'strangers' entering their community. Indeed, I have had to abide by these 'rules' to protect my own safety.
When initially meeting with a leader, I explain who is financing the research, the research methods used and my personal background. I have found that some individuals may not want to participate in research that is sponsored by the government or if the researcher is working directly with the government. Generally, this is because residents may believe that if they reveal sensitive information, such as if friends or family members are involved in illegal activities, that it will be shared with police or other government organisations. I also take time to clearly explain the objectives of the research. One participant told me of how a researcher in the past did not clearly explain the research objectives, which made her feel uncomfortable and a sense of deception. Another participant explained how they told a researcher everything they knew about a topic and then felt betrayed after believing they divulged 'too much' information and didn't feel comfortable speaking with other researchers in the future.
Initial meetings with leaders also offer an opportunity to answer questions and for them to explain their motivations for participating in the research. Some leaders want academic recognition while others want to see some positive change in their neighbourhood, believing their participation will see this occur. Some just want their experiences to be recognised and heard, and generally find the interview process to be www.crimejusticejournal.com IJCJ&SD 95 2020 9(3) cathartic, as it might be the first time they have shared their story. Others, who seek only monetary compensation for their participation, are generally avoided in my research.
Obtaining Reliable Data I analyse how different media sources report incidents of violence in the city to ensure I have the most valid and reliable data about the realities of city violence. In Medellín, some newspapers present information in a scandalous manner (amarillista) and others are more factual. Some newspapers have a conservative bias and others are more liberal. Often, the residents from marginalised neighbourhoods read the more scandalous newspapers. One recent example of a headline was, 'They found the head of a dismembered woman', with the by-line, 'They found the torso on Thursday'. This same incident might not have even been reported in the more conservative newspapers. As a researcher, I need to crosscheck data to ensure that I have the most valid 'reality' of violence in the city.

Challenges of Researching Violence
While I am an insider and have grown up around violence, it is often difficult to separate my role as a researcher and my identity as a resident of Medellín. For example, in one focus group in a marginalised neighbourhood, I asked participants about their experiences with security and trust. In the middle of discussions, a señora ran out crying. I followed her, asking her why she was so upset. In tears, she told me that in the same focus group were two boys she knew had killed her son. These boys were sitting across from her, laughing and smiling. I reminded the señora that she was free to leave the focus group at any time, but she explained how she needed the money (COP$18,000 or USD$6), which was offered for participation in the research. These realities of the violence are often difficult to comprehend both as a researcher and as an insider.
As an insider, I have struggled to balance adhering to research ethics principles and protecting my participants. For example, some do not feel comfortable signing consent forms and having their interview recorded. Others have contacted me months, or even years, after their interview, frantically asking me to destroy the recording because they could not sleep at night knowing they had told me 'too much' and felt at danger. Whereas an outsider can take a plane and return to their home country, I am unable to escape this complex reality.

Reflections from Insider and Outsider Researchers
Upon considering both insider and outsider perspectives, we now return to the third-person narrative to synthesise the four themes identified by the researchers. The first theme is the importance of recognising one's own identity and how this affects the data collected. Both researchers discussed how they needed to consider their physicality and how participants perceived them. Although the outsider used her identity to establish rapport with participants, it also meant she was unable to 'blend in' with the locals, thus, risking participants' safety when talking with someone so visibly foreign. Whereas the insider had more commonality with respondents, given his experiences with violence, his identity meant he had to mitigate more risks than the outsider. Further, he felt more responsibility in ensuring that the findings were communicated to his participants, given his closeness to them.
The second theme is obtaining and maintaining trust with participants. While trust is important in all qualitative research, the literature acknowledges its particular value when studying violence (Goldstein 2014;Moser and McIlwaine 1999;World Bank 2011). Both researchers discussed the importance of using snowball sampling for recruiting participants and gaining 'recommendations' (thus, obtaining their trust).
The outsider discussed how, given her identity, she was able to maintain and acquire participants' trust because she was only 'visiting' the city and so if they felt they had revealed 'too much' , their stories would be safe. Conversely, participants revealed to the insider their fear of what the researcher might do with their data, particularly if that information ended up in the hands of local authorities.
www.crimejusticejournal.com IJCJ&SD 96 2020 9(3) The third theme regards fragmentary data and researchers' need to crosscheck information for reliability, especially in environments with high levels of violence (Barakat et al. 2002). While the insider had local knowledge about different media sources, it was important to acknowledge their individual demographics.
In contrast, the outsider struggled with comprehending what the 'truth' was and how residents may have different 'truths' as to the reality of violence in the city. This insight shows the importance of researchers of urban violence conducting fieldwork to obtain the most valid and reliable reality of violence from local perspectives.
The final theme concerns the challenges of collecting data in an environment with high levels of violence and conflict. Both researchers had to mitigate situational and ambient risks when collecting data (Lee 1995). Whereas the outsider had to balance perceived and real risks following the assassination of a potential participant, the insider had to balance his role as a resident of the city and as a researcher. This included adhering to ethics principles and protecting the safety of his participants. The challenges presented provide important insight for qualitative researchers investigating violence in other Latin American cities.
This article contributes to the methodological and urban violence literature by emphasising the importance of collaboration between Global North and Global South researchers to ensure the most valid and reliable reality of urban violence in Latin America is presented. As acknowledged by Hogg, Scott and Sozzo (2017), Global South and Global North researchers should collaborate in their efforts to not only further understand violence on a global level, but also to find solutions that acknowledge the voices of participants. However, as presented in this study, there are structural and democratic imbalances in the academic community that do limit this collaboration. To address this, we offer four recommendations to future researchers and senior members of the academic community. First, researchers of violence should ensure they take collaborative approaches between academia and local communities, such as the recent participatory research project by Pearce and Abello-Colak (2019) and local researchers in various Mexican cities. Second, international conferences should be encouraged to offer discounted rates to scholars of the Global South, as the Latin American Studies Association does, and/or scholarships for Global South researchers, recently introduced for the Crime, Justice and Social Democracy Conference. These conferences should not only be offered in multiple languages, such as Spanish and English, but also have specific panels on the methodological challenges researchers face when studying violence. Third, journal editors should offer discounted subscriptions to Global South institutions and provide editorial assistance with the translation of articles to English. Finally, Global North researchers should ensure that as part of their research projects, whether written or verbalised, the findings are presented in the language of their research participants. This final recommendation will assist in protecting the integrity of research participants in the Global South.

Conclusion
Researchers face a range of challenges when collecting data in environments with high levels of violence and conflict. These challenges range from data collection methods to maintaining and obtaining trust with research participants. This article has contributed to the methodological and urban violence literature by providing insights from insider and outsider researchers collecting qualitative data in Medellín, Colombia. Their reflections showed that, especially in environments with increased volatility, insiders and outsiders must acknowledge how their physicality plays a role in their potential access to participants. Both insider and outsider researchers need to adopt different mechanisms for obtaining and maintaining trust with participants. While outsiders are free to leave this complex environment, insiders remain. This article also emphasised that both insiders from the Global South and outsiders from the Global North are responsible for collaborating in their efforts to address some of the structural, economic and democratic privileges facing the academic world to ensure that the most accurate and valid reality of urban violence in Latin America is communicated. It is only by sharing these realities that policymakers and scholars of urban violence will be provided with some of the tools needed to address the rising levels of violence in Latin America.