Abstract
Cannabis Social Clubs are typically non-profit associations established by adult users of cannabis, which produce and distribute cannabis among their members. Such associations can be found in several European countries and beyond, but with exception of Uruguay, they are illegal or otherwise operate at the margins of domestic drug control legislation, at risk for law enforcement detection. We have conducted several individual and collaborative studies on Cannabis Social Clubs and their key actors (e.g. Cannabis Social Clubs leaders, users affiliated as members, cannabis growers, other stakeholders), primarily drawing on qualitative methods (e.g. ethnographic fieldwork, interviews, qualitative media analyses) but also employing a quantitative approach (e.g. online surveys). In this article, we reflect on the research experiences of the authors in studying Cannabis Social Clubs, providing insights for future research in this area and within criminology and socio-legal studies. In particular, we aim to examine our approaches with regard to the recruitment of participants and research design, identifying good practices, but also discussing what the less successful strategies were. We draw on research conducted since 2014 in different countries, with different legal frameworks, which provides an opportunity for a comparative and more in-depth critical consideration of what might be helpful ways of reaching and researching hard-to-reach populations.
Background
Criminological and socio-legal research into hard-to-reach populations
The study of illicit activities, clandestine, or deviant behaviors and subcultures has been a central focus of criminological and socio-legal research. Otherwise hidden from view or misunderstood, this body of research has sought to “make sense” of those experiences and meanings, often from the perspective of the researched populations (Adler, 1990; Carlson et al., 2009). In this tradition, there have been a number of noteworthy ethnographic studies among illicit drug users, producers, and dealers (Adler, 1985; Becker, 1963; Bourgois, 1995; Potter, 2010; Zaitch, 2002). But other, more quantitative, research approaches have also been employed to the study of these phenomena (Barratt et al., 2015; Bouchard, 2007; Bouchard and Tremblay, 2005; for a review of primarily survey-based drug research, please see Miller and Sønderlund, 2010; Potter et al., 2015). An initial challenge, both practical and methodological, common to this research, relates to the hidden nature of the populations being studied (Rhodes, 2000). These are, for a variety of reasons, hard-to-reach groups.
A population may be considered hard-to-reach due to its physical or social location (e.g. this is the case for groups in remote geographical areas or for groups of social elites; Ellard-Gray et al., 2015; Marpsat and Razafindratsima, 2010). In addition, vulnerable or stigmatized groups are also considered hard-to-reach populations, because of reasons such as fear of losing financial support, lack of trust in institutions, or fear of misuse of research results (Ellard-Gray et al., 2015; Marpsat and Razafindratsima, 2010). Heckathorn (1997) summarized it as follows: “there exist strong privacy concerns, because membership involves stigmatized or illegal behavior, leading individuals to refuse to cooperate, or give unreliable answers to protect their privacy” (p. 174). For these hidden populations, Heckathorn (1997) pointed also to incomplete or absent sampling parameters as a typical challenge to do research among hard-to-reach groups: “no sampling frame exists, so the size and boundaries of the population are unknown” (p. 174).
Researching hard-to-reach populations can bring along additional challenges throughout the different stages of a research project. At the outset, it may be difficult to identify, and therefore, recruit potential participants within that population (Ellard-Gray et al., 2015; Faugier and Sargeant, 1997). There may be perceived and/or real risks associated with participation in the study for hard-to-reach/vulnerable groups too: potential participants may fear the research is useless, that it may not be helpful for their community or result in heightened legal, social, physical, or psychological risks for the individuals taking part in the study (Barratt et al., 2007; Ellard-Gray et al., 2015). Ensuring participants’ engagement for a long period of time may also prove difficult due to specific characteristics of vulnerable groups (Ellard-Gray et al., 2015). This has led researchers to devise strategies and research designs more suited to reach and research these populations, which we address throughout the article, in light of our own research experiences. What is particular to our contribution is that we were able to make a collective reflection drawing on the lessons from multiple research experiences, in different legal scenarios, but focusing on the same phenomenon and therefore engaging with similar populations. That allowed us to take a more holistic view of the issues at stake. In documenting these, we hope to provide insights that may be carried across to the study of other populations and contribute to laying the ground for further innovative thinking in developing appropriate recruitment and methodological approaches for the study of hard-to-reach populations.
Introducing Cannabis Social Clubs: a hard-to-reach group?
Cannabis Social Clubs (CSCs or Clubs, hereafter) are typically non-profit associations formed by adult cannabis users. Most of these associations advocate for cannabis reform and implement a diverse repertoire of action: collective cannabis cultivation, competitions among cannabis growers, study meetings and workshops on cannabis-related issues, protests, informative campaigns, proposals for alternative legislation, concerts, and other social events (Alvarez et al., 2016; Arocena and Aguiar, 2017; Calafat et al., 2000; Marín, 2009; Pettitt-Schieber, 2012). Key among these actions is the supply of cannabis to the members of these associations (Decorte et al., 2017; Jansseune et al., 2019; Pardal, 2018a), 1 which is not without risks for those involved. Indeed, we should note that Uruguay remains the only country to date to have (since 2013) enacted a detailed nationwide legal framework recognizing and regulating CSCs and the way in which they can produce and distribute cannabis to their members (Queirolo et al., 2016). In most other jurisdictions where CSCs are present the cultivation and/or distribution of cannabis for recreational purposes remains prohibited, albeit law enforcement priorities and practices may differ from country to country (Arana and Montañés, 2011; Decorte et al., 2017; Pardal et al., 2020).
While some CSCs, as it is the case with other activist groups or social movement organizations, may see participation in a research project as a way to “make their voices heard in a different arena” (Balsinger and Lambelet, 2014: 156; Klein and Potter, 2018), not all will share this view, and some may prefer to keep their existence and activities away from outsiders’ eyes. There are also indications that CSC members may be primarily interested in joining a CSC to secure access to cannabis, and engagement in activism or in other CSC activities may be secondary for some of them (Pardal & Decorte, 2018; Pardal et al., 2019). Furthermore, at least in part, due to the legal constraints experienced by most of these associations, many close down following law enforcement interventions or due to fear of detection, making it particularly difficult to keep track of their presence, activities, or affiliated members (Pardal, 2018b). Some of the research in this area has also pointed to the co-existence of several types of CSCs, with different degrees of engagement in activism and a different positioning in relation to whether (and the extent to which) they are overt about what they do (Pardal, 2018a; Jansseune et al., 2019). In addition, both CSCs active in contexts where no CSC-regulation has been enacted, as well as those operating in breach of existing CSC-regulation, may be cautious about sharing information about their (illegal) activities. Alternatively, they may perceive no potential gains in participating in the study. As a result, CSCs may remain hidden or uninterested in taking part in a research project.
This article’s goal
In this article, we reflect on the research experiences of the authors in studying CSCs, providing insights for research in this area and within criminology and socio-legal studies more broadly. Bearing in mind the ongoing discussions concerning cannabis policy reform and the importance of conducting research in both pre- as post-legalization scenarios (Kilmer and Pacula, 2016), it seems timely to draw lessons that may inform those future efforts. In particular, we aim to examine our approaches with regard to the recruitment of participants and research design, identifying good practices but also discussing what appear to have been the less successful strategies. We draw on research conducted in different countries, with different legal frameworks, which provides an opportunity for a comparative and more in-depth critical consideration of what might be helpful ways of reaching hard-to-reach populations.
Our experiences in studying Cannabis Social Clubs
We have been conducting research into CSCs since 2014 both as part of individual and collaborative research projects. In Table 1, we offer an overview of these studies. Throughout the article, we will refer to each of them using a code system (S1-S8), as indicated in Table 1, to facilitate the reading when drawing our methodological reflections. We have been able to learn more about CSCs in Belgium, Spain, the United Kingdom, Uruguay, and in 10 other European countries. 2 With the exception of Uruguay, in all other countries, the cultivation and distribution of cannabis through CSCs are not regulated by law. At the outset of these studies, we did not have a list of potential participants: not of CSCs, nor the users (or cannabis growers) affiliated with these associations. With regard to S8, we could draw on the official list of legal CSCs published by the national agency (Instituto de Regulación y Control del Cannabis or IRCCA) responsible for the oversight of the new cannabis law in the country (IRCCA, 2019). However, that list only includes information concerning the name of the registered CSC and the region in which it is based: no contact details or address are publicly made available.
Overview of CSC studies conducted by the authors.
CSC: Cannabis Social Clubs.
Sources: ¹Decorte, 2015; ²Pardal, (2018a, 2018b, 2018c); Pardal & Decorte, (2018); ³Pardal et al., (2020); 4Decorte et al., (2017); 5Jansseune et al., (2019); 6Bone & de Hoedt, (2018); Bone, (2019); 7Queirolo et al., (2016); 8Pardal et al., (2019).
The sample size should be interpreted with caution as (1) the (limited) available estimates suggest a varying number of CSCs per country and (2) in some cases, the researchers were not interested in reaching the largest number possible of CSCs (e.g. S6) or preferred to include different sources/voices (e.g. S2, S4, S8).
This study comprised all 28 European Union member states. However, only CSCs from the following 13 countries participated: Austria, Belgium, Czech Republic, Germany, Hungary, Ireland, Italy, Poland, Romania, Slovenia, Spain, the Netherlands, and the United Kingdom. Our own complementary searches and contacts with activists in the field suggest that, at the time of data collection, of the remaining 15 member states, 10 did not have CSCs.
We only refer here to the data collected in relation to CSCs in Spain.
A first note on research ethics
Ensuring the protection of research participants is a major duty of every researcher (Creswell, 2009). This is however not always a straightforward task and in fact much of the criminological and socio-legal research can be particularly challenging in this regard seeing as it is often interested in studying illegal, clandestine, or otherwise hidden activities—which could be subject to some form of repression (Lowman and Palys, 2001; Vander Laenen and O’Gorman, 2016). Studying CSCs and publishing the findings of that research may expose participants in a way that could result in organizational and/or personal harms. At the same time, privileging the voices of people who use cannabis and incorporating these perspectives into the evidence base could empower participants, and contribute to more informed and effective drug policies. In our CSC research, we took a number of steps to prevent or minimize harms for research participants.
Depending on the policies of the institutions we are affiliated with, we followed the procedures requested for ethical approval of our research plans (S2, S3). These typically involved preparing a relatively detailed research plan, as well as strategies for guaranteeing informed consent (including project information sheets and informed consent forms) and a data management plan. The ethics approval process at the University of Leicester was particularly comprehensive, given both the sensitive nature of drug policy research and the breadth of the study undertaken across Europe (S3). Alongside providing information sheets, informed consent forms and detailed invitation letters, we were required to provide a publicly available research protocol. In this protocol, we clarified our legal duties in relation to confidentiality, anonymity, and data storage; particularly since our legal duties varied across different national jurisdictions.
Beyond fulfilling this more formal requirement, we sought contact with gatekeepers or with some of the more central research participants (often CSC leaders) to discuss not only the objectives and design of our studies (and where the funding for those was coming from), but also to get their views on the more appropriate ways to engage with the CSCs and the broader community while conducting the studies (S1, S2, S7, and S8). For instance, in both S2 and S8, the research teams agreed not to have direct contact with the cannabis users affiliated with the CSCs (with regard to their participation in an online survey) but instead all communication was channeled through the CSC representatives—we sent out invitation letters, information sheets, and other materials to the CSCs which then shared it with their members. While this was a more burdensome communication strategy, it was the most agreeable option for the CSCs and their members. However, and as we explain later, this was not how we interacted with all participants and in fact, we developed a close contact with a core group of members of the community in most studies.
We were also careful in considering the extent to which confidentiality could be granted to study participants and discussed this with them. For example, in S7, we eliminated CSCs’ names from public presentations and academic publications in order to keep their anonymity. In S2, we noted that at an individual level no personal identifiers would be included in our research outputs, for instance, but that at an organizational level (i.e. regarding CSCs per se), we could not exclude some degree of identification—even if the actual names of the organizations were not disclosed (cfr. Snyder, 2002). In S6, we took a different approach as this was an autoethnography co-written by the participant (Greg de Hoedt). In any case, and before the start of each new data collection phase participants received tailored information about the study and how the data collected would be treated by the research team (in line with standard informed consent procedures).
During subsequent phases of the research, we sought to ensure the specific agreements made with regard to the treatment of data were met. That meant for instance, anonymizing interview transcripts or removing any Internet protocol (IP) data from our archives—or as noted earlier, ensuring that no individual identifiers were included in research outputs. Even so, when concerned that particular individuals could still be identified—and especially during the course of S2, where police investigations were taking place at the time involving some of the CSCs, we took additional action. For instance, when reporting on CSC-affiliated cannabis growers we tried to present the data in a more aggregated form, and omitted specific details that were not essential for the analysis of the issues at stake but that could be incriminating should those participants be identified (cfr. Adler and Adler, 2002; Wiles et al., 2008). We also shared preliminary drafts of the analysis with participants (S1, S2) in which any citations were highlighted in order to allow participants to check whether any sensitive information was being disclosed that could place them in a more vulnerable position (cfr. Snyder, 2002).
Conceptual fuzziness
Research into CSCs is further complicated by the fact that there is no consensual understanding of what a CSC is (or ought to be). There is no common agreed upon normative or shared legal definition (Decorte & Pardal, 2020), 3 there are different interpretations among the researchers studying the phenomenon, and perhaps, most importantly there are certainly disagreements among the community too. Some will emphasize CSCs’ activist role and strike to develop a diverse repertoire of protest (e.g. organizing marches, lobbying with politicians, etc.). Others will focus on CSCs as being cannabis suppliers, and even among these there will be disagreements as to how cannabis production should be organized (e.g. collectively by members of the association vs by external producers, etc.) and the extent to which CSCs may or may not seek to generate financial gains. These are just a few of the differences that have been identified to date (Decorte, 2015; Jansseune et al., 2019; Pardal, 2018a). Even in Uruguay, where the legal framework, to some extent, determines the way CSCs are functioning, there are different positions with regard to its implementation and to what the ideal features of CSCs may be (Queirolo et al., 2016; Pardal et al., 2019). In addition, CSCs operating outside of that legal framework may share a very different vision of the model.
In practice, this fuzziness had at least two broad repercussions to our research. First, to avoid restricting our analyses to a particular kind of CSC, we did not develop any pre-definition of what we would consider as constituting a CSC for the purposes of our studies. Rather, we drew on CSCs own conceptualization as such as a sufficient criterion for inclusion in the study (meaning that if an organization considered itself a CSC, then it was eligible for participation in the study). Even so, some uncertainty remained: for example, during S3 there were a few instances in which the research team had to reiterate the inclusion criteria to potential participants who were unsure about their eligibility to take part in the study. Second, throughout the studies, we remained attentive to any attempts or suggestions from participants to protect the association of the term “CSC” from some of the variants of the model. For instance, one of the stakeholders consulted in S3 argued that “to put all the models under the umbrella of CSC could be unproductive for the science and for the activism.” While we certainly noted the divergent views among the community with regard to their self-representation and goals, we did not limit the study to any particular view.
Mixing online and offline recruitment and communication strategies
Previous cannabis studies (in particular, among cannabis growers) highlighted the importance of engaging with potential participants through websites and online forums, as well as through social media (e.g. Facebook), (Decorte et al., 2019; Barratt et al., 2015; Potter, 2017). For some of our CSC studies, we created Facebook pages or websites (S2, S3, S8). In doing so, we wanted to be able to be both transparent about the projects and tried to identify and interact, online, with potential study participants. In the websites dedicated to the projects, we typically included information about the goals and methods of the study, about the research team and researchers’ affiliations—including also a contact form or the email addresses of the researchers. In three of the projects (S2, S3, S8), a link to the online survey was also included in the respective websites and/or Facebook pages.
In our experience, these were useful complementary tools to disseminate information about the study to a larger audience and especially to reach out and maintain some contact with CSC representatives and/or members throughout the projects. 4 Doing so required actual interaction though. For instance, in S5, we initially made an attempt to recruit CSC representatives online, based on the information provided on ‘Weedmaps’—a website providing the contact details of cannabis dispensaries, CSCs, and coffee shops worldwide. We were able to retrieve 100 email addresses of CSCs in Barcelona through that website. All were contacted, but only two CSC representatives responded and agreed to participate in an interview. This confirms the need to have meaningful interactions with online communities, rather than just posting advertisements of the study (for instance, in online forums) or “spamming” potential participants with unsolicited emails. Being vouched by members of those communities beforehand is also recommended (Barratt and Lenton, 2010; Barratt et al., 2015; Potter, 2017). With regard to Facebook, we had mixed experiences. In the context of S2, Facebook was a useful tool to publicly share information regarding the study (e.g. the launch of the online survey, the release of new publications from the project or relevant to the topic), and to find and “friend” or follow CSC pages or other cannabis activist organizations (see also Bhutta, 2012). It was also a way to exchange messages, in private, with some of the CSCs and in multiple instances we received updates regarding CSCs’ events or issues through this channel. In S8, we also tried this approach but were not able to generate the same amount of interaction. In particular, we paid for Facebook adds to reach CSC members and increase our survey response rate, but it did not seem to work. This could be explained in at least two ways. First, by law, CSCs in Uruguay are not allowed to be public, they cannot advertise or promote their activities, and as a result, they have less online presence than in other countries. Second, this might also be because we did not have a developed online network prior to the survey. In that sense, the use of social media seems to be more effective for research when it can build up on a preexisting online network. The experience in S8 differs from S2, where most CSCs had websites, Facebook, and YouTube accounts (Pardal & Tieberghien, 2018).
Nevertheless, primarily we followed more traditional offline routes of recruitment: we contacted and/or interviewed participants in previous studies in related topics, known cannabis activists, grow shop owners, or other stakeholders (in the legal side of the market) who referred us to other potential research participants (S1, S2, S3, S4, S5, S8). For instance, in the two studies in Spain (S4, S5), where CSCs are often represented by CSC federations or individual lawyers, we first interviewed a number of such stakeholders, learning important insights about the CSCs, and other contextual elements as well as, in some cases, gaining further access to the CSCs. Our strategy here was thus to start by approaching the more visible actors as a way to learn more about the community and through referrals within that social structure reach our target (sub)population (Heckathorn, 2011; Jacques and Wright, 2008). In Belgium and Spain, where the researchers had previously conducted studies into related topics (e.g. small-scale cannabis cultivation, domestic cannabis policy) 5 and thus had more “name recognition” among the target population, entry in the field and the early establishment of contacts with gatekeepers was facilitated (in line with Barratt et al., 2015). 6 Some of us also participated and spoke at events organized by CSCs or other cannabis activists (e.g. a podcast about drug policy in the United Kingdom, a debate during a cannabis fair in Uruguay, conferences or drug policy reform debates organized by activists, etc.), or invited CSC representatives to speak at our own academic events (e.g. Belgian CSCs were represented at a two-country conference on cannabis cultivation). By contributing to or inviting members of the community to these events, not only were we able to get to know other potential participants, but to some extent it also helped in shifting the power dynamics and building trust—a point we return to in the next section. For two of the projects listed in Table 1 (S2 and S8), we also prepared different dissemination materials (e.g. posters, flyers, QR cards) which we handed out at both grow shops and cannabis events or to the CSCs already identified and participating in the studies (in order to motivate members to fill in an online survey).
In both online and offline descriptions of the studies, we were careful to avoid stigmatizing language, narrow labels, and explicitly clarified that the research team did not hold any negative preconceptions of cannabis use or cultivation, of CSCs, their members, or their activities. Similarly, the literature in this area has highlighted the need to make the recruitment process less intimidating to further enhance the likelihood that potential participants accede to take part in the research projects (Decorte et al., 2019; Ellard-Gray et al., 2015; Sandberg and Copes, 2013). On one occasion, however (S3), the way we described the research project in the participant information sheet was considered as potentially leading by the ethics committee reviewing our application. We often needed to traverse a fine line between being “on their side” to increase participation and ensure that we met our ethical obligations as researchers (Becker, 1967).
Spending time in the field
Creating opportunities for dialogue with study participants and gaining their respect and trust not only facilitates participation (and the establishment of referral chains) but also has the potential to give the researcher a more comprehensive knowledge-base to interpret the research findings (Barratt et al., 2015; Ellard-Gray et al., 2015; Faugier and Sargeant, 1997; Potter, 2017). One of our studies (S2) had a strong(er) ethnographic component and we spent close to 2 years in the field (also, to some extent, in S4 and S5). During that time, beyond maintaining regular contact with representatives of the CSCs through social media, email, and phone, and multiple conversations in person, we were also present and made observations at different moments of activity of the CSCs: at protests and marches organized by the CSCs, internal meetings of the CSCs, growers’ competitions (so-called Cannabis Cups), at court sessions concerning CSC representatives or other affiliated individuals (e.g. cannabis growers), and so on. This was a gradual process; initially, while granted entry to the field, there was some reticence from a number of study participants. For instance, earlier on, a participant told the researcher that they thought the researcher did not (yet) understand the CSC model(s) and movement enough to talk about the topic at a public policy event. Gradually, that disbelief was overcome, and we became trusted outsiders (Bucerius, 2013). Although we do not claim to have immersed fully or for an extended period in the daily-life of those settings, the frequent “visits” as non-participant observers were insightful sources of data, and were also instrumental in building rapport with the gatekeepers (Bachman and Schutt, 2008; Bryman, 2012; Zaitch et al., 2016).
In addition, there exists a growing body of literature within drug policy scholarship which recognizes the value of a more inclusive, participatory approach to drug policy research. Ritter et al. (2018) highlight the importance of creating space and opportunity for the co-production of knowledge, by focusing on the process of participation, and scrambling the divisions between experts and laypersons, in order for new knowledge to emerge to identify drug policy problems and solutions. To avoid tokenistic consultation, Askew and Bone (2019) similarly reflect on the participatory process itself through adopting a sociocognitive approach (SCA) when analyzing their participant interviews on drug policy reform. Spending more time in the field with CSCs could likewise facilitate deeper insights into the clubs’ values and ideologies and the extent to which CSCs and their members internalize and/or challenge prohibition ideology. Indeed, asking CSCs to choose their preferred policy approach could be reductive/prescriptive since Askew and Bone (2019) found that the perspectives of people who use drugs on drug policy and the law can be diverse, complex, and that clarity surrounding regulatory options is needed. Since drug policy making invariably involves questions of values as well as evidence (MacCoun and Reuter, 2001), a meaningful participatory approach would require more time spent in the field (as was the case in S6, for instance). This could involve interweaving drug policy discussion with personal narratives, which could shape the public’s perception of cannabis and subsequently influence policy (Askew and Bone, 2019). For instance, in S6, an in-depth, unstructured interview was conducted by Bone interviewing Greg de Hoedt—the Chair of the United Kingdom’s Cannabis Social Clubs (a CSC Federation)—the subsequent article was co-edited and co-created by both authors to input de Hoedt’s experiential knowledge into the “expert” evidence base (Bone & de Hoedt, 2018). Elevating the voices of CSC members creates space and opportunity for the co-production of knowledge, to help legitimize drug policy outcomes (Bone & de Hoedt, 2018). Contrastingly, in S3, taking into account its geographical scope (all 28 EU member states), we did not have the means to truly develop an offline relationship with potential participants in most countries, and had to rely on an online survey as the main data collection instrument. This brought along difficulties in terms of both reaching potential participants, and gaining their trust or motivation to participate in the study. For instance, in one of the phone calls to invite participation, a CSC representative asked us if we could discuss the study in person and indicated that it would be too risky for the CSC to share the kind of information we were interested in online. In retrospect, we could have perhaps tried to develop in depth case studies in at least a few of the countries under analysis, prior to launching the survey. Indeed, with exception of S3, in all of the other studies, we typically spent some time in the field and conducted qualitative interviews with key figures of the community (CSC leaders, cannabis growers, etc.). An alternative possibility to gain access to CSCs and its members would be to run a Respondent Driven Sample (RDS). RDS is a method originally developed by Heckathorn (1997), and it is regarded as the gold standard sampling method for hard-to-reach populations. It is based on snowball recruitment performed by participants, not the research team. Participants are asked to recruit other subjects for the study, and in that way, hidden populations may be easier to convince to participate. This method was used in a study among frequent cannabis users and it might also be of relevance for future CSC studies (Boidi et al., 2016). 7 Moreover, with a view to being able to triangulate the data collected with other sources, complementing, and clarifying the findings, in most of our studies, we attempted to rely on multiple methods of data collection, including less invasive approaches for the target population (e.g. by relying on open sources such as media coverage or on other documentary sources; Creswell, 2009; Creswell and Clark, 2011).
. . . but avoiding the battlefield
While building a closer relationship with (potential) study participants or gatekeepers is important to the implementation of research projects (Dunlap and Johnson, 1998), it may also shorten the distance between researchers and the research participants/gatekeepers. The earlier Chicago School guidelines for fieldwork emphasize precisely the need for keeping some detachment from study participants (but the value of immersion in the phenomenon has since also been recognized; Adler, 1990; Jacques and Wright, 2008). The more participatory approaches we discussed earlier could indeed affect the researchers’ impartiality and/or be used to harness a particular research/reform agenda. Furthermore, continued research efforts among a community (for instance, we have remained in contact with key CSC individuals since 2014) can also impact the relationship between the researcher and the community. For us, conducting research among the CSCs and their broader community demanded caution in the way we interacted with each of the study participants and how we were perceived by them. Indeed, in the course of our studies, we often encountered cliques and disputes within the CSC movement, and it became evident as the research projects advanced that these groups may have specific vested interests. For instance in S2 (and to some extent also in S1 and S8), we found only a few instances of collaboration among CSCs, and there was distrust among the different CSC leaders as well as conflictive relationships (Pardal, 2018b). One of the interviewees in that study actually pointed to the existence of a “war on drugs between the Clubs” (Pardal, 2018b: 85). During fieldwork, accusations of theft (of cannabis plants), or of foul play among different individuals participating in the cannabis movement were communicated to us. While this is an issue commonly experienced in social movement research (Blee and Taylor, 2002), it could have had negative repercussions for the implementation of the study. It thus required a careful management from our side; we tried to avoid being aligned with a particular faction of the movement throughout the study (Balsinger and Lambelet, 2014; Beyens et al., 2016). For example, whenever contacted by a CSC representative who wished to tell us about recent events in the movement or a particular concern, we welcomed them in our offices at the University, indicating that this was a courtesy we would extend to all research participants, and that we would not interfere or support one CSC versus another.
In our experience, being associated with specific gatekeepers or groups can also be negatively perceived by the broader target population. For example, in S8, we built contacts with the official state agency overseeing the implementation of the legislation (IRCCA). Although we also reached CSCs through other (unofficial) networks and referrals, this agency sent out several invitations, in our name, to all the registered CSCs to participate in our research project. On reflection, though this was a necessary channel for us at the time, it is plausible that some CSCs may have refrained from participating because they did not fully adhere to the official codes of functioning established by IRCCA, for fearing that the research data would be shared with that state agency, and/or just generally associating the research team with the “other side,” that is, with the controlling institutions. 8
Legal CSCs: easier to reach, harder to motivate?
Most of our CSC research involved collecting data directly or indirectly related to activities (i.e. cultivation and distribution of cannabis) which are not legal and remain subject to criminal prosecution in the respective jurisdictions (EMCDDA, 2016). Our two studies in Uruguay (S7 and S8) were exceptional because that country has legalized and regulated CSCs since 2013 (Aguiar and Musto, 2015; Arocena and Aguiar, 2017), and we focused primarily on the practices of CSCs operating in adherence to that new cannabis law. One could expect that some of the difficulties associated with recruiting participants and avoiding withdrawal during the study would be less salient in the Uruguayan context, assuming that legalization and regulation of the cannabis market (and CSCs, in particular) was paired with some degree of de-stigmatization or normalization. Nevertheless, ensuring participation in those studies remained challenging. For instance, within S8, only representatives of 15 out of the 93 registered legal CSCs at the time agreed to participate. Although this comparison may be blurred to some extent by multiple factors (e.g. relationship of the research team with the community, recruitment strategy, research design), in our CSC studies in jurisdictions where their activities have not been legalized, CSC representatives seemed more eager to be engaged. In those cases, although there may be some fear of talking about criminal activities, there is also a strong interest in making the (cannabis) movement more visible in an attempt to gain recognition and societal support for the legal changes this community is striving for. Participation in research projects was used by some groups as a way of legitimizing their activities; by publicizing their “collaboration” with the University with which the research team was affiliated, emphasizing that public funding had been allocated to study their associations/movement, or even mobilizing some of our publications in the context of court cases. In S3, for instance, some of the participants expressed also appreciation of the research project in the final feedback question of the survey, writing “thank you for the work you do” (see also Decorte et al., 2019).
It can be argued that in a country where those community claims have generally already been granted (i.e. where a recreational cannabis market or the CSC model have been legalized), the eventual gains from participation might be seen as negligible or not worthy. This apparent research apathy may also be associated with a fear of being caught doing certain activities that are outside of the legal framework, or a more general avoidance of being controlled by the governmental agency that is in charge of implementing the law. In the field of cannabis studies, where in recent years several jurisdictions have taken steps or are considering policies that move away from prohibition, this could present a significant hurdle as researchers will seek to draw lessons on the implementation of those policies and perhaps encounter uninterested populations. It will be important to reflect on what may be adequate incentives or compensation(s) for participants in such research. In our studies, we did not offer any financial compensation, and although this is a relatively common practice, 9 there are opposing views on whether it is an appropriate and fair research practice, particularly for research with vulnerable groups (Collins et al., 2017; Fry et al., 2006).
Reflections for future research
Having presented and discussed our research experiences in the studies we conducted into CSCs, it is important now to consider whether more general lessons can be drawn that may be useful to other criminological and socio-legal research—as well as for future CSC research.
We assumed that our target population (in the case, CSC representatives, cannabis users, and cannabis growers), given its (un)known features and the ephemerous or volatile life-span of the associations, could be characterized as “hidden” or “hard-to-reach.” While this was the case in practice, it would be reductive to describe them exclusively as “hidden.” In fact, we often perceived the community as mainly being very “vulnerable” (for prosecution, and therefore, distrustful or careful), “stigmatized” (but at the same time holding clear convictions and engaging in activism), as “cooperative” but also “uninterested” or “apathetic.” Gaining a deep awareness of the community is thus key not only to securing access to the field and to maintaining that engagement throughout the research projects but, most importantly, to better contextualize and understand the complex and perhaps shifting nature of the issues under analysis. Having that kind of insight into the community will also be crucial to more carefully appreciate how the community and the various groups perceive themselves and each other, as well as any sensitivities that may underline those relationships (e.g. being aware of conflicts, tensions, or disagreements between individuals participating or informing the research projects). Building trust is of course a key element in that process. In our view, spending time in the field, online and/or offline (as proposed also by Potter, 2017), with meaningful interaction with the community can be an effective strategy to that effect. If the researcher and the community have an online presence, there may be benefits in sustaining some form of virtual contact, complementing the real life interaction (Barratt and Lenton, 2010; Miller and Sønderlund, 2010; Potter, 2017; Temple and Brown, 2011). While it may be challenging (given the typical funding and time pressures on research), allowing sufficient time within the design of research projects to participate in activities organized by the community, prior and during data collection, is also recommended (Ellard-Gray et al., 2015). Learning from other actors with links to the community (or from other sources) can also help shed light on the issues under analysis, while reducing the burden imposed on the community itself. In terms of research design, ethnographic methods have been considered particularly well-suited to the study of hidden groups (Carlson et al., 2009). This is not to exclude other possible recruitment or research approaches though—for instance, RDS has been successfully applied to the study of hidden or vulnerable populations, in particular within the drugs field (Boidi et al., 2016; Suárez H et al., 2014).
Legalization of an activity or behavior does not necessarily render that population more visible or willing to participate in research. On the one hand, the activity, behavior, or the community may still be perceived as being stigmatized, even after the passage of new legislation. Besides, some groups may continue to operate outside of the new legal framework and thus remain vulnerable to legal sanctions. On the other hand, the target population may have grown tired of contributing to research projects or feel “over-researched” over time (Clark, 2008), especially if there also seems to be little added value for their participation. Legal reforms may induce some degree of optical illusion for researchers—who may assume engaging with the community at stake may become easier per se. Our experience suggests that there may be new challenges in such scenarios and invites a reflection on what may be adequate incentives or ways to make future research more appealing to potential participants. In the case of CSC studies, it might be worth considering alternatives to monetary incentives as, beyond the due ethical considerations (for a discussion, see for instance, Fry et al., 2006), giving a small compensation for participation would be unlikely to substantially trigger or retain participation in light of what is known about the socio-demographic profile of the users affiliated with these associations (Pardal & Decorte, 2018; Parés et al., 2019). Perhaps it may be worth discussing with current participants whether providing training on harm reduction or educational initiatives on drug policy reform models would be more relevant contributions from the research team to the community and vice versa—to facilitate a more inclusive participatory approach and the co-production of knowledge. Meaningful participatory approaches could shift the power dynamics in drug policy research, by privileging the voices of people who use cannabis. Such a shift could increase the willingness of CSCs to participate as well. As the discussion around cannabis reform intensifies and other jurisdictions may take steps to legalize various supply models for cannabis, it will be important that researchers think carefully about how to ensure that we can continue to learn from these communities.
Footnotes
Acknowledgements
The authors would like to thank all the CSC representatives, growers, members, and all other stakeholders for their participation and assistance in their research projects. They thank also Charlotte de Kock for the helpful feedback on an earlier version of this article.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Dr Mafalda Pardal is supported by a post-doctoral fellowship by the Research Council of Ghent University. She received also research grants by the Research Foundation Flanders in the context of the studies mentioned in this article, namely FWO grant nos G.0A85.15N and V409318N. Dr Melissa Bone would like to thank the University of Leicester for supporting a period of study leave from September 2018 to January 2019 which helped her contribution to this research. Dr Rosario Queirolo, Eliana Álvarez, and Lorena Repetto were supported by OSF Grant OR2016-27307 while doing the research on CSCs in Uruguay.
