Abstract
This article explores the interplay between online and offline approaches in criminological ethnography. Criminology has come to embrace online research: in addition to offering numerous research benefits generic to the social sciences, the Internet offers solutions to various problems specific to active offender research. Furthermore, as many types of criminal or deviant behaviour increasingly have online aspects, so engaging in online research becomes both valid and vital to any meaningful ethnography. However, online approaches should be treated with caution: they are subject to their own limitations, and to rely on online methods as an alternative to traditional approaches can be as problematic as failing to embrace online research at all. Drawing on my experiences researching cannabis cultivation, I demonstrate some of the ways that offline and online methods complement one another. Online methods were useful in expanding my own study beyond the normal constraints of ethnography by generating a larger and more varied sample and providing access to more data than traditional ethnographic approaches. They were also essential for exploring the various online aspects of cannabis cultivation. But offline methods proved invaluable in accessing and recruiting respondents online and in providing the experience essential to participating in – and understanding – cultivation-related online interactions. Both approaches revealed findings not identified by the other, and research in each environment helped with understanding experiences and observations in the other. I argue that while there are clear strengths in online approaches to criminological ethnography, certain pitfalls arise when online techniques are used without employing face-to-face research as well. Triangulation of online and offline methods can enhance the understanding of many human behaviours, but may be particularly useful in overcoming the difficulties inherent in criminological ethnography. For many (although by no means all) criminological topics, online methods can usefully enhance, but not replace, traditional ethnographic techniques.
Introduction
There is a strong tradition within criminology of conducting ethnographic research with offender populations. 1 Critical criminologists, from the Chicago school onwards, have centralised qualitative, interpretivist methods as a necessary counterpoint to a mainstream administrative criminology characterised (and arguably crippled) by a devotion to quantitative positivism. However, criminological research of this kind has never been easy: methodological, ethical, legal and safety concerns abound and the competitive funding environment of contemporary academia is increasingly alien to the demands of ethnography.
As elsewhere in social science research, online approaches offer solutions to many of the challenges facing criminological ethnography. Generic advantages include enabling access to a larger number of respondents, handling greater volumes of data with ease, economic savings and temporal and geographical flexibility. Advantages pertinent to studying crime and deviance include ease of identifying and contacting hidden populations, providing (at least the illusion of) privacy and anonymity when discussing stigmatised activities, and separating the researcher from the immediate proximity of criminal acts or dangerous situations. What is more, many types of criminal and deviant behaviour increasingly have an online presence whether they manifest online (i.e. cybercrimes), are facilitated by Internet technologies (Décary-Hétu and Aldridge, 2015) or are (merely) subject to discussion in virtual social spaces. To approach this from a different angle, social interaction now takes place online as much for those whom we might identify as ‘criminal’ as for any other contemporary social group. 2 Either way, online approaches are not just valid, but arguably vital to any meaningful understanding of much contemporary criminality: virtual ethnography is justifiable on epistemological as well as utilitarian grounds.
However, Internet approaches should be treated with caution: both online research tools and online research sites (Barratt, 2011; Hine, 2000; Illingworth, 2001; cf. Markham, 2007, 2011) are subject to their own limitations; cybercrimes aside, criminal acts still actually occur in the ‘real’ world, and offenders and victims (like everyone else) still predominantly reside and interact offline. While there is clearly an argument for online aspects to ethnographic research, these should be a supplement rather than an alternative to traditional face-to-face methods (Bryman, 2012; Korf, 2015).
The aim of this article is to demonstrate some of the ways in which offline and online ethnographic methods can – and should – complement each other, particularly when researching deviant activities or hidden populations. I do this by first discussing the history and role of ethnography in criminology and the challenges inherent therein. I then explore how Internet methods can be – and have been – utilised in response to some of these challenges. Next, I outline my own experiences of using face-to-face and online methods in studying cannabis growing and cannabis growers. Examples from my research demonstrate various ways in which online and offline methods enhance one another. Online methods can be useful in expanding a study beyond the normal constraints of (criminological) ethnography by increasing sample size and heterogeneity and by providing access to considerably more data. Offline methods, however, are an important precursor to online research success by enabling identification of and access to relevant online communities and information sources. Offline experience as an accompaniment to online research also enhances the ability to participate in – and understand – online communication about criminal activities, especially where familiarity with specific terminology, technology or cultural reference points is required. Both online and offline approaches can reveal findings that may have remained hidden to the other, and both can enhance understandings and interpretations of behaviours, ideas and interactions encountered in the other context.
In short, triangulating online and offline methods enhances both, with the knowledge to be gained from a combined study greater than the sum of its component parts. The examples from my own research support the conclusion that online approaches to criminological ethnography have the potential to greatly enhance our understanding of contemporary crime and deviance, but need to be treated with care as they can be subject to various pitfalls, particularly if seen as an alternative to traditional methods. However, triangulation of online and offline methods can clearly enhance the understanding of many human behaviours and may be particularly useful in overcoming the difficulties inherent in criminological ethnography.
The ethnographic tradition in criminology: from street to screen
Ethnographic approaches to the study of crime can be traced back at least as far as Henry Mayhew’s studies of the 19th-century urban poor in London (Hobbs, 2001). However, it was within the Chicago School that such methods were placed firmly at the centre of an empirical sociology of deviance, with Robert Park’s famous instruction to his students to ‘go get the seat of your pants dirty in real research’ (Becker, cited in McKinney, 1966: 71). Especially suited to interpretivist epistemologies, ethnographic methods have remained a favourite of critical criminologists and have seen a resurgence since the 1990s particularly (but by no means exclusively) under the banner of cultural criminology (e.g. Ferrell, 2009; Ferrell et al., 2015; Kane, 2004).
The argument for ethnographic research has centred on providing an important counterpoint to the official picture of crime as painted by criminal justice statistics, media portrayals and popular and political discourses (Bryman, 2012; Ferrell, 2009; Hobbs, 2001). Proponents of ethnography seek balance to the positivist-reductionist explanations for crime generated by quantitative methods: ethnography gives voice to those normally portrayed as offenders in mainstream criminology and enables a deeper understanding of the realities and meanings of those groups and behaviours that come to be labelled as ‘criminal’ (see especially Ferrell, 2009; Young, 2011). Realist commentators urge caution around ethnographies’ tendencies to over-sympathise with criminals, over-romanticise crime and reverse, rather than neutralise, the very biases they seek to counter, but still recognise the importance of ethnography in enabling us to uncover ‘truths’ about crime that are hidden from other methodological approaches (Matthews, 2013). Regardless of ideological positions and ‘taking sides’ (Becker, 1967; Cowburn et al., 2013), it is widely recognised that ‘investigative field research (Douglas, 1976), with emphasis on direct personal observation, interaction, and experience, is the only way to acquire accurate knowledge about deviant behaviour’ (Adler, 1985: 11). Relevant to my own area of research, it has been argued that ethnographic methods ‘are especially necessary for studying groups such as drug dealers and smugglers because the highly illegal nature of their occupation makes them secretive, deceitful, mistrustful, and paranoid’ (Adler, 1985: 11).
Ethnography may be a particularly good way to research transgression, but it still faces the core challenge that ‘most deviants would not choose to advertise themselves’ (Downes and Rock, 1995: 28). Unless known to criminal justice (or other) agencies, criminals are not readily identifiable – and even if we can find them, they may not be willing to openly discuss the aspects of their lives that we are interested in. Faced with this problem of access – not just of identifying suitable respondents but also of ‘getting in and getting on’ (Monaghan, 2002: 409) when we do – ethnographers of deviance have had to rely on a combination of serendipity and hard work to get into the position to gather data. Serendipity occurs when researchers find they already have access to a deviant population, such as through existing social or professional networks (e.g. Howard Becker’s (1963) role as a jazz musician gave him access to a population of cannabis smokers; Jason Ditton’s (1976) employment in a bakery allowed him to study workplace pilfering) or socio-geographic proximity (e.g. Patricia Adler (1985) and her husband discovering they were neighbours to an upper-level drug smuggler; Dick Hobbs (1986) conducting research in and among the community he grew up in). Such ‘foot-in-the-door’ or ‘on-the-doorstep’ forms of access (Potter, 2010a) – different variations of what is often referred to as ‘privileged access’ (e.g. Pearson, 1993) – may make it easier for the researcher to develop rapport, but even with (and especially without) this pre-existing relationship, encouraging respondents to trust researchers – to relax in participant-observation situations and to open up in interviews – requires time and patience. This is true, to an extent, of all ethnography, but particularly true when the population being studied may wish to keep themselves and their activities hidden to avoid the criminal justice repercussions or social stigma that might otherwise go with being identified as a law-breaker. Of course, covert research can avoid the need to overcome suspicions and fears of being researched, but this carries its own challenges (Bryman, 2012; see Calvey, 2013, for a defence of covert ethnography in criminology). For the insider employing covert approaches, access may be relatively straight forward, but there are ethical issues – including personal safety concerns if ‘outed’. For the covert outsider, these same issues remain, and the researcher still needs to work at being accepted by the group under study in the first place.
Access is only part of the challenge facing the criminological ethnographer. Even if a target population is identified and opens up to the researcher, there are still questions of validity and reliability – particularly the real possibility that deviant populations may tend towards dishonesty if they do feel the need to keep their activities hidden – and of generalisability of findings. While ongoing participant observation and interviews with multiple participants allow for certain cross-checking of facts within the group being studied, there remains some risk that the group as a whole keeps some things hidden from the researcher and no way to be sure how representative that group is of any wider population. Standard scientific approaches to addressing these problems include increasing the size and representativeness of samples and employing multi-level triangulation (Denzin, 1978). But ethnographies of deviance do not lend themselves to these approaches – with both the access problem and the time-intensive nature of ethnography, ethnographic studies focusing on multiple populations for triangulation purposes are rare indeed (although see Dreher, 1982, for an example pertinent to my own research topic). In any case, sample sizes within ethnography – as with qualitative methods more generally – are usually small compared to the standards of quantitative approaches, and there is the well-established observation that, by definition, nobody knows what a representative sample of a hidden population would even look like.
Online methods as a solution
Online methods offer potential (at least partial) solutions to many of the problems inherent in researching active offenders, and criminology (along with other social sciences) has firmly embraced online research – recognising cyberspace as both a valid research site and a convenient research medium. Arguments in favour of online research in criminology include those common to all social sciences – from the mundane arguments of practicality (e.g. ease and cost of data collection) to the theoretical and empirical arguments of the significance of online space – as both information source and communication medium – in the late-modern world. But there are features of cyberspace that are particularly relevant to researching crime, criminality and deviance. Online communities exist for almost any activity or specialist interest one can think of (and many more that one can’t), no matter how unusual, immoral or illegal. These can be found, contacted, monitored and even joined by the researcher – providing information about these topics, (potential) access to online community members and vast amounts of existing data in the form of message boards, discussion forums, Facebook group pages and so on. Online communication can offer (perceived) anonymity to research subjects, 3 and the lack of physical proximity – and even identity – between researcher and respondent minimises risks of physical harm or of directly witnessing illegal acts. Hence, trust, in both directions, may be more easily established. The counter-argument is that trust is more easily built through face-to-face interactions, and while personal experience would suggest this is often true, a researcher must still convince a respondent to meet personally in the first instance. Providing the option for either online or offline interactions allows the researcher to be flexible and to accommodate the preferences of the respondent, a point which further supports the overall argument that utilising both online and offline approaches is better than just using either alone. Online forums also allow for the cross-referencing of contributions by any given individual so that while we may not be sure that they are honest or reliable, we can at least check for consistency in both style and content of what they say.
Overall, it is important to recognise that questions over validity and reliability of data remain with online methods. It is impossible to be sure of the representativeness of a sample recruited online, just as it is with traditional face-to-face methods of researching hidden or deviant populations: not all members of a target population will be present and active in online forums, and the sub-set that are willing to participate in research may be further unrepresentative in some way. But combining online and offline recruitment strategies can increase both the number and heterogeneity of respondents recruited, although the inherent challenge of overall representativeness remains.
Furthermore, if an increasing amount of crime-related communication, and an increasing proportion of a criminal’s social life, occurs online, then engagement with information sources and discussion forums related to criminal activity becomes an essential component of any research attempting to elicit a holistic understanding of the offender and their behaviour. Elsewhere in the social sciences, online ethnography (virtual ethnography, cyber ethnography, netnography, etc. – see Bryman, 2012: 451, for a discussion on nomenclature) has become a valid method reflecting the increased role that online interaction has in everyday life (Bryman, 2012; Markham, 2007, 2011). The role of the Internet in planning and discussing crime – predominantly through message boards, discussion forums and newsgroups (in the earlier days of the Internet), and Facebook, Twitter and an increasing array of other social media platforms – has been recognised by criminologists since the late 1990s (Mann and Sutton, 1998) and reflected in the proliferation of studies utilising such online forums as data sources and research sites (see Holt, 2015, for an overview of ‘Qualitative criminology in online spaces’).
Online drug research, in particular, has grown rapidly in recent years (Boothroyd and Lewis, 2016; Potter and Chatwin, 2011) and epitomises both the strengths and weaknesses of online methods in criminology. On one hand, it helps overcome the practical difficulties inherent in researching hidden populations (Potter and Chatwin, 2011) and reflects the multiple roles the Internet has in supporting and enabling this area of deviant activity (e.g. in providing information about drugs, providing forums for users to share experiences, allowing both professional and peer-to-peer advice around dosage and methods of use, dealing with bad (and good) experiences, and coping with overdose, addiction or other problematic drug-related behaviour, allowing for the organisation and discussion of events where drug taking occurs and even providing a platform for drug supply) (see, for example, Barratt, 2011; Boothroyd and Lewis, 2016; Wouters and Fountain, 2015). On the other, regardless of the role of the Internet, drug production, drug supply and drug taking are ultimately things that occur offline – and as such, any online-only method will only partially reveal the realities of contemporary drug use and supply (Korf, 2015). In drug research, as elsewhere, there is concern that the desk-based research options offered by cyber research ultimately fall short of the aim of understanding the non-cyber realities of drug use, drug markets and the actors that participate in them, and that in embracing the benefits of online methods, there is a tendency to gloss over the weaknesses.
Researching cannabis growers
The core of this article draws on over 15 years of experience utilising both online and offline methods to study cannabis cultivation in the United Kingdom and elsewhere. An interest in the role of domestic cultivation in the UK cannabis market grew out of my MA dissertation, where small-scale growers selling home-grown cannabis to their friends cropped up as a finding in a broader study of retail-level drug dealing among and around a student population (Potter, 2000). Cannabis cultivation – this previously unstudied part of the UK drug market – seemed fertile ground for a PhD project, and I spent the next 6 years conducting ethnographic research that included a strong online component (Potter, 2008, 2010a, 2010b; Potter and Dann, 2005). Since completing the doctorate, I have been involved in three distinct projects looking further at cannabis cultivation and distribution: an online survey examining regular cannabis users’ experiences obtaining and supplying cannabis (Chatwin and Potter, 2015; Potter and Chatwin, 2011, 2012), an international online survey of small-scale cannabis growers (Barratt et al., 2012, 2015; Potter et al., 2015; see also http://worldwideweed.nl) and an ongoing project 4 looking at the interplay between ‘criminal’ and ‘medical’ identities and motivations of cannabis growers in the United Kingdom. These projects have all utilised online approaches to recruitment, including participation in cannabis-themed discussion forums, Facebook groups and other social media, which have generated qualitative data alongside survey and interview responses. Although only the PhD was formally construed as an ethnography, interviews and participant observation (both online and offline) have been carried on alongside these other projects as and when opportunities have arisen (e.g. Potter, 2011). Experiences from the formal and informal research across the entire 15+ years have informed the current article, with the case studies and examples used in the following sections drawn from fieldnotes, interviews and various online forums recorded, conducted and scrutinised across that period.
The interplay between online and offline ethnography
Many of the strengths of online research, particularly as related to researching hidden populations such as drug users and dealers, have already been established and discussed previously (see above and, for example, Potter and Chatwin, 2011): my own experiences generally support these observations, and the general benefits of online research are not dwelt upon here. Instead, I wish to discuss three broad and overlapping areas where online and offline research enhance each other, but particularly where the latter enables and improves the former. These areas relate to access and recruitment and the role of offline networking in opening online gates, holistic understanding and the importance of real-world experiences in understanding data collected online (and vice versa) and triangulation between online and offline approaches and respective sample populations.
Access and recruitment: offline networking and online gatekeeping
In my MA dissertation (Potter, 2000), I described six of my respondents who grew cannabis for personal use and to share with or sell to friends. Having long known people who grew cannabis and were involved in social or social-commercial supply (Hough et al., 2003) of their own produce, it did not seem to me a particularly novel finding, but my dissertation supervisor suggested I should explore this further for my PhD as this aspect of the UK cannabis market had not been previously researched. Far be it from me to ignore such sage advice, nor the opportunity to take advantage of my friends (for many of my respondents for the MA, including all those who reported growing, were my friends); I conceived a project that aimed to understand the who, how and why of cannabis cultivation. Ethnography seemed the obvious way forward, and I engaged my existing contacts (cannabis dealers who bought from growers as well as those who grew themselves) – my foot-in-the-door access – as informants and as gatekeepers who could help me gain access to other growers. I also contacted owners and staff of local grow-shops and the local police to gain other perspectives on cannabis cultivation in the city where I lived and made direct contact with other growers myself through word-of-mouth. But I quickly hit the sampling problem discussed above.
I wished to be able to say something about cannabis cultivation in the United Kingdom in general, but my initial access to a group of growers was very much dependent on existing social contacts, and respondents recruited through snowballing tended to reside in the same geographical area and share similar socio-demographic characteristics to those contacts I already had. And although I managed to recruit small numbers of respondents in other parts of the United Kingdom, the opportunity to do proper ethnographic research outside the city I was living in, or even beyond the groups I had gained access to through snowballing, was not available. Trying to conduct a full-scale ethnographic study elsewhere – for triangulation purposes – was beyond the time, financial and social capital resources were available to me.
Online research had occurred to me, and I had identified two UK-based online forums 5 dedicated specifically to cannabis cultivation that I hoped to use not just as information sources but as means to recruit respondents for interviews. I was still working out a strategy for introducing myself and my research when an interview with a cannabis grower I met at a festival (where I was helping on some other drug research by handing out questionnaires) provided what turned out to be an ideal opportunity for experimentation with online methods: my new contact was a senior moderator on one of these sites.
The two online forums identified seemed very similar in many respects – both claimed 10,000+ members, of which maybe 10% actively contributed to online discussions and a few hundred seemed to account for the bulk of posts. Both had discussion boards devoted to a wide range of topics and sub-topics of interest to growers – the mechanics of indoor cultivation, the advantages and disadvantages of growing different strains, methods of preparing the end product, tips for avoiding detection, scientific research and political activism related to cannabis and so on. Both were open to anyone to browse (and hence conduct passive, covert research) and were free and easy to join, with membership allowing one to contribute to discussions and to send personal messages to other members. The difference, for me, was that on one site I had an inside contact – a gatekeeper – to help me with my recruitment.
I approached both forums in broadly the same way. In both cases, there was a dedicated sub-forum for new members to introduce themselves. I posted here explaining who I was and what my research project was about and asking for people willing to answer a series of questions via email or other online methods. I included a link to my online profile at my university, which included a photo of me (complete with the waste-length dreadlocks I had at the time) and mentioned my previous involvement with some UK-based cannabis campaign and drug research groups. Initial responses on both sites were very similar. One or two people expressed interest in participating, but wanted assurances. Others were convinced I was a cop and that any response to my questions would put not just the respondent but the entire ‘community’ at risk. The overall tone was one of urging people to exercise extreme caution before volunteering to participate – or advising outright to not participate at all. Cannabis growers can be a suspicious bunch!
It was at this point that my experiences of the two sites diverged. Very quickly after my post, and the initial responses to it, my moderator friend added his own response to say he had met me in person, that I was what I claimed to be and that I was trustworthy and on their side. He even vouched for me as being a smoker (we had shared a spliff at the festival where we met) and a grower (we had discussed my own limited experiences here – a point I will return to later). Instantly, the tone of conversation changed: if he said I was alright, then that changed everything. Some forum members volunteered to take part in the research after this intervention. Others still said no, but were polite and supportive and cited concerns over the security of online communication in general rather than doubts about my own trustworthiness (somewhat ironic, given their willingness otherwise to actively participate in the online community). A few who had been vociferous in their doubts and suspicions offered apologies, but hoped I understood their initial concerns.
Meanwhile, on the other site, suspicion remained strong. Posters suggested I was an undercover cop – one even commented that my own online profile was written in the first person (most of my colleagues had third-person entries), which must reflect the fact that nobody else at the university had been willing to write a profile for me (because I wasn’t really an academic …), and suggested that my dreadlocks were an obvious and over-zealous attempt to appear legit (which both amused and bemused me: they’d taken a number of years to grow). Another suggested that if people did wish to participate, they should not do so online (lest email or IP addresses be traced), but should ask for the questions to be made available, print them off and return the completed interview by post (although not from their local post box, of course). They even suggested that gloves should be worn when handling the paper, and that neither the stamp nor the envelop should be licked, to avoid passing on either fingerprints or DNA! The difference was well illustrated by results within the first week: on one site, I had 20+ volunteers to take part in detailed interviews and dozens more willing to take part in online surveys and polls; on the other, I had only two volunteers for interview (one of whom was a member of both sites).
The main lesson here was obvious: online communities of offenders are as suspicious, if not more so, as their offline counterparts, and as with traditional access to hidden populations, an introduction by a trusted gatekeeper helps. However, what was interesting was that methods usually considered as helpful in gaining access – such as being identifiable as a (partial) insider (e.g. through my hair style and discussing my own experiences with cannabis) or through demonstrating academic credentials (e.g. through my online profile at my institution) – were taken as points of further suspicion in the online environment. Further ventures into online research – when trying to recruit survey and interview respondents for later projects – reinforced this lesson: approaching online communities and forums without introductions always involved dealing (with greater or lesser amounts of success) with suspicion, distrust, paranoia and (on occasion) outright hostility; being able to make reference to real-world events (such as cannabis-related social or political meetings I had attended) or mentioning high-profile campaigners I had met made things easier (having met Howard Marks, the notorious British cannabis smuggler, on multiple occasions was particularly helpful here); and being introduced or vouched for online by people I had met offline notably increased acceptance and, hence, recruitment.
Holistic understanding
Having offline connections to enable online recruitment was only one way in which pre-occurring offline research enhanced my ability to research online. In addition to recruiting people for interview outside of the discussion forum environment, I also strove to participate in forums to further inform my research. Hine (2000) argued in one of the first examples of online ethnography that lurking, rather than fully participating, leaves the researcher short of any full understanding of online communities. Participating in discussion forums meant I could get data from people who had not volunteered for interview and that the information and social interactions within the discussion forums could also be more effectively used as data in its own right. Here, again, offline research proved invaluable in maximising the utility of these data. In short, it meant that I knew what people were talking about in their online discussions. Cannabis cultivation is a real-world activity – and while it can be very simple (put seed in soil, add water and sunlight, wait), growing cannabis well can be a technically complicated process that requires sophisticated horticultural knowledge (Potter, 2010a). Discussions drew heavily on both scientific and slang terminology and referred to many physical aspects of cultivation that were not always reproducible in the text-only format of message boards. Even pictures – widely used in some sub-forums, Facebook groups and photo-based social media like Instagram, particularly when growers were showing off their plants and produce – only go so far in reproducing the complex reality of the growing experience. A few examples of cultivation-related issues and experiences that cannot be fully explained or appreciated in discussion forums illustrate the point:
Many cannabis growers take pride in their knowledge of different strains of the plant and skills needed to be a successful grower – from the basic division of Indica and Sativa (and the less-well-known Americana and Ruderalis) varieties to more nuanced details of cross-breeds, feminised seeds, sexing young plants (males produce less tetrahydrocannabinol (THC) and pollination of females inhibits their productivity, hence males are removed) and so on. Seeing photos provides some knowledge of the differences; watching different plants grow gave me a level of understanding that enabled me to participate knowledgeably in both online and offline discussions – but depended on my physical participation with growers and their growing operations (grow-ops) over an extended period.
The complexities of high-tech cultivation techniques (lights, hydroponic equipment, preparing and administering nutrients, etc.), especially in (often) small, cramped growing spaces hidden within private houses, must be seen to be appreciated: I have seen dozens of grow-ops, but am still regularly impressed at the complexity and ingenuity of individual set-ups – and at the elegant simplicity of others.
Cannabis is extremely pungent, and growers spend a lot of time discussing techniques to minimise the smell (and hence the risk of detection). Visiting grow rooms (and, for that matter, simply walking the streets of some UK cities where cultivation is common) demonstrates this in a way that online participation can never replicate. Having a grow-op opened and experiencing the smell emerging where previously it was undetected shows both the effectiveness and the necessity of the use of carbon-filters, fans and other smell-mitigation techniques. Field notes from early in my PhD record my surprise at having an attic hatch opened to reveal a 15-plant grow-op (which produced about 100 ounces (2.8 kg) of cannabis netting the growers around £10,000): although the house itself smelt faintly of cannabis smoke, the overwhelming smell of 15 mature, high-potency plants had not been detectable even from directly below the attic room. As recently as February 2017, and despite having developed a particularly keen sense of smell with regards to cannabis in the intervening years, I noted again how a pungent six-plant grow-op went unnoticed until revealed even though I had been conducting an interview with the grower for an hour in the room immediately below.
Growers often have a bond with their plants that is hard to explain in words – they may express joy, pride, even love, in online forums, but facial expressions and physical gestures say much more. One female grower told me how she would dance, naked, around her plants every morning. A partial re-enactment (clothed) was something that could never have been satisfactorily conveyed online (even a webcam would not have allowed for me to be the centrepiece of the dance – occupying the space where the plants would be), but reading growers’ online descriptions of their emotional involvement with their plants took on a new meaning for me after face-to-face encounters like this.
Fungus, pests, light pollution, scorching and poor use of nutrients can all seriously damage a plant (and hence its drug-productivity). Seeing these things, the methods used and time and effort invested to try to reduce them and the reactions (initial frustration and long-term despondency) to them were far more evocative than text-based reports. I even had two plants growing hydroponically in my own room at one point (set up by one of my initial friend-contacts featured in the original MA thesis) and witnessed first-hand the effects of how playing computer games late at night led to light pollution notably impacting the development of one plant (the one closer to the open wardrobe door). This experiment also provided first-hand experience of the anxieties of being caught: I will never forget the day I had to hide quietly in my room, with the door locked, while my house-mates dealt with an unannounced visit from the landlord two floors below.
Harvest – chopping down the plants, removing the buds from the less-desirable vegetative matter, and drying and curing these – and the social aspects of this (listening to music, drinking beer, showing off particularly good results to friends, sharing the end product, rolling and smoking the first spliff from the latest crop) are a four-dimensional, multi-sensory and emotional event. The reaction to an unexpected knock at the door, or passing police sirens, during this process is also something that cannot be replicated online!
All these examples reflect common experiences for cannabis growers – and, hence, common topics of online discussion. These real-world ethnographic experiences did not just provide insights into aspects of cannabis cultivation that were only partially available in the online environment; they enabled me to participate in – and understand – online discussions that alluded to such issues without me having to constantly ask for explanations (which may not have been possible in text format anyway). Being able to talk about such issues from an informed perspective also aided recruitment in later research – for example, initial wariness over our use of the word ‘skunk’ (often used as a short-hand for all strong varieties of home-grown cannabis in the UK media, but used only to refer to a specific strain by connoisseurs and experienced growers) in a survey question was overcome when I was able to demonstrate a full understanding of these nuances in an online discussion (see Potter and Chatwin, 2012).
Although offline experiences improving my ability to engage with – and understand – online discussions were more common, the reverse was also true: when growers discussed how they learnt to grow, or found out about different varieties of cannabis, or methods of growing, or how to deal with problems (like parasites, smell or light pollution), they often referred to the online community. Participation in these online forums meant that I understood this better – how numerous growers share their time and experience to talk through problems and solutions, willing to help less-experienced growers with patience, humour and good grace. Furthermore, online cannabis communities often led to offline relationships – growers would communicate online, but often meet up offline for social events, political campaigning or to exchange tips and knowledge. Numerous respondents reported real-world friendships that were initiated online or of developing online contacts with people they would meet to buy, sell or otherwise exchange cannabis products, growing equipment or cuttings with. At the same time, disagreements, arguments and ‘trolling’ also occurred online and were sometimes discussed later in the offline environment. Engaging solely in either online or offline research would have only given a partial picture and, hence, a partial understanding of both the practical and social aspects of cannabis cultivation.
Triangulation
It should already be apparent that utilising online and offline approaches improved my research both by increasing the available sample size and by giving insights into the overall cannabis growing experience that would not have become obvious with just one or other strand of research. However, there is an extra dimension to report, which harks back to the original aim of engaging with online research. I ended up with, effectively, two broad sample populations: those who were (primarily) part of my traditional real-world ethnography, and those who were (primarily) part of the online ethnography. The latter were drawn from across the United Kingdom; the former were largely concentrated in the conurbation in which I resided during my PhD fieldwork. This meant it was possible to explore similarities and differences between the two.
Most noticeable were the similarities – it was possible to assert with some certainty that the types of people involved in cannabis cultivation, their motives for and attitudes to growing, and the methods they employed in cultivation, harvest and distribution were largely the same across the country, although important differences could also be seen. Geographically, it seemed that there was some variation, for example. Most obviously, growers living in rural areas were more likely to grow outside, while those in urban areas predominantly utilised indoor techniques. People in different parts of the country perceived the risk of detection – or of severity of punishment should they be detected – differently, reflecting (perceived) differences in the priorities of local police forces and courts, or public attitudes. Perhaps more interestingly, people in different parts of the country have also reported different levels of concern over being threatened, intimidated or otherwise victimised by (other) criminals. Related, my online research did not produce many (but did produce some) respondents who were involved in – or willing to admit being involved in – large-scale commercial cultivation, behaviour that would likely be treated more harshly by the courts as profit-oriented drug dealing. I struggled to recruit growers of this type in my face-to-face ethnography as well, but did have more success with offline methods than online here.
There were also notable differences in how people learnt about growing – and found information and advice to help them deal with challenges such as infestations or disease. Although more growers seem to engage with online information sources in recent years than when I started my research – a reflection of the development and expansion of online technologies and social media in general as well as the establishment of online cannabis communities – there remain a number of growers who do not engage with online information sources or social media at all (or, at least, not in relation to their cultivation activities): some cite the perceived risk of detection related to official agencies monitoring online communication, others (particularly, but not exclusively, older growers) simply do not engage with the Internet all that much, if at all. Some of these Internet-averse growers engage in real-world networks of growers (such as the cooperative growers discussed in Potter, 2010a) who offer advice and support and who club together to harvest – and sell – their crops, whereas many of those engaged in online communities reported these as opposed to real-world contacts as their primary sources of advice and support. Just as some growers reported online engagement as potentially risky, so others felt that having offline contacts know about their activities was too dangerous and so only resorted to online contacts and forums for advice and support.
Discussion
Having outlined the utility of combining on- and offline methods in my own research, let me now turn to the question as to how useful these methodological reflections might be to criminological ethnography more broadly.
Clearly ‘online’ elements play an increasingly important part in most people’s everyday social lives: this is one of the defining characteristics of the contemporary social world. This is as true for those people who may be labelled as ‘offenders’ or ‘deviants’ as it is for everybody else (the idea that such groups are fundamentally different from ‘normal’ people has long been debunked in criminology). As such, ethnographic methods – which seek to understand people and behaviours holistically – need to not just acknowledge but embrace this aspect of people’s lives. And while it is worth noting that the distinction between online and offline, or ‘virtual’ and ‘real’, domains is increasingly blurred in the hyper-connected world of the 21st century to the point where seeking to distinguish between the two may seem artificial (most of us carry smart phones around with us, and we are only just beginning to realise the full implications of Web 3.0 and the Internet of things – see, for example, Crang and Graham, 2007; Dodge and Kitchin, 2005; Kluitenberg, 2006), this only strengthens the argument that ethnography needs to integrate online and offline elements to remain effective, or even relevant. However, it is also clear that while populations we are interested in studying will (usually) occupy both online and offline space, it is also clear that the aspects of their lives we are interested in as criminologists (i.e. their crimes and deviant behaviours) may not always be the subject of online communities – or discussed openly online – in the same way that cannabis cultivation is.
Online methods are only suitable when the research topic or population has an online presence (although most things are at least mentioned online somewhere). Online ethnographic methods are only suitable where a relevant online community (broadly construed) exists – and is accessible to the researcher (whether covertly or overtly). This is likely to be the case for any deviant behaviour facilitated by online communication, but particularly so for those types of offending behaviour that are part of a group identity – what we might think of as ‘lifestyle’ crimes or the sorts of people and behaviours normally targeted by cultural criminology (use of the Internet being, essentially, a cultural phenomenon). In addition to cannabis cultivation, drug use clearly fits into this category (hence the proliferation of online drugs research discussed earlier). With the emergence of cryptomarkets, drug dealing (and other illegal markets) could also be targeted for online ethnographic research, but developing parallel offline research may be difficult or inappropriate in these cases: integrating online and offline approaches can only apply where both online and offline elements are possible and meaningful. Hate crime has been researched using both offline (e.g. Fielding, 1981) and online ethnographic techniques (e.g. Pollock, 2010), albeit separately, and Pearson (2012) has integrated both approaches in researching football hooliganism – these provide not just examples but potential templates for integrating online and offline ethnography. At the same time, there are activities and groups less amenable to this integrated approach. Crimes like theft may be less likely to be subject to online communities, although an early example of online criminological research focused precisely on forms of thieving (Mann and Sutton, 1998). News stories regularly report on online networks of paedophiles, terrorists and computer hackers; in these examples, the offline ethnographic element may not be possible.
There are general lessons to apply about how to adopt this approach when there are both on- and offline communities that are both identifiable and accessible for ethnographic research. Pearson (2012), like myself, spent time initially doing traditional ethnography – and it was this that both demonstrated the role of online interactions in the population/offence being studied and facilitated both access to and understanding of the online aspects of the phenomenon. Although initial online research may increasingly be the way to identify groups and possible respondents, offline work with members of these groups opens up the possibility of using the online as a research site in its own right. Then, insights gleaned from online research feed back, in turn, to a greater understanding of the offline environment, and so on. Thus, continuously moving between online and offline improves access to and understanding of both – and a greater understanding of the population or behaviour as a whole.
Conclusion
Using Internet approaches to overcome the access difficulties associated with active offender research is now well established within criminology: the aim here was not to champion online methods per se, but to warn against either ignoring online methods completely or, more pertinently, engaging in online methods without suitable accompanying offline research. As both deviant behaviour itself and the social lives of offenders (along with almost other aspects of the late-modern social world) have an increasing online presence, online ethnography seems more justifiable. Add to this the generic advantages of online research as being less resource intensive and more flexible than traditional methods, so it makes sense that online ethnography is increasingly attractive within criminology. However, most deviant activities themselves (cybercrimes excepted) take place in the ‘real’ world, and most criminals – like most other people – still spend large chunks of their lives offline. As such, and as I hope I have demonstrated, much that is of interest to the criminological ethnographer still depends on traditional, offline research – and even when the advantages of online methods, particularly around increasing sample sizes and available data, do seem too good to resist, offline research can be a vital component to maximising online research possibilities. As Korf (2015) has said, ‘… it is not always a matter of choice between them, but rather a combination of online and offline methods that offers the best chances for social drug research’ (p. 12) – and this is true of research into other types of crime and deviance as well. The correct balance of online and offline methods will, of course, vary – for some types of research (such as cybercrimes themselves, or the role of discussion forums in organising terrorism or hate crimes), online methods may take priority (although the offline lives of those involved should not be forgotten), and for others (e.g. the behaviours of groups with limited Internet access, such as the homeless), offline methods may be a priority. Furthermore, economic, practical and safety concerns may also influence decisions over which methods to use, and some crimes (like murder) may not be subject to group identity or any but the most private discussion either online or offline, and therefore are not suitable for ethnographic research in either environment. But in the 21st-century, many criminals and deviants, like most everyone else, have both online and offline presence. Ethnography, as a method that seeks a holistic understanding of the lived experience, needs to reflect this.
Footnotes
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
One of the research projects informing this article was supported by British Academy/Leverhulme trust small grant small grant ref: SG132364.