Qualitative research methods within the addictions

Authors


Joanne Neale
School of Health and Social Care
Oxford Brookes University
Jack Straw's Lane
Marston
Oxford OX3 0FL
UK
Tel: 01865 482600
Fax: 01865 485297
E-mail: jneale@brookes.ac.uk

BACKGROUND

In 1998, Fountain & Griffiths conducted a content analysis on papers published during 1995–96 in three leading international drug publications: Addiction, Drug and Alcohol Review and Addiction Research (Fountain & Griffiths 1998). Of the 291 papers scrutinized, only 17 (6%) reported on studies that had wholly or partially used qualitative methods. In Addiction, the proportion (at 2%) was even lower. Addiction published just three qualitative papers in 2004: Amos et al. (2004); Furst et al. (2004); and Rhodes, Davis & Judd (2004). Such marginality prompts uncomfortable questions. Is qualitative addiction research only a minority interest? Are qualitative analyses somehow less informative or less valuable than quantitative investigations? Or do some addiction journals directly or indirectly militate against the publication of qualitative research?

It was with interest, then, that we received an invitation to write a review of qualitative methods for Addiction. Our brief was to produce a statement that would explain what qualitative research is, highlight its strengths and weaknesses, and outline some of the important principles and processes involved in its undertaking. The offer provided reassuring evidence that Addiction does welcome qualitative work and recognizes its importance. In responding to our task, we hope that we have written a paper that will especially interest and assist newcomers to the qualitative field. For others, our intention is not to offend by reporting what may sometimes seem obvious. The aim is rather to offer a reminder that we need to see a high standard of qualitative work appearing more often in journals such as Addiction.

WHAT IS QUALITATIVE RESEARCH?

Qualitative research has been the historical bedrock of a number of social science disciplines, such as anthropology and history. However, interest in the approach has grown in recent decades as more researchers in basic disciplines and applied fields have adopted a qualitative paradigm (Miles & Huberman 1994, p. 1). Within the addictions, qualitative studies have been occurring regularly since the 1920s. In fact, Feldman & Aldrich (1990) suggest that Herodotus’ fifth-century account of marijuana use in Scythian purification rites is a very early example of a qualitative ethnography. Over the years, qualitative methods have provided drug researchers with valuable techniques of scientific inquiry. In return, drug taking has provided qualitative investigators with an ideal arena in which to apply and develop their methods (Fountain & Griffiths 1998).

Although defining qualitative research is not straightforward (see, for example: Denzin & Lincoln 1994; Lambert, Ashery & Needle 1995), it is generally accepted that qualitative researchers study phenomena in their natural settings and their aim is to understand how people account for, take action and otherwise manage their day-to-day situations (Miles & Huberman 1994). As Rhodes (2000) has argued, qualitative addiction research seeks both to describe the social meanings that participants attach to drug use and the social processes by which such meanings are created, reinforced and reproduced. In particular, qualitative research focuses on exploring the details of people's lived experiences and on appreciating why drug-using behaviours occur and how they are understood in different contexts among different social groups (Rhodes 1995).

Frequently, qualitative research and quantitative research are described in oppositional terms (Martin & Stenner 2004). Qualitative work is perceived as falling within a ‘realist’ tradition, utilizing data that are ‘deep’ and ‘rich’ to explore ‘subjective understandings’ and ‘values’. In contrast, quantitative work is associated with a more ‘positivist’ epistemology, employing ‘hard data’ to focus on ‘objectivity’ and ‘fact’ (ibid.). Important differences between the two methods are evident. For example, qualitative studies are typically less structured than quantitative ones and tend to take more time to complete. They generally yield much larger volumes of data from a smaller number of individuals. Additionally, analysis is more explicitly interpretive, creative and personal (Walker 1985).

Despite this, qualitative and quantitative studies can usefully be undertaken together and it is therefore better to see them as complementary rather than adversarial approaches (McKeganey 1995). For example, qualitative work can be performed as a preliminary to quantitative research. This might involve using qualitative techniques to provide background information on issues and concepts to be investigated in a subsequent survey. Qualitative methods can also supplement quantitative work. Here, qualitative data might be used either to validate findings from a quantitative study (‘triangulation’) or to help build a wider picture of a particular phenomenon under investigation. Thirdly, qualitative research can complement quantitative work, for example by exploring complex phenomena or areas not amenable to quantitative inquiry (Pope & Mays 1995).

DIVERSE TECHNIQUES AND DATA SOURCES

In practice, qualitative research is an umbrella term that is used to describe a variety of methodological approaches. Space does not permit an exhaustive account of what might be involved, but a brief review of some of the main qualitative techniques and sources of data that have been employed by drug and alcohol researchers in recent years will illustrate some of the diversity of the qualitative paradigm.

Techniques

Ethnography has played a particularly important role in qualitative addiction research. The ethnographer seeks to describe the lives of a group of people in a way that is faithful to how they see it themselves (McNeill 1990, p. 64). The intention is to understand the underlying patterns of behaviour and meanings of a culture—including the tacit rules, symbols and rituals that guide group norms (Geertz 1973; Agar 1986, 1996). Ethnographers depend heavily on observation and, in some cases, complete or partial integration into the society being studied. They also often conduct in-depth interviews with community members and examine materials or artefacts that are relevant to people's everyday lives.

In the field of alcohol research, there is a very long tradition of ethnographic studies that have mostly been conducted by anthropologists working in remote regions of the world (SIRC 1998). For example, Lemert examined the relationship between drinking, alienation and cultural change among North-west coast Native Americans in the 1950s (Lemert 1958). More recently, Eber (2000) explored female identities, roles, relationships and sources of power in the context of women and drinking in a Mexican community. Ethnographies of illicit drug use have predominantly been conducted in the United States and one of the most often-quoted is a study of marijuana users by the sociologist, Howard Becker (1953). Becker showed how social setting or context is a significant determinant in how individuals become marijuana users, as well as why they continue or desist from its use.

Other qualitative techniques (such as themed interviews, focus groups and open-ended questions inserted into quantitative surveys) are conducted during one-off contact with research subjects and have the advantage of being less time-consuming and cheaper to conduct than a full ethnographic study. Equally, their flexibility enables them to generate good-quality data on a wide range of issues from a wide range of individuals. For example, Nygaard et al. (2003) interviewed young people who had been involved in drink driving episodes about their experiences of drink driving and riding with drink drivers. Beich, Gannik & Malterud (2002) interviewed general practitioners to explore the suitability of a screening-based intervention for excessive alcohol use, and Lawrie et al. (2003) interviewed pharmacy customers to ascertain their views and experiences of using pharmacies that provided needle exchange services.

A further qualitative technique that builds upon the interview approach and has been used by addiction researchers is narrative research. Here, those being studied are encouraged to tell their stories and these are then analysed in terms of their form (e.g. structure of the plot) and content (e.g. what happened). For example, Hilte (2000) analysed women's recovery from psychoactive substance abuse by examining the main characters in their accounts and by exploring how characters’ actions impacted on subsequent events. Not dissimilarly, McKeganey et al. (1995) and Hughes (1998) have used personal stories but combined them with projective techniques in the form of vignettes and extended vignettes. These involve presenting study participants with a character and setting and asking them for responses to particular events or moral dilemmas.

Further to the above, qualitative investigators have used both community engagement and action research to give those being studied a participatory role in the research process. In these approaches, local people may help to design the study, assist in conducting the research and contribute to feeding the results back into local action. Both qualitative and quantitative components can be involved. For example, local black and minority ethnic community organizations across England have been involved recently in undertaking their own needs assessment projects relating to drug prevention, education and treatment issues (Winters & Patel 2003). In Scotland, meanwhile, a study of community responses to drug use in rural Lanarkshire resulted in local people initiating a breakfast club and homework club for children living in an area badly affected by illegal drug use and problem drinking (Parkin et al. 2001).

Lastly, rapid assessment methods provide a good example of qualitative and quantitative approaches being used together. Rapid assessment methods have been developed to exploit diverse data sources and produce quick, practically orientatated assessments in areas where epidemiological systems are poor (Vingoe 2000). The first rapid assessments occurred during the United Nations Drug Control Programme missions to Vietnam and Lebanon (Ingold 1994; Power 1996). More recently, Fitch et al. (2003) have conducted a rapid assessment of drug and alcohol use in Jersey. This study involved analysis of existing data sources, key stakeholder interviews/focus groups, a community-recruited survey and prevalence estimation of problem drug users. Its aim was to inform the development of strategy performance indicators and to make recommendations for change in existing prevention, law enforcement and surveillance activities.

Data sources

From the above, it is clear that qualitative techniques involve collecting and analysing one or more types of data. Moreover, just as qualitative and quantitative methods can profitably be used together, so two or more qualitative techniques and data sources can be used within one study to complement, supplement and validate each other. The main types of qualitative data are observations, interviews and documentary materials. A useful summary of some of the respective advantages and disadvantages of each of these is provided by Creswell (2003, pp. 186–187).

Observations can be gathered in various ways. For example, researchers may engage in roles varying from a non-participant to a complete participant. Equally, they may take fieldnotes on the behaviour and activities of individuals in an unstructured or semistructured way. These notes can then be kept in the form of a diary, an unfolding narrative or even personal letters. Bourgois’ direct observations of the lived experiences of homeless heroin users in San Francisco provide an important example of the power and value of this type of material (Bourgois 1998a,b).

Interview data are usually collected in person or sometimes by telephone. The interview format may be structured, semistructured or unstructured and can involve one interviewer and one respondent (one-to-one interviewing) or be conducted in small groups (focus groups). Because the researcher cannot be expected to remember all the responses given to their questions, interviews are generally audio-recorded and transcribed verbatim. It is the interview transcripts that are later analysed, increasingly with the assistance of computer software packages. Neale (2002) provides a detailed account of analysing semistructured interview data from a study of non-fatal drug overdose using the computer program WinmaxPro.

Documentary materials can include newspaper articles, books, reports, casenotes and audio-visual and digital media. Cooper (2004) analysed 297 medical articles on opiate addiction in order to explore the relationship between US health professionals’ theories of opiate addiction's aetiology and contemporaneous perceptions of opiate addicts’ race/ethnicity, social class and gender. Gilman (1998) used Usenet news groups on the internet to gather information about routes into heroin addiction, and Taylor et al. (2004) analysed video data produced by filming drug injectors in their homes and other venues to explore how injecting practices place individuals at risk of HCV infection and the implications of this for harm reduction.

WHY USE QUALITATIVE METHODS?

Having introduced something of the nature and diversity of qualitative research, it seems important to review why it might be used. Fundamentally, it has been argued that quantitative methods are too narrow and too reductionist for studying the social world (Martin & Stenner 2004). Human beings are purposeful, goal-seeking, feeling, meaning-attributing and meaning-responding creatures (Hughes 1976, p. 24). Consequently, it is necessary to understand how they perceive and interpret their environment if their behaviour is ever to be interpreted usefully. Qualitative investigations facilitate this by enabling the researcher to understand drinking and drug use from participants’ perspectives.

Because of their capacity to explore and explain human behaviour, qualitative methods have proved very valuable in demystifying drug and alcohol use and replacing stereotypes and myths about addiction with more accurate information that reflects the daily reality of substance users’ lives. In so doing, ethnographic research has humanized the image of the drug taker and drinker, improved general understanding of the everyday context of drug use and drinking, and facilitated the development of socio-cultural perspectives on substance consumption. This has resulted ultimately in important new theory, policy and practice.

Thus, ethnographic research conducted during the 1950s and 1960s questioned the prevailing medical assumption that drug addiction was a disease and drug users were passive, anxious and inadequate people. Instead, it showed that drug addiction was a social experience that needed to be understood and that drug users were normal self-determining individuals who were participating in a life-style that just happened to be deviant (Faupel 1991; Stephens 1991). Similarly, anthropologists have questioned the high level of alcohol problems associated commonly with tribal and village cultures, arguing that while drunkenness may be common, alcohol addiction is actually rare (Leland 1976; Heath 1982). Furthermore, the preoccupation with problem drinking detracts from its more pleasurable and routinized aspects (Levine 1981).

In terms of formulating new policy and practice, qualitative methods were employed in the 1990s in response to the public health imperative to reduce HIV and other blood-borne infections linked to injecting drug use. This was in recognition of the fact that effective interventions needed to be based on an understanding of the complex risk behaviours and social factors involved in drug taking (Leshner 1998). Accordingly, various qualitative techniques were used to examine injecting practices and especially the social and cultural dimensions of needle sharing (cf. McKeganey & Barnard 1992; Grund 1993). The rich data resulting from this work were then used to inform the expansion of crucial harm reduction services.

Qualitative research has also been making contributions to the expanding literature on drug treatment and its effectiveness by exploring users’ views and experiences of service provision. Here, the use of qualitative methods is underpinned by the belief that treatment is most likely to be effective when individuals feel positive about and satisfied with the support that they receive. For example, Neale (1998) investigated drug users’ opinions of drug workers and other service providers while Copeland (1997) examined the barriers drug users face when attempting to access help. In the alcohol field, Thom et al. (1992) used open questions and free discussion to generate qualitative material on patients’ perceptions and experiences of clinic and general practice-based care for drinking problems. Meanwhile, Lock (2004) conducted focus groups to explore patients’ views and experiences of alcohol and brief intervention in primary health care.

Further to the above, qualitative studies have always been invaluable in accessing ‘hidden’ or ‘hard-to-reach’ populations. The illegality of much drug-related activity, users’ erratic life-styles and/or the social stigma associated with substance misuse mean that addiction research often concerns hidden populations (Power 1995). Qualitative techniques (such as interviewing or observing venues where drugs are sold, bought and consumed) increase the addiction researcher's knowledge of hidden and hard-to-reach groups. Equally, feelings of trust and rapport emerging during fieldwork can enable qualitative researchers to permeate the complex social networks that link substance users and their associates together. In addition, qualitative researchers have developed innovative sampling techniques to minimize the potential for bias in their data. For example, snowball sampling is a carefully staged process that involves the recruitment of new respondents through referrals from earlier respondents (see Kemmesies (2000) for a full description).

As indicated previously, qualitative research is also important in identifying emerging trends in drug taking. According to Clatts et al. (2002), institutionally derived sources of data used to forecast changes in patterns of drug abuse typically omit a number of hidden populations. In contrast, ethnographic research can identify and track new and hidden drug-using groups as well as new and hidden drug-taking practices. One reason for this is that qualitative researchers are encouraged to go where the data lead them (Belgrave, Zablotsky & Guadagno 2002). Furthermore, qualitative data can preserve the chronological flow of social processes, so helping to identify precisely which events lead to which consequences (Miles & Huberman 1994). Addiction researchers have, in recent years, exploited this strength by using qualitative methods to explore both the increase in drug prevalence among young people as well as the reasons underpinning this growth (Fountain & Griffiths 1998).

Finally, qualitative techniques are highly effective in researching sensitive and/or illegal activities—hence substance use and misuse in general, but particularly very sensitive drug-related issues. Providing evidence of this, qualitative researchers have regularly elicited detailed, personal information on drug dealing and other drug-related criminal behaviour, injecting practices, parenting while drink- or drug-dependent and high-risk sexual activity. Again, feelings of trust and rapport arising between the researcher and research participants facilitate discussions about intimate information. Similarly, qualitative researchers’ sensitivity to the social and cultural specificity of their study population can foster an awareness and empathy that encourages those being researched to disclose their vulnerabilities (Allen 2002).

SOME GENERAL LIMITATIONS

Alongside the many strengths of qualitative work, there are inevitable weaknesses. As Martin & Stenner (2004) point out, qualitative techniques promise ecological validity at the cost of the classic scientific requirements of reproducibility and generalizability. In other words, more opportunistic sampling techniques and the small numbers of individuals participating in qualitative work mean that findings can never be generalized nor portrayed as representative of drug users in general. Breadth and scale are sacrificed in order to obtain a deeper and more contextualized understanding of peoples’ lives and experiences.

While this loss of scope is an acceptable compromise for those committed to a qualitative paradigm, other issues can provoke greater discomfort. One recurrent problem is the lack of consensus about what exactly qualitative research is, compounded by the variety of approaches—each with its own rules, aims and logic—that can be adopted (Hammersley 1990). The lack of clearly formulated procedures and conventions—particularly in respect of data analysis—means that qualitative researchers do not only need to explain and justify their techniques to those wedded to quantitative methodologies. They also need to explain and justify themselves to other qualitative researchers.

A further weakness of qualitative research is the potential for bias and misrepresentation. The integrity and honesty of drug users’ responses to interview questions are often doubted, along with their ability to recall events and emotions. Such concerns are, of course, equally relevant to those conducting quantitative surveys. Indeed, the trust and rapport fostered during qualitative research can help respondents to give more accurate accounts. That notwithstanding, it is still the case that interviews and observation intrude as foreign elements into the social setting they would describe. Therefore, the particular dynamics of any given researcher/respondent interaction can create as well as measure attitudes (Webb et al. 1966). For example, the respondent may simply report those beliefs and behaviours that he or she feels the researcher wishes to hear.

Finally, qualitative methods have some very practical limitations. Data collection, processing and coding all tend to be labour-intensive and thus demanding from a resource point of view. Similarly, analysing the vast amounts of material produced requires a high degree of organizational skill. Even 40 in-depth interviews or a few days of observation can produce such a large volume of information that it is easy for the unskilled and/or inexperienced researcher to become dangerously adrift on a seemingly endless sea of unstructured materials.

ADDRESSING THE PRACTICALITIES

Because qualitative data commonly concern aspects of everyday life, there is a dangerous assumption that anyone and everyone can do qualitative research with little or no prior training. In this section, we highlight what we consider to be some important features of undertaking good qualitative work within the addictions. For simplicity we have separated the processes of planning, data-gathering, analysis and dissemination. In reality, we appreciate that these activities tend to occur simultaneously. We also recognize that many of the ingredients of a good qualitative study mirror those we would advocate for quantitative research. We none the less include them here for coherence.

Planning

Before selecting and piloting a research method or methods, the literature should have been reviewed and one or more clearly defined research questions or hypotheses formulated. If the aim of the study is to achieve reproducibility and generalizability, qualitative methods should be rejected in favour of suitable quantitative techniques. If ecological validity and practical and personal relevance are sought, the next stage is to decide on the type/s and amount of qualitative data to be collected. While large sample sizes and rigorous sampling techniques tend to be inappropriate for qualitative work, materials and respondents must still be selected objectively and sufficient data should be collected to ensure that all aspects of the research issue are explored (Morton-Williams 1985).

As with quantitative research, a parallel planning stage involves thinking through and addressing any potential ethical issues. Obtaining formal ethical approval—perhaps from a university or NHS Local Research Ethics Committee (LREC)—is usually advised, but can take weeks or months. This can result in lost time, which is both frustrating and costly. If formal ethical approval is needed this should be sought as soon as is practically possible. Even if formal ethical approval is not required, procedures should be developed for dealing with issues relating to consent, confidentiality, respondent identification and payment/compensation (for an interesting discussion on paying drug users to participate in research see McKeganey, 2001).

As a final precursor to data collection, access to the data field must be negotiated. This involves the researcher contacting those who will be directly or indirectly involved in the research: drug users, professionals and members of drug users’ family and friendship networks, etc. To convince others that participation in the study is worthwhile, the researcher has to be clear about what exactly is being investigated and why. Sometimes this can be explained successfully in one brief telephone conversation. On other occasions, multiple strategies may be required (for example, phone calls, letters, meetings, posters and information sheets).

Data collection

For data collection to progress smoothly, researchers need to be confident about using their chosen research methods and happy to answer any questions regarding the study. They should also enter the field fully equipped—with recording devices working, consent forms prepared and information sheets packed. For personal safety reasons, addiction researchers are wise to carry an identity card and mobile phone. Similarly, they should notify colleagues of their approximate whereabouts and expected times of return.

Alongside these very practical issues, researchers have to be alert to the more subtle dynamics of gathering qualitative data. For example, they are advised to dress so that study participants feel at ease and, if interviewing or focus groups are involved, they need to consider where and when these will occur and even how people will sit so that eye contact and observation of facial expressions are permitted without forcing constant confrontation (Morton-Williams 1985). Researchers must also be alert to participants’ non-verbal cues and ensure that they do not themselves react in ways that could influence others’ responses (for example, by expressing surprise or horror).

In order to generate the most reliable data, the subjects of qualitative research must be reassured that their anonymity will be protected at all times and that they can speak and behave honestly and openly (Jones 1985a). To this end, the researcher must construct a system of coding respondents so that each interview can be documented without any real names or places being disclosed. In addition, the researcher must win the trust and confidence of all study participants through regular reassurances of confidentiality and a non-judgemental attitude. While this is an important aspect of any research, it is critical when investigating drug users who may have very real concerns about the illegality, danger and stigma associated with much of their behaviour.

In practice, of course, the fieldwork encounter can never be planned completely. Humans behave in diverse and unpredictable ways and qualitative researchers must adapt their style to the particular individual/s they are with and to the shifts and developments occurring during the interactions experienced (Jones 1985a). This requires flexibility and accommodation, alongside an ability to remain focused on the stated research topic. A competent qualitative researcher will follow-up interesting leads and seek clarifications when issues seem unclear; yet they will always maintain a basic level of control over the fieldwork situation.

Frequently, however, the effects of drink or drugs complicate the addiction researcher's task. Subjects’ comprehension may be limited, their memory confused and their powers of expression poor. Sometimes, they will convey very little. Sometimes, they will talk at great length. Occasionally, it may be necessary to terminate an interview because a respondent is too intoxicated to talk coherently. At other times, the repetitive respondent may need to be guided gently on to another question. Any participant who becomes upset should be allowed time for composure and given the opportunity to withdraw from the study. Researchers should also be in a position to provide information on local sources of professional support if this is needed.

In order to provide a robust description of the sample, as much information as possible must be kept on people who were approached but refused to participate in the study and on people who began to participate but withdrew part-way through. Researchers should also write detailed fieldnotes immediately after any data collection episode. Such fieldnotes will include subjective opinions and contextual information to supplement all of the other data collected. Those who help with any aspect of data gathering must always be thanked for their time and effort. Similarly, they need be told what will happen to the research findings and given feedback when this is desired.

Analysis

Qualitative research is generally based on the analysis documentary materials, interview transcripts and detailed fieldnotes. Because these data tend to be unstructured, the researcher must impose order and structure on them in order to generate meaning and significance (Morton-Williams 1985). Increasingly, this process of ordering and structuring is being undertaken with the assistance of computer software packages such as NVivo, MAXqda, Atlas.ti, Nud*ist, HyperRESEARCH and Qualrus. These packages do not ‘analyse’ the data for the researcher. They simply enable materials to be ordered and sorted more quickly and systematically than is generally possible by hand.

The analysis itself is a highly personal activity (Jones 1985b). The technique/s adopted within any particular study will probably relate to the nature and amount of data collected, the aim/s of the project, the analytical skills and interests of the researcher, as well as the time and resources available. There is plenty of scope for flexibility. Despite this, procedures must always be systematic, rigorous and complete. All the material collected should be reviewed and re-reviewed thoroughly. If the data are analysed superficially or randomly, only random and superficial (rather than valid and meaningful) findings can ever result.

There are various accepted techniques of methodically analysing qualitative data, such as thematic analysis, constant comparative method, analytical induction, narrative analysis, grounded theory, deviant cases, content analysis and framework. While it is not possible to examine each of these here, the following are common processes: (i) affixing codes to the data; (ii) noting down reflections; (iii) sorting and sifting through the material to identify important phrases, patterns, and themes; (iv) isolating emergent patterns, commonalities and differences; (v) producing a small set of generalizations that cover the consistencies discerned in the data; and (vi) confronting those generalizations with a formalized body of knowledge in the form of constructs or theories (Miles & Huberman 1994).

According to Miles & Huberman (1994), these various processes are underpinned by three concurrent activities: (i) data reduction (selecting, focusing, simplifying, abstracting and transforming raw data, as well as making decisions about which parts will provide the initial focus); (ii) data display (organizing the information, for example by assembling it into matrices, graphs, networks and charts); and (iii) conclusion drawing/verification (interpreting the data by noting regularities, patterns, explanations, causal flows and propositions, as well as testing the provisional conclusions for their plausibility and confirmability).

Dissemination

As findings emerge, researchers will wish to identify suitable audiences for dissemination. Written research outputs might appear in reports, books, practitioner magazines and peer-reviewed journals. Given the wide range of possible audiences and publications that might be targeted, it would be odd to find Addiction selected on every occasion. Indeed, this would suggest a lack of awareness of all the other high quality drug, health, social science and policy journals in circulation.

Although Addiction does not stipulate a maximum number of words, conciseness (as is the case with many scientific journals) is prioritized. Because of their focus on description and explanation, qualitative papers tend to be longer than quantitative ones. Their greater length may be one reason why it can be difficult to get them published in some journals. It is not impossible to write up qualitative work succinctly, but it does require a clear focus and a willingness to omit some of the hard-earned contextual detail.

As with quantitative studies, there are a number of important conventions in writing up qualitative work. For example, authors should relate their findings to the existing literature, provide a clear description of the research setting, discuss the sampling strategy adopted, comment on the research methods employed and acknowledge any obvious study limitations. Equally, they should explain the analytical strategies employed. All too often, authors blithely state that the data were analysed using a particular approach (most commonly grounded theory) without providing any indication of how this was achieved. At the very least, researchers should document the processes through which the themes, concepts and categories they identify were derived from their raw materials.

Two final issues are specific to writing up qualitative work. The first is a tendency to over-rely on the use of quotations. Verbatim comments are an ideal way of bringing to life a particular point or issue. However, they should be used sparingly to illustrate key findings. Papers that string together quotation after quotation with little comment in between suggest that the data have not been analysed rigorously. The second issue relates to anonymity. Qualitative researchers who have promised their research subjects confidentiality must remember that this extends into publication. Because qualitative research is often based on detailed information collected from a small number of individuals in a particular location, identities can be revealed inadvertently. Due care must therefore be taken to ensure that privacy is never violated nor reputations harmed.

CONCLUSIONS

This review began with three rather troubling questions. Is qualitative addiction research only a minority interest? Are qualitative analyses somehow less informative or less valuable than quantitative investigations? and do some addiction journals militate directly or indirectly against the publication of qualitative research? We return to these questions below.

Even in this brief review, we have identified numerous examples of qualitative addiction research. The examples cited are illustrative rather than representative, but reveal that there is both a long history and a strong, contemporary presence of international qualitative work within the addiction field. Qualitative addiction research is, in other words, mainstream. That said, we have revealed how conducting qualitative work is generally labour and resource intensive. Sufficient time and money must be allocated to all the key stages of negotiating access, securing ethical approval, recording and transcribing materials in full, and systematically and rigorously analysing all data collected. Cutting corners invariably results in under funded and/or overly ambitious programmes of work that yield unreliable and invalid findings.

In short, qualitative research—and particularly ethnographies involving lengthy periods of time collecting very detailed data from one particular drug-using subculture—cannot be seen as a cheaper, quicker or easier alternative to undertaking a quantitative inquiry. On the contrary, qualitative techniques often appear economically uncompetitive at the side of their quantitative counterparts. Furthermore, quantitative studies tend to be more suited to producing the kinds of ‘hard’ statistical data demanded in today's climate of measurable targets and performance indicators. As a result qualitative studies are often not the preferred option of research funders, policy-makers and service planners/commissioners. Given that they are less likely than quantitative research to be undertaken in the first place, it is not surprising to find that they are poorly represented in the published literature.

Despite this, we hope that we have shown that the strengths of qualitative addiction research are many and diverse. They have helped us to understand and demystify drug taking, dispel unhelpful myths and stereotypes about drug users, build and develop theories of addiction and formulate and evaluate drug policy and practice. They have also had particular advantages in studying hidden and hard-to-reach groups, identifying emerging trends in drug consumption and researching particularly sensitive drug issues. In addition, qualitative techniques have played an important role in complementing quantitative research by helping to interpret, illuminate, illustrate and qualify empirically determined statistical relationships (Walker 1985, p. 22).

We conclude that qualitative analyses are no less informative or valuable than other forms of investigation. They simply answer different questions and in different ways. However, we emphasize that the success of qualitative research is heavily dependent upon the expertise—but also the interpersonal skills—of those undertaking the work. Good qualitative researchers will understand how to select the techniques and data sources that are most appropriate to answering the question/s being posed. Equally, they will know how to put participants at their ease, establish a sense of trust and rapport and probe and respond flexibly to interesting and unexpected issues without losing sight of the stated research aims and objectives.

Like all research techniques, qualitative methods have limitations. To minimize these, qualitative researchers must always sample study participants carefully, be wary of invoking socially desirable responses, never generalize from their findings and disseminate their data in ways that are accessible to as wide an audience as possible. They must also provide detailed, but jargon-free accounts of their methods of data collection and analysis. Collectively, these strategies should help to ensure that qualitative work is given the credibility and support it deserves by research funders, policy-makers, practitioners, academics (of all methodological persuasions) and journal editors.

Finally, we turn to the question of whether some addiction journals directly or indirectly inhibit the publication of qualitative research. A charge of direct discrimination could be made if journals had an explicit policy of not publishing qualitative work or if editors/reviewers consciously or subconsciously dismissed qualitative research as inadequate or inferior. Indirect discrimination might occur if journals demanded such a formulaic structure that it was virtually impossible to write up qualitative studies in the required format. Without adequate information (qualitative or quantitative) on these issues, we can offer no answers. In any case, the proposed question is unhelpful—being likely only to aggravate those whom qualitative researchers would be better advised to convince.

Instead, we pose two challenges. First, those who are undertaking qualitative addiction research should have confidence in the scientific rigor and value of their methods and not hesitate in writing up their data for any journal that will reach their target audience. Secondly, addiction journals should adopt policies and practices that will potentially encourage more qualitative submissions. This might include a statement welcoming articles based on qualitative research, not being too prescriptive about the structure and length of papers, ensuring that qualitative work is always reviewed by those with qualitative expertise, commissioning publications from those known to be undertaking valuable qualitative investigations and setting an approximate per annum quota of journal space for qualitative contributions.

Acknowledgements

The authors wish to acknowledge Griffith Edwards for commissioning this review and three anonymous referees for their helpful comments on an earlier draft.

Ancillary