Harnessing expertise: involving peer interviewers in qualitative research with hard‐to‐reach populations

Abstract

The use of peer interviewers with privileged access to a particular population group, which is difficult to reach via more conventional methods, has been acknowledged in recent research. This paper explores a number of key issues relating to the employment of peer interviewers by reflecting on a project designed to explore the views and experiences of parents who use illegal drugs. The project presented the research team with a number of challenges. These included the need to provide on‐going support for the interviewers, a sense of distance felt by the researchers from the raw data they collected, and the difficulties of gaining from the skills and experiences of peer interviewers without exploiting their labour. The paper also explores the advantages of involving peer interviewers closely in research work and reflects on the nature and boundaries of expert knowledge that can become evident in such collaborations. The need for a certain amount of flexibility over the roles and domains of control that lay experts and researchers traditionally inhabit is suggested. In conclusion, it is argued that the involvement of peer interviewers in research can be a valuable means of enhancing our knowledge and understanding of a variety of population groups who tend to live beyond the gaze of more orthodox researchers.

Introduction

The lives of illicit drug users tend to be concealed by a veil of ignorance, on the other side of which ministers, policy‐makers and service providers struggle to develop consistent strategies designed to improve public health, uphold community safety and prevent the perpetration of crime. Drug users are brought to public attention mostly in a negative context, for example when they give birth to `junkie babies', abandon syringes in local parks, steal to support their habits, or their children overdose on methadone kept in the family home.

The publication of the UK Government's 10‐year drug strategy in the late 1990s ¹ coincided with a growing recognition that the number of people using illicit drugs is rising ² and an acknowledgement that our understanding of the lifestyles of drug users, and particularly those beyond the reach of formal support services, is relatively limited. It seems likely therefore that pressure will increase in relation to the commissioning of research which enhances our knowledge of these and other hard‐to‐reach population groups. However, in view of the fact that illicit drug users are both committing a criminal act and are stigmatized as belonging to a social `underclass', gathering this information is by no means an easy task.

One way of trying to overcome these problems has been to work with `peer' or `privileged access' interviewers as data collectors. These are people who have ready access to a population group that is not accorded to more traditional researchers. ³ As far as drugs research is concerned, this can include anyone who has close contact with, and is trusted by, members of the local drug‐using communities. The employment of peer interviewers has been used, for example, to explore the management of risk in relation to injecting behaviour, ⁴ compare the characteristics of drug users who are using services with those who are not, ⁵ evaluate the success of drug‐related health education interventions, ⁶ and assist in the social mapping of drug‐using communities. ⁷

The advantages of involving stable and former drug users as health educators have also been acknowledged in a wide variety of initiatives aimed at reaching groups of users who would not normally access formal services. Their insight can prove to be a valuable research resource in exploring the views and needs of a wide range of drug users, although this has been known to create tensions for the peer educator/field worker who encounters conflict in managing such a dual role. ⁷

Our own experience of working with peer interviewers presented a number of challenges to us as researchers that are considered in this paper. These included the on‐going support we needed to provide for the interviewers, our sense of distance from the raw data they collected, and our concerns about working with and benefiting from their skills and experiences without exploiting their labour. It also highlighted the potential advantages of involving peer interviewers closely in local research, and raised a number of issues around the traditional divide between `expert' and `lay' knowledge in such work.

Eliciting the views of drug‐using parents: background to the fieldwork

The needs of parents who use drugs have been of growing concern to health and social services for a variety of reasons. Stories of the accidental and sometimes fatal methadone poisoning of young children, perceptions about drug‐using parents being less likely to use services than their childless counterparts, a lack of information about why these parents should fail to access services, and a lack of evidence relating to successful models of provision, have all contributed to the debate.

Although national guidelines were being drawn up by the Local Government Drugs Forum and the Standing Conference on Drug Abuse ⁸ during 1997, one multi‐agency drug action team (DAT) decided to explore the local picture in more detail. We were commissioned to interview drug‐using parents living within the DAT's catchment area, which included one health authority but two local authorities each with its own, and very different, organizational and cultural history and model of service delivery. Members of the DAT were particularly keen for us to talk to non‐users of services as well as to those who were already in treatment.

In light of the problems already outlined we realized that as orthodox researchers we would find this a difficult, if not impossible, task. Fortunately, one of the two community drug teams (CDTs) in the target area had a tradition of involving volunteers or `user representatives' (both stable and former drug users) in the provision of services at the CDT itself, at local drop‐ins and on a mobile needle‐exchange scheme. Some of these volunteers had also been recruited to help administer local questionnaire surveys. We knew they would have access to respondents we could never reach and this, together with their previous experience of local research (albeit predominantly quantitative and highly structured) and the fact that they had already undergone a recruitment process, persuaded us to invite two of them to take part in the project.

Finding two similar interviewers in the other CDT area was a much lengthier process, as it had no such tradition of working with local drug users. Here we had to rely on providers to suggest people who might both be willing and have the potential skills to work as peer interviewers. Because they would come into close contact with more chaotic drug users these people had to be considered capable of controlling their own addiction, a concern which has been highlighted by other researchers in this area. ⁹

The fieldwork itself consisted of 52 interviews with drug‐using parents, some with couples but the majority with individual users. We asked the peer interviewers to use their own networks and contacts to identify parents using illegal drugs for participation in the study, although they were instructed not to include people who were using only cannabis in isolation. The interviewers came from different parts of the drug action team area, and so some geographical spread was expected in their contacts. We carried out most of the interviews with parents who were in touch with local services ourselves, accessing them through a range of key workers and the four peer interviewers. However, we relied on the peer interviewers to access and interview nearly all the former and non‐service users who took part in the research. The interview sample was more or less equally divided between female and male respondents.

This was approached as an exploratory, qualitative study. The aim was not to obtain a statistically `representative' sample of interviewees, but to explore the relatively unknown terrain of the day‐to‐day experiences of drug‐using parents and their families and to see how, if at all, external support from formal or informal sources impacted on their lives. We therefore decided to use a fairly open interview schedule, rather than a structured questionnaire, since it was considered a more appropriate and effective means of obtaining insight into the views and experiences of a relatively hidden population.

Training and debriefing

All four peer interviewers received training from us prior to starting the fieldwork. To a certain extent this training was tailored according to their individual needs and previous experiences, and was carried out individually or in pairs. For the two interviewers who had not been involved in research before, the training sessions included an introduction to the study, an explanation of research interviewing techniques, a demonstration and a role‐play with feedback and discussion. All the interviewers were asked to tape‐record interviews wherever possible and shown how to use the tape recorders. However, only one of the interviewers used the tape recorder at all, the others saying that respondents had refused to be recorded. This had implications for the research that are discussed in the following section.

Whilst allowing for the time it would take to recruit and train the peer interviewers in open interviewing techniques, we underestimated the amount of on‐going support and supervision that would be necessary to ensure the data collection process ran more or less according to the objectives and time‐scales agreed with the DAT. The peer interviewers were individually debriefed after every two or three interviews, both to ensure that as much information as possible had been recorded in their notes and to allow the interviewers to raise any problems or concerns. These debriefing sessions had advantages for both sides: they acted as a support mechanism for the interviewers and helped them to feel more involved in the project, and they gave us a chance to develop a relationship of trust with the interviewers and reflect on the data as the research progressed. They also provided a valuable opportunity for the researchers to learn something of the language and culture of the respondents from the peer interviewers. The debriefing sessions may have been lengthier and more frequent than we had anticipated, but they were crucial to the success of the project.

Supporting the interviewers

While this debriefing went some way towards providing the mechanisms needed to support the interview process, working in this area raised additional issues about supporting the peer interviewers themselves. One of the biggest concerns was that they should not have their stable drug use or drug‐free status threatened by being placed in close contact with potentially more chaotic users. It was important therefore that the interviewers were considered capable of controlling their own drug use, but it was equally important that we maintained relatively close contact with their key workers throughout the project. We ourselves were unqualified to respond to any slide towards problematic behaviour by the interviewers and were wary of setting ourselves up in a quasi‐counselling role in the event of any such problems.

This issue of support at two different levels was one recognized by the peer interviewers themselves. In a final debriefing and discussion session they said they had felt adequately supervized by the research team on the whole but realized they were `on their own' when interviewing, which was the time of greatest risk to their drug‐using status. They favoured some form of peer support system for dealing with this situation in the future, with more experienced interviewers acting as mentors for the newer recruits. They also felt that regular interviewer team meetings as the project progressed would have offered a useful forum for information‐sharing and problem‐solving, as well as a means of on‐going peer support.

Working with `second hand' data

The sense of distance from the raw data that was created by our adoption of a `back seat' in the interviews with non‐users of services, was a source of anxiety. As already mentioned the peer interviewers had been asked to use tape recorders during the interviews, but they claimed that most people would not agree to talk if they felt their anonymity was being in any way compromized. This obviously raised concerns in relation to quality control; we had not expected to have to rely on `second hand' accounts of the interviews to such an extent. In this context we could have followed the example of other researchers who have insisted on the tape recording of interviews; Griffiths et al., for example, found the argument that taping was to check whether the interview had been carried out correctly, if at all, proved persuasive with most of their respondents. ³ However, given that many of our interviewees had chosen to hide their status as drug‐using parents from families, neighbours and service providers, their reluctance to commit their stories to tape was not that surprising.

These concerns were compounded by the fact that the interviewers and respondents shared a language and lifestyle which was relatively alien to us, and which meant that some areas of common knowledge and experience discussed during the interviews were not recorded in the interview notes. It was only during the debriefing sessions that these notes could be supplemented by data that we saw as relevant and important but the interviewers tended to regard as `old hat'. Of course this is an issue in other areas of research too, where interviewer and interviewee share an experience to such a degree that the views are not made explicit in the data.

Given such concerns, in future we would probably try to make more use of peer interviewers as `finders', asking them to access respondents through their personal networks who would be prepared to be interviewed by us as `outsiders' because we had been introduced and vouched for by someone they trusted.

Empowerment or exploitation?

While it was hoped that this was a potentially empowering piece of work which built on and valued the knowledge, experiences and skills of the peer interviewers, we were aware of the fine line between involving and empowering people on the one hand, and exploiting their labour and expertise on the other. Two of the interviewers had already spoken of their frustration at feeling `used' by both the community drug team and the local drug reference group, which they felt had benefited from their contact networks, views and experiences as volunteer workers, but had offered little or nothing in return in the way of recognition, remuneration or a sense of ownership of the work.

To counteract this, an attempt was made to involve the interviewers throughout the research, from the initial design of the interview schedule to discussion of the main themes emerging. A fee per interview satisfactorily completed, paid after each interview debriefing session, was negotiated, and payment was also made at an agreed hourly rate for attendance at the training sessions.

With hindsight the peer interviewers' involvement was probably limited by the fact that we had no formal contract with them. While this may have given the interviewers a greater degree of freedom in relation to any other commitments they had, it meant that their time with us, apart from the formal training and interviewing time, was given freely. It also meant that they had no traditional employment rights or benefits. The extent to which this latter issue could be considered exploitative is a moot point, however. Many areas of research work require a staff input like this for short periods of time only, and in reality the peer interviewers may not have wanted a more formal and potentially more binding contract with us. ³

What did the project gain?

In research terms, the most important advantage of working with peer interviewers was that we were able to obtain the accounts of a hidden population group that it would have been difficult, if not impossible, for us to reach in any other way. In addition, the interviewers were able to collect the data in a relatively short space of time from a diverse range of users. The interviewers themselves came from different backgrounds and geographical areas and so moved within, and accessed respondents from, different networks and communities. As other studies have pointed out, data validity can be improved by multiplying the types of social networks and milieus in which peer interviewers recruit their respondents. ^{9 , 10} Rapid data collection is also an important advantage for researchers responding to the needs of health and local authorities for relevant information from hard‐to‐reach groups in a reasonably short period of time.

It must be emphasized, however, that the benefits of involving people in research who have privileged access to hard‐to‐reach groups such as illegal drug users are only realized if there is some investment in developing their roles in the first place. Our research depended very much on the fact that the peer interviewers from one local authority were already established as mediators between drug users in the locality and the community drug team. Another important factor in the research process itself was the investment of time in ensuring a relationship of trust between the researchers and the interviewers.

Notwithstanding such benefits, the project did raise a number of questions about the traditional divide between `lay' and `expert' knowledge and experiences. While we would readily acknowledge the importance of the privileged access to respondents that they provided, the interviewers also claimed a certain expertise as field‐workers in a domain (that is, in interviewing techniques) that we tended to feel was our own territory. As in any research which involves lay experts, negotiating and recognizing the boundaries we both inhabited and the knowledge and skills we could share with one another was an important issue to confront.

Negotiating areas of expertise

The interviewers claimed that they could develop a rapport with the respondents which someone who had never had problems with drug use, even an experienced researcher, would be unable to reproduce. They also felt they shared experiences which enabled them to `read' situations essential to understanding the nature of the encounter, to recognize how and when to probe further into certain aspects of respondents' lives as drug users, and to enter into a discourse shaped by a common language which would be `strange' to most researchers.

The final point would probably have to be conceded. The street jargon respondents used was quite opaque to us at times, particularly at the start of the interview process, and doubtless hid opportunities to explore particular aspects of the lives of some of the parents we interviewed; opportunities which the peer interviewers would probably have recognized. However, the remaining claims raise issues about the strength of the argument that matching respondents with interviewers who have an insider status in their world will provide `true' accounts which cannot be accessed by other people.

The assumption that the true account of a person's views and experiences exists suggests that within each individual there is one unique story which somehow needs to be `unlocked'. The interviewer's skills would therefore be to hold the correct verbal keys to release each unique account. However, as has been argued in relation to the issue of using black or white interviewers with black respondents, for example, people may give very different accounts to different interviewers and reinterpret their experiences according to who they are talking to, but this does not mean that one account is less valid than any other. ¹¹

It does, however, emphasize the importance of recognizing the ways in which different interviewers may influence the responses given. In relation to drug use there is evidence to suggest that respondents are more likely to play down their addiction with their peers than they are with other researchers. ¹² The need to maintain dignity and display `honour' plays a crucial role in shaping their accounts. In a similar context in their study of street workers, McKeganey and Barnard note that the closeness of relationships between interviewers and respondents may constrain or influence the kind of information the interviewers are offered. ¹³ However, this should not be taken to mean that these accounts are invalid as a consequence.

Respondents may also talk to outsiders about certain issues they would have considered it inappropriate to raise with one of their peers. In our own interviews with drug‐using parents for example, ¹⁴ issues such as depression, suicidal feelings and domestic violence arose that were not apparent in the data collected by the peer interviewers. However, this may also have reflected the fact that we are female, while three of the four peer interviewers were male.

Some reflections

All these issues serve to emphasize the range of knowledge and skills that peer interviewers and traditional researchers could share with one another, suggesting the value of retaining a certain amount of flexibility over the roles and domains of control that they both occupy. As researchers we were forced to relinquish a degree of control over an area of expertise that we felt we owned, but we also benefited enormously from the knowledge and expertise the interviewers themselves possessed as lay experts in the field of drug use.

Reflecting on the research process involving peer interviewers provoked ideas about the possible development of a more co‐operative and egalitarian approach. In practical terms this could entail joint training and debriefing sessions leading up to, during and following the undertaking of fieldwork. This would help to ensure interviewers with privileged access to peoples and cultures that are relatively inaccessible to more orthodox researchers are appreciated for the different, but equally valuable, skills they bring to the research process. At the same time researchers would be able to share their theoretical and methodological ideas with peer interviewers in a more open and engaged context.

Of course this calls into question other issues in relation to the pay and conditions of peer interviewers and the nature of contracts between commissioning agencies and research organizations that need to be confronted. Whatever decisions are reached, however, should be informed by an understanding that the peer interviewers are undertaking the work as an integral part of the research team on the basis of their relevant skills and expertise.

What this project appeared to demonstrate is that the employment of peer interviewers with privileged access to `hidden' populations can prove to be a valuable means of enhancing our knowledge and understanding of people belonging to a diversity of cultures and communities. This knowledge could in turn help to inform public and policy debates about how we construct the problems facing us and work to meet public health and social welfare needs. In acknowledging their worth there must, however, be a recognition of the skills that peer interviewers themselves bring to the research process, and an understanding that some of these skills will challenge a domain currently occupied by more traditional researchers. This paper makes some attempt to explore the essence of such skills within the context of other debates about the nature of lay expertise in qualitative research.

Acknowledgements

Thanks are due to: the Salford & Trafford Drug Action Team which commissioned the research on which this paper is based; to all of the users and service providers who agreed to be interviewed as the project progressed; and especially to the four peer interviewers, without whose involvement the research would not have been completed, nor this paper written.