Issue15	Winter 1996

Social Research Update is published quarterly by the Department of Sociology, University of Surrey, Guildford GU2 7XH, England. Subscriptions for the hardcopy version are free to researchers with addresses in the UK. Apply by email to sru@soc.surrey.ac.uk.

Collecting data through joint interviews

Hilary Arksey is a Research Fellow in the Social Policy Research Unit at the University of York. She is currently working on a research project examining (informal) carers' perspectives of hospital discharge procedures for young adults with physical and complex disabilities. Her main research interests are in the area of sociology of scientific knowledge and the sociology of health and illness. Her doctoral research used the disputed work related health problem repetitive strain injury as a case study around which to discuss the interactions and influences between lay and expert groups in the construction of medical scientific knowledge. She is currently producing a book version of the thesis.

Key points

• Joint interviewing involves one researcher talking to two people together, for the purposes of collecting information about how the pair perceive the same event.
• Data produced from a joint interview is qualitatively different from that obtained in a sole interview.
• The technique has implications for response rates, organisational matters and finance.
• Problems are likely to arise in relation to one respondent dominating the other; there is an underlying risk of provoking friction and disagreement between the interviewees.
• Potential benefits include the generation of more comprehensive data, eliciting shared and/or dissimilar understandings, and giving a voice to those who might otherwise be silent.

While recent qualitative research methods textbooks discuss interviewing at some length, joint interviewing is seldom mentioned. This neglect is surprising since it is a method adopted in studies dating back to the 1970s (Pahl and Pahl 1971; Edgell 1980). Drawing on research texts and documented empirical studies, this Update examines joint interviewing and comments on implications of the technique in terms of the research, the researched and the researchers.

What is joint interviewing?

The first point to consider is just what is meant by the term 'joint interview'. Is it one interviewer questioning two respondents; two interviewers questioning one respondent; or two interviewers questioning two respondents? What about an interview intended to be conducted with only one person, but which in the event takes place in the presence of someone else who then makes a contribution (Drummond and Mason 1990; Mullen 1990)? If a researcher poses more or less the same questions to a husband and wife within individual interviews and then constructs some sort of joint account by reconciling the two versions (Seymour, Dix and Eardley 1995), what status does this hybrid narrative merit?

It is clear that 'joint interviewing' could cover many different situations; this article considers it to be when one researcher interviews two people together. In many instances, joint interviewing takes place in combination with separate interviews (see next section) and so this Update also attends to issues likely to have a bearing upon this mixed strategy.

The interview context

Joint interviewing is commonly used in the context of inquiries involving people in marital relationships or living as couples (McKee and O'Brien 1983; Mason 1989). Specific areas of study include labour-market decision-making in low income (Jordan et al 1992) and higher income (Jordan, Redley and James 1994) households, finance (Pahl 1989; Huby and Dix 1992; Eardley and Corden 1996) and unemployment (Ritchie 1990). Families constitute a further research site; in recent years, studies using this interview method have explored changing generational attitudes to health enhancing behaviours (Backett 1990a, b), the experiences of non-sexist child-rearing (Statham 1986) and what it is like to live with a disabled child (Voysey 1975; Baruch 1981). In addition, joint interviewing is employed in projects looking at illness and disability where the respondents generally consist of the care recipient and the carer -- usually, but not always, a spouse (Radley 1988; Gerhardt 1991; Parker 1993; Shakespeare 1993; Baldock and Ungerson 1994).

It often happens that joint interviews are used as part of a multi-interview research design, incorporating both individual and joint conversations. Backett's (1990a,b) study into health beliefs and behaviours in middle-class families, for example, involved three rounds of interviews with parents spread over 18 months. Men and women were first interviewed separately, then jointly, and finally both separately and jointly. An ancilliary study focusing on the children of these families completed the fieldwork. Other projects are less complex in design and use the joint interviewing technique in conjunction with just one separate discussion (held either consecutively or simultaneously) with each of the individuals concerned.

One benefit of combining joint and individual interviews is that in principle this provides a sense of equity for both participants (Brannen 1988). This dimension is especially important for female respondents; one cause for concern noted by a number of researchers (Pahl and Pahl 1971; Jordan et al 1992; McKee and O'Brien 1983) is that women, who in an individual interview are talkative, are inhibited during the associated joint meeting. According to the evidence, it tends to be men more than women who are the 'scene-stealers' (Shakespeare 1993); in the context of a joint interview, men are more likely to be overbearing, to interrupt or speak on behalf of the other (Jordan et al 1992).

Insofar as unplanned joint interviews are concerned, there is clearly an issue about whether or not to continue with the interview given the aims of the research. If efforts to persuade the 'gatecrasher' to leave the room fail, the situation may nevertheless be turned to advantage. In their study of the experiences of people with diabetes, Drummond and Mason (1990:40) used the interview as an opportunity to obtain reports of what it was like to 'live with a diabetic'.

Advantages and disadvantages

On the positive side, joint interviewing helps to establish rapport and an atmosphere of confidence (Edgell 1980); reveals the different kinds of knowledge held by each person (Seymour, Dix and Eardley 1995); and, produces more complete data as interviewees fill in each other's gaps and memory lapses (Seymour, Dix and Eardley 1995). In addition, researchers (Pahl 1989) claim joint interviewing offers insights into the interactions and nature of (power) relationships between couples through observation of non-verbal modes of communication. The technique enables the voices of those who might otherwise be silent to be heard, for example when a carer acts as a 'prompt' or 'translator' on behalf of a person with disabilities (Shakespeare 1993).

If joint interviews are held after separate ones, factual data can be cross-checked (Morgan 1988) and inconsistencies investigated and clarified (Jordan et al 1992). For some (Pahl 1989; Backett 1990a, b), this combined design has an edge in the sense that it brings into relief disparities and areas of tension between couples. Their different perceptions are more likely to be lost sight of if one person has been left out of the study and their views inferred from the data given by their partner.

Some of the disadvantages of joint interviewing have already been hinted at, in particular one person dominating the other. Other drawbacks are also reported. The occasion has the potential for stirring up antagonisms and conflicts of interests. Radley (1988), for instance, documents an exchange between a husband, claiming to be recovering after coronary artery by-pass graft surgery, and his wife, who instead doubts the success of the operation. Having eventually admitted to the return of his angina symptoms, the couple then had problems comprehending each other's needs. This disharmony was manifest in the last interview of a series of five, which ended with the pair falling out and the wife giving way to tears. Related is Pahl's (1989:61) comment that in her study researchers 'went away feeling very anxious about [their] responsibility for post-interview marital rows'.

Despite these sorts of recorded anxieties, research texts are often reluctant to address how best to terminate interviews. According to Seymour, Dix and Eardley (1995), difficulties are more likely to arise in cases of joint interviews than individual ones. In their experience, if both types of meeting are being held, it is preferable to conduct the joint one first. This interviewing sequence has less chance of disrupting the procedures researchers use to conclude their discussions, which might well consist of giving something back to the respondent and/or agreeing prerequisites for publication of the research material (Laslett and Rapoport 1975).

Data characteristics

In light of the above observations, a key question to address concerns the quality of the data emerging from a joint interview, a point little explored in authors' accounts of their work. One exception, however, is Huby and Dix (1992) who explicitly reflect on the quality and quantity of information obtained through structured interviewing of both partners together. Following evaluation of the piloting exercise, the researchers reached the conclusion that '[i]n neither quantitative nor qualitative terms did interviewing couples produce significantly better data than did interviewing named respondents only' (Huby and Dix 1992:141). This judgement, supplemented first by the belief that respondents frequently did not concentrate as well when there were two present, and secondly by the desire not to cause dissension in the household, prompted Huby and Dix to address the main part of their investigative work to individuals rather than couples.

A further point to bear in mind is that the nature of joint interview data is qualitatively different from that obtained during one-to-one conversations. This partly reflects the fact that the latter is an individual reconstruction of events, opinions and the like whereas the former consists of a jointly constructed picture (Seymour, Dix and Eardley 1995). There is a sense in which couples exhibit rhetorical practices of 'jointness', in other words they deliberately attempt to generate a unified image or reality (Jordan et al 1992) although not always successfully (Jordan, Redley and James 1994). This sort of 'collaboration' or 'sharing' supports the idea that joint interviews produce more consensual data (Morgan 1988), and that overt conflict during the course of an interview is the exception rather than the rule (Jordan et al 1992). Implicitly, it appears that people are trying to present themselves in a favourable light, a notion which is consistent with the likelihood that respondents are presenting their 'public' accounts of socially acceptable behaviour (Cornwell 1984). As part of this, it has been suggested that individuals use the joint interview occasion to legitimate or justify their actions (Radley and Billig 1996), or alternatively to normalise their situation (Voysey 1975).

Practical considerations

There are various practical difficulties which are directly pertinent to a joint interview research design. It seems that this sort of technique is associated with a low response rate. Pahl (1989), for example, notes that a comparatively high proportion of couples contacted declined to be involved with her project, in part reflecting that the focus of study was the sensitive issue of money, but also that both spouses were to be involved. While one partner would agree to be interviewed, their partner refused. Men were more likely than women to be unwilling to participate (Pahl 1989; Jordan et al 1992).

Setting up joint interviews can be a long drawn out process, especially if they are conducted in combination with separate interviews (Pahl and Pahl 1971; Pahl 1989).

Finally, it is important not to lose sight of any financial implications, although again, this issue is likely to be more significant when the research design comprises both joint and individual interviews. The areas where increased expenditure may be expected include: the training and time commitments of the interviewer(s); fieldwork activities; transcription and coding of data; and, data analysis and interpretation. Such extra resourcing has the effect of putting implementation of this sort of technique out of reach of many researchers, in particular those working within the limits of postgraduate funding.

Joint Interviewing from the perspective of the researcher

To complete this review of the various dimensions of joint interviewing, it is appropriate to consider what the technique is like from the point of view of the researcher(s). Overall, it appears there are particular problems associated with the strategy, ones which are compounded in those instances where both joint and separate interviews are undertaken. The research interviewer in Pahl and Pahl's study (1971), for example, describes how exhausted she felt after talking to husband and wife for anything up to eight or nine hours (and no doubt the respondents were equally tired). The occasional loss of concentration meant that questions were not always followed up satisfactorily. Stresses and pressures, albeit for a different reason, were similarly experienced by the sociologist conducting the interviews for Jordan et al (1992). Among others, these related to tensions and points of difference between the couple which were exposed during the course of the separate and/or joint interviews, manifestations which in turn gave rise to unanticipated ethical dilemmas.

Other commentators (Kadushin 1972; McKee and O'Brien 1983) document feelings of divided loyalties and a clash of allegiances, as well as highlighting the difficulties of obtaining a relatively equal level of commitment and response from both participants. However, the other side of the coin in terms of the latter point is that the interviewer might have the opportunity to sit back and observe interaction between respondents.

Conclusion

As with any research, the choice about what method to use for collecting data is influenced by theoretical and conceptual issues, as well as practical considerations. This article has focused on the strengths and weaknesses, the advantages and disadvantages of joint interviewing. While the approach has certain strong points, it is important to bear in mind that joint interviews can be difficult to organise and conduct. Clearly, there is a need to think very carefully beforehand as to how appropriate it is to use this particular method.

References

Social Research Update is published by:

Edited by Nigel Gilbert.

Winter 1996 © University of Surrey

Permission is granted to reproduce this issue of Social Research Update provided that no charge is made other than for the cost of reproduction and this panel acknowledging copyright is included with all copies.