Tag Archives: Fieldwork

Treating interviews as ethnographic encounters

I have emerged from weeks and months of blog silence in order to think about a question posed on Twitter this morning, in which @MatthewHanchard asked: “Should I be worried if my semi-structured interviews only seem to last about 30 mins if all Q’s were covered (planned for 90 mins) #phdchat”. This not only interested me, and brought back to mind my own early struggles with interviews, but also seemed to require more than 160 characters to give a proper answer. So here goes.

First off, a caveat: I do not know sufficient details of research design, methodology, research questions etc. to give a  precise answer to Matthew’s question. It could well be that 30 minutes is enough. I was interviewing a GP yesterday who had very kindly agreed to see me between clinics. I turned off the tape recorder at about 35 minutes, in which time he had given me a very concise but informative set of responses, and offered me a distinct perspective on my research questions – not bad considering I am in the last week of data collection. I also think that almost regardless of how much time he had available, I might not have been able to delve much deeper than the story he first offered me. Sometimes you just find yourself in front of someone who for whatever reason cannot, or will not, take you any further down a particular path; they may resist or even become irritated if you carry on attempting to push them further and you risk losing the interview altogether. 30 minutes of thoughtful, if circumscribed, description is certainly better than an hour of terse one-word responses, disinterested agreement, or, worst of all, a terminated interview.

The context of the research is also very important. The research I am conducting at the moment is a service evaluation. The main output will be a 10,000 word report back to the trust who commissioned the research. Of this report, the first 2 pages in which we will give a set of key messages and recommendations will be the only 2 pages read by the majority of people who see it (admittedly not the most exciting piece of research I’ve ever done, but don’t knock it until you’ve experienced the post-doc research job market). If I had radical theoretical insights based on innovative methodologies on my immediate research horizons, then I might not be particularly satisfied with a 30 minute interview. But I do not. I need actionable findings situated within the immediate policy context in which they are going to be spread. So, an interview from a senior local clinician full of concise, to-the-point, quotable responses was just the ticket.

I am assuming that Matthew did not come away completely satisfied with his interview encounter. As a PhD candidate he will have aspirations to break new ground in some regard (to do so is after all the fundamental requirement of the PhD: contribution to knowledge). And so 30 minutes feels a bit thin. The fact that he is planning to spend 90 minutes per interview suggests to me that Matthew either has a very long interview schedule, or that he wishes to delve deep into his questions. 90 minutes is a long time to try and hold your own and your participant’s attention. As Matthew has said he is using a semi-structured interview schedule, I am assuming he is taking a qualitative approach. The challenge with the ever-popular ‘semi-structured’ interview, is achieving a balance of open/closed structure that works for you and your research questions, while allowing the participant the space to find their own narrative. Your interests and theirs may be wildly divergent. When you ask a particular question you do not have any idea what it might spark. This is of course the fascination of exploratory, qualitative research. But it means you can find yourself listening to a very interesting story that has strayed far away from your research intentions, and it takes considerable skill to gently shepherd a stray back to the path, without them feeling, well, like a sheep.

These kind of skills come with experience, confidence and comfort with the interview situation. But you can also do yourself many favours by paying close attention to your research schedule, and building in some contingency to it. Regardless of your methodological framework, I think there is an excellent case for thinking about the interview situation as an ethnographic encounter. I’m using ethnography here less as a prescribed set of methods and principles, and more as a set of aspirations based around accessing a contextualised descriptive account of your participant’s story, in their own language. There is a great challenge to doing this in the interview context, which is brief and performative – by which I mean there are quite formal sets of cultural rules and conventions shaping the interview context. These conventions can act as a barrier to the establishment of trust and empathy between researcher and participant. I like to do what I can to break down some of these power relations before the actual event of the interview: ‘gaining access’ in ethnographic research means much more than gaining permission to physically access a particular group or organisation, rather, it is an ongoing concern, an always temporary agreement between researcher and participant which requires ongoing maintenance and sometimes repair. How exactly you will attempt to negotiate access in your own research will be a very situated and subjective concern. The point is you only have rights to ‘be there and ask that’ as long as the person you are asking gives you that right. This will not likely be something that is always negotiated openly, so paying keen attention to social cues, your participant’s level of comfort etc. is very important. This is not just about creating a ‘nice’ research context, it is about having confidence in the veracity of the story you have been told. All stories are constructed, and so gaining a sense of trust is about guiding the interpretations you will generate out of these stories.

The fundamental building block of analysis in ethnographic research is description, and the insight it gives you into context. So, it is essential to allow breathing space in interviews. I often start this right at the beginning, those gentle, open questions at the beginning of the interview. You might not want to just say ‘tell me about yourself’, it should still be focussed on your research questions, but the emphasis should be on giving participants the chance to address you in their own words, to begin to build their narrative, their approach to or experience with the things you are interested in. When I have been aspiring to something as long as a 90 minute interview, as much as a 1/3 of that time might be spent drawing this story out – it is very unlikely that will come in one 30 minute rush, rather you have to follow-up responses, probe for more description, get a fuller context – it is this context that will allow you to make more precise judgements of the responses you get to more structured questions later on.

I have used the word ‘participant’ throughout this post. This is often a very euphemistic term in interview research, implying a degree of participation which is in fact not permitted by the closed, circumscribed, power-laden manner in which interviews can be experienced. As I’ve said, interviews are highly structured by convention, and so it requires extra work on the part of the researcher to break down some of this. It is very important that words like ‘participant’ finds some meaning in the manner in which you go about arranging, accessing, conducting and analysing interviews. Not just in the interests of conducting more ethically legitimate research, but also in providing you the researcher with some confidence in your interpretation of what has been said – regardless of how long it took to say it.


Small Change/Big Change

This post presents a response to my last post, which talked about routines, and I think probably gave them a little bit of a hard time. Yes, they can be a little dehumanising, and we can be far too dependent on them, and external agencies can use our patterned behaviour against us…but, they are also very powerful little strategies which can be put to good use. If we use them reflectively, perhaps we can avoid many of the pitfalls I talked about previously.

One of the things I have come to learn about critique in general is that it feels a bit disingenuous to simply stand at the sidelines pointing out all the implications and consequences of everyone else’s action. At some point you have to get your own hands dirty, whether that be making recommendations to practitioners, contributing to policy debates, or even just having an honest conversation with someone you’ve done some research with.

I don’t know if i’ll ever find it easy – nailing my colours to the mast, while simultaneously being able to see the potential problems with each new solution. Nevertheless, it is necessary if you wish to make a contribution to a debate, if you wish to engage with practitioners and policy makers in their languages, and if you wish to have some kind of impact. Bringing about change in practice is one of the biggest challenges we face in our work. So often, the change you really want to be able to make is transformational, not this piecemeal, do things slightly differently, change one or two procedures, but the big, system changing, relation changing, heavy change. I don’t know whether lots of small changes add up to a big change in institutional terms. I think in some cases they can do. After all one of the greatest sources of power is alliance, so collecting enough small things together might prompt change in much bigger things.

I used to have a job with the National Institute of Health and Clinical Excellence (NICE). They produce guidelines which are non-statutory but make up the official benchmarking for clinical practice in the NHS. I was employed by one of the NICE guideline development groups as a service user representative. This was never a role I managed to find any great comfort with. Some of the reasons for this was to do with my own hang ups about what or who I was supposed to be representing, and what exactly it was that made me representative of a particular group. But all sorts of other things were just to do with the a priori constraints that existed and limited the possibilities that could be achieved through such a guideline. For example, when making assessments about which treatments to recommend for particular conditions, we were constrained by the existing health technology assessments, which used a health economics measure of efficacy called a QUALY. Basically, this is a numeric tool combining the cost of a treatment with its reported improvements of quality of life. We had the calculation and its method explained to us – it was problematic to say the least. QUALYs are seen as significant to the fourth decimal point, and yet the calculation that goes into them is hopelessly vague and subjective. But regardless of how many of the group offered their own anecdotal best practice, if it did not fit the QUALY hierarchy then it could not be recommended.

And yet, there were things about the NICE process where there was a certain amount of compromise available between the group of individuals producing the guideline, and the institutional codes which constrained us. We managed to persuade the college that was funding the guideline to commission some primary research the analysis of which was included in the guideline. This is the first time that this has ever been the case with a guideline, which ordinarily is more like a systematic review of existing research. I think it strengthened the guideline greatly, made it appear more human, and will have given cause to stop and reflect for any practitioner reading it. I suppose that this was a fairly small, but nevertheless significant change that we prompted, and I suppose if it became the norm within a guideline process then that could be considered a bigger change. Yet, there were many accepted methods and languages that we found no give in at all.

I’m approaching the stage in my current project where we will be required to write a report back to the funders, including our recommendations. Research recommendations are often the least interesting bit of a report. They can be so dry, stripped of life, taken out of context, superficial, preachy and obvious sounding. You can imagine people reading them and going ‘duh! I already know that!’, and yet they are difficult to flesh out, give context too, make thicker, because they are intended to be snappy, direct, practice focussed, headlines.

Another project I am involved in found a creative solution to this issue. We wrote the report in as ‘lively’ style as we felt able, nevertheless we had a huge amount of really rich research material that would and could never make it into this rather dry 40,000 word document. So, the PI decided to use the remaining dissemination budget to commission a writer to write a script based on our fieldnotes. This had never been a stated intention of the project, when we were researching it we never had any sense that our work was going to be dramatised – and thus we never ‘went looking for’ dramatic moments. It has been a fascinating process to be involved in though.

Roughly two years later the script has been written, workshopped, well received by stakeholders, and we are now working towards a full production of the play in June 2013. One of the most exciting things to have happened is that the healthcare trust with whom we conducted the original research have remained engaged in this unexpected follow-up – to the extent that they and two other trusts in the area have already reserved 1000 seats across 12 performances of the play for their staff to attend. This feels like double impact – both a novel and creative means to disseminate research findings, and a largely new theatre audience, drawn from some of the most poorly paid frontline staff in the NHS. That three healthcare trusts – with all the recent policy flux and budget slashing – are prepared to release something like 150 staff at a time to come and see a piece of theatre…that certainly defies expectation, perhaps represents a certain amount of transformation in itself, and if we are able to present something to challenge as well as entertain then the ripples could spread much further.

I’ll certainly be blogging more on this process, so watch this space.


Whose line is it anyway?

Research interviews are a strange world in my experience. It was quite a long time before I had any faith in them at all as research methods, and longer still until I gained any comfort in the role of interviewer. I’m getting there on both counts.

The current project I”m working on is almost entirely based on interviews, around 100 of them in total, each lasting around 90 minutes. That’s going to be a hell of a lot of words. But what do they actually mean? Each piece of data is really just an improvised statement, prompted by a question which the respondent may never have given any thought to previously, all contained in a completely manufactured situation – the interview – which might appear contained, in a room, on a tape, but in fact is very much part of the fabric of the world that it enters, it’s just getting a handle on all the different meanings of the varying performances one encounters in the interview situation which can make your head spin. An enigma machine might be handy.

Is this how Clive Anderson felt after each sweaty, incomprehensible, possibly brilliant performance from Tony Slattery? His response seemed, generally, to give the points to Ryan Stiles.

I think performance is a good way to begin to try and understand the things said and done in interviews though. It’s a metaphor which has found its way across the arts and social sciences with a variety of uses, though all linked – from the conventional notion of actors on a stage with a script, to the regulative performances of teachers when Ofsted are in town, to the (sub, semi, un?) conscious performance of different identities shaped by different social situations. With interviews, we’re looking at elements of all three of these.

One of my current projects is an interview based study exploring the knowledge and identities of managers in the NHS. Not the most popular group of people in the world, which is one of the things that makes them interesting to conduct research with.

My most recent interview was with a woman in her 40s, for the sake of a handle we’ll call her P, who had climbed the ranks of her chosen occupation very successfully, was very friendly, confident, candid, and generous with her time. Sounds like the perfect interview right? Well, it was a very interesting interview, however, I have never felt so policed by an interviewee before. P spoke about the things she wanted to speak about. If a question interested her – or she could move it on to territory that did interest her then she gave some great responses. If the question didn’t interest her or moved outside her comfort/interest zone she simply blanked them, moved the conversation on to something else, or on more than one occasion treated them not as questions but as statements – she just nodded, smiled, maybe just turned up the corner of her mouth or shifted her eyebrow slightly, and I was forced to move on. Given her seniority in the organisation, I am not at all surprised that she is able manipulate a situation to her advantage – I probably gave her very little challenge in this regard.

The tricky question is how to make sense of all this context when looking at the interview later as a piece of data, to be compared, contrasted and categorised with 90 others.

My interview style tends towards allowing the interviewee to follow their own nose as much as possible. Obviously sometimes you need to try and pull things back to your own research interests, otherwise you can be left with a very interesting but completely unwieldy set of data which won’t lend itself to any kind of meaningful system of analysis. There is a sense in which this interview drifted much further from the schedule than I should have allowed. However, I think the responses were heartfelt and honest – which can be two of the greatest unknowns in the interview situation. I found this viewpoint strengthened when P commented, after I had switched the tape recorder off, that she was glad that this would all be anonymised, because she had felt comfortable speaking honestly and openly. If I had stuck more rigidly to the schedule I think I would have got a series of fairly brief, disengaged responses, and actually I think she would have become bored and made moves to terminate the interview earlier. As it was I heard a storm of opinions about P’s everyday work, the decisions she had made in getting to the position she was in, the styles of management and leadership she found inspiring and the things she tried to do to be a role model to others, there was also some very critical opinions about the organisation she was working in and the system of healthcare as a whole. And despite the fact that, like almost everyone I have interviewed for this project, she complained about the lack of hours in the day, she gave me almost 2 hours of her time and seemed not to tire or resent this at any point.

All this stuff was going round my head on my train back home after the interview. I had felt utterly stunned on coming out of the interview, it had been an assault on the senses. In some ways it had been a challenge to my own sense of authority as the interviewer – a role which implies, according to conventional accounts of research, that the interviewer does the steering, and takes control. While I think this conventional account is fundamentally flawed, I think there are still expectations hanging in this kind of situation, that an interviewer is going to come in and ask questions and the interviewee is going to answer them. This can make interviewees very nervous – props like a tape recorder often don’t help, they are all part of the performance, which comes with a set of inherited meanings – being on stage, being watched, judged, making mistakes, forgetting your lines. Sometimes you can work and work at getting what you think might be an unconsidered response – not in the sense that you want someone to be careless, but in the sense that you want someone to not over construct their performance, systematically censoring out anything they don’t want seen. None of this was the case with P. She said exactly what she wanted, sometimes considered, sometimes completely off the cuff, sometimes she would even ironically apologise for an opinion – it all seemed to suggest that she understood the rules, and was happily playing with them in whatever way she wanted.

So, I’m thinking maybe becoming the audience in this particular performance might not have been a bad thing. But every different stage seems to demand a different performance, and still i’m plagued by the eternal, cliched, questions.

What does it all mean?