I was watching a long interview with Kazuo Ishiguro the other night – as you do if you live in the UK and still have some good (high) cultural television – and something he said leapt out at me. I can’t vouch for these being the exact words he uttered but they were something like …

“We are all unreliable narrators of our lives, particularly to ourselves. Whether it is in conversation, Facebook or a letter, we all need interpretation.”

Yes! Yes! I thought to myself. If only some thesis writers understood this. So I rushed to write Ishiguro’s words down before I lost them. But even if I don’t have the words down entirely accurately, I do think I’ve got the point he was making. And getting at what you think is the meaning, although not necessarily with 100% accuracy is not unrelated to the reason why thesis writers need to think carefully before they claim to have written a text in which ‘participants speak for themselves’.

So why did I think Ishiguro’s words were so pertinent to researchers?

Well, one of the big problems examiners often see in the thesis is the over-use of participants’ words. This usually takes the form of selections from interview transcripts presented in huge indented blocks on the page. Or worse still, over several pages. Sometimes it’s a personal narrative the researcher has constructed from a summary of what someone has said. The story stands alone. like a tree on a hill. There is little by way of introduction or commentary. Or it is someone’s life story, a précis of what the researcher was told by a project participant. The story is presented as if it were ‘truth’.

When a researcher does this kind of data dumping it’s usually because they either want to provide ‘evidence’ of the veracity of their claims, or more commonly that ‘they want to let the person tell their own story’. The first of these ambitions is laudable, and the second somewhat problematic.

Let’s start with the first goal – producing ‘evidence’. Producing ‘evidence’ through the use of people’s words is important if you are doing qualitative research. Let’s say a researcher wants to establish that there is something common and important in what her respondents have said. This can be achieved in a number of ways. The researcher needs to establish that there is a typical view – this requires her to do something more than reproducing a single quotation. A single quotation does nothing to signify commonality. What is needed is a way of communicating all of the particular observations/comments. One option is to present a representative quotation that is backed up… Appendices are useful places to present summaries of themes together with a sample transcript that shows how the themes were developed. Another option might be to put summaries of themes into the actual text, explain and show why a particular piece of data has been selected as being ‘typical’ or ‘atypical’. These are not the only ways to achieve the goal of using data to evidence a claim – the trick is to do more than simply dump the data.

The second aspiration – to let people tell stories in their own words – is pretty problematic. Let’s take the example of a researcher working with an interview or an oral history. The transcribed interview is something produced in a particular location, in response to particular questions posed by the researcher, at a particular time. These then are already ‘produced’ words, they haven’t emerged from everyday life, but a research context. But then these words are converted from an embodied conversation into text, via further selecting and editing. Changing spoken words to text loses gestural and visual information, just for starters. A transcript is subject to a lot of manipulation before the researcher even starts to think about analysis, let alone choosing quotations for their text. There is nothing pure about this at all.

But this doesn’t mean that there aren’t very good reasons for including research participants’ words, despite all of the processing. The researcher might want to show something of the person’s mode of explanation, their choice of words and examples, the representation they made at the time of an event or of an opinion. Indeed the research itself might be about these very things. That’s good and fine – just let’s not imagine that this use is somehow unfiltered and unadulterated – an ‘authentic voice’.

So how to reconcile these two goals – show some ‘evidence’ of what people said, and at the same time deal with the ways in which the sayings were ‘produced’?

Well, some research traditions do require that researchers make explicit how their data was produced, selected and then presented in a text. Ethnographers, narrative researchers, researchers working with portraiture who aim to offer more literary forms of research texts usually address their ‘authoring’ process in methodology sections. However, this is not always the case with some other qualitative research – interview based work for instance. The question of how the material was selected and presented in the text is often glossed over as if it is self evident – and this presents a difficulty. The data is dealt with as if it is ‘authentic’, a clear window on participants’ reality.

However, dealing with the manufactured nature of data requires even more than acknowledging the role of the researcher in participants’ texts. It needs more than a discussion in the methodology section. Something also has to happen with the way in which the data itself is presented in the actual thesis text.

Let’s go back to Ishiguro and his remark that everyone is an unreliable narrator of their own lives.

Researchers too need to be concerned with the unreliability of research participants. But the way of researchers is not that of fiction writers. In a novel, a character’s speech is presented as if it is being heard by the reader. This is the fiction writer’s maxim of ‘showing not telling’. However, as Ishiguro’s comment makes clear, the fiction writer’s intention is that the reader will respond to the test. The reader will interpret the words that are spoken. This is what the novel writer wants to happen. The act of imaginative interpretation is the way in which a reader makes sense of fiction – they are invited by the writer to imagine the words being said, to not only hear them but also to respond to them perhaps empathetically, perhaps critically.

However, the researcher’s job is not only to show but also to tell. The researcher is the first interpreter of the material they have generated. That very unreliability is what makes it imperative for researchers to take what they are told as something other than a ‘fact’. The researcher must interpret. Why did this person say this? How is this framed by their life circumstances, being who they are, where they are, and when they are? What might the person not see about their situation? What might they not want to say to the researcher? And they have to present that interpretation to the thesis reader. They can’t just dump the data there for the thesis reader to interpret for themselves.

Qualitative researchers need to do more than simply put other people’s words on a page verbatim.(It’s the same principle as working with literatures and avoiding the quote dump – the researcher needs to tell the reader what use they have made of quotations.) Researchers have to say how they have understood the unreliable narratives, the ones they have manufactured with their participants. Researchers can’t rely on thesis readers to do their interpretive work for them. With the exception of some forms of arts-based research, in which artistic forms of interpretation are offered, researchers working with qualitative material do well to offer their thesis reader the interpretation they have made of their material together with the material itself. They thus make their interpretation available for scrutiny and interrogation. The reader can see the selected data, and they can see how it has been analysed and to what end.

Making the analysis explicit, and avoiding the claim to authenticity with the associated data dump, is particularly important in doctoral research, where the examiner is looking to see whether the researcher is capable of critical interrogation of data. The examiner wants to see that the researcher can conduct a piece of research thoroughly and with integrity. Researcher integrity includes the way in which data has been handled from production to its final form in the text.

And this is much, much more than snipping out bits of transcripts and stringing the words of unreliable narrators together.