Cul-de-sacs and Narrative Data Analysis – A Less Than Straightforward Journey

The Qualitative Report, Aug 2018

This article focuses on the methodological journey I took as a novice narrative inquirer, particularly regarding data analysis, for my doctoral data; a journey characterised by floundering, meandering, wrong turns and cul-de-sacs. It explains the initially overwhelming process of moving from collecting “data” to constructing the narratives of five postgraduate international students, challenges faced as well as lessons learned. Despite its complexities, narrative data analysis enables colour and emotion to be added to research. This article continues to add to a somewhat meagre research literature about how to move from collecting “data” to constructing narratives.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

Cul-de-sacs and Narrative Data Analysis – A Less Than Straightforward Journey

James, G. ( ul-de-sacs and Narrative D ata A nalysis - A Less Qualitative Report Gwyneth James Comparative Historical Methodologies Commons Follow this and additional works at; http; //nsuworks; nova; edu/tqr - C Creative Commons License This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 4.0 License. Acknowledgements The author thanks two previous reviewers for their helpful feedback on this article. This how to article is available in The Qualitative Report: Cul-de-sacs and Narrative Data Analysis – A Less Than Straightforward Journey Gwyneth James University of Hertfordshire, United Kingdom This article focuses on the methodological journey I took as a novice narrative inquirer, particularly regarding data analysis, for my doctoral data; a journey characterised by floundering, meandering, wrong turns and cul-de-sacs. It explains the initially overwhelming process of moving from collecting “data” to constructing the narratives of five postgraduate international students, challenges faced as well as lessons learned. Despite its complexities, narrative data analysis enables colour and emotion to be added to research. This article continues to add to a somewhat meagre research literature about how to move from collecting “data” to constructing narratives. Keywords: Narrative Data Analysis, Narrative Inquiry Introduction As opportunities have arisen in conferences, classes and conversations, I have enthused about the wonders of using narrative inquiry for research. It is an approach which is both “a method of investigation and a mode of representation” (Craig, 2012, p. 91) and which I employed for my doctoral research, where I was aiming to understand in depth the lived experiences of five female postgraduates, all international students, whom I had been teaching over a period of one academic year. Narrative inquiry has been described as a “fluid form of research” (Craig, 2012, p. 91) as it unfolds in response to what is encountered during the research process rather than “being driven by predetermined research principles” (ibid.). So, it is individuals’ experiences which are foregrounded, not at the expense of theory, as some conversations I have had have intimated, but instead recognising that “people and their stories do not always ‘fit’ the theory” (Trahar, 2013, p. xiv) . My research was influenced by constructionist and poststructuralist philosophers, psychologists, sociologists and narrative inquirers (Clandinin & Connelly, 2000; Lawler, 2002; Polkinghorne, 1995; Riessman, 2008) with the philosophical basis of narrative inquiry for me underpinned by a Deweyan view of experience. In this view experience is seen as relational, continuous and social (see Clandinin, 2013, for a more in-depth discussion). Individuals cannot be understood in isolation as they are, arguably, always in relationships. Dewey’s notion of continuity is one which particularly resonates with me (i.e., that all our experiences develop from and lead to further experiences). In that vein, neither my research participants nor myself are “blank sheets” (Sawir, 2005, p. 270) ; they arrived in the UK from various countries in Latin America with a past, with considerable educational experiences and with professional experiences and their lives are now continuing after graduating. In turn, I was born in the UK but was brought up bilingual and according to my mother, only started speaking English when I went to kindergarten at the age of three. I have also had considerable educational experiences here having done my BA and MA in England and Northern Ireland respectively; my doctorate was also done in England. Professionally my experience has encompassed both jobs here in the UK and in Southeast Asia, so as Clandinin and Connelly (2000, p. 64) astutely observe: “their lives do not begin the day we arrive nor do they end as we leave. Their lives continue.” The same is, of course, true of my life. In the course of my exhortation of narrative inquiry, what inevitably surfaces at some point is the question of how I analysed my data, particularly considering narrative inquiry is such a different form of qualitative research (e.g., Riessman, 2010) . Even now, post doctorate where I researched and wrote a narrative inquiry (see James 2014) I find that this is not a question to which a simple answer can be given. My own frustration with exactly how to analyse my data is still prevalent because of the diversity and complexity of narrative inquiry and of this approach. But encouragingly I have discovered that I am not alone. In discussion with other narrative inquirers it has become clear that more detailed accounts are needed of how data were narratively analysed in published research using narrative inquiry, as it currently seems to be severely lacking. In 2004 McCormack wrote of her paper “… [it] fills a gap in a research literature that is largely silent about what to do after researchers have transcribed their interview conversations” (p. 319) and today, there is still not much more that has filled that gap. In this article, therefore, I try and respond to the seemingly straightforward question of how I analysed my data by describing both the cul-de-sacs I ended up in as well as the steps and stages I went through in the hope of presenting a process which other researchers may wish to ponder on as they answer the frequently overwhelming post data collection question of “so, what next?” My own process of analysis is not reproduced here as a blueprint to be strictly adhered to; rather it is intended to give ideas and perhaps function as a springboard for other research which requires a narrative analysis of data. The way in which I have written it is also an attempt to show that although my journey was less than straightforward this did not mean I was doing anything “wrong”; perhaps it was just that I needed to discover for myself the twists and turns and messiness that seem to accompany (novice) narrative inquirers on their journey. The First Stage: Collecting Field Texts Clandinin and Connelly (e.g., 2000, p. 82) refer to data as “field texts” and I will use the two terms interchangeably throughout. My field texts were collected over an academic year and comprised three research conversations with each student averaging an hour each, emails which we exchanged throughout the year, conversations that “just happen” (Trahar, 2006, p. 122) – with both my participants and with colleagues when discussing narrative inquiry – as well as my extensive readings around what narrative inquiry is and “how to do it” and reflections on those readings. Once I had this vast amount of rich information, how was I to go about constructing and co-constructing (the latter with the participants’ input) and analysing it as narratives? Were the construction and analysis separate stages, even? How on earth was I going to make the move from raw data to co-constructed narrative? What would this process look like? Where could I go for help? What should I add? What should I omit? In previous research I had done for my MA I had followed a highly structured, linear (step-by-step) approach to analysing my applied linguistics interview and focus group data. But for this doctoral research I was looking for actual narrative examples, narrative frameworks which would guide me safely through what I was beginning to see was a minefield of alternatives. I struggled to locate these clear accounts of “how to do narrative analysis” in order to approach this daunting task and initially what I found was others who empathised with me in the maze of approaches to take (e.g., Squire et al., 2008) . This is not what I had been expecting and so my journey through the maze continued. At some point of course I had to take a deep breath and simply take the plunge, stop getting tied up in knots and going round in circles and approach this huge task squarely in the eye. I write about my insecurities here in this way because I do not wish to give the impression that the narratives were constructed with ease. A comment made to me by someone who had both read them and heard me talk about how I had written them was that I had portrayed them like the image of “a swan on a lake”: what is seen in the written narratives (i.e., above the surface) is serene, calm and ordered but what was heard/inferred as I spoke about writing them exhibited furious, frenetic and frustrating activity below that surface. I had inadvertently portrayed them as having been effortlessly written, that is, that their construction and analysis was simple and straightforward as I tried to find the most appropriate way to share with my readers the depth and richness of the data I had been given. I was disappointed that I had not made this clear, as it was unintentional on my part (or perhaps I hadn’t wished to show the messiness of the process in a formal piece of writing?) Hence what follows is a response to try and rectify that by being (more) transparent in the struggle I had. Before I continue outlining my somewhat tortuous journey of discovery in this regard, let me briefly explain how I understand narrative data analysis, again doing so to illustrate the complexities and challenges I came across during this circuitous route of moving from raw data to constructed narratives and how analysis fitted in. A Slight Detour: What is Narrative Data Analysis? Kim (2016, p. 190) suggests that along with interpretation, narrative data analysis is “an act of finding narrative meaning” (emphasis in original). In this sense, she defines narrative meaning as “a meaning-finding act through which we attempt to elicit implications for a better understanding of human existence” (ibid.) So narrative data analysis is fundamentally about searching for the meaning(s) and making sense of an experience or experiences which we – and in a narrative inquiry that includes the researcher – have had. And yet, as Kim (ibid.) goes on to say, “meaning is not tangible, nor static, thus it is not easily grasped” and neither do we have direct access to the meaning which someone else gives to their experience. This of course leads to various issues vital for consideration in a narrative inquiry, such as those of narrative interviewing, re-presentation, trustworthiness, time and memory (for further discussion of these see James, 2014) . Polkinghorne’s (1995) distinction between two types of narrative inquiry has been really useful to me and many others in beginning to tease out the different approaches and understandings of analysing data narratively. He talks about “analysis of narratives” and “narrative analysis”. As a crude definition, the former is research which uses stories as data and the latter is the use of storytelling to analyse data and present findings. So narrative analysis here is “ the stories are produced and what we can learn from them” (Barkhuizen et al., 2014, p. 3, emphasis added) . The analysis of narratives relies on one of Bruner’s (1986 ) two ways of knowing (i.e., paradigmatic, with the other way being narrative). Simply put, paradigmatic knowing is “more recent and associated with development of rational thinking” (Barkhuizen et al., 2014, p. 1) , whereas narrative is “both older and more deeply rooted in everyday thinking” (ibid.). So paradigmatic thinking prefers certainty, whereas narrative accepts ambiguity (Peters, 2005) , which goes back to the point made earlier about it being a fluid form of research. It is arguably the latter type of knowing which can help us navigate our way through life and understand how ambiguous and complex our lives are, and this can only ever be partial, situated and temporal. The analysis of narratives, then, attends to the general, looking for common themes across stories. Its goal therefore is to identify “general themes and patterns... [thereby minimising] ambiguity” (Kim, 2016, p. 196) . In contrast, narrative analysis as understood here is attending to the particular and is done by looking at the data as a coherent whole. The narratives are created and co-constructed by integrating certain events into a temporally organised whole with a thematic thread (plot). Note my use of the word “certain” which itself constitutes an analysis or interpretation of sorts and which I will come back to. The purpose of this way of analysing data is “to help the reader understand why and how things happened in the way they did, and why and how our participants acted in the way they did” (Kim, 2016, p. 197) . Barone (2007, p. 456) calls this a “narrative construction”; it is a “recasting of data into a storied form [which] is more accurately described as an act of textual arrangement than of analysis” (ibid.). While I agree in part, this construction was for me also an in-depth analysis of the data as I had to interact with it on a deeper level to begin the meaning-making process for each participant. The collection of narrative data is inextricably linked with its analysis; it was a less than linear process as I was constantly discovering key related issues (e.g., that of time), and I did not wait until I had all my data before starting analysis. It was more an iterative process, one which is “interpretive at every stage” (Josselson, 2006, p. 4) . In creating a narrative construction, then, I am also necessarily analysing by making the decision as to what to include as I cannot possibly include everything. But an end point has to be reached amidst all the intricacies and complexities and so let me now return to my journey. I had my field texts so how did I now construct the narratives? The Second Stage: From Raw Data (Field Texts) to Narrative Construction – Or: “Forays into Narrative Analysis” In what follows, I explain my initial forays into narrative analysis and then focus on how I decided to create the five narratives in my doctoral research to then “analyse” them in the more traditional sense of extrapolating and discussing certain issues emerging from the five narratives with what I found in the literature. It is worth bearing in mind that in using the term narrative analysis below I am intertwining both the construction and the analysis of the narratives, recognising that while I was constructing I was analysing and vice versa. Foray #1 My very first attempt at a narrative analysis was an assignment for the taught part of my doctorate written the year before I started collecting data for my actual doctoral research. It was an exploratory attempt at doing a narrative analysis of an extract from my Frenchspeaking postgraduate student’s (Lucie, a pseudonym) personal oral narrative, dealing with her transition from leaving a prestigious professional career in Canada to coming to the UK to undertake a postgraduate degree and her reasons for doing so. My analysis was based on Riessman (1993 , 2008) as a model for structural analysis of the extract. This type of analysis is fundamentally looking at how narratives are organised and while it is interested in content, form is paramount (Riessman, 2008) . Structural analysis is based on a sociolinguistic approach to oral narratives using clauses and was originally developed by Labov and Waletzky (1967) , further developed by Labov in 1972 , where each clause is given a function depending on which elements of a narrative it is portraying (see Table 1). Summary of the story Time, place, characters Sequence of events Narrator comments on meaning and communicates emotions Outcome Ending the story and returning it to the present This appealed to my applied linguistics background and former research in sociolinguistics. To begin the analysis, I followed Riessman’s (2008, p. 89) exemplar by asking: (1) how is this story put together? and (2) how are the structural elements arranged? I listened many times to the 10-minute extract of Lucie answering my question, “Can you tell me a bit more about your transitions in any particular area or every area of your life?” and following Riessman’s (1989 ) example, I divided the narrative into clauses and then further subdivided it into four shorter transcripts, primarily when Lucie used a conjunction (e.g., and, so) to return to the main point of her narrative. It was an interesting approach to take in analysing part of an oral narrative, but felt rather clinical to me and not particularly creative. Considering I like structure and “how to…” guidelines and frameworks, my reaction surprised me. Part of my feedback on the assignment echoed much of my own thoughts as I reflected on what I had learned from the exercise: “You take on a narrative inquiry approach for the first time and what arises is a valuable reflective piece on some of the problems that beset the application of an ‘off-the-shelf’ model of analysis…” I also realised that employing Labov and Waletzky’s model for oral story was not transferable to the interview discourse which I had conducted with Lucie, ending up with my applying the model beyond its original purpose. Such a closed, structural rendition of a two-way interview based on Conversation Analysis (which I had used to great effect in my MA) did not here sufficiently capture the meaning making I was interested in. Foray #2 What follows is based on what I wrote in James (2014) . My second foray was a year later when I practised analysing the data from my pilot study (January to March 2012). For this study I had interviewed six of my students, all Latin American female postgraduates. Additionally they had sent weekly or fortnightly emails describing in varying degrees how they were progressing both academically and nonacademically. I also ran two focus groups but with different students to the aforementioned six. My aim was twofold: firstly to practise both a more narrative style of interviewing, one where I “follow participants down their trails” (Riessman, 2008, p. 24, emphasis in original) , and secondly I wanted to practise analysing what they told me, rather than conduct a narrative inquiry per se because I simply did not have sufficient time at my disposal for this very relational methodology (Clandinin, 2013) . I began by transcribing five of the six interviews myself. Briefly, and not wishing to gloss over the act of transcription as it is not the focus here, I do not see transcripts as “providing ‘objective’ accounts of recorded data” (Davidson, 2009, p. 41) but rather as an act of re-presentation, an “aid” to help in my analysis. For the first three participants I inserted my own thoughts as speech bubbles into the transcript when what I read and heard resonated with me, in effect asking questions “after the event.” The emails were not explicitly used during my reading but only as extra information was required. By the fourth participant’s analysis I was becoming a little jaded with the format of inserting comments so I experimented with something different. Based on Alan Bennett’s Talking Heads I decided to construct the data as a three-act monologue, using the participant’s own words from both her interview and her emails. She then read this through but did not suggest any changes to what I had written. In fact, she kindly agreed to record it for me to use in a workshop I was doing that summer on narrative inquiry. Although I enjoyed writing the monologue I had to reduce it considerably in length to a 6-minute monologue, so for the fifth participant I went, unintentionally, to the other extreme and engaged in what became a thorough but very lengthy analysis using Gilligan et al.’s (2006 ) The Listening Guide (LG). In this guide, interview transcripts are listened to four times (and in my case I also read the emails) so that the researcher can focus on distinct aspects of the participant’s experience. I also needed to be “present” in the listening, reflecting the key notion in narrative inquiry of reflexivity. In this analysis the focus is on listening rather than reading, although I did both as I adapted the method to suit my research purposes (see also Mauthner & Doucet, 1998; Woodcock, 2010) . It seemed to “tick all the boxes” and I did find the LG a fascinating way to analyse the data because not only did it help me make sense of the complexity of what was being said on the surface but it also gave me the security of a concrete set of guidelines. So why did I not employ this framework for my actual doctoral study? Ironically, although I could adapt the LG to suit my research purposes, I felt it was too rigid with too many steps and stages for my research context. Coming to the end of that one analysis I realised I had written 12,000 words, which meant that practicality had become an issue. I needed something more manageable for myself and my readers. By the time I came to the sixth participant I had run out of steam and out of time, which left me worrying about the ethics of having collected data but not using it. After all, data collected need to be adequate but not excessive in order to protect the participants and to ensure that principles of good practice, developed from the Data Protection Act, are adhered to (e.g., Bond, 2009; Wood, 2005) . A positive consequence of this, however, is that it did help me in choosing only five participants for my actual study. Despite my second attempt at constructing and analysing my data, I was still unsure about how to do it for my actual doctoral study. In analysing part of Lucie’s oral narrative I felt that the creativity was lacking; in analysing the pilot data I felt that something was missing; I did not feel that I was really “getting” it (i.e., how to do the analysis). There seemed to be more rather than fewer gaps in my “how to do narrative analysis...: knowledge. And I had even more data to contend with than on both previous occasions. The Third Stage: Decision Time Admittedly by this point I was becoming increasingly frustrated with the lack of guidelines and concrete examples available. I had found only a handful of clear models in the literature (e.g., Blumenreich, 2004) but they were not entirely suitable for my own research context. The gap in my knowledge of “how to analyse” still remained. I kept, if patchily, a research journal throughout my five years of part-time study, and related to this gap I’d written: “this is all quite overwhelming” and “how am I going to analyse all this?” If I sound like a stuck record in repeating this point that is intentional as it aptly illustrates my frustration that this was not a straightforward process of data analysis; remember the image of the swan? Although I felt like I was returning to the drawing board I started searching for anything that had been written and published on how to analyse narrative data, even at one point simply googling “narrative analysis”. This just led me down more maze-like paths, so I decided to start with what I had read right at the beginning, namely revisiting Riessman’s (2008) four forms of analysis. I began by combining two of those: thematic and dialogic/performance. The former focuses on the content (i.e., what is said) while the latter focuses on storytelling as an activity and is interested in the “who, when and why” (Riessman, 2008, p. 105) . Dialogic/performance analysis looks specifically at the dialogue between speakers, how stories are performed within both their given contexts and their broader historical and cultural contexts. Stories do not “fall from the sky...they are composed and received within contexts – interactional, historical, institutional, and discursive...[and so are] social artifacts, telling us as much about society and culture as they do about a person or group” (Riessman, 2008, p. 105) . During the telling different roles are taken on by participants which leads to shifts in positioning of the “performer.” Hence, storytelling includes what occurred both before and after the speaker speaks (Bamberg, 2006) . Dialogic/performance analysis also allows the researcher’s own role and voice to be heard, so becoming “an active [and reflexive] presence in the text” (Riessman, 2008, p. 105) . While this gave me a clear framework it still did not give me my much sought after and increasingly elusive “how to analyse” guidelines. Since then I have read Kim (2016) who clearly warns against giving “how to” methods, because these methods would end up being pigeonholed, “causing you to search for where and how you can fit your data analysis into one particular method” (p. 195). Instead, she advocates “exploiting the idea of surprise and curiosity, creating a space where aims can be worked out, allowing room for less-familiar possibilities and playing with new ideas” (ibid.). Of course this now makes sense, but at the time I felt I was going one step forward and then two steps back. Eventually, and with deadlines looming, I discovered Hunter’s (2010) article which focused on those challenges she faced in narratively analysing her PhD data. What was particularly “comforting” as well as helpful was her detail in outlining what she did to prepare and then write up her data for analysis. At last – some answers. And yet as I started along her lines I quickly realised, as Kim (2016) cautions, that I couldn’t fit my data into her method, so I adapted the latter. Preparing the Ground In total over the academic year I had had three conversations with each of the five participants at the end of each term. Initially I decided to transcribe the first set of conversations myself and although I did, I found it a complex process because of my previous, extensive experience transcribing using Conversation Analysis. How much detail should I put into the transcriptions? After all, as a narrative inquiry and not a sociolinguistic approach did I need to transcribe the “how” as well as the “what” which is so integral to a CA approach? What about fillers (e.g., hmm...) and pause lengths, even approximate ones? Should laughter be added? What about tone and intonation? Should stress on certain words and phrases be highlighted? Transcription is both labour intensive and time consuming, and although I know it helps me familiarise myself with the data, although I know it is necessarily an interpretive practice (Mishler, 1991; Green et al., 1997) , I not only wanted to avoid “reifying” transcripts (Riessman, 2008, p. 26) but was also teaching full-time so I had the remaining two sets of conversations transcribed professionally which released me from many of these questions. Admittedly I did not brief the transcriber but left any decisions to her as a professional academic transcriber. Prior to the first conversation I had also emailed each participant a “background questionnaire” with biographical questions amongst others which I could use in my first interview (e.g., “why did you come to the UK to study?” and “thinking of your first term, how would you describe your experiences using adjectives?”) and this information was included in my notes on my first listening. To illustrate the remaining steps I have used extracts from field texts in constructing Lucile’s narrative (for Lucile’s whole1 narrative please see James, 2014) . Step 1: First Listening My construction/analysis process was iterative in that I did not wait until the year was over before beginning it but rather started after each conversation had been transcribed. Additionally I was working full-time and studying part-time and needed therefore to utilise my time judiciously, the first time listening to the transcription and making notes on what struck me in the content. As well as writing my own questions and responses in note form to what was said (which were then emailed to the participant or used as the starting point for the next 1 Although this sounds like a contradiction in terms as narratives are never complete, I am using the term “whole” here simply to mean the narrative in its entirety from the academic year. conversation) I also attempted to use the participant’s own language to “code” the data into themes and sub-themes and noted “quotable quotes” (Hunter, 2010, p. 50) . This was not coding as per the more traditional understanding but rather an attempt to impose some sense of order on the data collected. I did not use the same “codes” for each participant as they are simply what emerged organically from each. The result of Step 1 was a “spidergram” of notes (see blue paper in Figure 1). Step 2: Second Listening An example to illustrate this stage is taken from Lucile’s first conversation with me in December 2012, which lasted 40 minutes. While listening this time, 2-3 months later, I read my notes from the first listening (on the blue paper – see Figure 1) and added to those, this time ending up with a handwritten set of notes for each participant (see Figure 1). These included the notes from the background questionnaire responses (on the left) and notes from the emails sent in the second term (top right) for me to try and get a more holistic picture of the participant before the second conversation. I also included follow-up questions for the next conversation on post-it notes. The result of Step 2 was a large set of handwritten notes (see Figure 1). Step 3: Third Listening My third listening of Lucile’s first conversation came a few months later still. Below is an extract from my notes, made on the computer this time. These six pages of notes were now focused along “thematic” lines and for the sake of some sense of organisation were structured around the questions I had asked during our conversation. This extract is based on what she said after my question which asked her to elaborate on the adjectives she chose to describe her first term: Lucile’s first term was really positive, which was not what I expected to hear. She describes it as: incredible outstanding amazing priceless demanding great Why wasn’t I expecting to hear this? I guess I just assumed, based on what I’d heard generally from students in previous years and some of the literature, that the first term in particular would always be hard. But I am starting at the beginning of my time getting to know her, and of course her past goes back way before London... enriching inspiring happy At the end of the notes I wrote: Lucile’s key phrases (verbatim): Everything is like a great experience; I’m really enjoying everything I like to attend the lectures; I really enjoy coming and discussing and questioning things I’m really exhausted / A big change / I have incredible role models The result of Step 3 was the beginning of emergent themes. Step 4: Summary Once I had done the above for all three conversations I wrote a lengthy summary for each interview from my notes, referring back to the transcript when necessary. I then wrote a summary of all three interviews, using the key themes which had emerged from the above listenings to guide me. In Lucile’s case the summary focused around three areas: her home country (i.e., family, education, role models); the UK university she was studying at (i.e., reason for choosing the UK, a memorable experience, essays, grades); outside the university (i.e., living in London, English culture, relaxing/free time, summing up her year) and finally her home country again (i.e., future career and her forthcoming marriage). I followed these four steps for each participant which explains why there were gaps between listenings. I tried to listen to each recording three times, but what I am compelled to admit here is that I did not adhere rigidly to this process; at times I only listened twice, choosing instead to read the transcript a third time. I realise that writing a summary could be construed as being reductionist, perhaps also untrue to narrative inquiry’s ethos as well as the complexity of who we are as human beings but this was not my intention; rather it was an attempt not only to make the data more manageable but to try and create a coherent and holistic picture, at each stage, of each participant. Throughout, I read and listened to the data, understanding all the while that “analysis and writing up are interwoven processes” (Hunter, 2010, p. 50) . Step 5: From Summary to Narrative Once I had the summaries I could then construct a rough draft of each narrative in a more discursive and arguably more creative form than the summaries, namely the “narrative construction” I mentioned earlier. Here I needed to remember the three “commonplaces [of] temporality, sociality, and place” (Clandinin, Pushor, & Or, 2007, p. 23) . Temporality means looking towards the past, present and future of both Lucile and the events she experienced. In her case, I chose to present her narrative as a blog and, while the dates are fictional, the events/experiences portrayed are chronological in terms of following her academic year. However, this linearity was not apparent during the data collection, partly because Lucile needed a considerable amount of guiding to elicit lengthier responses, and partly because we flitted back and forth across certain topics during each interview and the three interviews as a whole and so they were at times repetitive (albeit from a varying perspective as time influenced the re-telling of certain experiences). Sociality refers to the social environment in which Lucile’s experiences were occurring: “these social conditions are understood, in part, in terms of cultural, social, institutional, familial and linguistic narratives” (Clandinin, 2013, p. 40) . They emphasise the fact that all her (and indeed our) experiences took place in specific contexts. Inherent in attending to sociality is the need to attend to inward, personal conditions, namely Lucile’s emotions and moral responses to her experiences. Sociality also encompasses the relationship between the researcher and the participant, most clearly exhibited in Lucile’s narrative as I was commenting directly on what she had said in her blog. The last commonplace, that of place, is attending to the physical space and boundaries within which Lucile’s experiences occur, whether that was at the university, in her accommodation, in London or at home in Uruguay. At this step I also included data from their background questionnaires along with various email exchanges sent throughout the year. In constructing the narratives I was aiming for something which would “glow with life” (Ely, 2007, p. 569) and in the spirit of a narrative inquiry I wanted to compel the reader to an emotional response in which “alternative readings and multiple interpretations” (Ellis & Bochner, 2000, p. 745) were possible. In the final narratives (see James, 2014) I therefore offer one possible interpretation, but in fact personal narratives long to be used rather than analysed; to be told and retold rather than theorised and settled; to offer lessons for further conversation rather than undebatable conclusions; and to substitute the companionship of intimate detail for the loneliness of abstracted facts. (Ellis & Bochner, 2000, p. 744) Although “narrative [does not reflect] ‘the’ reality, with the help of the reader, narrative creates a version of reality” (Ely, 2007, p. 571) . I presented the narratives through the format of a student magazine, inspired by Ely (2007, p. 569) who wrote that “people must want to read what we wrote, must want to stay”. She also emphasised that researchers have a responsibility to honour the stories of those who provided us with the information and I wanted to do that. So, what would make me stay? I thought of a student magazine because my participants were students but also because I could be creative in the way in which I re-presented each narrative (e.g., a blog in Lucile’s case). At the beginning of each narrative I briefly explained my rationale for why I chose the format I did: Lucile’s was the first narrative I constructed, and I decided on a blog as I wanted to be able to insert my own thoughts and comments throughout, some of which were spoken aloud at the time and some of which were added retrospectively while constructing her narrative. Initially I had planned to incorporate the literature in the narratives, and the “comment” allowed me to do that. After writing the first draft, however, I decided to allow the narrative to speak for itself and include the literature in a later chapter. The blog was structured chronologically according to Lucile’s academic year although the dates do not accurately correspond to the events described. The other formats I used were a newspaper article, a series of emails home, a series of personal journal entries and a semi-fictional story (see James, 2014, as space does not permit me to further exemplify this here) . As far as possible, I used my participants’ words verbatim because I wanted them to be situated at the centre of my research, not me (Bold, 2012) . So I became “off centre,” not removed entirely. Four of the five participants kindly read through their narratives and emailed me brief reflections. Lucile for example emailed: I loved to read what you wrote. I think that you understood what I was experiencing... Thank you for sending me the document, it was great to read my feelings and to remember my experience told by myself in the past (email communication, 26 May 2014). I liked the creativity and variety that these presentational forms afforded, enabling me to communicate “different meanings and emphases via their different conventions. They help us to tell the stories of our research. A shower of possibilities” (Ely, 2007, p. 572.) . Briefly, the final part of this process was the more traditionally understood “analysis and discussion,” where I used the narratives to locate themes that emerged both within and across the five narratives. This was not an attempt at reductionism or seeking to draw generalisations; rather it was an attempt to see what the narratives highlighted both individually and corporately. I had decided, somewhat unconventionally but to me in keeping with the creativity which narrative inquiry affords, to combine the more traditional “literature review and discussion” section regarding these key themes and present them after the narratives as “Letters to the Editor”, in keeping with the format of the student magazine which I had used to portray the five narratives. I wanted to use the format of “letters” to give me the freedom to write each (of five) letters from a differing viewpoint, and I wrote: Letters to the Editor are those which praise, clarify and refute (among other things) what has already been written elsewhere. In writing these letters, I am incorporating what the literature says regarding issues of student transition as well as reflecting on this. The five letters highlight key themes/issues emerging primarily from the narratives, some of which converge and some of which diverge from what already exists in the literature… I have written these five “letters” from a variety of perspectives so that I can use a different voice and therefore style in each; in doing so I hope to foreground matters which vary in their significance (Ely, 2007) according to each writer. (James, 2014, p. 132) The first letter, for example, was entitled Educational transitions – an overview and was written by A Researcher. This allowed me to reflect the style of a more traditional and impersonal literature review (see James, 2014, for the entire chapter). Lessons Learned from Each “Foray” In this paper I have attempted to describe, partly in the form of my own personal narrative, the struggles and joys, twists and turns of analysing my doctoral data narratively. This has included specific practical steps I took along the way, whether these were fruitful and led somewhere or whether they led me down a cul-de-sac, causing me to retrace my steps and start again. The thematic/dialogic analytical framework I eventually decided on, having come almost full circle, seemed to suit my aim in focusing both on the content as well as the act of storytelling. Combining the what, who, when and why more or less mirrored Clandinin and Connelly’s (2000) three commonplaces of temporality, sociality and place. It also allowed me as the researcher to be a reflexive presence. The concrete steps and stages for this framework, however, came from Hunter (2010) , allowing me to adapt her method and “play with new ideas” (Kim, 2016, p. 195) to suit my own research context. In my first foray, structural analysis, while interesting from a linguistic perspective, felt rather clinical and lacked creativity. It was nevertheless a clear lesson in how taking an “offthe-shelf” model of analysis was beset with problems when applied to a unique personal narrative. I also realised that I was using the model for a purpose for which it had never been intended, namely an oral narrative, but one originating from within the confines of an interview rather than in spontaneous conversation. Then in my second foray, for my pilot data collection, I transcribed five of the six interviews and for the first three I inserted speech bubbles representing my own thoughts as I read and listened to the interview. But this quickly became repetitive and risked turning a potentially rich source of data into something humdrum because I was simply going through the motions. Attempting to interject some variety for the fourth interview, therefore, I constructed a three-act monologue using both the interview and emails, which the participant then kindly recorded for me. The downside of this method was having to reduce it considerably to last no longer than six minutes, otherwise I felt I would risk losing the interest of those in the workshop I had planned to play it to. My final method at this stage was to follow Gilligan et al.’s (2006 ) Listening Guide, listening four times to the interview, each time focusing on a different aspect of the experience. Although this was a concrete set of guidelines to follow, the resulting analysis was too long and the practicality of doing this for five participants then became problematic. Summary of My Unique Contribution to the Data Analysis Method My own unique contribution to the method I finally used focused heavily on both listening to and reading the interview and other data (see Appendix 1 for a tabular representation). In some senses I was not methodical in how I approached the analysis of each interview because sometimes I listened and sometimes I read, sometimes both, making notes in the margins of the transcriptions. This was somewhat ironic, as I had started my journey wanting something rigid and methodical and is a clear illustration of how narrative inquirers do not emerge from their research unchanged. My method represents both a mix of other methods I tried and tested as well as a subjective approach to this whole process: I embodied fluidity in not being restricted to one method, using parts of others to suit my own research context. In the process I devised something which both suits me and reminds me of Pick’n’mix sweets, but not because I intentionally set out to do this; rather, I have done what Kim (2016, p. 195) among others advocates, namely "exploiting the idea of surprise and curiosity, creating a space…allowing room…[and] playing with new ideas” instead of forcing my data and its analysis to fit one method, bringing to mind the metaphor of trying to fit a square peg into a round hole. Playing with new ideas was particularly evident in my use of a student magazine to highlight and utilise a variety of discourse genres for the narratives, a refreshingly creative way for me of presenting data and re-presenting the five participants’ experiences. A cautionary note needs to be made here, though, namely that just because I used this method for this research does not mean that I will follow it rigidly by using it in its current form (see Appendix 1) for future research. Each narrative inquiry data analysis must suit its own unique and particular context to itself avoid becoming an “off-the-shelf model.” In this way methods of narrative data analysis are fluid and organic, each of which should perhaps be presented as a process for other researchers to ponder on and can be used as guidelines or as springboards for creativity, not as a blueprint to be replicated exactly. Re-presenting someone else’s – and my own – experience in such a way is not a simple or straightforward process. As Hunter (2010, p. 50) writes, it “needs to be done with respect and humility” and to do justice to my five participants who so generously gave of themselves, narrative inquiry was the most appropriate methodology through which to approach this considerable task. Perhaps the most important lesson I have learned is that there is no one way to analyse data narratively, and to have approached it in such a manner was both naïve and erroneous. This may seem an obvious realisation but it is also part of developing ‘the maturity and experience that are required to be a good narrative researcher’ (Kim, 2016, p. 2) . Kim also writes that it takes years to develop these characteristics and that she is “still working at it” (ibid.), as am I. Diversity characterises narrative inquiry and should not be something we shy away from because of its concomitant complexity; rather it is something to be celebrated. Appendix 1: “my” method Step 1: first listening/reading 2: second listening/reading 3: third listening/reading 4: summary Process/Features  made notes on what struck me in the content/conversation  wrote my own questions and responses in note form to what was said (used as the starting point for the next conversation in March)  used participant’s own language to “code” data into themes and sub-themes  noted “quotable quotes” (Hunter, 2010, p. 50)  Result: “spidergram” type notes of the conversation (on blue paper)  Done 2-3 months later (just before the second research conversation)  Read my notes from the first listening (on blue paper) and added to those + notes from background questionnaire + emails + follow-up questions for the next conversation on post-its  Result: handwritten set of notes (see image)  Again, done 2-3 months later (before the third and final research conversation)  Notes made on the computer this time focused along “thematic” lines; structured around the questions asked during the conversation  Selected key phrases (in Lucile’s own words), for example, Everything is like a great experience; I’m really exhausted…  Result: beginnings of emerging themes  Repeated the above for all three conversations, then wrote a lengthy summary for each interview from my notes, referring back to the transcript when necessary (used the key themes which had emerged from the above listenings/readings to guide me)  Lucile (3 key areas): - her home country (i.e., family, education, role models) - the UK university she was studying at (i.e., reason for choosing the UK, a memorable experience, essays, grades) - outside the university (i.e., living in London, English culture, relaxing/free time, summing up her year) - her home country again (i.e., future career and forthcoming marriage) 5: from summary to narrative (construction) “narrative construction” (Barone, 2007, p. 456) , that is, rough draft of the narrative in a more discursive (and arguably more creative) form than the summaries compelling the reader to an emotional response in which “alternative readings and multiple interpretations” (Ellis & Bochner, 2000, p. 745) are possible student magazine (blog, newspaper article, series of emails home, series of personal journal entries, a semi-fictional story) co-constructed narratives (4 of 5 read through and emailed me back reflections) Author Note Dr. Gwyneth James is a Senior Lecturer in the School of Humanities at the University of Hertfordshire. Her research interests are using narrative inquiry and autoethnography to research student experience, teacher resilience and the relationship between language, identity, culture and music. Correspondence regarding this article can be addressed directly to . The author thanks two previous reviewers for their helpful feedback on this article. Copyright 2017: Gwyneth James and Nova Southeastern University. Bamberg , M. ( 2006 ). Stories: Big or small . Why do we care? Narrative Inquiry , 16 ( 1 ), 139 - 147 . Barkhuizen , G. , Benson , P. , & Chik , A. ( 2014 ) Narrative inquiry in language teaching and learning research . New York, NY: Routledge. Barone , T. ( 2007 ). A return to the gold standard? Questioning the future of narrative construction as educational research . Qualitative Inquiry , 13 ( 4 ), 454 - 470 . Blumenreich , M. ( 2004 ). Avoiding the pitfalls of “conventional” narrative research: Using poststructural theory to guide the creation of narratives of children with HIV . Qualitative Research , 4 ( 1 ), 77 - 90 . Bold , C. ( 2012 ). Using narrative in research . London, UK: Sage. Bond , T. ( October 9 , 2009 ). Values, ethics and educational research . Lecture presented at Bristol University: Bruner , J. ( 1986 ). Actual minds, possible worlds . Cambridge, MA: Harvard University Press. Clandinin , D. J. ( 2013 ). Engaging in narrative inquiry . Abingdon, UK: Routledge. Clandinin , D. J. , & Connelly , F. M. ( 2000 ). Narrative inquiry- Experience and story in qualitative research . San Francisco, CA: Jossey-Bass. Clandinin , D. J. , Pushor , D. , & Or , A. M. ( 2007 ). Navigating sites for narrative inquiry . Journal of Teacher Education , 58 ( 1 ), 21 - 35 . Craig , C. J. ( 2012 ). “Butterfly under a pin”: An emergent teacher image and mandated curriculum reform . The Journal of Educational Research , 105 , 90 - 101 . Davidson , C. ( 2009 ). Transcription: Imperatives for qualitative research . International Journal of Qualitative Methods , 8 ( 2 ), 36 - 52 . Ellis , C. , & Bochner , A. ( 2000 ). Autoethnography, personal narrative, reflexivity . In N. K. Denzin & Y. S. Lincoln (Eds.), Handbook of qualitative research (2nd ed ., pp. 733 - 768 ). Thousand Oaks, CA: Sage. Ely , M. ( 2007 ). In-forming re-presentations . In D. J. Clandinin (Ed.), Handbook of narrative inquiry - Mapping a methodology (pp . 567 - 598 ). Thousand Oaks, CA: Sage. Gilligan , C. , Spencer , R. , Weinberg , M. K. , & Bertsch , T. ( 2006 ). On the listening guide -A voice-centred relational method . In S. N. Hesse-Biber & P. Leavy (Eds.), Emergent methods in social research (pp. 253 - 271 ). Thousand Oaks, CA: Sage. Green , J. , Franquiz , M. , & Dixon , C. ( 1997 ). The myth of the objective transcript: Transcribing as a situated act . TESOL Quarterly , 31 ( 1 ), 172 - 176 . Hunter , S. V. ( 2010 ). Analysing and representing narrative data: The long and winding road . Current Narratives , 1 ( 2 ), 44 - 54 . James , G. ( 2014 ). Being a student again: A narrative inquiry into the transition experiences of Latin American postgraduates at a UK university (Unpublished doctoral dissertation ). University of Bristol, Bristol, United Kingdom. Josselson , R. ( 2006 ). Narrative research and the challenge of accumulating knowledge . Narrative Inquiry , 16 ( 1 ), 3 - 10 . Kim , J. ( 2016 ). Understanding narrative inquiry . Thousand Oaks, CA: Sage. Labov , W. ( 1972 ). Language in the inner city: Studies in the Black English vernacular . Philadelphia, PA: University of Pennsylvania Press. Labov , W. , & Waletzky , J. ( 1967 ). Narrative analysis: Oral versions of personal experience . In J. Helm (Ed.), Essays on the verbal and visual arts (pp. 12 - 44 ). Seattle, WA: University of Washington Press. Lawler , S. ( 2002 ). Narrative in social research . In T. May (Ed.), Qualitative research in action (pp. 242 - 258 ). London, UK: Sage. Mauthner , N. , & Doucet , A. ( 1998 ). Reflections on a voice-centred relational method . In J. Ribbens & R. Edwards (Eds.) , Feminist dilemmas in qualitative research (pp. 119 - 146 ). London, UK: Sage. McCormack , C. ( 2004 ). Storying stories: A narrative approach to in-depth interview conversations . International Journal of Social Research Methodology , 7 ( 3 ), 219 - 236 . Mishler , E. G. ( 1991 ). Representing discourse: The rhetoric of transcription . Journal of Narrative and Life History , 1 ( 4 ), 255 - 280 . Peters , E. ( 2005 ). Ways of knowing . Available from Polkinghorne , D. E. ( 1995 ). Narrative configuration as qualitative analysis . In J. A. Hatch & R. Wisniewski (Eds.) , Life history and narrative (pp. 5 - 25 ). London, UK: Falmer Press. Riessman , C. K. ( 1989 ). Life events, meaning and narrative: The case of infidelity and divorce . Social Science & Medicine , 29 ( 6 ), 743 - 751 . Riessman , C. K. ( 1993 ). Narrative analysis . Newbury Park, CA: Sage. Riessman , C. K. ( 2008 ). Narrative methods for the human sciences . Thousand Oaks , CA: Sage. Riessman , C. K. ( 2010 ). What's different about narrative inquiry? Cases, categories and contexts . In D. Silverman (Ed.), Qualitative research (3rd ed., pp. 310 - 330 ). London, UK: Sage. Sawir , E. ( 2005 ). Language difficulties of international students in Australia: The effects of prior learning experience . International Education Journal , 6 ( 5 ), 567 - 580 . Squire , C. , Andrews , M. , & Tamboukou , M. ( 2008 ). What is narrative research? In M. Andrews, C. Squire , & M. Tamboukou (Eds.), Doing narrative research (pp. 1 - 19 ). London, UK: Sage. Trahar , S. ( 2006 ). Roads less travelled: Stories of learning and teaching in a multicultural higher education environment (Unpublished doctoral dissertation ). University of Bristol, Bristol, United Kingdom. Trahar , S. ( 2013 ). Preface . In S. Trahar (Ed.), Contextualising narrative inquiry: Developing methodological approaches for local contexts (pp. xi-xxi) . Abingdon, UK: Routledge. Wood , C. ( 2005 ). Data protection issues in educational research . In K. Sheehy, M. Nind , J. Rix , & K. Simmons (Eds.), Ethics and research in inclusive education: Values into practice (pp. 242 - 250 ). Abingdon, UK: Routledge. Woodcock , C. ( 2010 ). The listening guide for coaching: Exploring qualitative, relational, voice-centred, evidence based methodology for coaches . Coaching: An International Journal of Theory, Research and Practice , 3 ( 2 ), 144 - 154 . place” (Clandinin, Pushor , & Or , 2007 , p. 23 ) James , G. ( 2017 ). Cul-de-sacs and narrative data analysis - A less than straightforward journey . The Qualitative Report , 22 ( 12 ), 3102 - 3117 . Retrieved from

This is a preview of a remote PDF:

Gwyneth James. Cul-de-sacs and Narrative Data Analysis – A Less Than Straightforward Journey, The Qualitative Report, 2018,