Narrative Smoothing in the Wild: A Pack Based Approach to Co-Constructing Narratives for Analysis

Abstract

This methodology paper introduces a collective, team-based approach to constructing narratives in narrative research. The goal of the larger study was to explore the pedagogical belief and practices of engineering faculty members. The newly formed team of researchers ranged from novices to experts in the field of qualitative research, and this space created a unique opportunity to reflect on and explore the co-construction of Cody’s narrative, the first narrative that the team constructed. The narrative was smoothed and constructed in a way that reduces some of the limitations inherent in narrative smoothing, through a deliberate and intentional negotiation process. We hope that this deeper exploration of our methods is helpful for other narrative researchers who are interested in team-based approaches to co-construction of narratives.

Introduction

No narrative can be presented without interpretation (Kim, 2016), a rigorous and guiding theory of capturing narratives include a clear set of strategies sensitive to the nuances of interpretive inference. The central act of presenting a narrative is narrative smoothing done by researchers which renders participants’ messy, scattered and apparently disjoined components into a logical and coherent narrative that is engaging the readers (Kim, 2016) and it is the pivotal component of story-telling. Narrative analysis is a qualitative approach to research that involves developing an understanding through stories.

This methodology paper is an emergent work related to a larger narrative research study, participants were asked to share their stories of how they became engineering educators at the time of interview and we conducted narrative smoothing post interviews to be able to represent their stories in an authentic way. Specifically, faculty who had demonstrated efforts towards shifting their teaching from traditional faculty-centered pedagogies to more student-centered pedagogies were the focus of this research. We sought to understand the journeys they each took to adopt novel teaching paradigms to foster student engagement and promote a novel culture of teaching, which in some cases was a step away from the norm. It was critical to understand the supports and barriers in these processes, the origin of the professors’ desires for changes and the process by which they became comfortable with the change. The process of changing the approaches towards teaching were authentically described in the narratives shared by these faculty participants.

We used narrative analysis to construct stories by integrating events and happenings into a temporally organized whole with a thematic thread, called the plot, which enables us to help the reader understand why and how things happened in the way that they did and why and how our participants acted in the way that they did (Kim, 2016). Presenting this meaning coherently and in a way that is engaging to the readers requires a refinement to the data obtained in its raw form. This involves a process of arrange the data to be more refined, coherent and organized, referred to as narrative smoothing. Sometimes narratives also demand to provide a whole context of a situation by researchers in the process of construction. Narrative smoothing is an essential, although somewhat controversial, part of narrative analysis that allows the story to maintain its metaphoric richness while being representative of the story as recounted by the teller.

It is not surprising that the narrative as delivered by the participant in the interview were shared non-linearly and in a form that is not ready for immediate dissemination (Bruce et al., 2016; Cain, Estefan, & Clandinin, 2013; Connelly & Clandinin, 2006; Sandelowski, 1991; Wertz, 2011). Before these narratives could be analyzed to develop insights and understandings of human meaning and experience, they needed to be constructed in a way that was organized temporally and coherently to make sense to the reader, in other words, the research team needed to smooth the narratives to tell the stories in a coherent manner. This paper focuses on the process of a team-based approach to co-constructing a narrative, which was a precursor to an analysis of the narratives. In this particular narrative we co-constructed Cody’s journey towards a student centric teaching in engineering.

Literature Review

Constructing a narrative becomes critical as narratives are never coherent and linear. Hence, narrative smoothing is an important process in the field of narrative inquiry. Narrative inquiry aims to understand the phenomenon of focus and mediate the reader’s understanding (Gracia, 2012). Garcia (2012) noted that the process of narrative smoothing is centered around five strategies: focus, omission, addition, appropriation, and transposition. Focus signifies the strategies to present the significant elements of a story while omitting (omission) elements of a narrative that is insignificant to the phenomenon; the addition enhances the narratives and provides a comprehensive picture; and transposition means utilizing a theme within a story to a different context to represent the meaning something that might have not been captured previously. Appropriation is a step where the researchers present some aspects resonating strongly with them through the storyline and create the magic of storytelling while “making it their own” (Gracia, 2012, p. 225). In all those steps the goal remains to represent a narrative aesthetically while honoring the voices of the participants. The stories, or narratives, are constructed and analyzed for patterns that emerge across participants (Clandinin, 2007; Creswell, 2013; Kim, 2016; Polkinghorne, 1995; Riessman, 2008; Wertz, 2011). Narratives are studied as “functional, occasioned and constitutive of identity” (Benwell & Stokoe, 2006, p. 137).

Narrative research is both the method and phenomena of study (Meier & Stremmel, 2010; Pinnegar & Danes, 2007), this adds complexity once narratives are uncovered as they can be analyzed and presented in several different forms such as structural narrative analysis (Labov, 1972), fictionalized representation (Clough, 2002) and narrative analysis (Polkinghorne, 1995). Burke (1966) developed the dramatistic pentad as a method to unpack people’s narratives through the five basic elements of drama (act, agent, agency, scene, and purpose). Narrative analysis is a qualitative approach to research that involves developing an understanding through stories. These stories, or narratives, are constructed and analyzed for patterns that emerge across participants (Clandinin, 2007; Creswell, 2013; Kim, 2016; Polkinghorne, 1995; Riessman, 2008; Wertz, 2011). Narratives are studied as “functional, occasioned and constitutive of identity” (Benwell & Stokoe, 2006, p. 137). Not as only reality, but as constructions that shape world around us and “teach us what worthy life is, what we should aspire to and what we should avoid” (Spector-Mersel, 2010, p. 208). Often requiring the embodiment of the ideas we communicate (Alsup, 2006), narratives are critical to understand people’s experiences, actions, knowledge, and ultimately, identity.

The narrative construction technique involves iterative work with reflective practices in order to build consensus and generate components that directly inform the process as noted by Smith (2016).

Objective

The purpose of this methodology paper was to introduce a collective, team-based approach to co-constructing narratives in narrative research.

Our research team consisted of a faculty member, Dr Nadia Kellam, who has expertise in narrative research methods; a postdoctoral researcher, Dr Brooke Coley, an engineer who at the time was new to qualitative research, and a postdoctoral researcher, Dr Audrey Boklage, who had recently completed a PhD in Curriculum and Instruction, Science Education, and who had experience in qualitative research methods in general, but not in narrative research methods specifically. Dr Debalina Maitra at the time worked as a postdoc with Dr Brooke Coley later went through the data and smoothed the narratives in a publishable format in the year 2023. This was our first experience as a research team, so throughout this process we were learning how to work together and how to optimally learning about each other. With the experiential diversity of our team, the initial dilemma was in identifying a clear method that articulated the narrative analysis process. Additionally, our plan to each co-contribute to the co-construction of these narratives further complicated the approach to analysis, while also providing an opportunity to critically discuss and reflect throughout the process of narrative smoothing. Carter et al. (2014) highlighted that researchers need to begin with their own story as they seek to understand the stories of others which indicates one’s positionality. In this article, we share our process and reflections of engaging in reflexive and dialogical storytelling as suggested by Carter et al. (2014).

This paper will focus on our process of developing a team, co-constructed narrative, the black box of narrative analysis which Polkinghorne (1995) describes as the “process by which happenings are drawn together and integrated into a temporally organized whole” (Polkinghorne, 1995, p. 5). Our primary objective was to provide an example for researchers interested in narrative methodology and/or team approaches to constructing narratives. In this paper we share our insights and challenges of learning to co-construct narratives. Also, this paper will capture aspects on social identity through analyzing the lived experiences and emotions and how this has steered the narrative to collating data on the lived experiences as highlighted by Adams (2021).

Data Sources

The data used for this methodology paper come from a larger study of engineering faculty from institutions across the United States. Faculty are at varying stages in their careers, from different engineering disciplines and diverse in demographics. These interviews were conducted with IRB approval and occurred in person, over the phone and with Skype. Interviews lasted between 29 and 84 minutes.

In this methods paper, our data included an exemplar participant’s interview transcript, email exchanges, memos from our research notebooks, and comments and exchanges within shared files. This set of data supported the development of our approach to co-constructing narratives within this larger research project. Of the 12 participants, Cody was selected as an example that would “challenge the research team in their process of converging on what to include and/or exclude” (Personal communication, August 15, 2015) from the richness of Cody’s data contained in the structural narrative. Cody was a natural storyteller and provided an abundance of information that would instigate challenges on what to keep when constructing a smoothed narrative.

The size of the team was n = 3 which had an impact on interpretation of Cody’s narratives and the nuances of what must be included in the story. The issues were not about maintaining the integrity of the participant rather the inclusion of what as most critical given the reality that all cannot be included.

Methods

In this study, participants were asked to describe their transition to a student-centered classroom. The first phase of the interview was focused on narration in which the interviewer encouraged story telling from the participant. The interviews then switched to a conversation phase in which the interviewer asked questions or provided prompts to encourage elaboration on ideas introduced in the first narration phase (Kim, 2016). The conversation was not entirely semi-structured in nature as we asked a lot of follow up questions to the faculty participants. The questions asked during the conversation were targeted towards their beliefs about students, performance driven pedagogy, understanding of student-centric pedagogy, and finally, how they became more student centered and inclusive inside the classrooms. The interview structure resulted in nonlinear narratives. This study focused on the collaborative process of narrative smoothing, an interpretive act and intermediate step to the development of a constructed narrative as suggested by Kim (2016).

In co-composing stories across time and place, a sharing of knowledge embedded in participants’ lives occurs as highlighted by Pino Gavidia and Adu (2022). In this work, three researchers converged perspectives using narrative smoothing to generate one co-constructed narrative. The following section details the steps taken across the team to co-construct the narrative.

Step 1: Familiarization of Transcripts

The first phase of our analysis involved familiarizing ourselves with the data. The most experienced member of the research team, Nadia, conducted the interview with Cody, which focused on his change in teaching practices towards a student-centered approach. Cody was deemed suitable for this initial narrative smoothing experience because of the length and complexity of his narrative. Specifically, Cody’s transcribed interview was almost 12,000 words and 43 double-spaced pages in length. Cody was a natural storyteller, and it was evident his story would require smoothing to be both chronological and coherent to the phenomenon of interest. In the case of Cody, there were many tangents and sub plot-lines in his personal story.

Audrey and Brooke individually listened to the audio recording of the interview and after read through the interview transcript at least twice to become familiar with the data. The first round of listening and reading was for familiarity purposes only, whereas the second round was executed with the purpose of recording initial impressions of the data for each researcher.

Step 2: Development of Coding Scheme and Application of Codes

After the aforementioned rounds of becoming familiar with the data, the first attempt at coding the interview transcript was conducted by the researchers in Dedoose ®, an online, cross-platform application for analyzing qualitative and mixed methods research. Each researcher set to Dedoose with the intent to code for goals and events in Cody’s story. An event was described as something that happened in his personal life or anything that related to his journey of becoming a teacher (e.g., deciding to pursue his PhD, being offered a fellowship, being awarded a grant, getting married). The researchers met several times to discuss the interpretation of the data.

There were different approaches assumed by the researchers in this phase and this resulted in the first discussion for clarifying divergent approaches to the analysis. Audrey initially used a general approach to code the events not specific to student-centric pedagogy. She wanted to capture Cody’s experiences overall. Brooke, on the other hand, added child codes to differentiate event types. In an email exchange following the first attempt at coding Cody’s story, Brooke’s response to Audrey was demonstrative of the pressure facing the novice and newest of the team to narrative smoothing, “As of now, I’m at line 781 [of code]...will try to get through the remainder by the end of the day. Don’t mind me if it’s too specific. As this is my first time, I was doing it in a way that made sense to me. Once we share, maybe I’ll get a better idea of how to run through [the data] for initial coding.” (Personal communication, September 3, 2015). At this point, both researchers were coding inductively with their foci on events, goals and aspirations. No codebook had yet been developed for data analysis as the intent was for the two researchers to first iteratively determine the focus of the co-constructed narrative. Brooke wanted to capture all the privileges and opportunities that Cody received during his journey. However, Audrey was more focused on Cody’s approaches. Brooke was adamant that all the opportunities Cody had to be critical in his trajectory, and therefore, not to be ignored.

The research team considered Schutze’s (1976) approach of narrative construction, for example, where the coding process is informed by structural description and analytical abstraction of the narrative. However, the team found the conversations of consideration to be both insightful and constructive and chose to proceed with an approach that sustained that dialogue. We coined this as iterative convergence described as the ongoing process of dialogue across varying perspectives until consensus is reached and decided to focus on agentive sequences in Cody’s career such as stories of his transformation and trajectories, as those related to Cody’s overarching goal of achieving a student-centered classroom through both philosophical and pedagogical changes in practice.

Step 3: Negotiation and Narrative Building

The initial coding for events and goals was the first point in the research process in which Brooke and Audrey had to negotiate what to include or exclude from the narrative, the first steps of narrative smoothing. Once each had their final set of codes, all team members then read each other’s codes and met to discuss their thoughts about what was and was not coded. The purpose of this meeting was to discuss the identified excerpts with the purpose of negotiating, consolidating and connecting the excerpts to later construct Cody’s narrative. We recognized this would be a critical and challenging step due to our varied positionalities. As stated in Arias López et al. (2023), we had to address the question of how to do collaborative research when, geopolitically and socially, team members came from “evidently asymmetric contexts and positions of privilege” (p, 6).

The iterative convergence played a huge part in our negotiation process. Each code was discussed until a consensus was reached to ensure reflexivity. These discussions were not cantankerous or contentious and instead served as a medium for team members to get familiar with each other’s positionality regarding the subject matter. These discussions also focused on how much context to include with each event as omitting context can also lead to a final narrative that is not representative of the original story (Spence, 1986).

Here, we situate an empirical example of our narrative smoothing process: the two coded transcripts were overlaid to identify where there were differences in thoughts, perspectives or approaches in analyzing the narrative. Brooke wanted to include the following event:

I applied to four schools…Cornell, of course, where I had gained the summer experience. I was accepted at all four and received funding from two. I do distinctly remember having a conversation with one of my professors at UNH. I remember him suggesting ‘go where there’s funding. You should be funded to go to grad school.’ That’s important, because that always stuck in my mind and that’s exactly what I tell undergraduate students here.

Audrey did not want to include this event as she felt it was detracting from the narrative. In her comments she thought only Cody’s “final selection of an institution and choice to pursue graduate school” worthy of inclusion. However, to Brooke, this was a critical influence of mentorship, which ultimately impacted his pathway to his journey. This passage fostered a critical discussion amongst the team and serves as a significant example of what often became the complexity of simple what would appear to be simple considerations.

This specific example made us hyperaware of our varying positionalities and pushed us to recognize the influence of such on our coding and interpretation of parts of Cody’s story. For Audrey, as a white woman, she was able to focus primarily on Cody’s philosophical and pedagogical approaches. However, from an intersectional identity, Brooke felt it imperative to consider and include the vast privileges and opportunities afforded to Cody. For her, the reality of his received support and direction could not be disentangled from his career trajectory and success. Absent the advice received, Cody’s path could have been drastically different. It was later decided through discussions as a research team that this excerpt would be included in the constructed narrative. This process of negotiation continued throughout this phase of narrative smoothing.

In continuing the process of narrative smoothing, Audrey and Brooke referred back to the initial research questions of the project with the purpose of keeping the events and goals critical to the narrative in relation to the project. The research questions of this project are the following:

1. How do engineering faculty describe their transitions from using teacher-centered to learner-centered teaching strategies?

2. How do faculty describe the barriers they face in transitioning to student-centered strategies and how do they persist in spite of these barriers?

3. What contextual and individual factors are revealed through these narratives that contribute to the success of transitioning from teacher-centered to student-centered teaching?

Items that had been coded as a goal or event were designated to one of three research questions. If there was no research question to which the information was related and/or could be designated, that was deemed grounds for the omission. This process was done independently by both Audrey and Brooke and then compared in a team meeting; events that were common across both of their analyses were included in the constructed narratives. Events that were not common were discussed until consensus was reached on how to treat the event.

Step 4: Constructing Narratives and Final Consensus

Next, both Audrey and Brooke independently constructed a narrative based on the codes aligned to the research questions. Schutze’s (1976) approach was discussed among the members in consideration of organizing the narrative. The first step in this process involved temporally ordering the events (Mishler, 1986, 1995), with the culmination of the narrative being the participant’s present-day, student-centered classroom. This process was a way of reconstructing the told story of the participants from the telling story of the researcher (Kim, 2016; Mishler, 1995) to construct a narrative that clearly communicated the participant’s journey towards a goal.

Team members combined events and added words to smooth the narrative with a secondary purpose of keeping the researcher’s voice out of the narrative. When the researcher’s voice was added, italics were used in the constructed narratives to denote researcher versus participant voice.

Upon reading the constructed narratives, Audrey questioned why Brooke had left out details of Cody’s lecture notes and textbook transition. Audrey found this to be an important part of his journey, and yet, at this point in the analysis it was found to not be included in the other researcher’s constructed narrative. As a result of this communication, the research team decided not to include the lecture notes in the final narrative as Audrey and Brooke both agreed they were not key to his journey towards a more student-centered classroom. This is critical to highlight as at the point of actually constructing the narrative, differences remain between what researchers are categorizing as critical that may be missed at earlier checkpoints in the analytical process.

Methodological Considerations and Implications

In the story of ‘doing’ narrative smoothing described above, we developed a process for smoothing narratives and focused many of our analytical examples on how we made decisions in regard to omitting parts of the participant’s story. In the end, we used the research questions from the broader study to help filter the events within the story. This decision had a significant impact on the final constructed narrative. In some ways this decision helped us ‘morph’ the narrative into one that is useful for our research project omitting the components that were not directly related. However, smoothing our narrative in this way may have inadvertently resulted in us changing the original essence of the story as told by the participant to be one that is helpful for the researchers in our broader research project. Narrative research is often framed by individualized and humanistic frameworks (Feely, 2019). For this current work, we sought to let the multiplicity of Cody’s academic career and teaching experiences flow.

We focused on the narratives centered on Cody’s career decisions and transitions. We collectively agreed to pay attention to Cody’s experience during graduate school and during his tenure. We had to omit the in-depth experiences of his undergraduate journey; however, we were intentional to include parts of his early life story that could have influenced his professional trajectory today. As an example, we collectively decided that his interactions with professors would be presented in the narrative. One of his professors suggested he apply for an undergraduate research experience, which served as his first exposure to research, the underpinning experience of mentorship that influenced him as mentor and professor. We suggest a consistent process of iterative readings of the data coupled with team discussion.

The team-based narrative smoothing process described in this paper also brings to light differences in researchers, their epistemologies, their inclinations of how they perceive and handle the data, and finally, the collective storytelling process. There were some instances where one member of the team was focused on what was said in the data while the other was focusing on what was missing from the story. In having a team-based narrative smoothing experience, it may be easier to simultaneously take both an interpretation of faith and interpretation of suspicion approach to engage in the analysis of the data in a deeper, more meaningful way. We highlight the importance of positionality and purpose of your final co-constructed narrative. In our case, positionality proved to greatly impact the interpretation of data significance for inclusion in the narrative. We also found it important to keep revisiting the purpose of the work, which made us continuously call back to the research questions of the study. Having natural storytellers like Cody, it’s easy to invest in the story overall and lose sight of what information matters. We constantly revisited the research questions to make sure the narrative remained contextualized to our target purpose.

Conclusion

In this paper, we described our first narrative smoothing endeavor as a new research team, which provided rich, focused, and open dialogues surrounding our process of narrative smoothing. Through documenting these processes of negotiation, selection, and omission of data to construct a narrative, we both improved our understandings of the importance of narrative smoothing and, hopefully, will provide others with some considerations as they begin doing narrative smoothing within their own research projects.

The most significant contribution of this work is advancing through the negotiation process with each researcher bringing a different perspective to the narrative building. In addition to team-based construction of narratives from interview data, this work also informs readers about the factors that influence the pedagogical decision-making of engineering faculty members.

Supplemental Material

Supplemental Material - Narrative Smoothing in the Wild: A Pack Based Approach to Co-Constructing Narratives for Analysis

Supplemental Material for Narrative Smoothing in the Wild: A Pack Based Approach to Co-Constructing Narratives for Analysis by Brooke Coley, Audrey Boklage, Nadia Kellam, and Debalina Maitra in International Journal of Qualitative Methods