Applying ATLAS.ti to a Nexus Analysis of Ethnolinguistic Labelling Practices
This Certified Trainer Case Study was written by Nathan A. Wendte, PhD (Independent Certified ATLAS.ti Professional Junior Trainer). We publish these articles to share the diversity of approaches to qualitative research. Although the ATLAS.ti Core Training Team publishes regular content that is endorsed by ATLAS.ti GmbH, we also wish to give voice and respect to independent trainers’ experiences as well, because qualitative research is so vast and flexible!
Nathan A. Wendte, PhD is a linguistic anthropologist who completed his PhD at Tulane University. His research focuses on discourses of language and ethnicity pertaining to identity and change, specifically related to Creole-identifying peoples of the U.S. Gulf South.
Linguistic anthropologists deal with a wide range of qualitative data, and their methodological approaches are as varied as the data they seek to analyze. In this article, I present the ways in which I used ATLAS.ti to conduct one stage of a nexus analysis of ethnolinguistic labelling practices among Creoles of the Gulf South region of the United States. First, I introduce my project’s research questions and the data I collected in order to answer them. Second, I give a brief summary of nexus analysis as a methodological approach (Scollon and Scollon 2004) and the stages of which it is comprised. Third, I provide examples of how I used ATLAS.ti to conduct specific tasks associated with my methodology. Lastly, I evaluate ATLAS.ti’s effectiveness as a tool for nexus analysis.
In late 2019, I successfully defended my PhD dissertation entitled “A Tale of Two Triangles: Ethnolinguistic Identity among Gulf South Creoles” (Wendte 2020). At the heart of my project was the issue of ethnolinguistic identity, the particular set of overlapping ethnic and linguistic traits that serves to identify groups and individuals. In order to investigate this, I conducted interviews with self-identified Creoles in Texas and Louisiana who were also self-professed Creole speakers. The central question in the interviews was this: What does Creole mean as a label for people and as a label for language? Participants responded in a wide variety of ways. I asked a general list of follow up questions that was tailored for the specific interview situation. This included basic sociolinguistic information as well as clarifications about each participant’s individual life-story. At the end of the interview, I asked participants to translate a number of English sentences into Creole. Taken together, the recorded sociolinguistic questionnaire and the translation task were my primary data. I wrote a general, chronological log for each interview and subsequently coded each for specific topics (explained in greater detail below).
Identity is a notoriously difficult thing to capture. It is a moving target that emerges through interaction and discourse. For this reason, I focused on the discursive act of labelling (and label definition) as a gateway to an understanding of identity. This required a methodology that foregrounded social activity as the starting point for analysis. In the next section, I introduce the methodology that I used to frame my project: nexus analysis. Then, I illustrate some of the ways in which I used ATLAS.ti within this methodological framework. Afterwards, I evaluate the effectiveness of ATLAS.ti as a support for nexus analysis.
In brief, nexus analysis prioritizes social practice (a repeated action that is recognized as being socially meaningful) as the basis for understanding social issues. One works outwards from the nexus of practice (the specific time and place where and when the action occurs) and looks for the components that influence the action and the effect that the action has on subsequent iterations of that same action. Nexus analysis was pioneered by Scollon and Scollon (2004). According to them, the methodological hallmark of nexus analysis is its abductive orientation (Scollon and Scollon 2009). This contrasts with both inductive and deductive approaches. An inductive methodology projects observations at the individual level onto the group, and a deductive methodology seeks group characteristics at the individual level. An abductive approach, by contrast, takes both inductive and deductive insights together and seeks relationships between them that best explain the data pertaining to a given social practice.
A nexus analysis takes place in three stages. First, one must engage the nexus. This involves identifying a social issue and a social practice through which that issue is foregrounded. Second, one must navigate the nexus. This involves analyzing the configuration of moving pieces that occupy the nexus where and when a discrete social actions occur. Third, one must change the nexus. This involves acting in some way within the nexus of practice in hopes of changing the trajectory of that social action for the better. The second stage of a nexus analysis—engaging the nexus—is what I will discuss in the remainder of this article.
Engaging the nexus with ATLAS.ti
Before proceeding, I want to take a slight detour on the subject of coding. Within my project, I had dozens of hours of interviews to sift through. Rather than transcribe the entirety of these interviews, I used a two-step process to segment my data into broad themes. The first step was chronological topic logging. In this step, I kept a minute-by-minute log of the main ideas being discussed at any given time during a particular interview. See below for an excerpt from a topic log of one of my interviews:
The second step was to code the topic logs for specific topics of interest. I created hierarchies of related codes in a topic code catalogue (see below) and then assigned each of these codes an alphanumeric identifier. Note that the topic log naturally creates one-minute sections that easily convert into quotations within ATLAS.ti once they are assigned a code (see righthand side of the above example).
I found this system very useful when the time came to identify sections of my data that I wanted to transcribe and analyze more closely at later stages. Working with these one-minute chunks was and efficient and helpful mechanism for exploring relationships between different ideas and relationships, which is the principle concern in the second stage of a nexus analysis—engaging.
One of the primary tasks associated with engaging the nexus is called mapping. In mapping, one traces the entry and exit of different factors influencing the nexus of practice. The three main categories of factors concern the historical body (e.g., individual participant experiences), interaction order (e.g., elements pertaining to the setting of a social action), and discourses in place (e.g., attitudes, ideologies, assumptions, etc.). Here are three examples of how I mapped influencing factors within the nexus of ethnolinguistic labelling.
Each participant has unique biographical traits that can be assumed to influence the way they apply ethnolinguistic labels. One trait I looked at was individual experiences of rural life associated with specific labels.
By analyzing these codes in a cooccurrence table, I concluded that “Creole” (represented by code I.A.1) is more likely to describe rural life (represented by code IV.H.7) than “Cajun” (I.A.2), “Frenchman” (I.A.3), or some other ethnic label (I.A.4). The quotation beneath the label captures this relationship.
The setting of each interview, though purposefully comparable, was always unique. The entry and exit of participants in the conversation was important as were physical characteristics of the interview space (e.g., homes, public libraries, etc.). I did not use ATLAS.ti to note these features, instead I kept track of it in the general log of each interview.
Discourses in place
These abstract ideas, assumptions, ideologies, etc. were meticulously coded so that I could see how they pattern with one another. Discourses in place tended to present themselves in bundles. A cooccurrence table helped me identify these bundles and examine how the discourses were related to one another. The following code cooccurrence table shows codes for different language ideologies along both the vertical and the horizontal axes.
The highlighted cell notes cooccurrences of “diversity & variation” (III.A.15) with “broken” (III.A.6). The quotation clarifies that some participants associate the differences within their language with the concept of linguistic brokenness or lack of purity.
Evaluating ATLAS.ti as a tool for nexus analysis
Although ATLAS.ti is not built specifically for nexus analysis, it does possess a number of tools that make it quite useful for navigating the nexus. In particular, cooccurrence tables can be a powerful means of identifying code clusters that may indicate significant underlying causes and effects of social practices. A necessary caveat when using interview data is that coding matters. Poorly coded transcriptions (or general logs, as in my case) are inherently less useful. If I were to begin my project from scratch, I would certainly make more use of networks. I have begun to use this tool to examine some of the finer-grained discourse data that I did not have space to treat in my dissertation, and I find it very helpful. I look forward to discovering more about my data as I continue to use ATLAS.ti.
This article introduced the methodology known as nexus analysis and evaluated the effectiveness of ATLAS.ti as a tool for conducting certain tasks associated with that methodology. ATLAS.ti’s capability to visualize and explore relationships between objects, ideas, and discourses makes it very suitable for researchers conducting nexus analysis and other forms of qualitative discourse analysis.
Scollon, Ron, and Suzie Wong Scollon. 2004. Nexus Analysis: Discourse and the Emerging Internet. London ; New York: Routledge.
———. 2009. “Breakthrough into Action.” Text & Talk 29 (3).
Wendte, Nathan A. 2020. “A Tale of Two Triangles: Ethnolinguistic Identity among Gulf South Creoles.” PhD, New Orleans, LA: Tulane University.