Integrating a Social Ecological Model With Focused Ethnography for Culturally Informed Health Behavior Research

The Exemplar Study: Context and Scoping

In the following section, we provide a guide for scoping a qualitative exploration of health behaviors among a cultural group by discussing the process we used to complete a focused ethnography of factors influencing eating behaviors among emerging adults in the U.S. Navy. The first author conducted the study for her dissertation, and the second author served as the methodology expert on the committee. The Uniformed Services University of the Health Sciences Human Research Protection Program (HRPP) Office approved the study as exempt human subject research under Category 2 as defined by 32 CFR 219.104(d)(2). All of the study participants provided verbal consent. We conducted the study during the height of the COVID-19 pandemic and faced several challenges related to the public health restrictions and competing priorities of the research population. Despite these unexpected obstacles, we collected rich data for analysis and interpretation, providing researchers, healthcare providers, and military decision-makers with foundational information for improving military personnel’s eating behaviors and health. The results of this study have been published elsewhere (Troncoso et al., 2023, 2024).

Defining the Culture and the Problem

The first step in our process was to determine the cultural group of interest, beginning with the broad culture of the U.S. military (hereafter military) and narrowing to the subcultural of early career enlisted Navy Sailors in the period of emerging adulthood (i.e., 18-25 years). The military is a culture with shared values, norms, behaviors, language, customs, and rituals distinct from the general U.S. civilian population (Meyer, 2015). The military has various subcultures, such as service (e.g., Navy and Army) and rank (e.g., officer and enlisted). Junior enlisted (early career) emerging adults in the Navy represent a unique subculture. They experience a critical period of transformation from adolescence to adulthood and are part of a culture that demands high levels of responsibility and self-sacrifice but with limited autonomy and agency. Unlike emerging adults who are full-time college students or slowly transitioning to increased responsibility, young service members are thrust into demanding jobs that may require leading others and working long hours in austere environments and unpleasant conditions.

Most young, early-career enlisted service members live in a controlled environment (i.e., barracks) with disproportionate access to unhealthy foods and a lack of access to cooking and food storage. Despite the need for active duty service members to have eating habits that promote optimal health and physical and mental performance, many enlisted service members have unhealthy eating behaviors ranging from not eating enough nutrient-dense foods, overeating processed foods, to disordered eating (Barlas et al., 2013; Chukwura et al., 2019; Gaviria & Ammerman, 2022). After a systematic review of the military and civilian literature, we identified a gap in research and a critical need to focus on the eating behaviors of junior enlisted Sailors in emerging adulthood.

Design

The philosophical assumptions undergirding this study were that there may be multiple realities of a phenomenon and that knowledge is constructed with subjective evidence, requiring the researcher to “get as close as possible to participants being studied” (Creswell & Poth, 2018, p. 21). Focused ethnography provided a systematic way to examine the contextual influences of eating behaviors from the participants’ perspectives. In contrast to the extensive fieldwork and global observations required in traditional ethnography, focused ethnography allowed us to focus on a specific phenomenon within a subculture, using interviews as our primary data collection method (Higginbottom et al., 2013). In addition to semi-structured interviews with participants, we conducted informal interviews and included focused observations, artifacts, and photo elicitation. We triangulated data sources and methods to understand the influences of junior enlisted Sailors’ eating behaviors in multiple contexts. Doing so helped us create an adequate cultural description, including explicit and tacit cultural knowledge (Spradley, 1979). For example, we could learn about influences within the physical environment by interviewing participants and their leaders, observing food availability and accessibility, identifying printed information and marketing, and reviewing policies related to the military nutrition environment.

Data Collection Process

We collected and analyzed data iteratively and recursively to allow for constant comparison, member checking, and triangulation of the data. What follows is a brief description of the data collection methods and analysis.

Semi-Structured Interviews

We developed an interview guide based on the literature review, conceptual framework, and research questions. We then piloted the initial guide with four Sailors and refined it with the expertise of an anthropologist with military knowledge and the second author. The interviews followed the method described by Spradley (1979), beginning with broad questions and ending with more focused questions. All interviews were audio recorded and transcribed verbatim using NVivo transcription software and were verified by the first author. The second author provided accuracy checks on random transcripts. For every interview, I (first author) recorded field notes and memos to capture reflections, analytical hunches, and inquiries for further explorations. Because analysis occurred in parallel, subsequent interview questions were more focused and served to examine and clarify assumptions.

Informal Interviews

Informal interviews are conversations with experts about the culture that further help to discover categories and meanings, identify values, and confirm or elucidate data gathered from semi-structured interviews, observations, or artifacts (Fetterman, 2010). I (first author) conducted informal interviews with commanding officers and senior leaders, middle managers, health promotion professionals, dieticians, and food services employees with knowledge and expertise in the military food environment.

Observations

As the first author, I conducted observations at varying levels of involvement, ranging from observer-as-participant to observer (Roper & Shapira, 2000) using methods described by Spradley (1980). The core purpose of these observations was to gain a deeper understanding of the food environment within the military context, specifically assessing the access to and availability of food options and identifying facilitators and barriers to healthy eating behaviors. I also focused on observing participants’ interactions with the food environment and documenting social situations involving food and eating behaviors. This observational data was critical for triangulating information gathered through interviews, documents, and policies, providing a more comprehensive picture.

The observation process involved moving from broad descriptive observations to more focused and selective observations, guided by Spradley’s (1980) dimensions of social situations. Regarding participant awareness of my role, this varied by setting. In small group settings like military office breakrooms or workspaces, where I was present in civilian attire, participants were aware of my dual role as both researcher and observer. However, in less formal public spaces such as military dining facilities, during walks around the bases, or while in on-base eating establishments or grocery stores, I did not overtly make it evident that I was conducting research. I maintained a presence as a patron in these settings to capture more natural interactions with the food environment and social dynamics.

In the beginning of the study, observations were global and descriptive and guided by a priori themes derived from the conceptual framework and nine major dimensions of social situations: (a) space (the physical place or places), (b) actor (the people involved), (c) activity (a set of related acts people do), (d) object (the physical things that are present), (e) act (single actions that people do), (f) event (a set of related activities that people carry out), (g) time (the sequencing that takes place over time), (h) goal (the things people are trying to accomplish), and (i) feeling (the emotions felt and expressed) (Spradley, 1980). During later observations, I examined social situations with knowledge gained from previous observations and ongoing interviews by using emerging themes and the nine dimensions previously mentioned to refine and clarify assumptions. For example, during observations, I asked and answered questions such as, “Who is involved in organizing and patronizing command potlucks?” and “What are the differences between Sailors who bring their lunch to work and Sailors who order food for lunch during work?”

Artifacts

Artifacts include different data sources, such as objects or documents (Creswell & Poth, 2018) and can assist in understanding explicit and tacit knowledge and provide context to data gathered from observations and interviews (Munhall, 2012). Artifacts in this study included signs or flyers related to food and eating at military bases, military and/or command policies and procedures, information on public and social media channels associated with the eating behaviors of junior enlisted Sailors, and data from the Healthy Eating Score-7 questionnaire (Shams-White et al., 2019).

Photo Elicitation

Photo elicitation involves the use of photographs—taken by participants or selected by the researcher—during the interview process to elicit deep and meaningful talk, provide visible data that otherwise may be invisible to the researcher, or reduce areas of misunderstandings (Harper, 2002; Lapenta, 2011). We used photo elicitation to supplement gaps in observations and gain insider information on places where participants eat and prepare food.

We asked participants to voluntarily provide photographs of places where they eat, elements that depicted barriers or facilitators to healthy eating behaviors, or food preparation and storage areas in their living or workspaces. Photographs did not include people. Photographs were discussed at the end of the interview using questions or statements such as, “Tell me about this photo.” Photographs were uploaded to NVivo and linked to the participant interview for analysis. Figures 2 and 3 are photographs provided by a participant and discussed during the interview.OPEN IN VIEWER

OPEN IN VIEWER

Figure 3.

Participant Photograph of Condiments and Beverage Storage in Barracks

Integration and Triangulation

We used multiple data sources and methods in conjunction to build a comprehensive understanding of the influences of Sailors’ eating behaviors. We then employed triangulation throughout the data collection and analysis phases by comparing and cross-referencing these diverse data streams. Specifically, we triangulated by data source (e.g., comparing perspectives from participants, leaders, and food service personnel), by method (e.g., examining how interview accounts aligned with or differed from observed behaviors), and by data type (e.g., integrating insights from interview transcripts, observational field notes, and policy documents). For instance, participants might discuss challenges accessing healthy food options in interviews; we would then compare this with observations of food availability in dining facilities and stores, and review relevant military policies. Photo elicitation provided visual data that either confirmed or added new dimensions to what was discussed in interviews and observed in the environment. This constant comparison across different forms of data allowed us to verify emerging themes, confirm or challenge initial assertions, and develop a more robust and accurate understanding of the multi-faceted influences on eating behaviors within the military subculture. This layered approach significantly enhanced the credibility and depth of our findings, ensuring they were well-grounded in the diverse realities of the study participants and their environment.

Data Analysis Process

Ethnographic data analysis is a continuous and iterative process, beginning with the first iteration of data collection (Spradley, 1979). We worked closely, meeting weekly or biweekly, during data collection and data analysis to ensure accuracy in the methodology, examined our potential biases, and evaluated the credibility of our findings. We followed the iterative and cyclic qualitative data analysis process described by Miles et al. (2020), which includes four concurrent processes for data analysis: data collection (discussed above), data condensation, data display, and drawing and verifying conclusions. We also used ethnographic analysis methods (i.e., domain analysis, taxonomic analysis, and componential analysis) to ensure cultural analysis and interpretation of the data (Spradley, 1979, 1980). These two methods were complementary and consistent with basic paradigms and assumptions of qualitative research methodology and ethnography.

Data Condensation

We organized the data according to data type (e.g., transcripts and field notes) and reviewed them several times to gain familiarity and a general sense of the data before organizing them into codes, categories, or themes. All interview transcripts and field notes were coded; we also coded artifacts (e.g., policy documents and photographs) where appropriate.

We conducted first and second cycle coding, which involves the initial process of labeling the data (first cycle) and then recoding, reorganizing, and reanalyzing data coded during the first cycle (second cycle) (Saldaña, 2016). During the first coding cycle, an eclectic combination of descriptive, a priori, in vivo, process, and value codes were applied to the qualitative data. We applied definitions to codes, wrote reflections of the codes, reviewed pertinent literature, kept an audit trail of code revisions and deletions, discussed emerging thoughts and ideas with members of the dissertation committee, and created a codebook.

Second cycle coding involved finalizing the codebook and making domain and taxonomic analyses (Spradley, 1979, 1980). Following Spradley’s methods, we conducted domain and taxonomic analyses to identify cover terms for potential domains and taxonomies. Cover terms are names for cultural categories, derived from either folk terms (participants’ own language) or analytical terms (researcher-defined terms when folk terms were not available). Domains are symbolic categories that encompass other categories referred to as taxonomies.

Our domain analysis process was iterative and guided by the research questions. We explored various semantic relationships within the data to better understand eating behaviors from the cultural perspective of junior enlisted Sailors. For example, we investigated semantic relationships such as “kinds of eating behaviors,” “reasons for skipping meals,” “ways leaders influence eating behaviors,” and “attributes of cooking in the barracks.” The first step involved identifying a relevant semantic relationship and then searching the data for terms that fit this relationship. After reviewing lists of folk terms from preliminary searches and rereading interview transcripts and field notes, we conducted detailed analysis on these semantic relationships, noting how they complemented many of the first-cycle codes. As we uncovered information within one domain, such as “kinds of eating behaviors,” this informed our exploration of subsequent, related domains, such as “dieting” and the “reasons for dieting.” This iterative process of uncovering and exploring interconnected domains, rooted in folk and analytical terms, provided a rich and culturally grounded description of the influences on eating behaviors. A list of the key domains explored in this study is provided in Table 1. Figure 4 Provides an excerpt illustrating our exploration of the “kinds of eating behaviors” domain, showcasing the identification of cover terms, folk terms, and included terms.

OPEN IN VIEWER

Table 1.

Domains for Factors Influencing Eating Behaviors Among Junior Enlisted Sailors in Non-deployed Settings

1. Kinds of eating behaviors

2. Kinds of sailors

3. Kinds of social networks

4. Kinds of barriers to healthy eating

5. Kinds of facilitators to healthy eating

6. Reasons for meal prepping

7. Reasons for snacking

8. Reasons for skipping meals

9. Reasons for eating fast food

10. Reasons for eating healthily

11. Reasons for eating unhealthily

12. Reasons for dieting

13. Reasons for drinking energy drinks and coffee

14. Reasons for “watching what you eat”

15. Reasons for sporadic eating

16. Reasons for disordered eating

17. Reasons for stress and emotional eating

18. Reasons for grabbing something quickly

19. Reasons for eating meals

20. Reasons for taking food to work

21. Ways peers influence eating behaviors

22. Ways leaders influence eating behaviors

23. Ways social networks influence eating behaviors

24. Ways workplace structure and systems

25. Ways living situation influence eating behaviors

26. Ways to learn about cooking and nutrition

27. Ways eating behaviors change during PRT season

28. Attributes of living in the barracks

29. Attributes of cooking in the barracks

30. Attributes of eating in boot camp

31. Attributes of eating in “A” school

32. Attributes of the galley

34. Places to get food

OPEN IN VIEWER

Figure 4.

Excerpt Illustration of the “Kinds of Eating Behaviors” Domain

The codebook contained: (a) the code; (b) the definition of the code based on the researcher’s analysis of the data and/or the literature; (c) an exemplary quote; (d) an indicator if it was an a priori (“P”) or In vivo (“Inv”) code; and (e) a classification label related to the presumed level of influence within the social ecological framework (e.g., individual factor or physical environment). An excerpt of the codebook is shown in Figure 5.OPEN IN VIEWER

Data Display

Data display involves organizing and assembling the data to provide analytic reflection and action and may include extended text, matrices, networks, and graphs (Miles et al., 2020). We used various types of data displays, including tables, matrixes, and visual models, to explore, describe, explain, and make predictions with the data. For example, we created matrixes to explore variations in eating behaviors and influences among groups, and visual models helped to make connections from sequence domains (e.g., stages in acculturating to military life/food environments). We also used displays when making componential analysis—“the systematic search for the attributes associated with cultural categories” (Spradley, 1980). We identified dimensions of contrasts between categories (e.g., Kinds of Sailors) and plotted these data into simple tables to identify nuances quickly and comprehensively within the culture related to eating behaviors.

Drawing and Verifying Conclusions

We employed various tactics to minimize bias, test or confirm meaning, and draw conclusions (Miles et al., 2020). These tactics included: (a) noting patterns and themes, (b) seeing plausibility, (c) clustering, (d) counting, (e) making contrasts and comparisons, (f) partitioning variables, and (g) subsuming particulars into the general. These tactics were employed in concert with, or as part of, the data condensation and display processes described above and during meetings. For example, noting patterns and themes and clustering occurred while reading and reflecting on the data, during first and second cycle coding, and identifying overarching themes. We employed counting when exploring how frequently specific issues were reported or the number of participants who mentioned a particular concept. Counting is not quantitative analysis but is important when making generalizations about the data, such as estimates on a particular recurrence or “significance” (Miles et al., 2020).

The final stages of drawing and verifying conclusions resulted in identifying cultural themes related to the specific aims and conceptual framework. As described in the section on data condensation, data were initially condensed into codes, domains, and categories. We identified themes while making contrasts and data comparisons, identifying stages in processes, and recognizing generalities of particulars. Data saturation, defined as reaching a point where no new salient information emerged from subsequent data collection or an intuitive feeling of learning nothing new, guided the decision to conclude recruitment. While an initial target number of participants was considered based on other qualitative studies of a similar nature, the final sample size was ultimately based on achieving this saturation, a determination made by the first author in consultation with the dissertation committee.

Ensuring Rigor and Trustworthiness

Establishing trustworthiness—the quality or goodness of qualitative research—is important for ensuring credibility, plausibility, and integrity of research findings (Ryan et al., 2007). To ensure credibility of findings, we engaged with the data extensively by reading and rereading transcripts, field notes, and artifacts to have an intimate understanding of the data and stay close to the data with descriptions and interpretations. Triangulation among methods and data sources produced confirming conclusions, and the research findings linked to the initial conceptual framework and other social and behavioral theories. I (first author) journaled, wrote memos, and discussed my perceptions, thoughts, and biases related to the study with the second author and dissertation committee members. Factual reporting and auditability of the research process and findings were important for ensuring dependability. The recursive and iterative processes of parallel data collection and analysis allowed for constant evaluation and comparison of emerging findings. We maintained an audit trail of all data collected, analytical reflections of the data, and the analysis process and meeting notes. Transferability—the fit of findings to other similar contexts and larger import (Miles et al., 2020)—rather than generalizability was the goal of this study. We presented findings that included detailed descriptions for readers to examine similarities and differences within their settings. Throughout the research process, a variety of readers—including researchers and service members from the Navy, Army, and Air Force— remarked on how consistent the findings were with their own experiences with junior enlisted personnel. Confirmability establishes that the conclusions and interpretations reached are coherent and supported by the data (Miles et al., 2020; Ryan et al., 2007). Confirmability is achieved when the other three criteria above are met or through the publication and dissemination of results. We invited participants and people familiar with the culture (e.g., Sailors not included in the study) to read drafts of written reports and provide feedback before presenting the findings in presentations, reports, and publications.

Reflections and Recommendations

Reflecting on our experience applying this integrated methodological process, a primary lesson learned was the critical importance of clearly defining the research problem and rigorously scoping the study from the outset. Focused ethnography yields rich and nuanced data. Staying focused on the predefined research questions and utilizing the conceptual framework as a guide was essential to managing the project within our timeline and resources. It required making deliberate decisions about which avenues to pursue in depth and which to acknowledge but set aside for future inquiry, a process facilitated by continuously referring back to our core aims and the first author heeding the advice of committee members.

Effective communication and the cultivation of strong rapport were also paramount throughout this process. Frequent and focused dialogue among the dissertation committee members was critical for timely progress. My (first author’s) regular interactions with gatekeepers were vital for accessing the population and navigating the complexities of conducting research during the COVID-19 pandemic. Building trust involved maintaining respectful curiosity, a teachable attitude, and flexibility. Furthermore, given my insider positionality, reflexive journaling and memoing were invaluable tools. These practices allowed me to record not only observations and analytical hunches but also my own perceptions, thoughts, and potential biases. Regularly discussing these reflections and emerging ideas with my mentor and other team members provided essential external checks, challenging my assumptions and helping to ensure the trustworthiness and credibility of our interpretations. This iterative process of ‘trusting but verifying’ through triangulation, reflexivity, and team discussion was fundamental to managing subjectivity.

For researchers undertaking similar ambitious qualitative projects, especially those working independently or with limited team members, carefully considering the scope is crucial. Based on our experience, limiting the study to one or two primary aims with a focused set of research questions, rather than three aims, may be more manageable. Using visual tools throughout the analysis process also proved highly beneficial. Tools like sticky notes for initial coding and concept mapping features within qualitative data analysis software helped us to visualize connections between ideas, organize complex relationships between domains and themes, and avoid getting overwhelmed by the volume of data. These practical strategies, combined with rigorous scoping and consistent reflexivity, were key to conducting in-depth qualitative research effectively and efficiently.