The Qualitative Report

Volume 26 Number 6 Special Section 19

6-22-2021

Conducting Thematic Analysis with Qualitative Data

Chad R. Lochmiller
Indiana University, clochmil@indiana.edu

Follow this and additional works at: https://nsuworks.nova.edu/tqr

Part of the Quantitative, Qualitative, Comparative, and Historical Methodologies Commons, and the
Social Statistics Commons

Recommended APA Citation

Lochmiller, C. R. (2021). Conducting Thematic Analysis with Qualitative Data. The Qualitative Report,
26(6), 2029-2044. https://doi.org/10.46743/2160-3715/2021.5008

This Special Section is brought to you for free and open access by the The Qualitative Report at NSUWorks. It has
been accepted for inclusion in The Qualitative Report by an authorized administrator of NSUWorks. For more
information, please contact nsuworks@nova.edu.
Conducting Thematic Analysis with Qualitative Data

Abstract
This article discusses one approach to conducting thematic analysis using structured qualitative data
collected from focus groups. Thematic analysis is one of the most used but often poorly defined
approaches in the qualitative research community. The method is principally concerned with the
identification of patterns which are then reported as researcher-generated themes. In this article, I use
data obtained from the Qualitative Data Repository to demonstrate how secondary qualitative data can be
analyzed to produce themes. I note the ways in which this process unfolds as well as how it differs from
other techniques.

Keywords
qualitative research, research methods, thematic analysis

Creative Commons License

This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 4.0 International
License.

This special section is available in The Qualitative Report: https://nsuworks.nova.edu/tqr/vol26/iss6/19

The Qualitative Report 2021 Volume 26, Number 6, 2029-2044
https://doi.org/10.46743/2160-3715/2021.5008

Conducting Thematic Analysis with Qualitative Data

Chad R. Lochmiller
Indiana University, Bloomington, Indiana, USA

This article discusses one approach to conducting thematic analysis using

structured qualitative data collected from focus groups. Thematic analysis is one
of the most used but often poorly defined approaches in the qualitative research
community. The method is principally concerned with the identification of
patterns which are then reported as researcher-generated themes. In this article,
I use data obtained from the Qualitative Data Repository to demonstrate how
secondary qualitative data can be analyzed to produce themes. I note the ways
in which this process unfolds as well as how it differs from other techniques.

Keywords: qualitative research, research methods, thematic analysis

Qualitative research is a diverse field that employs a variety of analytic techniques to

produce an understanding of rich datasets. Among the more common techniques used by
qualitative researchers, thematic analysis involves the identification of recurring patterns that
are presented by researchers as overarching statements or themes. Webster’s Dictionary
conveniently defines a theme as a unifying or dominant idea, and this captures well the essence
of a theme in qualitative research. In this article, I discuss my approach to thematic analysis.
For this discussion, I draw from a common data set retrieved from the Qualitative Data
Repository at Syracuse University. In this article, I begin by discussing the literature related to
thematic analysis. Next, I describe how I typically undertake thematic analysis. My approach
may differ from other scholars who use this methodology or who have sought to describe this
methodology in the research literature. Indeed, for a contrasting perspective, scholars might
wish to review Braun and Clarke’s (2006) well-cited approach. Finally, I offer reflections on
the analysis I present which explain the affordances and limitations of thematic analysis.
In comparison with other analytic strategies used by qualitative researchers, thematic
analysis is one of the most ambiguously defined analytic techniques in the qualitative research
field. Scholars have reached limited agreement about how thematic analysis should be
performed (Attride-Stirling, 2001; Boyatzis, 1998; Tuckett, 2005). As Braun and Clarke (2006)
have previously asserted, many argue that thematic analysis is “poorly demarcated” and “rarely
acknowledged” in qualitative research (p. 77). Indeed, one need only review the methodology
sections of published research articles or qualitative dissertations to find that there is significant
ambiguity about how to identify themes from qualitative data. Braun and Clarke (2006) offer
what is perhaps one of the clearest definitions in the methodological literature. They define
thematic analysis as “a method for identifying, analyzing and reporting patterns (themes)
within data” (p. 79). This definition presumes that an analyst produces a generalized
understanding of coded data based on the recurring application of codes and the patterns
associated with those codes. The more frequently codes appear in the data set, the more likely
it is that the analyst will state that code as the basis of a theme. While thematic analysis may
consider coding frequencies in the production of themes, this analytic approach should not be
treated as a purely quantitative exercise. Rather, it often requires the analyst to move beyond
frequencies to ascertain the salience and meaning of the underlying dataset. This reflects what
is traditionally thought of as the iterative nature of qualitative research. Finally, though I treat
2030 The Qualitative Report 2021

thematic analysis as a discreet analytic approach in this article, it clearly has the capacity to
propel readers toward more sophisticated analytic techniques. Indeed, Braun and Clarke (2006)
have noted that this analytic approach serves as the foundation for many types of qualitative
analysis, which includes some types of content analysis (Krippendorff, 2018; Vaismoradi,
Turunen, & Bondas, 2013), ethnography (Hammersley, 2006) and thematic narrative analysis
(Riessman, 1993), among many others.

Orientation to Data and Research Participants

Broadly, thematic analysis is consistent with Denzin and Lincoln’s (2011)

conceptualization of qualitative research, which holds that “qualitative researchers study things
in their natural settings, attempting to make sense of, or to interpret, phenomena in terms of the
meanings people bring to them” (p. 3). When focused on interviews or focus groups, thematic
analysis presumes that the participant’s recollections have value that merits exploration,
synthesis and intensive description. When focused on documents or other retrieved artifacts,
thematic analysis presumes that the recorded information is an accurate reflection of the reality
that existed at that time and therefore should be treated with the same degree of confidence as
participant’s spoken recollections. This orientation presumes that the researcher sees the data
reported by the participants or retrieved in documents as convincing. Thematic analysis is thus
predicated upon a high degree of confidence about the reliability and trustworthiness of the
information reported. The aim of thematic analysis is, then, to consider how the reported
information addresses a specific research question or invites a new conceptual or theoretical
understanding. In doing so, thematic analysis speaks to the data in a large sense and less often
delves into the very detailed nuances of an individual’s experiences.
The analyst’s orientation to thematic analysis inherently depends on the specific
research questions they use to guide their study. In my view, thematic analysis is a
fundamentally question-driven exercise that depends on clearly articulated lines of inquiry to
frame the scholar’s interpretation of the data. Braun and Clarke (2006) observe that thematic
analysis can be oriented so that it “reports experiences, meanings and the reality of participants”
(p. 81). This has implications for the kinds of questions that an analyst might address using
thematic analysis and interpretation. These questions often begin with “what” or “why” and
less frequently include questions that begin with “how.” When oriented in this manner,
thematic analysis summarizes what the participants see as valuable and is principally concerned
with producing a descriptive account of the participants’ understanding. A thematic analysis
might focus on similarities or differences in repeated patterns found within the dataset. For
example, Dodson, Baker, and Bost (2019) conducted a qualitative research study of 10 nurse
practitioners who routinely prescribed medication within a clinical setting. The study sought to
assess perceptions of trustworthiness surrounding clinical decision support tools and mobile
applications. The authors produced four themes, which related to the use of pharmacists to help
with current prescribing practices, reliance on the electronic medical record (EMR) as a clinical
decision support tool, lack of mobile application use in clinical practice, and desire for
affordable and reliable mobile application resources. These themes described broadly the nurse
practitioners’ perceptions about the trustworthiness of clinical decision support tools and the
potential affordances of mobile applications.
Alternatively, Braun and Clarke (2006) note that thematic analysis can also be a
“constructionist method, which examines the ways in which events, realities, meanings,
experiences and so on are the effects of a range of discourses operating within society” (p. 81).
When framed in this way, the constructionist orientation seeks to apply the analyst’s
interpretation as rooted in a conceptual or theoretical framework. As such, the research moves
from simply describing what participants report to offering an interpretation of the patterns that
Chad R. Lochmiller 2031

the analyst has observed. Galvin, Suominen, Morgan, O’Connell, and Smith (2015) conducted
qualitative interviews with 12 mental health nursing students to explore how stress during their
training experience impacted them and to articulate what potentially caused their stress to
increase. The researchers identified “unreasonable demands” that arose during clinical blocks
as well as negative staff perceptions both related to increased stress among the 12 mental health
nursing students. Unlike the study conducted by Dodson, Baker, and Bost (2019), which largely
recounted participants’ experiences, this study sought to explain why a particular phenomenon
existed based on the repeated statements of participants in the study. As such, the study
provides an instructive example demonstrating how a theoretical idea can be introduced and
validated through the identification of repeated patterns in data.
Thematic analysis thus can be descriptive, explanatory, and/or critical in nature.
Thematic analysis enables scholars to define and describe what a participant’s reality is using
their own written or spoken account. This orientation summarizes what participants report and
therefore aggregates these understandings into identifiable patterns. As an explanatory tool,
thematic analysis can be used to infer meaning about experiences, perspectives, or belief
systems through the lens of a particular conceptual or theoretical framework. This approach
involves considering how the patterns found within data depict particular conceptual or
theoretical ideas. This approach requires the analyst to match patterns to a specific theoretical
or conceptual explanation. Thus, it is often undertaken when the analyst’s coding scheme is
defined a priori. Finally, as a critical analytic tool, thematic analysis can be used to identify
persistent gaps in the reported experiences of participants. These gaps are often representative
of recurring patterns which point to the existence of oppression, discrimination, or an
imbalance of power. The themes produced thus seek to elevate these patterns for the purpose
of scholarly interrogation and to potentially address these issues in practice, policy, or through
the development of enhanced social awareness.

Components of Thematic Analysis

Although scholars approach thematic analysis differently, I have found that this analytic
technique customarily includes three components: individual codes, categories, and researcher-
produced themes. Codes and coding practices have been described at length in the qualitative
methods literature (Saldaña, 2015). Codes are, quite simply, “a word or short phrase that
symbolically assigns a summative, salient, essence-capturing, and/or evocative attribute for a
portion of language-based or visual data” (Saldaña, 2015, p. 4). Codes might be applied to a
single word, sentence, paragraph, or visualized portion of qualitative data. This data might
include interview transcripts, observational notes, researcher- or participant-generated journals
or notebooks, documents, open-ended survey responses, images (e.g., photographs or
drawings), video, websites or blog posts, written correspondence (e.g., emails, letters, etc.), or
published academic literature (e.g., books, articles, technical reports). Codes thus connect the
researcher’s analysis with the data s/he collected (Charmaz & Mitchell, 2001). In thematic
analysis, codes have a particularly powerful role as it is from individual coded segments that a
researcher’s identification of patterns begins. Individual codes often produce a sense of the data
that then informs how the analyst assigns value to different perspectives, experiences, or
recollections. Typically, in my own research, I tend to code large sections of text as opposed
to individual words. This tends to capture the essence of the experience that the participant
reported.
In my view, thematic statements ultimately depend on the development of robust
categories as codes are insufficient to produce themes. As Riessman (2011) observed, coding
causes “detail and specificity to slip away in favor of general statements about the phenomenon
of interest” (p. 311). Coding means that interviews and other data sources are “fractured” into
2032 The Qualitative Report 2021

smaller analytic bits that become the basis for “generalizations about human processes that
hold across individual participants” (Riessman, 2011, p. 311). Indeed, as Saldaña (2015)
observed, “a code can sometimes summarize, distill, or condense data, not simply reduce them”
(p. 5, emphasis in original). Thematic analysis is thus dependent upon re-assembling these
fragments into constructs that contribute to the larger theme. This re-assembly is encapsulated
in the identification of robust categories that speak to the essence of the underlying codes.
These categories define (un)related and (dis)similar patterns in codes that the efficiency of
themes do not allow a researcher to report. Categories thus represent the first step toward the
identification of patterns. Saldaña (2015) defines a pattern as “repetitive, regular, or consistent
occurrences of action/data that appear more than twice” (p. 5). Patterns explain
(inter)relationships in ways that allow researchers to state commonalities across their
observations. These patterns serve as the basis for themes and are operationalized within the
data analysis as categories. Categories represent emergent patterns that the analyst aggregates
to develop thematic statements. The aggregation allows the analyst to put different pieces of
the data into conversation with one another. For instance, two categories which point to a
discrepancy might be encapsulated in a thematic statement as conflict or differentiation.
Before proceeding, it bears noting that the function of categories is one of the more
opaque aspects of analysis and it is especially so in thematic analysis. For some, categories
may feel like an unnecessary step. If patterns exist at the coding level, why bother establishing
categorical descriptions since they are ultimately encapsulated within broader thematic
statements? Since they are not visible to the reader, why are these important? Certainly, these
are valid questions and they merit consideration by the analyst. In thematic analysis, it is
certainly possible for researchers to identify and report on patterns derived from individual
codes. This seems particularly common in content analysis (Vaismoradi et al., 2013), where
individual coding frequencies ultimately yield the production of findings. However, in case
study research or ethnography, it is my opinion that this approach weakens the development of
robust themes as it reduces the analyst’s ability to “build up” the theme in a transparent and
meaningful way. In effect, it leaves an element of mystery between the individual codes and
the themes. In doing so, it raises legitimate questions about the transparency of the analysis and
the trustworthiness of the analyst’s claims. For instance, how do you know that the theme refers
to more than an isolated case or single example drawn from the data?
Thus, in my view, developing categories is an essential step in the development of a
theme as they contribute to the analysts’ ability to communicate the substance of the theme.
Namely, they help scholars articulate the various theoretical or conceptual assumptions about
the data that they made during the analytic process. In my own work, I use categories to
populate sub-headings to walk the reader through my interpretation of the data so that they
understand how I produced my themes. This practice allows the reader to follow my
interpretations as well as see how the data relates. I find this also supports the writing process
as the researcher can construct sections of the manuscript/final research report around
underlying categories which provides a logical and clear flow.
Finally, thematic analysis involves as a final step the production of broad overarching
statements that describe what is happening in the underlying data. Thematic statements embody
the same qualities as a well-known song in that they describe the essence of the data in a
compelling way. For example, “enabling actions in healthcare systems for improved patient
care” might be offered as a theme describing a combination of actions taken in a healthcare
system to improve the quality of patient care. Hypothetically, it would follow that categories
such as “frontline nursing practices,” “physician’s bedside relationships,” and “quality of
communication” might broadly support this theme and, as mentioned in the previous
paragraph, be used to build-up the theme so that the reader understands how it was produced.
Themes thus serve as an invitation to explore more deeply a particular aspect of the dataset.
Chad R. Lochmiller 2033

Data Sources

To complete the analysis presented in this article, I used a shared dataset from a research
study that explored postnatal care referral behavior by Traditional Birth Attendants (TBAs) in
Nigeria, including the perceived factors that may deter or promote referrals to skilled health
workers. I acquired the data from the Qualitative Data Repository at Syracuse University. The
dataset was provided to me and I was not part of the original study’s conceptualization or data
collection. The data includes three transcripts from focus groups conducted with hospital-based
health workers, TBAs, and Nigerian women who delivered their child at home with the
assistance of a TBA. The focus groups were conducted at a federal teaching hospital and range
in length from 64 to 87 minutes. A research team conducted each of the focus groups. I did not
participate as a member of the original research team. Indeed, a significant limitation of this
exercise is that I was unaware of the motivation for the study.
Importantly, the focus groups employed a structured format wherein researchers asked
participants a series of open-ended questions followed by more precisely focused probes. This
format works especially well in a thematic analysis as the structure of the conversation lends
itself to grouping ideas, concepts, or issues more easily. This structure also allows researchers
to compare responses across participant perspectives. For example, in this data, I found it
helpful to compare the responses from health workers, TBAs, and Nigerian women to
determine how they collectively and individually identified differences in the quality of care
provided in hospital- and home-based settings.

Analytic Process

For this illustrative example, my approach involved three phases: setup, analysis, and
interpretation. I discuss each of the phases below providing examples from the same dataset
where appropriate. Notably, these phases subsume some of the elements of Braun and Clarke’s
(2006) approach for both simplicity of illustration, as well as to better support novice
researchers. They also mirror the process proposed in the introductory research text by
Lochmiller and Lester (2017).

Setup Phase

Before completing a thematic analysis, I always invest considerable time setting up my

data for the analytic efforts I plan to undertake. This practice involves inventorying the dataset
in preparation for analysis as well as setting up files so that they are well-organized in my
computer assisted qualitative data analysis software (CAQDAS) package. I begin by
transcribing audio or video files, deleting identifiable references, and uploading these files to a
CAQDAS package. At this stage, I also convert my field notes to Microsoft Word (if they were
previously handwritten) and scan any paper-based artifacts to Adobe PDF file format. I also
download and save any Adobe PDF documents from research participants or online searches
that I intent to use in my analysis. Finally, I adopt a consistent file naming protocol for each
file and record the information about each file in a Google or Microsoft Excel spreadsheet. I
use this spreadsheet to track the file name, and include a brief description of the data, variables
pertinent to the data, and the date I collected the information. For interviews or focus groups, I
always record the length of the conversation and the number of participants involved. While
this step may seem trivial on a dataset with only three focus group transcripts, it is especially
important in a study with a large dataset that includes multiple interviews or focus groups.
Indeed, I have found in my own practice that a poorly organized dataset slows the analytic
process and creates more work in the end.
2034 The Qualitative Report 2021

An important consideration in the setup phase is whether to establish all of the codes
before the analysis begins. Thematic analysis does not assume that the analyst fully develops
his or her coding scheme. In my own practice, I tend to make this decision based on the
theoretical or conceptual assumptions that guide the study as well as the complexity of the
dataset. In complex studies involving large amounts of data, I typically establish a simple
descriptive coding scheme that helps organize the dataset by participant, setting, and other
important characteristics. For instance, in this illustrative study, I established descriptive codes
that noted the participant’s role (e.g., TBA, Healthcare Worker, Patient, Researcher). I also
created descriptive codes that linked to the topics for the questions (e.g., Methods of Treatment,
Quality of Care, etc.). These codes helped prepare the dataset to make comparisons by
participant perspective. Admittedly, this was a relatively simplistic dataset and so did not
require much preparation or setup. In a more complicated undertaking, this process might
involve significantly more effort. This approach was also appropriate as the present study did
not use an explicit theoretical or conceptual framework. In more complex studies, I would have
included more robust descriptive codes that would have allowed more precise and in-depth
comparisons.
In studies where an explicit conceptual or theoretical framework is used, I often develop
a coding scheme that relates to concepts in the literature. For instance, had this study considered
how TBAs and health workers learn from each other, I could have framed this study using
concepts from organizational theory, such as a community of practice (Lave & Wenger, 1991;
Wenger, 1998; Wenger, McDermott, & Synder, 2002). This concept, which has been applied
to health research (Li et al., 2009), would have allowed me to explore how TBAs and health
workers engage each other for the purpose of learning and socialization that occurs between
novice and experienced health providers. This framework would have allowed me to identify
new health strategies that health workers and TBAs adopted as they became familiar with the
normative expectations of the local health community. The coding schemes would have
focused on behaviors associated with participation in a community of practice, such as situated
learning (Lave & Wenger, 1991) or identity development (Wenger, 1998). Alternatively, I
could have focused on the roles that members of a health community assumed, such as leaders,
champions, or facilitators (Wenger et al., 2002). This would have shifted my focus to exploring
the roles that health workers and TBAs assume relative to their patients. Ultimately, the themes
produced through the analysis would have referenced theoretical or conceptual ideas drawn
from this perspective.

Analysis Phase

After setting up the data, thematic analysis begins by reviewing the dataset to become
familiar with the central ideas, concepts, or experiences described by research participants,
recorded in observation notes, or referenced in artifacts. Braun and Clarke (2006) describe
thematic analysis using a model that includes six phases. While their approach certainly
informs my own, it does not completely define it nor does it fully inform how I approach
thematic analysis.
My approach begins by familiarizing myself with data. This represents the first phase
of the approach by Braun and Clarke (2006). At this stage, I seek to become familiar with the
data but try to avoid imposing an interpretation on the dataset. To this end, I often make
informal jottings in the margins or use the memo function in my CAQDAS package to capture
any initial reflections about the participants, data, or circumstances that shaped the completion
of the interview or focus group. In some places, I may highlight specific words, phrases, or
participant recollections that I feel could be potentially responsive to my research questions
Chad R. Lochmiller 2035

and therefore worth noting for later analysis. It is important to ask basic descriptive questions
at this stage in order to become familiar with the dataset. These questions might include:

• What is happening?
• What is the participating saying?
• What key points or ideas are they expressing?
• What points do they appear to agree or disagree about?
• What perspectives are (dis)similar?
• What experiences do they hold in common?
• What experiences are being described?

Familiarity with the dataset is key when conducting a thematic analysis as it allows the
researcher to begin identifying initial patterns that may serve as the basis for categories. Thus,
I use these descriptive questions to familiarize myself with the data and try to avoid making
unnecessary inferences about the participants’ perspective or ideas.
Similar to Braun and Clarke (2006), my next step is to develop an initial set of codes.
Saldaña (2015) refers to this as first cycle coding. In thematic analysis, the first cycle of coding
is designed to partially reduce the data set to identify initial ideas that may be further developed
or excluded as the analysis proceeds and as patterns become more or less visible. Thus, a
portion of this work might have already been completed when setting up the dataset. Unlike
other approaches to qualitative research, the aim of thematic analysis is rarely to code word-
by-word or line-by-line. Rather, thematic analysis seeks to discern the gestalt of the data and
therefore examines (un)related and (dis)confirming points or perspectives as evidenced in
coding patterns. The aim of the first cycle of coding is to identify meaningful and potentially
relevant passages of text that will be analyzed further in subsequent coding cycles. It is
important to recognize that not all passages will be equally valuable, indeed some may not be
considered as the analysis unfolds. In the sample dataset, I included descriptive codes such as
“health risks,” “example of pain,” “hospital practice,” “TBA practice,” “financial cost,” and
“use of medication” among others. These codes describe experiences, perspectives, or
reflections offered by the focus group participants and thus begin to illuminate their primary
concerns.
The third phase of Braun and Clarke’s (2006) model emphasizes the search for themes.
Their particular description does not fully align with my own approach. Indeed, in my own
practice, I rarely find themes at this stage and so recommend that researchers progress through
a second cycle of coding as recommended by Saldaña (2015). In the second coding cycle, I
typically focus on previously coded passages and strive to produce categories that broadly–but
still incompletely–define related portion of the dataset. In this cycle, I pay particular attention
to codes that appear (un)related as their relationship (or lack of it) serves as the basis for the
identification of patterns. For example, in the current data set, I noted that “roles” and “provider
authority” appeared somewhat frequently in the focus groups transcripts obtained from health
workers and TBAs. Upon further review, I found that participants often discussed their
respective status, boundaries of their professional knowledge, and their perceived authority to
provide care in specific (usually high risk) situations. Given the frequency of the pattern, as
well as the number of references, I opted to produce a category at this stage which referred to
these segments as “authority to provide care.” Thus, in the sample dataset, I assigned this
categorically related code to any statement that appeared related. I use these codes to identify
and select salient passages or quotes for inclusion in the final report. I generally repeat this
process until I have identified a sufficient number of categories to be able to articulate a theme.
Indeed, at this stage, I enter the final step in my analysis, which involves reviewing possible
themes and beginning to articulate them in a meaningful way.
2036 The Qualitative Report 2021

Interpretive Phase

The final step in my analysis process involves identifying (dis)connections across

categories in order to produce themes. I call this the interpretive phase because it describes a
much more intentional and sophisticated kind of analytic work. Broadly, Braun and Clarke
(2006) refer to this as phases four and five of their model and it essentially describes the
researcher’s efforts to review, define, and name themes. To be fully transparent, I acknowledge
that there are no hard rules about this process nor guidance that can inform the decision-making
process. Rather, what typically guides my thinking at this stage is the extent to which a theme
is substantiated by categories as I find that a sufficient number of categories ensure that the
theme can be “built-up” for the reader. The questions I ask at this stage often include:

• How do the categories support the development of the theme?

• To what extent is the theme supported by the perspectives of multiple participants?
• What areas of agreement, disagreement, similarity, or difference does the theme
include?
• Which quotations or examples offer the most compelling support for the theme?
• How should I substantiate the theme using quotations and/or primary data sources?

While not comprehensive, these questions provide opportunities to explore the

relationships that I find most commonly result in the production of a theme. In my own work,
I typically report themes with a large number of categories in the final manuscript as these
categories allow me to demonstrate to the reader how the theme came to exist and to illuminate
any analytic decisions that led to the production of the thematic statement. At this stage of the
analysis, I also refer back to my research questions to determine how the theme(s) respond (or
not) to them.

Illustrating the Product of Thematic Analysis with Findings

To effectively communicate the results of a thematic analysis, it is important to

structure findings in relation to themes and to substantiate these statements using examples
drawn from the dataset. As shown in Figure 1, the presentation of data includes a combination
of statements referring to specific themes and a series of related categories and/or sub-
categories derived from the dataset. Thematic statements capture the essence of the data, while
categories offer greater detail from the participant’s observations or reflections. This structure
has two distinct advantages. First, it allows the empirical argument to be developed around the
themes reported from the data. Second, it produces a robust understanding of the dataset by
demonstrating how various underlying data elements relate to the theme(s). Below, I illustrate
the structure of thematic findings using the results of my analysis of the shared dataset.
In the sample dataset, I found one prominent theme across each of the three focus group
transcripts. This theme, which I describe as “Defining a Standard of Maternity Care,” refers to
the ways in which the participants each described their own perspectives about the maternity
care that they felt most appropriate or valuable given their perspectives. This theme broadly
captures the discussion presented in the three focus groups and explains the relationships
among the participants’ perspectives. Related to this theme, I found that much of the discussion
focused on the importance of clinical practice, descriptions of the care process, and a
discussion about provider roles and authority. These perspectives were especially prominent
among health workers and TBAs; yet, they also corresponded with the statements offered by
women who participated in a focus group. To these categorical descriptors, I also noted some
underlying points that appeared frequently and thus served as sub-categories that merit
Chad R. Lochmiller 2037

particular discussion. This included, for example, orientation to patient and access to treatment
as an important part of the care process. I also noted that provider boundaries, methods of
treatment, and opportunities for collaboration appeared to be centrally related to the discussion
of provider roles and authority.

Figure 1
Illustration of How Categories Shape a Theme

Drawing from the structure described above, I next present an illustrative reporting of
themes derived from the shared dataset. This section illustrates how I typically approach the
development of a findings section and may be especially instructive to researchers who are
unfamiliar with how to produce and share findings produced from a thematic analysis. For this
exercise, I used a single research question to inform my interpretation: How do health workers,
traditional birth attendants, and patients describe maternity care in different medical settings?

Theme 1. Defining a Standard of Maternity Care

Participants in each of the focus groups discussed the care women required for safe
childbirth and how care should be provided given their respective clinical settings. This
conversation invited negotiation between the focus group participants about an appropriate
standard of care that reflected the woman’s specific health needs, issues that arose before,
during, and after childbirth, as well as the provider’s own professional perspectives. Broadly
stated, the data suggest that participants were attempting to define a standard of maternity care
that was appropriate given the local health conditions and individual patient needs. Healthcare
workers and TBAs both described the importance of prompt and thorough medical care during
childbirth, though differed in their individual assessment about when such care could be
provided outside an institutionalized medical setting, such as a regional hospital or medical
clinical. Responses within the focus groups suggest that all participants were concerned about
how to provide medical care to women who experienced complications during their pregnancy
or immediately following childbirth, though the responses differed in their emphasis on the
importance of pre-natal and post-natal care. Indeed, I found that much of the discussion focused
2038 The Qualitative Report 2021

on the quality of care as defined by clinical practice or resources available in the medical
setting. This included how the provider oriented to the patient and what level of access they
had to emergency medical treatment. In addition, participants also discussed different provider
roles and authority noting that this influenced the methods of treatment available and suggests
the need for cross-provider collaboration to ensure improved health outcomes. I discuss each
of these points below.

Comparisons of Provider Quality and Healthcare Resources

The women who participated in the focus groups discussed the quality of care provided
by the health center and the resources available in a comprehensive medical setting using
explicit comparisons drawn from their experience. The women generally saw the quality of
care provided by the health center as being superior to the quality of care provided by a TBA,
though their assessment was not universal. At times, women disagreed that TBAs could not
provide care that was as effective or safe. Women who saw care provided at the health center
as being superior to the TBA based their perspectives on access to advanced medical services
and rationalized the quality of care based on the potential dangers associated with childbirth
(e.g., bleeding, infection swelling legs, etc.). As one woman reflected, “It is better to go to the
health center because they will take good care of you.” Her assessment was based on the health
center’s ability to provide safe pre-natal and post-natal medical services. Another woman
offered a stronger rationalization noting that women who received care at the health center had
access to better medical equipment. She stated, “It is good to go to the health center because
they can take good care of you more than the TBA, those women do not have all the equipment
to take care of any problem that arises.” The comments collectively suggest that the quality of
care was related to the perceived quality of the health center’s infrastructure and resources.
Indeed, access to medical equipment appeared to be a significant factor that explained whether
the women saw the care provided by a TBA as safe.
Healthcare workers offered a more nuanced description of the differences in the quality
of care, with many emphasizing the availability of sterile care practices as a predictor of
successful patient outcomes. Health workers repeatedly observed that they did not see TBAs
employing sterile practices during pre-natal, delivery, and post-natal interactions with women.
For instance, one healthcare worker observed that TBAs did not use gloves during delivery or
sterile razors. As she reported, “There was a time they were sharing mama kits and delivery
kits,” a reference to medical supplies used during childbirth that should not be shared from
patient to patient due to the potential for infection, including HIV. Another health worker
observed that TBAs should be given clear guidance from health workers about appropriate
sterile practices. As she noted, “they should use marking touch for the women during delivery
and not to use carton or mat, we should encourage them to use gloves and to maintain general
hygiene during delivery. They should use Dettol and Jik in sterilizing the place they use for the
delivery.” For healthcare workers, infection control was thus viewed as a significant marker of
quality of healthcare, particularly for women who were HIV-positive at the time of their
delivery. Indeed, healthcare workers observed that TBAs often did not exercise the appropriate
precautions to prevent spreading HIV from mother to baby. In one instance, a healthcare
worker found that a baby was positive for HIV while the mother and father were both negative.
The healthcare worker reported that she “found out that the baby was contaminated at the place
of delivery, for there were actually two women that delivered there that same day.” As the
healthcare worker learned, a “TBA used the same tools she used for the HIV positive mother
to deliver the other woman, and the baby contaminated it from there.”
Both health workers and TBAs commented about the availability of drugs and
immunizations as another resource that highlighted a difference in the quality of care provided.
Chad R. Lochmiller 2039

As one TBA reported, “You cannot compare the care and services of a health facility with ours,
because the drugs given to the women there have dose, say for morning, afternoon and night.”
In contrast, the TBA described herbs that she used with women who could be “given a full cup
of the concoction to finish, and you cannot tell if that will make her loose blood or gain blood.”
While coming from different perspectives, both health workers and TBAs pointed to the
availability of drugs and immunizations in the healthcare facility as one reason the care
provided in that setting might be perceived as superior to home-based care provided by a TBA.
Indeed, health workers openly observed that having access to drugs (i.e., antibiotics and pain
medications) and immunizations was an important resource for caring for women during and
after childbirth. They cited emergency situations where access to these resources improved the
chances of both a successful birth and successful recovery for the patient.
The TBAs offered a different perspective about the reason(s) that a woman might
pursue care in a health center, emphasizing the degree of risk associated with the woman’s
pregnancy. Similar to the health workers, the TBAs placed particular emphasis on accessing
health centers when their own resources were insufficient to address a woman’s health needs.
The TBAs appeared acutely aware that the quality of care could not be compared. As one noted,
“you cannot compare the services of the health center with that of the TBAs, because the TBAs
use their bare hands, but the health center has the facilities and equipment needed to provide
adequate services for the women.” Another TBA reported that she will not “make the woman
experience the pains of labor in my house when I see that things are getting complicated, that
is just suffering the woman.” Consideration for the patient’s health was a major consideration
for TBAs. They associated high risk deliveries with a greater likelihood of severe
complications and thus sought to reduce these circumstances by engaging the support of the
health center. Their responses suggest that the TBA defined the quality of care they could
provide in relation to the risk or potential risk of the patient’s health condition.

Access to Treatment Resources

The participants noted the importance of access to treatment resources as a factor in

determining whether they considered the care they received to be of high quality. These
resources included sterile supplies, immunizations, antibiotics, and pain medications in
addition clinical facilities that allowed the prompt intervention in case of a medical emergency.
These were viewed as particularly important because they increased the likelihood of a
successful delivery with fewer complications. As one woman noted, “It is good to go to the
health center because they can take good care of you more than the TBA, those women do not
have all the equipment to take care of any problem that arises.” Another woman noted that,
“There is this pain that women usually have after delivery, it is necessary that a woman goes
there to be injected as that pain can lead to bleeding.” In both statements, the women pointed
to the resources available at the health center as a primary marker for the quality of care that
they would receive. This perspective appeared to align with the views of health workers who
participated in the focus groups, as well. As one health worker noted,

It is important that the women come [to the health center] for postnatal care
because there are some conditions that the women will be into that will require
the attention of the health workers in the health facility. There was this woman
that came for antenatal care in the health center but she later went to the TBA
to deliver, she later had some complications there and they brought her here,
after examining her she was looking so pale and had to blood. So, I had to refer
her out, when they got there, they had to do blood transfusion immediately. If
not for that, the woman would have lost her life like that.
2040 The Qualitative Report 2021

The health worker’s comments are illustrative and point to the legitimate health risks that a
health center or hospital is equipped to address. Further, it provides evidence about the real
risks that many women face during childbirth. One of the main reasons that care provided by
the health center was viewed as being high quality was that it offered services that appeared to
the women and healthcare workers to minimize risks to the mother and baby.

Personalized and Responsive Practice

Although the health center infrastructure was an important quality consideration for all
of the participants, the women collectively spoke to the importance of relationships as a marker
for quality care. For women who viewed relationships as particularly important, many spoke
to the importance of securing care from a TBA. Indeed, the women spoke about the fear and
unnecessary trauma associated with receiving care from a health center. Their statements
suggested that care within this setting was less personalized than care provided in their home
and the absence of such a connection contributed to their fear of childbirth. The women
collectively perceived that the hospital staff paid less attention to their patients and found that
their time was often distributed to other patients. This understanding seemed linked to the
degree to which a provider offered one-on-one attention to the patient during all phases of the
childbirth experience. As one woman reflected, “The reason most women do not like going to
the hospital is that a health worker there can walk pass you even while you are there shouting
in pains.” Another woman observed that the health faculty “will leave you in the pains and will
pay attention to you only when the baby has engaged.” This approach led the women to view
the hospital as lacking compassion. As one observed, she felt that the hospital allowed women
to have “unnecessary tears” when experiencing childbirth. This perspective represented one of
the implicit critiques of hospital-based care.
Many of the women suggested that the hospital’s approach differed from the care
provided by a TBA. The TBA was seen as being more responsive to a woman and better
equipped to calm her during childbirth. As one woman reported, “the TBA will hold you and
console you while in pains till you deliver your baby.” Another woman stated that the TBA
would “pet the woman” during childbirth. This personalized care was viewed as an important
difference between health centers and TBAs in that it helped the women feel “confident that
we will deliver safely.” Indeed, in relation to the standard of care expected by women, it seemed
that personalized attention and individual support was an essential standard that the hospital
and health workers were unable to match. While these were expressed concerns by the women,
the health workers did not share this perspective. Indeed, according to them, the hospital was
well-suited to provide compassionate and individualized care but within the context of a more
sterile and medically skilled setting.

Provider Roles and Medical Authority

One of the questions that appeared often in the focus groups related to which providers
had authority to decide what maternity care for the women should be. Indeed, I noted that
categories often related the position of health care workers to the position of TBAs. This
discussion was less substantial than the discussion focused on healthcare quality but did suggest
that the participants perceived that health workers and TBAs had differing degrees of authority
in the medical community. Focus group participants seemed to position health care workers as
having greater authority with regard to their ability to support safe childbirth practices as well
as provide adequate medical guidance to mothers. As such, I found that these workers were
often afforded more respect by participants in the study and tended to be viewed as having
Chad R. Lochmiller 2041

more specialized knowledge about childbirth and maternal health. I found that much of the
discussion focused on the superior knowledge and skills of the healthcare providers, as well as
the more advanced methods of treatment.

Emphasis on the Importance of Knowledge and Skills

Health workers emphasized the importance of medical skill frequently and suggested
that TBAs had less knowledge than a health worker. This description seemed grounded in
references to the kinds of practices that a health worker believed, based on their training, led
to superior health outcomes for the mother and baby. For example, health workers stressed the
need for pre-natal and post-natal care as well as the kind of clinical practices used during
childbirth. As one health worker noted, “I am of the opinion that [the women] should receive
post-natal care at least for twenty-four hours so as to be sure that they are strong. We should
not be in a haste to send them home. It is also good that the TBAs know this.” The health
worker described a situation in which a woman who had recently given birth collapsed after
being allowed to walk without assistance. Another health worker observed,

I am also of the opinion that the women should not leave immediately after
delivery, because I have also had a personal experience in my health center, the
woman delivered at a TBA's home but the baby’s navel was not protected. So,
there was a woman that went to see the new baby and observed that the baby
was not looking alright. When the baby was examined, they found out that the
baby's navel was not well protected, and they had to come to meet us at the
health center, before they could get to us, the baby died.

In both instances, the health worker privileged their opinion and cited best practices that the
TBA was not following. Many of their comments were often rooted in worst case scenarios
that described situations in which best practices had not been followed and thus the woman or
her baby became seriously ill. The TBAs largely deferred to the health workers as having more
significant training and therefore more knowledge about certain health situations. As one TBA
noted, “if there is something that is confusing, or that I don't understand, I will take the woman
to the health facility so that a health worker will examine her and say where the problem is
from.”

Methods of Advanced Treatment

A secondary pattern in the data highlighted differences in methods of treatment used

by health workers and TBAs. Health workers noted important practices in each phase of
childbirth and cited examples demonstrating that TBAs did not consistently adhere to these
practices. As one health worker noted, “the TBA is supposed to tell the woman to go and
immunize her child after two or three days of delivery, then she can come back for checkup
after six weeks of delivery.” This practice seemed to be a common expectation among health
workers. At the postnatal visit, health workers described conducting a routine assessment of
the woman’s health. As one described, this visit includes “a checkup on their body system and
also [an examination of] their physical appearance to know if she is healthy.” During this visit,
the health worker will also, “listen to her to know if she has any complaints concerning her
body… [and] advise her on family planning, if she is interested.” Another health worker offered
a more expansive discussion of a postnatal visit. They noted that the postnatal visit. According
to this health worker,
2042 The Qualitative Report 2021

we call them inside the facility, place them on the couch, examine their eyes
and palpate their tummy and ask them whether they are having pains there or
on any part of their body. You ask them other questions regarding their health
generally. The woman will tell you certain things in the course of the discussion.
You can also ask her about the family and how they are faring.

This quote demonstrates the basic clinical practice that health workers suggested healthcare
providers should follow when assessing a woman’s health in the clinical setting. As such, the
quote demonstrates what health workers believed should occur and therefore illustrated what
was likely not occurring in conversations between TBAs and their patients.

Reflecting on the Structure of a Thematic Analysis

The sections above represent my basic approach to producing a thematic analysis and
provides an illustrative example of how I craft findings. As the example illustrates, the hallmark
of an effectively developed thematic analysis involves making clear linkages between the
overarching theme (e.g., Importance of Clinical Practice) and secondary themes that
substantiate, strengthen, and further illuminate it. In my own research, I use this common
structure and argue that it should be used by other qualitative researchers to help guide their
readers through the various nuances of a theme. Notably, a mistake common to conducting a
thematic analysis is to present each theme without substantial depth. Such an approach misses
an opportunity to educate the reader about the construction of each theme and to make explicit
connections to the underlying data. While the sample above is limited in scope (i.e., it only
included three focus groups), it still demonstrates how I typically approach connecting
participants’ perspectives, as well as seek to represent areas of potential convergence and
divergence in the data.
The short illustration also demonstrates how participants’ roles or perspectives can be
used to enhance thematic interpretation. I use a similar approach when using interview data or
primary documents. When developing a theme, I seek to illustrate themes using quotations that
highlight areas of similarity or potentially demonstrate areas of disagreement or divergence. I
find that quotes that are similar reinforce the strength of the theme. This is especially helpful
when seeking to argue in favor of a particular perspective. I tend to point to areas of
disagreement or divergence when a theme is intended to illuminate an area of disagreement
among participants. This strategy is particularly helpful when coding suggests inconsistent
agreement among participants or when categories suggest that there are differing
understandings about an issue, concern, and/or experience. In the illustrative example above, I
found this approach particularly helpful in unpacking the differences between health workers
and TBAs about the conditions of the health facility and practices that should be used to care
for women.
A final, but important point, reflects thematic analysis’ ability to potentially reduce
voluminous data into clearly articulated thematic statements. Themes resonate with readers in
much the same way that a song does with an avid radio listener. It evokes a particular way of
thinking about the data and allows readers to recall what has been presented and its relation to
their own practice. While other types of qualitative analysis may offer greater depth or
opportunities for theoretical generation, I find that thematic analysis is particularly useful when
a scholar seeks to define a sense of the data in broad strokes and/or to communicate the gestalt
of the dataset to the reader. In both of the nursing examples I described above, the readers can
recall nurse practitioners’ perspectives about the use of technology (Dodson et al., 2019), as
well as the stressful experiences of nursing students (Galvin et al., 2015). Indeed, thematic
Chad R. Lochmiller 2043

analysis’ emphasis on broad understanding make it a useful tool for summarizing qualitative
data.
While thematic analysis is useful, it does have important limitations. First, when done
without rigorous coding or careful consideration of underlying categories, thematic analysis
can give the appearance that the researcher engaged lightly in their analytic effort. There are,
unfortunately, many examples of thematic analysis in the published literature which serve to
illustrate this point. These studies often lack compelling evidence that the researcher has
developed the theme from underlying categories and/or cannot demonstrate how they made
specific connections to the data. Second, thematic analysis is very dependent on well-
formulated research questions. When conducting a thematic analysis, it is important that a
researcher’s questions give opportunities to refer to specific patterns in the data. In my work,
this often means developing “what” or “how” research questions. Such questions prompt
description and/or elaboration of underlying patterns and thus draw upon thematic analysis’
strengths. Third, making explicit connections to a theoretical or conceptual framework can be
difficult in a thematic analysis. In my own work, I prefer to use an a priori coding scheme when
my research involves a particular framework, as this approach allows me to consider how
conceptual or theoretical ideas might “speak” to patterns identified in the data.
In sum, while thematic analysis has some important limitations, it remains a very
straightforward method for identifying patterns in qualitative data and producing themes that
describe them. Indeed, for practice-based researchers, as well as those who are new to
qualitative research, this approach is one that can be undertaken without significant training.
For novice researchers, thematic analysis is often the first step toward learning about and using
more sophisticated analytic methods (Braun & Clarke, 2006). Indeed, it serves as an important
foundation for understanding other more sophisticated qualitative analytic techniques (Lester
et al., 2020).

References

Attride-Stirling, J. (2001). Thematic networks: An analytic tool for qualitative research.

Qualitative Research, 1(3), 385-405.
Boyatzis, R. E. (1998). Transforming qualitative information: Thematic analysis and code
development. Sage.
Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research
in Psychology, 3(2), 77-101.
Charmaz, K., & Mitchell, R. G. (2001). Grounded theory in ethnography. In P. Atkinson, A.
Coffey, S. Delamont, J. Lofland, & L. Lofland (Eds.), Handbook of ethnography (pp.
160-174). Sage.
Denzin, N. K., & Lincoln, Y. S. (2011), The discipline and practice of qualitative research. In
N. K. Denzin & Y. S. Lincoln (Eds.), The Sage handbook of qualitative research (pp.
1-19). Sage.
Dodson, C. H., Baker, E., & Bost, K. (2019). Thematic analysis of nurse practitioners use of
clinical decision support tools and clinical mobile apps for prescriptive purposes.
Journal of the American Association of Nurse Practitioners, 31(9), 522-526.
Galvin, J., Suominen, E., Morgan, C., O’Connell, E. J., & Smith, A. P. (2015). Mental health
nursing students’ experiences of stress during training: A thematic analysis of
qualitative interviews. Journal of Psychiatric Mental Health Nursing, 22(10), 773-783.
Hammersley, M. (2006). Ethnography: Problems and prospects. Ethnography and Education,
1(1), 3-14.
Krippendorff, K. (2018). Content analysis: An introduction to its methodology. Sage.
Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation.
2044 The Qualitative Report 2021

Cambridge University Press.

Lester, J. N., Cho, Y., & Lochmiller, C. R. (2020). Learning to do qualitative data analysis: A
starting point. Human Resource Development Review, 19(1), 94-106.
Li, L. C., Grimshaw, J. M., Nielsen, C., Judd, M., Coyte, P. C., & Graham, I. D. (2009).
Evolution of Wenger's concept of community of practice. Implementation Science,
4(1), 11.
Lochmiller, C. R., & Lester, J. N. (2017). An introduction to educational research: Connecting
methods to practice. Sage.
Riessman, C. K. (1993). Narrative analysis. Qualitative Research Methods, 30. Sage.
Riessman, C. K. (2011). What’s different about narrative inquiry? Cases, categories, and
contexts. In D. Silverman (Ed.), Qualitative research (3rd ed., pp. 310-330). Sage.
Saldaña, J. (2015). The coding manual for qualitative researchers. Sage.
Tuckett, A. G. (2005). Applying thematic analysis theory to practice: A researcher’s
experience. Contemporary Nurse, 19(1-2), 75-87.
Vaismoradi, M., Turunen, H., & Bondas, T. (2013). Content analysis and thematic analysis:
Implications for conducting a qualitative descriptive study. Nursing & Health Sciences,
15(3), 398-405.
Wenger, E. (1998). Communities of practice: Learning as a social system. Systems Thinker,
9(5), 2-3.
Wenger, E., McDermott, R. A., & Snyder, W. (2002). Cultivating communities of practice: A
guide to managing knowledge. Harvard Business Press.

Author Note

Chad R. Lochmiller is an associate professor of educational leadership at Indiana

University Bloomington. His research focuses on educational leadership and policy, with a
particular interest in instructional leadership, supervision, and improvement. He is the editor
of Complementary Research Methods for Educational Leadership & Policy Studies. Please
direct correspondence to clochmil@indiana.edu.

Copyright 2021: Chad R. Lochmiller and Nova Southeastern University.

Article Citation

Lochmiller, L. R. (2021). Conducting a thematic analysis with qualitative data. The Qualitative
Report, 26(6), 2029-2044. https://doi.org/10.46743/2160-3715/2021.5008