1INTRODUCTION TO APPLIEDTHEMATIC ANALYSISUnsatisfied with the limitations imposed by any one particular martial art, BruceLee developed his own composite fighting style, which he called “Jeet Kune Do”(the way of the intercepting fist). Jeet Kune Do is not a novel set of fighting techniques, but rather a more focused style of combat that synthesizes the most usefultechniques from numerous fighting arts. For Lee, this was an emancipatoryendeavor that allowed practitioners of Jeet Kune Do to choose from a wide rangeof techniques and employ the most appropriate ones for a given objective. InLee’s words:I have not invented a "new style," composite, modified or otherwise that is setwithin distinct form as apart from "this" method or "that" method. On the contrary,I hope to free my followers from clinging to styles, patterns, or molds. . . [A] JeetKune Do man who says Jeet Kune Do is exclusively Jeet Kune Do is still hung upon his self-closing resistance, in this case anchored down to reactionary pattern, andnaturally is still bound by another modified pattern and can move within its limits.He has not digested the simple fact that truth exists outside all molds; pattern andawareness [are] never exclusive. (Lee, 1971, p. 24)Qualitative research is analogous in many ways to martial arts. Approachesto qualitative data collection and analysis are numerous, representing a diverserange of epistemological, theoretical, and disciplinary perspectives. Yet mostresearchers, throughout their career, cling to one style with which they arefamiliar and comfortable, to the exclusion (and often disparagement) of allothers. In the spirit of Jeet Kune Do, we feel that good data analysis (and researchdesign, for that matter) combines appropriate elements and techniques fromacross traditions and epistemological perspectives. In our view, the theoreticalor philosophical foundation provides a framework for inquiry, but it is the datacollection and analysis processes and the outcome of those processes that are3

4APPLIED THEMATIC ANALYSISparamount. In other words, “We need a way to argue what we know based onthe process by which we came to know it” (Agar, 1996, p. 13). From such aperspective, it does not make sense to exclude a particular technique becauseof personal discomfort with it, or misconceptions about or prejudices regardinghow and why it might be used. We are reminded here of Russ Bernard’s (2005)adage that “methods belong to all of us” (p. 2). Eschewing a compartmentalizedview of qualitative research and data analysis is the underlying theme of thisbook and the analytic process we describe. We call this process Applied ThematicAnalysis (ATA). Briefly put, ATA is a type of inductive analysis of qualitativedata that can involve multiple analytic techniques. Below, we situate ATA withinthe qualitative data analysis literature to help both frame the process and providea rationale for the name we have given it.Before defining our process, we first lay out the overall rationale for the bookas well as provide the reader with a sense of what this book does, and does not,cover. As noted in the preface, we have written this book in response to a perceived need for a published volume that gives researchers a practical frameworkfor carrying out an inductive thematic analysis on the most common forms ofqualitative data. Although we cover some of the theoretical underpinnings ofqualitative research, this book is primarily about process and providing researchers usable tools to carry out rigorous qualitative data analysis in commonlyencountered research contexts. To this end, we wanted to keep the content asfocused as possible and present readers with what we believe to be the mostefficient, yet rigorous, analytic techniques. We begin from the point of havingqualitative data in hand, and therefore do not address research design or datacollection strategies.We refer above to the “most common forms” of qualitative data. By this, wemean data generated through in-depth interviews, focus groups, or field observations (i.e., textual field notes). We recognize that qualitative data can be generatedthrough other activities such as open-ended questions on a survey, free-listing andother semistructured elicitation tasks, or visual data collection techniques. Thesemethods are all useful and appropriate for certain types of research objectives;however, they are not commonly used methods in the broadly defined world ofqualitative research.This book, then, is intended for the researcher, student, or other interestedparty who has been tasked with analyzing, and making sense of, a set of fieldnotes or transcripts from focus groups or in-depth interviews. How does one goabout thematically analyzing these types of data in a systematic way that resultsin credible answers to the research questions and objectives embedded within astudy? Helping readers meet this challenge is the fundamental purpose of thisbook. Note that the process we delineate can also be used to analyze free-flowingtext from secondary data sources, such as in document analysis. But to keep thisbook simultaneously concise and broadly appealing, the examples and exercisesprovided are from studies employing the more traditional qualitative datacollection techniques.

CHAPTER 1. INTRODUCTION TO APPLIED THEMATIC ANALYSIS5DEFINING QUALITATIVE RESEARCHBefore talking about process, we should first define what we mean by “qualitative research,” since the definition influences how we characterize qualitativedata analysis, the data items to be used in our analysis, and the types of analyseswe perform on our data. Many existing definitions are constrained by a dichotomous typology that contrasts qualitative and quantitative research or assumes aparticular epistemological foundation. Another common descriptive practice is tolist attributes as if they are exclusive or necessary features of qualitative research.These types of characterizations exist despite the fact that the attributes listed are:(a) not always present in qualitative inquiry and (b) can also be true of quantitative research (Guest & MacQueen, 2008). A simple Google search of “qualitativeresearch” and “definition” will bring up a host of examples, from websites andresearch methods course syllabi. For example, the Online Dictionary of the SocialSciences (n.d.) defines qualitative research as follows:Research using methods such as participant observation or case studies which resultin a narrative, descriptive account of a setting or practice. Sociologists using thesemethods typically reject positivism and adopt a form of interpretive sociology.Compare that to Denzin and Lincoln’s definition:Qualitative research is a situated activity that locates the observer in the world. Itconsists of a set of interpretive, material practices that makes the world visible.These practices transform the world. They turn the world into a series of representations, including field notes, interviews, conversations, photographs, recordings, andmemos to the self. At this level, qualitative research involves an interpretive, naturalistic approach to the world. This means that qualitative researchers study thingsin their natural settings, attempting to make sense of, or to interpret, phenomena interms of the meanings people bring to them. (2005, p. 3)Of particular note in these definitions is the joint emphasis on a philosophicalstance and a particular structuring of the analytic results as interpretive. Theinterpretive approach is generally set in contrast to a positivist approach, andindeed, for many the two are incompatible. Quantitative research methods aregenerally difficult to reconcile with an interpretive approach, while qualitativemethods provide considerable room for an interpretive inquiry. From this, manythen conclude that qualitative research methods are difficult to reconcile with apositivist approach. This is not true. It is what you do with qualitative data, andnot the methods themselves, that define whether you are engaged in a researchendeavor that is interpretive, positivist, or hybrid of the two.We prefer the simpler and more functional definition of qualitative research asoffered by Nkwi, Nyamongo, and Ryan (2001): “Qualitative research involves anyresearch that uses data that do not indicate ordinal values” (p. 1). The focus in thislatter definition is on the data generated and/or used in qualitative inquiry—that

6APPLIED THEMATIC ANALYSISis, text, images, and sounds. Essentially, the data in qualitative research are nonnumeric and less structured data than those generated through quantitatively oriented inquiry, because the data collection process itself is less structured, moreflexible, and inductive. An outcome-oriented definition such as that proposed byNkwi and colleagues avoids unnecessary and inaccurate generalizations anddichotomous positioning of qualitative research with respect to its quantitativecounterpart. It allows for the inclusion of many different kinds of data collectionstrategies and analysis techniques (which we describe later) as well as the plethoraof theoretical frameworks associated with qualitative research.Exclusion of specific data collection or analysis methods from the definitionalso paves the way for a more refined view of qualitative data analysis, one thatdistinguishes between the data themselves and the analyses performed on data. AsBernard (1996) notes, many researchers fail to make this distinction, madegraphically apparent in Figure 1.1.Figure 1.1Qualitative and Quantitative Data Analyses (adapted from Bernard, 1996)Type of DataQuantitativeType of AnalysisQualitativeQualitative(text, pictures, sounds)Quantitative(ordinal, interval, ratio)A) Interpretation of meaning in text orimagesB) Interpretation of patterns in numericdataItem of Analysis - images, sounds, text (sizeand precision of unit varies with technique)Item of Analysis - graphs, diagramsExamples- Grounded Theory- Cultural Models- Hermeneutics- Ethnographic MappingExamples- Epidemic Curves- Social Network GraphsC) Statistical and mathematical analysisof textD) Statistical and mathematical analysisof numbersItem of Analysis - numeric data (e.g.,similaritymatrices); well-defined, small units of text(e.g.,frequencies, truth tables)Item of Analysis – numeric data (e.g.,ordinal, interval, ratio)Examples- Content Analysis- Pile Sorts- Free Listing- Cluster Analysis- Chi SquareExamples- Correlation Measures(e.g.,regression)- Comparison of Means(e.g.,ANOVA)

CHAPTER 1. INTRODUCTION TO APPLIED THEMATIC ANALYSIS7Making the simple distinction between data type and the type of procedureused to analyze data broadens the range of “qualitative research” and opens upan additional category of analytical procedures that other conceptual frameworksexclude (Guest, 2005). Most definitions of qualitative research include only thetop left quadrant of the figure and miss an entire group of analytic strategiesavailable to them—that is, those that utilize quantitative analytic procedures onqualitative data (lower left quadrant). Throughout this book, we try to emphasizethe complementarity of both types of analytic procedures on the left side ofFigure 1.1 and downplay any antithesis between the two.ANALYTIC PURPOSEThe design and plan for a particular analysis depends a lot on the generalapproach taken and the type of outcome expected—the analytic purpose. In thisbook, we focus on inductive analyses, which primarily have a descriptive andexploratory orientation. Although confirmatory approaches to qualitative dataanalysis certainly exist, they are employed less often in social/behavioral researchthan inductive, exploratory analyses. We provide a summary of the differencesbetween the two approaches in Table 1.1. Further reading on how to do confirmatory qualitative research using a thematic approach, also known as classic contentanalysis, can be found in several comprehensive works, including Krippendorf(2004), Weber (1990), and Neuendorf (2001).Table 1.1 Summary of Differences Between Exploratory and ConfirmatoryApproaches to Qualitative Data AnalysisExploratory (“content-driven”)Confirmatory (“hypothesis-driven”) For example, asks: “What do x peoplethink about y?” For example, hypothesizes: “x peoplethink z about y” Specific codes/analytic categoriesNOT predetermined Specific codes/analytic categoriespredetermined Codes derived from the data Codes generated from hypotheses Data usually generated Typically uses existing data Most often uses purposive sampling Generally employs random sampling More common approach Less common approachThe main difference between the two approaches is that for an exploratorystudy, the researcher carefully reads and rereads the data, looking for key words,trends, themes, or ideas in the data that will help outline the analysis, before any

8APPLIED THEMATIC ANALYSISanalysis takes place. By contrast, a confirmatory, hypothesis-driven study isguided by specific ideas or hypotheses the researcher wants to assess. Theresearcher may still closely read the data prior to analysis, but his analysis categories have been determined, a priori, without consideration of the data.Objectives are also formulated differently: Research questions are better suitedto exploratory research while hypotheses better capture objectives of a confirmatory nature. Other differences between the two approaches relate to sampling and data sources. Exploratory studies generally are based onnonprobabilistic samples of research participants and generate primary data.Conversely, confirmatory studies typically employ probabilistic sampling strategies to select text from existing sources.The distinction between inductively exploring data versus assessing hypotheseswith data are made clear above. But here we wish to emphasize that while exploratory approaches to qualitative analysis are not specifically designed to confirmhypotheses, this does not mean that they are atheoretical