○내용분석과 주제분석 : 질적연구 수행 차원에서의 함의

내용분석과 주제분석 : 질적연구 수행 차원에서의 함의

Content analysis and thematic analysis: Implications for conducting a qualitative descriptive study

Mojtaba Vaismoradi, PhD, MScN, BScN,1,3 Hannele Turunen, PhD, RN2 and Terese Bondas, PhD, RN2,3

1College of Human and Health Sciences, Swansea University, Swansea, UK, 2Department of Nursing Science, Kuopio

Campus, University of Eastern Finland, Kuopio, Finland and 3Faculty of Professional Studies, University of Nordland,

Bodø, Norway

Introduction

In health care, qualitative methodologies 

• aim to explore complex phenomena encountered by nurses, other providers, policy makers, and patients (Denzin & Lincoln, 2000; Sandelowski & Barroso, 2003a; Tong et al., 2007). 
• The philosophy and the basic principles of methodologies, study aims and questions, and designs and data gathering criteria provide key differences between qualitative and quantitative methodologies (Ayres, 2007a). 
• A belief in multiple realities, a commitment to identifying an approach to in-depth understanding of the phenomena, a commitment to participants' viewpoints, conducting inquiries with the minimum disruption to the natural context of the phenomenon, and reporting findings in a literary style rich in participant commentaries are the main characteristics of qualitative methodologies (Streubert Speziale & Carpenter, 2007).

Qualitative methodologies consist of the philosophical perspectives, assumptions, postulates, and approaches that researchers employ to render their work open to analysis, critique, replication, repetition, and/or adaptation and to choose research methods. In this respect, qualitative methodologies refer to research approaches as the tools with which researchers design their studies, and collect and analyse their data (Given, 2008). Qualitative methodologies are not a single research approach, but different epistemological perspectives and pluralism have created a range of “approaches” such as grounded theory, phenomenology, ethnography, action research, narrative analysis, and discourse analysis.

Qualitative research in the field of health has, at times, been undertaken without identification of the specific methodology used. The term “approach” is used in this article to differentiate it from the narrower term “methods.” This indicates a coherent epistemological viewpoint about the nature of enquiry, the kind of knowledge discovered or produced, and the kind of strategies that are consistent with this (Giorgi, 1970; Holloway & Todres, 2005).

Qualitative approaches share a similar goal in that they seek to arrive at an understanding of a particular phenomenon from the perspective of those experiencing it. Therefore, the researcher needs to determine which research approach can answer their research questions (Streubert Speziale & Carpenter, 2007). There is a considerable overlap among available qualitative approaches in terms of methods, procedures, and techniques. Such an overlap of epistemological, aesthetic, ethical, and procedural concerns can encourage a generic view of qualitative research, considering it a “family” approach in which the similarities are more important than the differences, and where the notion of flexibility becomes an important value and quest. However, there is another point of view, concerned with how such flexibility can lead to inconsistency and a lack of coherence (Holloway & Todres, 2003). It should not be forgotten that consumers of research assess the quality of evidence offered in a study by evaluating the conceptual and methodological decisions the researchers have made. Therefore, the researcher needs to make good decisions to produce evidence of the highest possible quality (Polit & Beck, 2003; Høye & Severinsson, 2007).

Definition of content analysis and thematic analysis

Content analysis is a general term for a number of different strategies used to analyse text (Powers & Knapp, 2006). It is a systematic coding and categorizing approach used for exploring large amounts of textual information unobtrusively to determine trends and patterns of words used, their frequency, their relationships, and the structures and discourses of communication (Mayring, 2000; Pope et al., 2006; Gbrich, 2007). The purpose of content analysis is to describe the characteristics of the document's content by examining who says what, to whom, and with what effect (Bloor & Wood, 2006). 

On the other hand, thematic analysis often is seen as a poorly branded method, in that it does not appear to exist as a named method of analysis in the same way that content analysis does. Thematic analysis as an independent qualitative descriptive approach is mainly described as “a method for identifying, analysing and reporting patterns (themes) within data” (Braun & Clarke, 2006: 79). It has also been introduced as a qualitative descriptive method that provides core skills to researchers for conducting many other forms of qualitative analysis. In this respect, qualitative researchers should become more familiar with thematic analysis as an independent and a reliable qualitative approach to analysis.

Aim and focus of data analysis

It seems that both content analysis and thematic analysis share the same aim of analytically examining narrative materials from life stories by breaking the text into relatively small units of content and submitting them to descriptive treatment (Sparker, 2005). 

• Both content and thematic analysis approaches are suitable for answering questions such as: what are the concerns of people about an event? What reasons do people have for using or not using a service or procedure? (Ayres, 2007b). 
• Content analysis is well-suited to analyse the multifaceted, important, and sensitive phenomena of nursing (Elo & Kyngäs, 2008; Vaismoradi et al., 2011). If conducting exploratory work in an area where not much is known, content analysis may be suitable for the simple reporting of common issues mentioned in data (Green & Thorogood, 2004). 
• It has been suggested that thematic analysis, as a flexible and useful research tool, provides a rich and detailed, yet complex, account of the data (Braun & Clarke, 2006). Clearly, thematic analysis involves the search for and identification of common threads that extend across an entire interview or set of interviews (DeSantis & Noel Ugarriza, 2000).

It should be noted that both approaches allow for a qualitative analysis of data. 

• By using content analysis, it is possible to analyse data qualitatively and at the same time quantify the data (Gbrich, 2007). Content analysis uses a descriptive approach in both coding of the data and its interpretation of quantitative counts of the codes (Downe-Wamboldt, 1992; Morgan, 1993). 
• Conversely, thematic analysis provides a purely qualitative, detailed, and nuanced account of data (Braun & Clarke, 2006).

Exploration of the data analysis process

Both content analysis and thematic analysis are used in nursing studies. Nevertheless, a scarcity of information about the process of data analysis in nursing literature has resulted in a diversity of perspectives on how the approaches are used in research practice (Braun & Clarke, 2006; Elo & Kyngäs, 2008). A unified and standard data analysis protocol is preferred to be implemented by all researchers, because different results may be produced if different protocols are followed (Gbrich, 2007).

Regarding the data analysis process, different research approaches can be compared based on aspects such as “description and interpretation,” “modalities of approaches,” “consideration of context of data,” “data analysis process,” and “evaluation of the analysis process.”

Description and interpretation

• When using content analysis, the primary aim is to describe the phenomenon in a conceptual form (Elo & Kyngäs, 2008). The content analyst views data as representations not of physical events but of texts, images, and expressions created to be seen, read, interpreted, and acted on for their meanings, and must therefore be analyzed with such uses in mind (Krippendorff, 2004). However, it has been claimed that content analysis in nursing research can be applied to various levels of interpretation (Graneheim & Lundman, 2004). 
• In contrast, thematic analysis applies minimal description to data sets, and interprets various aspects of the research topic (Braun & Clarke, 2006).

Modalities of approaches

The current application of both content analysis and thematic analysis similarly, is associated with two modalities: inductive and deductive. 

• Inductive content analysis and thematic analysis is used in cases where there are no previous studies dealing with the phenomenon, and therefore the coded categories are derived directly from the text data (Hsieh & Shannon, 2005). 
• A deductive approach is useful if the general aim of thematic analysis and content analysis is to test a previous theory in a different situation, or to compare categories at different periods (Hsieh & Shannon, 2005; Elo & Kyngäs, 2008). This form tends to provide a less rich description of the data overall, and a more detailed analysis of some aspect of the data (Braun & Clarke, 2006). 
• It should be noted that both the approaches may begin with a theory about the target phenomenon or a framework for collecting or analysing data, but that does not mean there is a commitment to stay within this theory or framework (Sandelowski, 2010). The question of whether a study needs to use an inductive or directed approach can be answered in both methods by matching the specific research purpose and the state of science in the area of interest to the appropriate analysis technique (Hsieh & Shannon, 2005).

Consideration of context of data

Every analysis requires a context within which the available texts are examined. The researcher must construct a world in which the texts make sense allowing them to answer research questions (Krippendorff, 2004). The researcher, who has a broader understanding of the context influencing the stories of the study participants, may develop a wider understanding of what is going on, in addition to the understanding that she or he may share with those participating in the research (Downe-Wamboldt, 1992). Both approaches provide researchers with a framework of analysis within which the context of data is apparent. 

• Certainly, content analysis makes sense of what is mediated between people including textual matter, symbols, messages, information, mass-media content, and technology supported social interactions (Krippendorff, 2004; Hsieh & Shannon, 2005). 
• On the other hand, thematic analysis is able to offer the systematic element characteristic of content analysis, and also permits the researcher to combine analysis of their meaning within their particular context (Loffe & Yardley, 2004).

If in content analysis only the frequency of codes is counted to find significant meanings in the text, there is the danger of missing the context (Morgan, 1993). Therefore, researchers employing content analysis are sometimes accused of removing meaning from its context. The problem is that a word or coding category may occur more frequently in the speech of one person or group of people than another for different reasons. Frequent occurrence could indicate greater importance, but it might simply reflect greater willingness or ability to talk at length about the topic (Loffe & Yardley, 2004; Shields & Twycross, 2008).

Data analysis process

Like other qualitative methods, gathering and analysing data are conducted concurrently in descriptive qualitative approaches, thus adding to the depth and quality of data analysis. However, it is also common to collect all the data before examining it to determine what it reveals (Chamberlain et al., 2004).

The process of data analysis in content analysis according to Elo and Kyngäs (2008), and in thematic analysis according to Braun and Clarke (2006) is shown in Table 1. According to the table,...

• the preparation phase in content analysis and the phase of familiarizing with data in thematic analysis are equivalent. In both phases, the researcher is expected to transcribe the interview, and obtain the sense of the whole through reading the transcripts several times. While the thematic analysis researcher is mainly advised to consider both latent and manifest content in data analysis, the content analyst can choose between manifest (developing categories) and latent contents (developing themes) before proceeding to the next stage of data analysis. 
• Open coding, collecting codes under potential subcategories/subthemes or categories/themes, and comparing the emerged coding's clusters together and in relation to the entire data set conprise the next stage of data analysis, which is named the organizing phase in content analysis. 
• The same set of analytical interventions used in content analysis is applied in thematic analysis under the classifications of generating initial codes, defining and naming themes, reviewing themes, and searching for themes. 
• The final stage of data analysis in both approaches is related to reporting the result of the previous stages. This stage is especially highlighted as the final opportunity of data analysis in thematic analysis. In addition, in both approaches, the creativity of the researcher for presenting the result in terms of a story line, a map, or model is encouraged.

It is noted that in both approaches, high quality data analysis depends on gathering high quality data. It is the responsibility of researchers to conduct data gathering in such a way that any complex data would be suitable to present interesting findings. After data gathering and transcribing and paying particular attention to respondents' emotions besides their behaviours, it is recommended that the data analyst immerses himself/herself in data in order to obtain the sense of the whole through reading and rereading (Polit & Beck, 2003).

As mentioned previously, there are many similarities between the processes of data analysis presented at the different stages. The terminology used during the data analysis process in the approaches is comparable and equivalent to each other (Table 1). Data corpus, data item, data extract, code, and theme in thematic analysis are equivalent in content analysis to the unit of analysis, meaning unit, condensed meaning unit, code, and category/theme, respectively (Graneheim & Lundman, 2004; Braun & Clarke, 2006; Elo & Kyngäs, 2008).

The final product of analysis, namely the tool for presenting findings, is much debated in both content and thematic analyses. At the most abstract level, emergence of the theme/themes can be considered to be the result or final product of data analysis. The term theme has been associated with many definitions and is used interchangeably with a vast number of other terms such as category, domain, unit of analysis, phase, process, consequence, and strategy (DeSantis & Noel Ugarriza, 2000). In this respect, there is considerable diversity in nursing and qualitative research literature associated with the identification of themes, the interpretation of the concept, and its function in data analysis (DeSantis & Noel Ugarriza, 2000). 

• A theme is defined as a coherent integration of the disparate pieces of data that constitute the findings (Sandelowski & Leeman, 2012). It captures something important about data in relation to the research question, and represents some level of response pattern or meaning within the data set (Braun & Clarke, 2006). 
• A pragmatic way to state the difference between a theme and a category is that 
• the latter refers mainly to a descriptive level of content and can thus be seen as an expression of the manifest content of the text, 
• whilst the former is the expression of the latent content (Graneheim & Lundman, 2004). Especially in thematic analysis, themes are usually quite abstract, and therefore difficult to identify (DeSantis & Noel Ugarriza, 2000; Spencer et al., 2003). 

• Furthermore, in thematic analysis the importance of a theme is not necessarily dependent on quantifiable measures, but rather on whether it captures something important in relation to the overall research question (Spencer et al., 2003; Braun & Clarke, 2006). The latter perspective is different from the current idea in content analysis, where it is possible to reach a theme based on the frequency of its occurrence in the text. This approach is objective, systematic, and concerned with the surface meaning of the document rather than hidden agenda (Bloor & Wood, 2006).

• One of the first decisions that should be taken when conducting content analysis is whether to concentrate analysis on the manifest or latent content of data. It is said that both manifest and latent content deal with interpretation, but the interpretation varies in depth and level of abstraction (Graneheim & Lundman, 2004; Powers & Knapp, 2006). 
• In contrast, thematic analysis incorporates both manifest and latent aspects. It means that the analysis of latent content of data is an inseparable part of the manifest analysis approach (Braun & Clarke, 2006).

Another characteristic of data analysis in thematic analysis is drawing a thematic map. This refers to the visual presentation of themes, codes, and their relationships, involving a detailed account and description of each theme, their criteria, exemplars and counter examples, and other similar details. As one part of data analysis, it helps with reviewing themes and achieving the aim of identifying coherent but distinctive themes (Ryan & Bernard, 2000; Braun & Clarke, 2006). It should not be forgotten that data analysis processes in both approaches are not linear, simply moving from one phase to another phase, but should be recursive with frequent reviews. In addition, the result should be the identification of a story, which the researcher tells about the data in relation to the research question or questions.

Evaluation of the analysis process

Evaluating the validity or rigour of a qualitative study requires reviewers to distinguish between researchers' errors during data analysis (Sandelowski & Barroso, 2003a). 

• One criticism that has been leveled in all qualitative approaches is that they lack the scientific rigour and credibility associated with traditionally accepted quantitative methods. It means that the quantitative inquiry is assumed to occur within a value-free framework and which rely on the measurement and analysis of causal relationships between variables (Horsburgh, 2003). 
• Scientific qualitative research must yield valid results, in the sense that the research effort is open for careful scrutiny and it should be possible for any resulting claims to be upheld in the face of independently available evidence (Krippendorff, 2004). As an unavoidable part of all qualitative approaches, both researchers and readers should be helped to look for alternative interpretations. 
• Credibility, dependability, confirmability, and transferability are the most common measures to achieve rigour in qualitative studies (Lincoln & Guba, 1985). 

Although the assessment of rigour in content and thematic analysis shares many similarities, some differences emphasize the separate and unique identities of each approach. 

• For instance, intercoder reliability (analogous to interrater reliability) refers to the extent to which more than one coder independently classifies material in the same way as peer researchers. It is commonly used in content analysis and has been introduced as a measure for improving the approach's reliability (Cavanagh, 1997). However, because of the pure qualitative nature of thematic analysis, peer checking of intercoder reliability is not always possible since there is scepticism about the value of such testing. It has been discussed that one researcher merely trains another to think as she or he does when looking at a fragment of text. Thus, the reliability check does not establish that codes are objective, and merely two people can apply the same subjective perspective to the text (Loffe & Yardley, 2004). 

As a practical way to improve rigour in both approaches, researchers are encouraged to maintain a personal research diary. As a word of caution, the status of these additional materials in relation to raw data or field notes is sometimes unclear, as is the way in which they are expected to contribute to any interpretation. A conscious decision is made to include and code personal memoranda alongside field notes, and the same coding scheme is used for both types of data (Ballinger et al., 2004; Rolfe, 2006). 

Finally, one of the best ways for judging the quality of findings is whether new insights into the studied phenomenon have been provided; if so, the study should have increased the understanding of particular phenomena or informed practical actions (Krippendorff, 2004).

---

Content analysis and thematic analysis: Implications for conducting a qualitative descriptive study

1. Mojtaba Vaismoradi PhD, MScN, BScN^1,3,
2. Hannele Turunen PhD, RN^2,* and
3. Terese Bondas PhD, RN^2,3

Article first published online: 11 MAR 2013

DOI: 10.1111/nhs.12048

© 2013 Wiley Publishing Asia Pty Ltd

Keywords:

• content analysis;
• nursing;
• qualitative descriptive research;
• thematic analysis

Abstract

Qualitative content analysis and thematic analysis are two commonly used approaches in data analysis of nursing research, but boundaries between the two have not been clearly specified. In other words, they are being used interchangeably and it seems difficult for the researcher to choose between them. In this respect, this paper describes and discusses the boundaries between qualitative content analysis and thematic analysis and presents implications to improve the consistency between the purpose of related studies and the method of data analyses. This is a discussion paper, comprising an analytical overview and discussion of the definitions, aims, philosophical background, data gathering, and analysis of content analysis and thematic analysis, and addressing their methodological subtleties. It is concluded that in spite of many similarities between the approaches, including cutting across data and searching for patterns and themes, their main difference lies in the opportunity for quantification of data. It means that measuring the frequency of different categories and themes is possible in content analysis with caution as a proxy for significance.