# Theoretical preambule

# 💡 What kind of analysis ?

In social science, semi-structured interviews are often the base material of research. Once done, it is time to analyze them. According to Van Campenhoudt et al. (2017), two types of analysis are possible: statistical analysis and content analysis. Corpus enables the second category of analysis. Content analysis offers a certain degree of depth and complexity.

Three main categories can be distinguished in content analysis depending on whether the examination focuses on elements of speech (thematic analysis), on its form (formal analysis) or on the relationships between constituent elements (structural analysis).

Corpus allows for in-depth thematic analysis since the tool makes it possible to highlight social representations from the examination of certain constituent elements of discourse. This is not only applicable to interviews but can also be mobilized for documentary analysis for example.

Thematic analysis

Thematic analysis is a method of identifying, analyzing, and presenting patterns (themes) in data. It helps organize and describe a set of data in detail (Braun and Clarke, 2006).

# 🔍 Before processing

# Data extraction : transcription

If you are going to analyze interviews, the first step is the transcription. "First and foremost, you need a consistent, quality material (in this case interviews) that is perfectly rendered and fully available for analysis. Regardless of what happens next, it will always be possible to come back to that base material and find your way around." (Van Campenhoudt et al., 2017).

In order to make the material analyzable, it is therefore necessary to transcribe it as faithfully as possible while being aware that switching from audio to writing also means the loss of certain elements such as the non-verbal.

# Choose your epistemological approach

What posture does the researcher take regarding his/her object of study and the analysis of his/her data?

Thematic analysis, by its flexibility, allows it to be adapted to several epistemological approaches although their results and objectives are different: within the framework of an essentialist / realistic approach or within the framework of a constructivist approach (Braun and Clarke , 2016). Research epistemology guides what you can say about your data and informs how you theorize meaning.

# Essentialist approach (a.k.a. realistic approach)

It is about theorizing motivations, experience and meaning in a direct way because a simple and largely unidirectional relationship is established between meaning, experience and language (language reflects and allows us to articulate meaning and experience). It assumes that there is an objective reality to be revealed on the basis of the words of the actors. Thematic analysis will help determine themes to illustrate this objective reality (Chapter 5).

# Constructivist approach

Meaning and experience are socially produced and reproduced, rather than inherent in individuals. Therefore, thematic analysis conducted in a constructivist approach cannot and does not seek to focus on motivation or individual psychologies -- but rather seeks to theorize the socio-cultural contexts and structural conditions, which allow individual narratives. It would be the interpretation of the actors and the meanings they give to this objective reality that influences their behavior in this same reality.

Constructivist approach

Whatever approach you choose, it is important to present it in a clear and transparent manner to your readers. "To report on one's posture is therefore to allow the reader to understand how the material was collected and analyzed." (Parotte and Fallon, 2020).

# 📌 Thematic analysis

How to capture the relevant empirical material?

# Different type of analysis

As Braun and Clarke (2016) point out, it is first and foremost important to determine the type of analysis you want to perform, and the arguments you want to show, in relation to your data set.

  1. A first possibility is to provide a rich thematic description of all the data. This type of analysis helps give the reader a holistic view of the predominant or important themes. In this case, the themes identified, coded and analyzed must be an exact reflection of the content of the data set. In such an analysis, some depth and complexity is necessarily lost, but a rich overall description is maintained. This method can be particularly useful when you are familiar with an area that is little studied or when you are working with participants whose opinions on the subject are not known.

  2. An alternative use of thematic analysis is to provide a more detailed and nuanced account of a particular theme, or group of themes, within the dataset. Rather, this method will be used when it comes to tackling a more specific question or a theme already previously studied and analyzed.

# What is a theme ?

A theme captures an important element of the data in relation to the research question, and represents a certain level of structured response or meaning in the dataset.

It provides a partial answer to the research question asked. Partly, this means that this answer is necessarily incomplete, because it is the analysis of all the themes identified that will allow the research question to be answered (Parotte and Fallon, 2020).

An important question to address before coding is: what counts as a theme, or what "size" should a theme be? There is no perfect or absolute answer to these questions. It is important that the researcher retains some flexibility in these decisions. The important thing is that the topic captures important information related to the research question. This means that one should not blindly rely on the repetition of something in an interview to determine whether it is a relevant topic. It is possible that an item appears only once or twice in an interview but provides particularly interesting information about the research question asked (Braun & Clarke, 2006; Parotte & Fallon, 2020).

The flexibility of thematic analysis is that it allows you to determine themes (and prevalence) in a number of ways. What is important is that there is a certain consistency in how you do it in a particular analysis (Braun & Clarke, 2006).

# Inductive versus deductive logic

According to Braun and Clarke (2016), there are two main ways of identifying themes.

# Inductive logic

The themes identified are strongly linked to the data itself. It is then a matter of coding the data without trying to fit them into a pre-existing coding framework, or into the researcher's analytical preconceptions. In this sense, this form of thematic analysis is driven by data.

# Deductive logic

The identification of themes is motivated by the theoretical or analytical interest of the researcher in the field concerned, and is therefore more explicitly oriented towards analysis. This form of thematic analysis tends to provide less rich description of the data as a whole, and more a detailed analysis of one aspect of the data.

Also, the choice between inductive analysis and theoretical analysis is about how you code the data and why you are doing it. You can either code for a very specific research question (which corresponds to the more theoretical approach) or the specific research question can evolve during the coding process (which corresponds to the inductive approach).

# Explicit versus interpretative constructivism

Another decision concerns the "level" at which the themes should be identified: at the semantic / explicit level, or at the latent / interpretative level. A thematic analysis generally focuses exclusively or primarily on a single level (Clarke and Braun, 2016).

# Themes derived from explicit (semantic) level

Themes are identified in the explicit or superficial meanings of the data. The analyst is not looking for anything other than what a participant has said or what has been written.

# Themes derived from interpretative (latent) level

The analysis goes beyond the semantic content of the data and begins to identify or examine the underlying ideas, assumptions and conceptualizations that are theorized to shape the semantic content of the data. Thus, for the latent thematic analysis, the development of the themes themselves involves a work of interpretation, and the analysis that is produced is not only a description, but is already theorized. It follows from this observation that in general, a latent analysis derives from a constructivist approach (Parotte and Fallon, 2020).

# Resources