Deepening: Analyzing & Coding – Kwalitatief Onderzoekscollectief GGZ

Analyzing & Coding

“Qualitative Research” is an umbrella term for a broad spectrum of methodologies and methods. Methods are the practical procedures that follow from a
methodology
, which is to say: a conception of what is true, how that truth can best be investigated, whether that truth is also everyone’s reality, how researchers should or can position themselves in it, how and where to collect data, and what quality is. That is, methodologies rely on a
ontology and epistemology
on research and reality (read, for example, Sharlene Nagy Hesse-Biber’s wonderful book on this subject)

Below is a "card index" of methods.

The explanation in the method cards is a starting point, via “read more” you click through to delve deeper, and in connection with the KOG you find people to learn and collaborate with.

Thematic Analysis Narrative Analysis Interpretative Phenomenological Analysis (IPA).Grounded Theory Realist Evaluation Concept Mapping Mixed Methods Meta-synthesis

Previous Next

Thematic Analysis

What is it?

Thematic Analysis (TA) is a method of analysis that allows you to code statements in texts
code
and thematically
organizes
. A theme is a pattern of meaning that recurs in different (parts of) texts in your dataset. Themes are often abstract and not clear on first reading, they connect the data, through an organizing story or concept. With TA you can bottom-up (inductive) analysis by developing codes from the data (data-driven) and forming a topic structure that can be used to answer a research question. You can also TA top-down (deductive) by developing a codebook from theory, and then analyzing the transcripts through the “lenses” of that codebook. In this form, your codes are fixed in advance, but you use the data to deepen and refine that code content, and develop a code tree that does justice to the data. This deductive form of TA also allows one to work hypothetically.

Why apply?

Using TA, you form a topic structure by systematically coding and organizing spoken and written text. This can be at the individual level (case) or group level (sample). TA is a flexible qualitative analysis technique that can be applied both inductively and deductively. A big advantage of TA is the systematic approach, where you keep track of all the steps and thus can justify your formed themes from the data.

How does it work?

Prior to coding, it is important to already know your dataset well (“familiarizing”) by reading through your transcripts and understanding the storylines and messages. It also helps to listen to the audio because sometimes the way you speak gives a just different nuance than a written text (think sarcasm, for example). Next, you start coding. This involves labeling all sentences in a (selection of a) transcript. Codes are thus a collection of content labels. You merge these codes into a codebook. You can then organize this list by grouping codes. An inductive approach to Thematic Analysis proceeds in three rounds of coding: open, axial, and selective coding. Open coding means labeling all sentences in a (selection of) the transcripts. A code stays as close to the data as possible and retains as much content and context as possible (i.e., rather a sentence than a word). In the second phase, the axial coding, merge your codes that match content under an existing code or a new umbrella code. You do this for all codes. In the third phase, the selective coding, form main themes based on the code structure by merging the subthemes from the axial phase. In this phase, you name main themes that help answer the research question. This may also omit topics that were discussed but not directly relevant to the research question. You write out the retained themes in a report, where it is possible that already writing some more refinement or nuance in your code structure. This iterative character (going back and forth between coding, organizing and writing) is allowed and encouraged. When writing out, it is important to report all phases of the analysis by systematically describing the subthemes for each main theme and supporting them with citations (citing the source, for traceability). Then by taking the message of main themes together, you formulate an answer to your research question.

Points of interest

- A theme is
  not
  same as a topic or a summary of different perspectives on a topic. For example, “negative care experiences” is not a theme, but only a categorization by topic. Make sure your theme is substantively worded and makes clear what is going on. For example, “care as autonomy constraint.
- Themes are not “discovered” or “found. Those terms imply that there are a truth hidden beneath the themes, and suggests that it does not matter which researcher does the research. This is not true: rather, in the TA we assume that the questions the researcher asks and the interest and experience the researcher has play a role in shaping the themes. To recognize this active role of the researcher, we speak of “forming,” “developing” or “distilling” themes.
- Because the researcher plays this role, it is important not to conduct a TA alone: you always work in a team, ideally a team composed of people with different backgrounds, views and interests (!) in the research. Seniority is also important here: the more ingrained you are in certain subject matter the faster you will read that back into your transcripts (for example, also in subtle statements), so to avoid selective is going to read and code, it is important to have less experienced or employed researchers on your team who can question and challenge your reading. These can be younger researchers (assistants, students), people from another field, or experts by experience/people you are researching (from member-checking to co-creation).
- The conscious and active role of the researcher described above is also called
  reflexive thematic analysis
  mentioned. If we assume that the people we study consciously and actively experience and make meaning of the world around them, we must not forget that we ourselves, from a researcher role, also do. Awareness of your own beliefs, motivations and interpretations is a prerequisite for the quality of your research. Even in hypothetical or top-down TA, this reflexive attitude is
  necessary
  to account for the quality of interpretations.
- You can concretize that reflexive attitude through peer review and logbook keeping. This allows you to detect blind spots and brackets (“brackets”) your beliefs: You are aware of them, and make sure they do not guide your interpretations. In the log, you keep track of your research questions, make explicit any hypotheses, expectations or presuppositions, describe any changes or refinements to them, and note any surprises or obstacles, and so on. This provides insight for yourself and others into the dynamic analysis process and its quality.

Continue Reading

Must Read

Good Practice

Read about the basic principle of reflexive qualitative research and how it differs from analysis that starts from a pre-developed codebook:

Quality of findings depends directly on the quality of the research process. Therefore, before you start, read about good practices and quality criteria in designing, conducting and interpreting qualitative research

Fine Examples

An and fine example of Reflexive Thematic Analysis can be found here and here

KOG Members with expertise in Thematic Analysis:

Narrative Analysis

What is it?

Narrative analysis is a specific form of qualitative analysis. What distinguishes narrative analysis is that a story is studied as a unit. This holistic principle means that different story elements are always studied in conjunction with each other. There is a focus on the form and content of the story as well as the intentions of the narrator and the context in which the story was created. Narrative analysis is often applied to the study of stories expressed in language, such as published stories of experience or low-structured interview data, but can also be used for visual narrative forms such as videos or works of art.

Why apply?

Narrative research fits well within a recovery-oriented, emancipatory research tradition. By maintaining the unity and coherence of one’s story, the the narrator as subject with a life history and intentions respected. At this piece explains in detail the ways in which the philosophical premises of the recovery paradigm and narrative paradigm correspond. Examples of shared principles include holism and context: Narrative analysis facilitates a holistic understanding of people’s struggles and desires, and the way their identities and experiences are shaped by the context in which they live.

How does it work?

Narrative research is a broad and multidisciplinary field. Therefore, there is no one (good) way to do narrative analysis. As with other methods, it is important that their elaboration be consistent with the research goal and the philosophical principles you use as a researcher. Narrative analysis always begins with reading and rereading stories to become familiar with them. The analysis is laborious in the sense that you have to delve into the specific details of the story in order to ultimately arrive at a better picture of the connection. Narrative analysis therefore lends itself well to case studies and small samples. Yet it is also possible to do narrative analysis with larger amounts of stories and shared plots, or identify patterns in meaning.

A great guideline for narrative analysis that incorporates both the personal, interpersonal and social context of the story is Murray and Sools‘ narrative analysis. In doing so, you first analyze various story elements within individual stories, then work toward describing and contextualizing shared plot structures in your dataset. Central to this analysis is identifying the break (“breach”) in the story. The break refers to tension in the narrator’s signification. You can identify those by seeing where different story elements clash. For example, the narrator may have an intention (e.g., gaining self-insight) that is not supported by the setting a person is in (acute psychiatry). Or the narrator may struggle with intentions that feel contradictory (staying stable AND volunteering).

Points of interest

Analyzing narrative data is not the same as narrative analysis. Following
Polkinghorne
we recommend speaking of “analysis of narratives” when identifying common themes in narrative data, and of “narrative analysis” when analyzing individual stories in depth and as a unit.
During your research process and in your reporting, explicate how you view stories (a reflection of lived experience, a temporal identity construction, a dialogue or form of social action?).
Pay attention to meta-narratives and counter-narratives. Individual stories are not only personal, but relate to the stories that are shared and dominant in a particular place and time and culture.

“In telling his or her story the narrator makes use of this socially embedded language. It is not fully the narrator’s story: its structure is conditioned by both the immediate presence of others and the dominant plot in society. Researchers often focus on one level of analysis to the neglect of the other. The challenge is to articulate the telling of illness narratives across these different levels.”

Murray, 2000
(p.344)

Continue Reading

Must Read

A nice methodological introduction is
this
handbook, in which Riessman elaborates on various forms of narrative analysis.

This accessible classic by Arthur W. Frank illustrates the added value of identifying story types. Frank shows what narrative forms are available in the West to talk about illness, and what the moral implications of these are.

Good Practice

Spektor-Mersel and Knaifel state in this review note that the holistic principle is obscured in much research presented as “narrative.

How that can be improved is explained step-by-step by Murray and Sools, in this article on narrative analysis.

Fine Examples

In this study, I applied narrative analysis to examine how people with psychoses make meaning of trauma.

This review by Llewellyn-Beardsley and colleagues on recovery narratives shows how to analyze narrative studies at the meta-level.

In this ethnography, Mattingly shows that you can also do narrative research based on observational research on healthcare practices.

KOG Members with expertise in Narrative Analysis:

Interpretative Phenomenological Analysis (IPA).

What is it?

Interpretative Phenomenological Analysis (IPA) departs from a
phenomenological
perspective on the world and meaning: when we look at the world we
understand
we view the world from a frame of meaning, a frame of reference. That frame of reference is personal (people understand and interpret based on previous experiences and learned and acquired meanings) and communal (we learn from each other about shared meanings, and assume that others also understand our own meanings without question). An important implication of phenomenological philosophy is that we never
neutral
can understand: all our interpretations are colored by previous experience
and
by shared frameworks or discourses. Therefore, in analysis of texts or spoken or written data, we look for how people understand phenomena. By looking at the patterns in how people experience (lived experience), we can describe phenomena without claiming ontological truth in them. What is important here is that we as researchers also Understanding phenomena from our frame of meaning. This implies that in the interpret of phenomena include not only the frames of meaning of the investigated but also those of the researcher. We call this
double hermeneutics
: meaning as an interaction between the lived experience of people we investigate, and the frameworks within which we as researchers interpret it and come to understand a phenomenon.

Why apply?

The IPA is a method of analysis in which we look not only at
what is said
(content), but also at
how
it is said (form, process), and by
whom
(agents).

How does it work?

Because IPA assumes
double hermeneutics
is coded both what those under investigation say and what the researcher thinks and notices in the process. When we code on paper, the sheet is divided into three columns: the transcript is in the middle, in the “right column” we code the statements in the transcript, and on the left side we note all the researcher’s comments and interpretations. If you code using software, you can use memos or annotations for the “left column” (be careful to keep those as data/codes too!) or split codebooks.

To code experiences of those surveyed, IPA uses the same form of coding as Thematic Analysis (TA): via
rounds
of open coding, axial coding, and selective coding, a code structure is developed that is both
bottom-up
and
top-down
can be (see
link to TA
). An important difference from TA is that IPA can also encode
how
people name their experience, and also what
not said
is said.

To code the researcher’s experience, it is important to report extensively to yourself what impressions and interpretations come to you through the process. For example, you notice through the process that your hypotheses change or nuance, that you develop a different view of a concept or theoretical term, or that you begin to ask yourself questions that you didn’t have before. This formation of interpretations is a hermeneutic process and plays an important role in shaping final themes. Moreover, this is a
iterative
process: as you adjust your interpretations as you go along, it is important to go back to previous transcripts and look at them with this updated perspective as well.

Points of interest

- In IPA, we do not assume that there is
  a
  truth, but rather we look for the
  lived experience
  Of people we examine. This also implies that our findings are not simply or directly generalizable are. Therefore, describe your sample in great detail so that your reader will be able to
  evaluate
  whether the findings may be representative of his area of interest.
- Because of the
  double hermeneutics
  describe the
  researchers
  in the same detail as the sample.
- Because for an IPA you need detailed and
  in-depth
  analyzes, fewer cases are often included. Experience-in-context is necessary to form and substantiate interpretations.
- Always keep an open mind
  what is not said
  , or on what people find difficult to pronounce. Because we look at lived experience in an IPA, we may also encounter processes, obstacles and prohibitions that prevent people from talking about things. This information can be very helpful in capturing the experience of a phenomenon
- In doing so, we also pay attention not only to
  similarities
  between people to describe patterns, but also on exceptions and changes. In fact, these can be very informative in delineating the lived experience of a phenomenon.

Continue Reading

Must Read

The basic text and must-read is the paper by Smith & Eatough (2007)

Good Practice

Interpretative Phenomenological Analysis – what it is not

Fine Examples

IPA used to really listen to the lived experience of people with delusions, by Feyearts et al. (2021)

IPA studies of how patient-participants experience participating in psychotherapy research, and how scoring questionnaires can put symptoms under pressure, by Truijens et al. (2019, 2023)

And IPA study zooming in on the experience of adolescents growing up with a mother with depression, by Van Parys et al. (2014)

KOG Members with expertise in Interpretative Phenomenological Analysis: