By: Suyasha Koirala, Research Coordinator (KUSMS). MPH (SDU Denmark). BPH (IOM Nepal)
Grounded theory is similar to thematic Analysis. Thematic Analysis is the process/method of identifying, analyzing, and reporting patterns or themes within qualitative data. It is a flexible method as any pre-existing theoretical framework does not bind it. The patterns in the data which are important are summarized and interpreted to make sense. It uses both inductive and deductive ways to approach research patterns. In summary, it deals with diverse subjects by means of interpretation.
Thematic analysis is the method to reflect both the reality and unravel the reality. However, you should be careful on whether you are summarizing and organizing the data or analyzing it to explain the research questions/objectives as many make the main interview questions their themes.
According to the Braun and Clarke (2006), there are two levels of themes;
- Semantic/explicit level: Analysis is done explicitly which is not done beyond what a participant had said, no hidden meanings are looked upon
- Latent/Interpretative level: Examines and analyses the underlined ideas, assumptions, ideologies, and concepts that can be used to explain the semantic content of the data.
There are two ways of analysis that depend on how and why you are coding the data;
- Top-down/theoretical thematic Analysis/deductive analysis: Its coding focus is mainly based on the research question(s). It provides a detailed analysis of some aspects of the data.
- Bottom-up/inductive analysis: It is data-driven and is similar to grounded theory. Neither the questions asked to the participants nor the researcher’s interest plays a role.
For example, in inductive analysis, if the researchers are conducting research on the sustainability of health promotion activities then they would read and re-read the data set for themes without going through others’ research papers. In theoretical analysis,
Before conducting analysis, you should have transcription/translation ready and clear research question(s).
In this article, we will follow Braun and Clarke’s (2006) 6-step framework.
Step 1: Become Familiar with the data
It is suggested to take data collection and data analysis in parallel. After the transcriptions are in your hand read and re-read them. Even though you yourselves have collected the data it is necessary to reread it to be familiar with the depth and breadth of the data. Make a note and write the early impression you had after reading the full transcriptions. A rough note and marking ideas for coding should be done for each transcription you read.
Step 2: Generate initial codes
In this step, you summarize the data in the form of codes (Semantic or latent). Codes can be data-driven or theory-driven. Theory-driven focus on research questions(s) where you search for the data which captures the objective of our research question(s). In the inductive analysis, every single line is coded. It also depends on whether you want a rich description of the data set or a detailed account of one particular aspect.
Coding can be closed or open coding. Open coding does not have pre-established codes and is developed during analysis while close coding has a pre-established coding scheme and is searched in the data.
The general idea of codes comes to the mind of the researcher while making rough notes of each transcription. As there will be some issues that come up repeatedly.
It can be done either manually (through hardcopies with pens and highlighters) or through data analytic software (Nvivo, ATLAS, webQDA, dedoose, MAXQDA, PROVALIS, HyperRESEARCH, Quirkos, RQDA). The website link for the software is in Appendix 1. You can also use Microsoft Excel to code and generate code.
Step 3: Search for themes
Similar codes that easily fit in are made into one potential theme and similarly, other themes are formed. There is no hard and fast rule to creating themes. One code can be associated with various other themes. If the codes do not fit into any themes, then a separate “miscellaneous” theme is made. However, themes do not reside in the data set, they reside in researchers’ heads, in their way of thinking and understanding the data. So, for the same data set the theme generated by the two researchers can be different. It’s the researchers’ decision to determine what the theme would be. There is no right and wrong method for determining the prevalence of theme in the data set but consistency is important.
At the end of this step, you have a few themes that describe the research question(s).
Step 4: Review themes
The preliminary themes that you have created in step 3 are reviewed and modified. Themes must be unique to each other and make sense. The codes that are kept under certain themes are looked upon as whether they are associated with the theme. The theme will also need consideration so as to know whether the theme works both within a single interview or across all interviews.
It involves two levels of reviewing and refining your themes;
Level one: You need to be sure that the codes entered in the particular theme appear to form a coherent pattern.
If it does not then, the theme can itself be problematic, for that you can make a new one. If codes do not fit, then you can find a new theme or fit it in an already existing theme or discard them from the analysis. Once satisfied with the theme and codes, move to level two.
Level two: You validate whether the theme generated is representative of the data set as a whole. In this step, you again read the whole data set for two purposes. The first is to check whether the themes capture the essence of the data set. The second is to code the missing data within themes. If you are able to make a thematic map, then you can move to the next step. If not, review and refine the code until a satisfactory thematic map is created.
Need to remember
- Do not try to fit too much into a theme
- Are there subthemes within themes?
- Are there other new themes within the data?
At the end of this phase, you should be able to tell the themes and overall story of the data.
Step 5: Define and name themes
This step begins after you have a satisfactory thematic map.You need to consider what the theme is about and how they are interrelated to each other in this step. If there are subthemes, how do they interact and relate to the main theme? You may have to define and refine the theme. The themes generated should not be too diverse or complex. It must represent why the data extracted for the theme is of interest to us. The analysis involves moving back and forth between the entire data set.
The individual theme should be analyzed in detail and the “story” needs to be identified that each theme tells. The story should be representative of the whole data set.
At the end of this step, you should be able to distinguish which is your theme and which is not and prepare a short story behind each theme. If you are not able to describe the scope and content of the theme in a couple of sentences, then further refinement is required. In this step, you start to give catchy and attractive names to the themes.
Step 6: Producing the report
Although writing up is at the end but writing should begin from step 1 from writing rough notes to potential coding and the entire coding/analysis process. This step officially begins when you have the full set of themes. It is important that the analysis provides a short, clear, logical, interesting, logical, non-repetitive, and interesting story within and across themes. All themes must have sufficient data to exhibit the prevalence of the theme without further complexing the theme. The theme must have sufficient examples of the issue. Writing must be beyond narrating the data, you should present the argument based on the research question/objectives. You should broaden your analysis from a descriptive to interpretative level (referring to similar research).
According to Braun and Clarke, the question you need to ask at the end of the analysis are;
- “What does this theme mean?”
- “What are the assumptions underpinning it?”
- “What are the implications of this theme?”
- What conditions are likely to have given rise to it?”
- “Why do people talk about this thing in this particular way (as opposed to other ways)?”
- What is the overall story the different themes reveal about the topic?”
- Maguire M, Delahunt B. Doing a thematic analysis: A practical, step-by-step guide for learning and teaching scholars. All Ireland Journal of Higher Education. 2017 Oct 31;9(3).
- Braun V, Clarke V. Using thematic analysis in psychology. Qualitative research in psychology. 2006 Jan 1;3(2):77-101.
- Alhojailan MI. Thematic analysis: A critical review of its process and evaluation. West east journal of social sciences. 2012 Oct 14;1(1):39-47.
- Taylor GW, Ussher JM. Making sense of S&M: A discourse analytic account. Sexualities. 2001 Aug;4(3):293-314.
- Karlsen MM, Gabrielsen AK, Falch AL, Stubberud DG. Intensive care nursing students’ perceptions of simulation for learning confirming communication skills: A descriptive qualitative study. Intensive and Critical Care Nursing. 2017 Oct 1;42:97-104.
- Gagnon LL, Roberge GD. Dissecting the journey: Nursing student experiences with collaboration during the group work process. Nurse education today. 2012 Nov 1;32(8):945-50.
|RQDA (Free Software)