How is qualitative data analysed?

The previous article focused on what is qualitative data and its benefits. This article focusses on techniques used to analyse qualitative data. The choice of technique depends upon various factors. This typically involves the research question, its theoretical foundation, and appropriateness of the techniques.

There are four primary techniques of qualitative data analysis:

  1. grounded theory,
  2. content or thematic analysis,
  3. coding and analysis using NVIVO,
  4. and framework analysis.

Grounded theory

Grounded theory is a data collection and analysis method that aims to develop logical theories. The main purpose of grounded theory is to develop a theory after thorough analysis. Therefore, each stage in its analysis works towards theoretical formations. It involves inductive strategies for analysis. The researcher starts with individual cases, incidents or experiences and proceeds by developing more abstract categories. It is an attempt to explain, understand and identify patterned relationships. It provides a systematic procedure for shaping and handling qualitative data. However, it can also be applied to quantitative data (Khan, 2014).

Grounded theory can be applied to researches such as interpersonal relations and the interplay between individual and larger social processes. This method of data analysis is useful to analyse socio-psychological aspects such as personal experiences, motivation, attraction, emotions, identity, interpersonal conflicts and co-operations (Anselm & Juliet, 2010).

Step 1: Coding qualitative data

Grounded theory coding process for qualitative analysis
Grounded theory coding process for qualitative analysis (Pozzebon, Petrini, de Melio & Garreau, 2011)

The first phase in the analysis process of grounded theory is coding. Under this step, the researcher gives codes to the data and then brings them in the abbreviated terms to explain the observations. It helps to conceptualize the basic processes that occur in the search setting. According to the grounded theory, there are three types of coding:

1. Open coding

Under this method, the data is broken analytically to gain better insights from the data. In open coding, different events are compared and labelled. Then the similar ones are grouped together in order to form categories and subcategories. For example, to note interactions between nurse and clients, which is directed towards providing comfort, labels such as “comfort work” can be used. This category can now be broken into various other dimensions. For instance, comfort work has a duration of service and manner in which it is carried out. This categorization provides precision to the grounded theory (Khandkar, 2011).

2. Axial coding

Axial coding relates to the subcategory and the relationship are tested against the data. Under this type of coding different categories are developed and look for their indication is a continuous process. Continuing with the above example where comfort was categorized, work is carried out in different manners. The next step would have been to scrutinize the data to determine what conditions gave rise to it.

3. Selective coding

Selective coding method emphasises on determining the core category that requires further explanation. Such core categories require descriptive details. This kind of coding mainly occurs in the latter stages of the study.

Step 2: Memo-writing

At this stage, gaps are identified in the codes. It involves defining the code, delineating and analysing the properties, the conditions under which changes exist, and demonstrating relationships.

Step 3: Theoretical sampling

The next step is to acquire more data to develop, refine and check the boundaries for causes and consequences. This helps to raise the conceptual level of the analysis through comparison between theory and practice. Identification of gaps and the questions may require to return to the same field setting for obtaining data in a different setting that possess the characteristics to test. This comparative analysis helps to build the theory.

Thematic or content analysis

Thematic analysis is another way of analysing qualitative data. It involves systematically identifying, organizing and offering insights into the pattern of data. It helps to observe and make sense of collective meanings and experiences. It also focuses on identifying and explaining the common way of understanding an issue (Crowe et al., 2015).

Thematic analysis process
Thematic analysis process (Attride-Stirling, 2001)

Step 1: Coding qualitative data

The first step involved in thematic analysis is to compile the data. This is done by converting the text into manageable and meaningful segments through coding. The researcher first devises the coding framework.

For simplicity, this stage can be further divided into the following:

  1. Devising a coding framework: There are multiple ways in which the framework can be developed. However, it is generally done on the basis of the theoretical interest of the researcher, or silent issues that are present in the text, or on the basis of both.
  2. Dissecting using coding framework: The codes are applied to the text data in order to break it into text segments which are meaningful and manageable. The codes set should not be interchangeable and redundant. Also, they should focus exclusively on the object of analysis.

Step 2: Identifying the themes

With the coding done in the first stage, the themes are extracted from the coded text in this stage. For this, the themes must be abstracted and refined.

  1. Abstract themes: This involves the extraction of salient, common or significant themes by observing the text within the context of the codes under which they have been classified. This step helps to uncover the underlying patterns and structure.
  2. Refine themes: This involves further refining of the selected themes that are specific enough to be discrete and broad enough to cover a set of idea. This reduces the text and makes it manageable. Since the aim is to summarize the text, close attention to concepts is required.

Step 3: Representing the themes

This step focuses on the arrangement and the representation of the identified themes from the last step.

Arrange the themes

In order to form the grouping network, the first step is to assemble the themes derived in the last step on the basis of similarities. The similarities relate to the content and theoretical background. There is no rule regarding how many themes should be there in a network. However, more than 15 will be too many to handle and less than 4 will not do justice to the data.

Representation of themes

Once the themes are prepared, the next step is to present them as nonhierarchical and web-like presentation. This helps to interpret the original text. It involves further abstraction of data for analysis. Since the thematic analysis is the tool and not the analysis itself, deeper meanings have to be derived from the data.

Summarize the themes

After exploring the themes, the next step is to summarize the main themes and patterns defining them. The goal is to study the patterns that emerge.

Interpret patterns

The main focus of this step is to underpin the original research question on the basis of the themes developed in the exploration of the text.

Coding qualitative data using NVIVO

Qualitative data is known for its subjectivity, richness and comprehensiveness. Thus, analysing qualitative data is generally a muddled and time-consuming process. Traditionally researchers used props such as colour pens in order to categorize the data. However, software such as NVIVO has eliminated this task, thus giving researchers more time to discover theories, themes and derive the conclusions. NVIVO has significantly improved the quality of research (Leech & Onwuegbuzie, 2011).

Coding in NVIVO using coder

Coding in NVIVO involves desegregation of textual data into manageable segments and categorizing them based on similarities and differences. Thus, grouping the similar ones in the respective nodes.

Making and using memos

While analysing the qualitative data, the reflective ideas, theories and concept arise as the data is observed. NVIVO provides flexibility to record ideas as they emerge in memos.

Creating attributes

Attributes refer to the characteristics such as age, marital status, ethnicity or the educational level. NVivo helps to assign the possible attributes to the nodes. The attributes can further be removed, added or arranged to assist the researcher while making the comparisons.

Using models to show the relationship

Models are an essential way to describe the relationships in qualitative studies. NVIVO provides a visual exploration of the relationships. This helps to create, label and connect the ideas.

References

Ashni Walia

Research consultant at Project Guru
Ashni is a master of Economics from Amity University. She has been an active
member of Enactus and has participated in the 12th sustainability summit. She was also associated with the YES Foundation during her master’s programme. Apart from her interest in research, she has a keen interest in music and
reading fiction.
Ashni Walia

Discuss