Philosophy: ontology and epistemology

It is impossible to consider the development of rigorous qualitative research without being cognisant of philosophical approaches. Epistemological and ontological positions include conceptions of subjects and subjectivities, and understandings of how knowledge and reality are constructed and produced (Mauthner and Doucet 2003).

A ‘positivist’ (quantitative) ontological perspective understands reality as being external to the social world, objective and measurable, with the researcher neutral (Dingwall and Staniland 2020). In contrast, ‘critical’, ‘constructionist’ or ‘interpretivist’ qualitative ontological perspectives understand reality as being constructed through the interaction of social actors, with the world subjective and relational and the researcher present within it (Dingwall and Staniland 2020).

Epistemology reflects a researcher’s understanding of what constitutes knowledge and how they gain it – through either experience (a posteriori knowledge) or investigation (a priori knowledge) (Howell 2013). If researchers are aware of their epistemological perspectives, they can consider how to conduct their research to derive knowledge. Researchers who believe reality is constructed through the interaction of social actors will acquire knowledge using interviews and other qualitative methods that enable them to understand people’s experiences.

Ontology and epistemology therefore cannot be separated and must be scrutinised to align optimally with the appropriate methodological approach (St Pierre 2021).

Methodology

It is important to be mindful of the distinction between a methodology and a method: the former is a set of diverse principles, procedures and frameworks while the latter is a tool for collecting data (Silverman 2021). Nurses and other allied health professionals use phenomenology, ethnography, grounded theory, case study and other methodologies as underlying epistemological philosophies for their research (Silverman 2021).

Method

Methodological choice has implications for the choice of method. Qualitative methods include interviews, focus groups and (ethnographic) observations. Researchers must choose methods for their research that align with their methodologies’ philosophies. For example, phenomenology seeks people’s experiences, so one-to-one interviews are the most appropriate method to use; a focus group would be inappropriate as it elicits the experiences and views of the whole group.

Analysis

Researchers must choose a technique to analyse their data that aligns with their study’s methodology and method. For example, if they are conducting a study that has a hermeneutic phenomenological methodology, they would analyse their data using hermeneutic phenomenological analysis (Dibley et al 2020). Likewise, a study using GT methodology, might use grounded theory analysis (GTA). Other data analysis techniques might be less prescriptive in terms of their allegiance to a specific methodology such as thematic analysis (TA) (Braun and Clarke 2021a).

1. Familiarisation

Familiarisation begins with data collection and continues through the development of data transcripts. Researchers then immerse themselves in their data by reading and re-reading the whole data corpus. To prevent selection bias, no undue attention is paid to one data point more than the others.

Familiarisation is not a passive endeavour. Instead, researchers actively and critically engage with the data by circling, underlining and highlighting – as demonstrated by this extract from the ethnographic study’s reflexive diary.

‘Ethnographic field notes were translated into transcripts and semi-structured interviews transcripts returning from the transcriber were read and re-read. Although a temptation existed to focus on certain events that had become “sticky in my mind” or interview narratives that struck an emotional chord, I methodically read through the data highlighting, circling and underlining text and jotting notes in the margins. On occasions, I struggle to remain on-task, [which prevents] me from giving equal attention to all the data…’

2. Coding

The circling, underlining and highlighting of familiarisation give way to more formal ‘coding’ approaches in the second step. Coding involves applying to a section of data a word or a label that encapsulates and summarises the data’s content. The resulting ‘codes’ help the researcher to look at the granular level of the data to reveal patterns and differences that the researcher can then explore.

Managing the data

The researcher can code data by hand or using software such as ATLAS.ti or NVivo. ATLAS.ti was used in the ethnographic study to store, manage and code the data. However, coding initially began by hand, with ER using highlighter pens and Post-it notes to initiate her immersion in the data in an active, tangible way (see Figure 2).

Figure 2.

Example of data coded by hand

The researcher’s choice of coding method will depend on the volume of the data they collect, their chosen methodology, the researcher’s technical expertise and whether they are coding independently or as part of a team. However, the use of analysis software has been criticised on two grounds (Mauthner and Doucet 2003):

• • It can create a distance between the researcher and their data.

• • Novice researchers may believe the software will analyse the data for them and add an air of scientific objectivity to a fundamentally subjective, interpretative process.

Codebooks

It can be particularly helpful when coding in teams to also create ‘codebooks’ to define the meanings of the codes (see Figure 3). However, Braun and Clarke’s (2021a) arguments against their use included:

Figure 3.

An example codebook

• • Codebooks are positivist in their orientation as they rely on predefined, deductive codes.

• • Researchers do not need to agree codes, as this can produce superficial themes and stifle creativity.

However, in our experience, a research team can develop a codebook inductively, growing it organically from the code list; the team also does not need a codebook to reach a consensus. In addition, we have found value in displaying different researchers’ interpretations, as it can create avenues for more nuanced, profound and rich themes to emerge.

3. Theme development

When a researcher is satisfied that they have achieved information power, they can begin the process of theme development. This is a messy, time-consuming challenge, yet ultimately is rewarding.

To start developing themes, the researcher brings together codes with similar shared meanings to create subthemes and then an over-arching theme. The same codes might fit within several themes or subthemes.

However, other codes may not fit neatly into any group. These outliers should not be forced into groups and should be discarded if they cannot contribute to the overall narrative (Braun and Clarke 2012).

There is no definitive number of themes that should be developed. However, dissertations, published papers and reports typically present three to six (Braun and Clarke 2012).

It is important when developing themes and subthemes that researchers build a coherent narrative that answers the research question in a meaningful way. Themes should be distinctive, with each contributing something important to the overall story, enabling researchers to tell a coherent and comprehensive narrative.

The theme’s size or the number of codes it comprises is not an indicator of its importance or value; rather, it is the meaning and the analytical interpretation of the themes that is important. For example, in the ethnographic study, the theme of ‘ghost stories and the haunted mind’ was derived from a small proportion of the interview data. However, it was conceptually important in explaining how ambulance crew members managed their emotions on the road as they moved across the city, ‘haunted’ by memories of traumatic past jobs.

Visual techniques can help to explore different layers of the data. Collaborative software can also assist in organising data by enabling researchers to place virtual Post-it notes in piles or to draw thematic maps (see Figure 4).

Figure 4.

Code mapping

4. Reviewing themes

Once provisional themes have been developed, it is vital that researchers confirm the themes represent the data and that the narratives developed are coherent, do not overlap and contribute to the overall story.

The researcher can begin this process by writing a summary of the theme and subthemes. If this is challenging, or subthemes are being incorporated into other themes, the themes require revising. This is an iterative process, in which researchers may need to re-examine how they have grouped codes.

Once a concise and bounded summary of the themes and their subthemes has been written, check the sizes of the themes. Interrogate disproportionately smaller themes, as these may be more appropriately represented as subthemes.

Once satisfied that themes and subthemes are internally consistent and appropriately supported by the raw data, researchers can confirm them by defining and naming them. You should also order the themes to create a coherent and logical narrative.

5. Confirming themes

In this penultimate phase, researchers will move beyond describing the data to analysing by bringing into dialogue with theory and the wider literature to think about what it means and why it is important. This is often a challenge but a spreadsheet can provide invaluable assistance in analysing and interpreting the data to produce theoretically informed interpretations.

An excel spreadsheet can be invaluable in assisting the analysis and protestation of the data. In the first column, codes can be placed, in the second column, the title of the subtheme speaking to the main theme is presented.

Example quotations supporting the subtheme should go in the third column, while the fourth column should include theory, policies and wider literature that align with or contradict the data.

The researcher can then sit and ponder the data in the spreadsheet, thinking critically about their meanings and interpretation. Finally, place the resulting analytical claims in the fifth column of the spreadsheet.

Table 1 shows an example spreadsheet from the ethnographic study that explores the theme of ‘emotional detachment’.

Table 1.

Moving from description to analysis

6. Dissemination

The final stage is disseminating your findings as a report, dissertation or publication. A challenge researchers may encounter here is how to choose which data to publish, particularly when the word count available is limited, such as in a peer-reviewed publication.