Qualitative data analysis is a varied field with numerous paradigms, methodologies, techniques and tools so it can be overwhelming to navigate the different ways data can be analysed and interpreted.
However, whatever data you’re working with, methodology and analytic approach you’re taking, or digital tools you’re using, there are five core high-level 'analytic activities' that we all engage in: integration, organisation, exploration, reflection, and interrogation.
Ann Lewins and I developed this framework for navigating these complex and interrelated issues for our teaching of Computer Assisted Qualitative Data AnalysiS (CAQDAS) many years ago - the original version of which we published in the first edition of our textbook Using software in qualitative research: A Step-by-Step Guide (Sage 2007). We have built on it in subsequent publications (Silver & Lewins, 2014; Woolf & Silver 2017) and are currently revising it again for the 3rd edition of our textbook.
The framework can be used to introduce researchers to the analytic activities involved in all qualitative and mixed-methods projects and to the potential of CAQDAS-packages for facilitating analysis and interpretation.
Checkout our video about Analytic Activities in qualitative and mixed-methods research https://bit.ly/QualAnalyticActivities
Commonality amongst the diversity: analytic activities
What we do when we analyse qualitative data and how we do it, is contingent on the analytic methods we employ. So, when doing a thematic analysis, we develop different analytic tasks than when we do a discourse analysis or a grounded theory, and so on. And so it follows, that we use digital tools to accomplish these analytic tasks in different ways, for different purposes, in different sequences etc.
Nevertheless, five high-level 'analytic activities' are commonly engaged in across qualitative and mixed methods approaches, which are useful for thinking through analytic design and when learning about and preparing for the use of CAQDAS-packages. They are high-level activities rather than specific techniques or tasks, and the extent, manner and sequence in which they are undertaken varies according to research objectives, methodologies and outcome requirements.
Figure 1. below summarises the common analytic activities. When teaching QDA this framework allows different analytic methods (what we plan to do, our strategies) to be discussed concretely and practically in the context of CAQDAS tools (how we plan to do it, our tactics).
Figure 1. High-level Analytic Activities involved in qualitative and mixed-methods analysis
and facilitated by CAQDAS-packages (adapted from Silver & Lewins 2014).
Integrate (combining parts into a whole)
Integration means bringing things together, or combining parts into a whole. Three aspects of integration need considering in every qualitative or mixed-methods project:
Integrating materials. Materials include not only primary or secondary data we collect or harvest, but any background or supplementary items relevant to the project. The types of materials we work with, their role in the project and their relationship to one another in the research design (e.g. sequential, parallel, etc.) inform how to set-up an analysis within a CAQDAS-package and which tools are appropriate to use for analysis and interpretation.
Integrating analyses. Qualitative data can be analysed in a variety of ways, and this needs to be planned. At the highest level considering whether to analyse data qualitatively, quantitatively or using a mixed or combined approach is useful. Analytic methods such as thematic analysis, Interpretive Phenomenological Analysis (IPA), narrative analysis, content analysis, discourse analysis etc. can all be facilitated using CAQDAS-packages. Some projects use one analytic method across all data, others use a different method for different data types, or look at the same data using multiple analytic methods. Again these decisions affect which CAQDAS-packages are most appropriate to facilitate parts of an analysis, and how their tools are harnessed.
Integrating contributions. When working collaboratively, it is also necessary to consider how team-members’ work will be integrated. Decisions about how to split work amongst the team and how and when it will be brought together have both human and technical dimensions and need to be planned for. Where you sit on the debate about reliability and consistency in qualitative analysis is crucial here. Consistency is also important when working as an individual, it’s just that the focus is consistency throughout an analysis, rather than between team-members. All CAQDAS packages allow the analytic contributions of different team-members to be assessed qualitatively, and some also provide quantitative (statistical) measures.
Organise (creating structures relating to objectives)
Organisation is about splitting things up according to what’s important and different about them to create structures related to the analysis objectives. Two aspects of organisation need consideration in qualitative and mixed-methods projects:
Facts. The factual characteristics about research materials and the units within the analysis need to be captured. These might be socio-demographic characteristics of participants, metadata pertaining to documents, or features of the interactions we observe. Some factual characteristics are known from the early stages and may be the basis upon which we sample. Others are identified as part of the analytic process. Either way, they need to be recorded so we can explore materials and interrogate our analysis in relation to these facts. CAQDAS-packages have powerful ways of capturing facts and using them to query data.
Ideas. The ideas we have about what is interesting and meaningful within qualitative data, and how we conceptualise, capture and connect them to represent that meaning is the second aspect of organisation in qualitative data analysis. There are three key ways that we capture ideas, each facilitated by CAQDAS-packages:
o Transcription is an important analytic act because decisions about what to transcribe and how transcripts are formatted have huge implications on what we can analyse and how we can go about it. CAQDAS-packages handle transcripts in different formats, with some enabling them to be synchronised with the corresponding audio/video recordings.
o Coding. Many analytic methods use coding as a primary means to capture conceptualisations of meaning. The many approaches to coding qualitative data are well supported by CAQDAS-packages (e.g. inductive/deductive, manual and automatic, conceptual/thematic/linguistic…and so on). In contrast to manual methods and digital tools not specifically designed for QDA (e.g. Word, Excel, etc.) coded data can be retrieved, reviewed, recoded, rethought at the click of a button and is never disconnected from the original source context.
o Linking. Sometimes coding is not the most appropriate way of capturing associations within data. For example, coding does not capture the relationship between data segments other than that they are an instance of what the code represents. Many CAQDAS-packages allow data segments to be linked to one another separately from coding processes and some allow those associations to be named, providing an additional way of capturing ideas and navigating data on that basis.
Explore (examining the inherent nature of data)
Exploration is about what’s inherently in qualitative materials at an explicit or surface-level. When organising data we work interpretively; deciding what is meaningful and how to capture that meaning is an interpretation. But it’s always important to also consider the inherent nature of the data. There are two aspects to consider:
Content refers to the words and phrases used in texts, and any other explicitly identified features of qualitative materials. Some analytic methods (e.g. content analysis and some forms of discourse and linguistic analysis), focus entirely or mainly at the explicit content level, in others content is just one dimension of interest, or is not relevant at all. CAQDAS-packages have tools to identify and capture content quickly and reliably, and these tools are also useful for familiarizing with large datasets even when the analytic focus moves beyond the surface level.
Structure. Many forms of qualitative data have structures inherent within them that are analytically important. For example, questions asked in a survey, topics discussed in an interview or focus-group, the structural elements of an interaction, and so on. These structures can often be captured within CAQDAS-packages quickly and reliably if data are prepared so they can be identified. Therefore thinking through what structure exists and whether it will be important to capture it is an important aspect of qualitative and mixed methods analyses.
Interrogation (following up what has been done)
Interrogation is about asking questions about research materials and the analytic work done with them to identify, compare and illustrate. Functionality within CAQDAS-packages designed to facilitate interrogation is incredibly powerful and offers a range of options not possible when using non-dedicated tools. This extends methodological possibilities and provides opportunities for illustrating the quality of an analysis in concrete ways.
Identify: All qualitative projects identify patterns, relationships and anomalies. What they are, how they manifest themselves within data and how you look for or reveal them differs according to research objectives and analytic method. CAQDAS-packages are particularly powerful for interrogation because they have a range of tools that reliably retrieve data based on how materials have been organised. Thinking about anticipated patterns, relationships and anomalies can help plan an analysis and provide direction.
Compare: On some level all projects have comparative dimensions. Some are explicitly comparative, for example when two or more individuals or cases, groups or subsets, or instances of a phenomenon are being studied. When working with small samples or homogenous groups comparisons are more conceptual. CAQDAS-packages enable comparisons to be made in relation to how data have been organised and explored, at any level of work or point in the process.
Illustrate: We also need to illustrate the outcome of our interrogations and depending on the analytic approach is accomplished by testing, validating and visualising. For example, when working deductively we may test assumptions, hypotheses or theories, when working inductively we will validate interpretations. CAQDAS-packages allow these tests and validations to be undertaken systematically and to be combined when working in multiple ways (e.g. both inductively and deductively). Visualisations are a key way that findings and processes are communicated and most CAQDAS-packages provide numerous visualisations based on the interrogations.
Reflect (considering carefully and deeply)
The essence of qualitative analysis is reflection, because throughout an analysis we reflect on everything we do, all of the time. This is why reflection is in the centre of the Analytic Activities diagram (Figure 1.). For example, the ideas we have that form the basis of an analysis, for example when coding, are an aspect of reflection, because we consider carefully and deeply what to code, when to code, how much to code, and so on. In addition, defining codes and amending those definitions as we review and refine our thinking throughout the analysis are explicitly reflective.
Reflection is inherent to the analytic process so it is important to plan both what to reflect on (our reflective focus) and how to capture those reflections so they are not lost, but can be included in the analysis (the tools we use to record our reflections).
The focus of reflections relates to what we reflect on. Broadly we reflect on the data themselves, the concepts we develop to understand the data, the processes we use, and interpretations we develop (which includes our role in developing interpretations, known as ‘reflexivity’ in qualitative methodology).
There are two key ways we capture our reflections
Writing. Rather than being a discrete stage of analysis, writing happens throughout. We write at different levels, for different purposes and about different things, from informal post-it note type musings to carefully constructed accounts for reports, articles and dissertations.
Mapping. Stepping away to consider connections on a high or conceptually abstract level is another way of capturing reflections. This includes mapping connections within and between data, mapping theories and explanations, and mapping findings. Most CAQDAS-packages include mapping tools that are designed to facilitate this work.
CAQDAS packages have several spaces in which to write and most also provide mapping tools (although these vary quite significantly between programs). Although writing and mapping can be undertaken using other tools – paper and pen or other digital tools – the benefit of capturing reflections within CAQDAS-packages is that what you write and map can be connected to the data that underlies your reflections. This ensures reflections are grounded in the process.
Iteration and flexibility in strategies and tactics
The Analytic Activities of Integration, Organisation, Exploration, Interrogation and Reflection are closely connected and the process of analysis is much more “messy” than a linear written description or a two-dimensional illustration can depict.
The reality of analysis is more interrelated than Figure 1. illustrates, due to the iterative and emergent processes involved. Many of the elements in the diagram are actually related to one another in a web of interconnectivity, and a three-dimensional diagram would better capture the associations. The way these activities are accomplished, in what sequence and to what extent, varies from project to project and is driven by practical and methodological considerations.
However, these analytic activities are relevant to all qualitative and mixed-methods projects, and provide a framework for introducing qualitative data analysis strategies (what we plan to do) and tactics (how we plan to do it) at a high level. The specifics of what happens in an analysis, and how analytic tasks are accomplished using manual tools, generic tools, or dedicated tools (CAQDAS-packages) are driven by the research design and analytic methods employed by each project. This framework offers a straightforward way to introduce the variety in approaches, to discuss iterative and emergent processes and plan an analysis.
For more information about the Analytic Activities framework and how it can be used to teach researcher to analyse and interpret findings, please feel free to contact me or refer to the following resources:
Silver C. & Lewins A (2014) Using Software in Qualitative Research: A Step-by-Step Guide, 2e. Sage Publications (3rd edition forthcoming)
Woolf, N & Silver, C. (2018) Qualitative analysis using ATLAS.ti / MAXQDA / NVivo : The Five-Level QDA Method, Routledge.
Youtube video: Analytic Activities in qualitative and mixed-methods research https://bit.ly/QualAnalyticActivities