Consultancy in Qualitative Data Analysis
The art and science of qualitative analysis
Qualitative analysis, or the art and science of analysing qualitative data qualitatively, has historically being overlooked in the teaching and publication of qualitative research methods. While dozens of books are available on the politics, ethics and philosophy of qualitative inquiry, and roughly a similar number has been published on data collection methods, there is very little literature on methods to analyse qualitative data. While PhD programmes typically offer general courses on qualitative research, these mostly cover ethics and data collection, and qualitative analysis — if covered at all — is paid scant attention. And this is only in departments that view qualitative research as a legitimate form of social inquiry — as otherwise, qualitative methods are simply ignored. Interestingly enough, this state of affairs radically contrasts with the well-established statistics courses available in graduate programmes, where different methods are taught at introductory and advanced levels. It is therefore hardly surprising that graduate students pursuing an academic career in qualitative research, as opposed to quantitative research, have gaps in their knowledge of qualitative data analysis.
Conducting qualitative analysis and reporting its process is, consequently, the weakest link in the research design of a qualitative study. Beyond the catch phrase ‘themes were identified’ (just put that phrase, followed by ‘qualitative’, in Google Scholar and you’ll see how many hits result!), discussions rarely explain how researchers actually analysed their data. Not only does this make it impossible for the reader to trace the basis for the reported findings, it is also erroneous to think that qualitative analysis is merely about identifying themes in a dataset. Theme identification is certainly part of the process but the main question qualitative researchers must answer is: What do you do after you identify themes? Unfortunately, the answer to that question is often a black box that readers are left to figure out for themselves! This is one of the reasons why qualitative analysis has rightly been criticised for being opaque and for lacking systematicity.
What qualitative software does and does not do
The other key problem associated with qualitative analysis is the myth that qualitative software will do the data analysis for you. But qualitative software programs merely process data, nothing more. No qualitative software is equipped, as yet, with in-built templates or buttons that will automatically apply grounded theory or a phenomenology and conveniently return an output ready for publication. I compare CAQDAS to a microwave oven: both enhance efficiency while keeping everything in one place. But microwaves, like CAQDAS, can also make a mess of what you intended to do, or worse, destroy it altogether if you do not go about things in the right way! This is why I always advise first mastering your method of analysis and only then applying it in the CAQDAS of your choice. The reasoning is simple: it may turn out that your method cannot be applied in a given CAQDAS due to technical limitations that — very often — mirror the epistemological stance of the developer. So, never buy a CAQDAS licence without first finding out that it will do what you want it to do, the way you want it done.
The above issues, along with others I discuss below, are the main reasons why my expertise in teaching, conducting and reporting qualitative analysis (with or without CAQDAS) is sought by universities and research centres. Consultancy typically takes the form of:
- Teaching the methods of qualitative data analysis
- Recommending the use of a particular analytical method for a given qualitative study
- Teaching approaches to coding qualitative data
- Teaching pattern-seeking techniques for qualitative data
- Supervising intercoder reliability with teams
Below I describe some of the most common questions and situations I deal with when working with qualitative researchers.
What method should I use to analyse my data?
In qualitative research, a lot of time is often invested in data collection and little time is given over to analysing data properly and thoroughly. Researchers typically end up with a pile of data but have no clear analytical strategy in mind and may not even know which method to use to analyse their data. The ‘grounded theory’ approach is often evoked as the method of choice, although, in most cases, the procedures are not implemented properly, or not at all, and no theory is developed. In such cases I go back through the research process to look at the research questions, research purpose and type of data collected, and then propose the most suitable qualitative data analysis method.
I have finished coding my data. What should I do next?
Coding is part of the data analysis process, but coding in itself is not analysis. The primary task of coding is to sort large chunks of data into categories that are assigned codes so that patterns can be sought and relationships uncovered in the data. If you have coded your data but wonder what you should do next, then a key component of your data analysis is missing. In this case I review the assumptions, hunches or hypotheses underpinning your research, show you how to confirm or falsify these and, most importantly, make sure that the coding you did allows this to be done.
I am coding my data again and again and again
In her article Going the distance: ‘closeness’ in qualitative data analysis software, Gilbert (2002) refers to the coding trap — repetitive waves of coding — and how it hinders and even delays analysis. If you’ve been there, you know that there are always good excuses for doing some more coding, recoding your data differently or endlessly changing your code structure. Do not fool yourself, however; such excuses bog you down in even more confusion and simply waste precious time, leaving you no nearer to your goal. In this situation, I would put the brake on the temptation to code or recode and propose what you should do next to conduct analysis.
I cannot see the bigger picture in my data
Focusing on every little detail of the data, the idiosyncrasies of a case or the uniqueness of an interview makes you lose sight of the bigger picture behind your data. Stepping back from such micro-level concerns and looking at your dataset from a bird’s eye perspective enables you to identify and pursue patterns and weave your study results. In such situation, I would teach you the ability to move back and forth between data and the bigger picture, which is a core skill in qualitative analysis that is seldom taught.
How can I substantiate my claims?
Claims are the highlights of a study, and presenting evidence to back up claims is as important as the claims themselves. The audit trail, introduced by Lincoln and Guba in their book Naturalistic Inquiry in 1995, proposes 38 different types of evidence researchers can present to support the results of their qualitative study. However, when researchers report qualitative findings, they largely fail to show the audit trail that enable results to be traced back to how the data were analysed. In this situation, I would review the procedures you used to generate your findings, and make these procedures transparent using notes, analytical structures, models, summaries or indexes.
I need help to set up my dataset in NVivo
NVivo is a powerful platform that supports qualitative analysis, but it comes with its own logic, jargon and technicalities. While YouTube tutorials are useful in helping you organise and code data, you will need to invest time and effort in deciphering how to use key NVivo analytical features that will enable you to work with cases and variables, create sub-sets, establish relationships, generate and test hypotheses and present results in evocative graphic displays. I teach two NVivo courses, one focusing on the software’s functionalities and the other blending its core functionalities with key qualitative analysis concepts. Depending on your needs, I would coach you in setting up your dataset and teach you some core concepts that should accompany the qualitative data analysis process.
In the last ten years, I have collaborated as an expert qualitative data analysis consultant in several international research projects. My participation has often involved providing bespoke courses on methods of qualitative analysis, coaching in the coding process, supervising intercoder reliability done with teams and providing training in the use of NVivo software. Some of the projects I have consulted for are described below.
Courses taught at
As a freelance methodologist, I train social scientists and humanitarian practitioners in qualitative analysis, decolonising research and participatory methodologies. I coach research teams, teach doctoral-level courses and I consult for humanitarian aid agencies worldwide.