The CAQDA Paradox: A divergence between research method and analytical tool

Abstract A wide range of software tools are available to assist researchers with the process of qualitative data analysis.These include tools that emphasise manual handling of data, (e.g. NVivo, Atlas.ti) and tools that provide someautomated analysis based on statistical properties of texts (e.g. Leximancer). These tools are enhancingresearch, making research activities less complex and tedious, and rendering the process more transparent andportable (Dohan et al. 1998; Welsh 2002; Andrew et al. 2007; Jones 2007). The use of these tools in publishedworks over the last five to ten years has become increasingly more evident. However, in many cases, thisincrease in frequency of use is also masking the actual method of research. Many researchers who use termslike “Data were analysed using NVivo” are using their chosen analytical package as a proxy for actualembedded methods of analysis. It is possible therefore that Computer-Assisted Qualitative Data Analysis(CAQDA) tools are becoming a substitute for actual, and perhaps valid, techniques for research, analysis anddiscovery. This paper investigates the extent of this problem, examining CAQDA based papers which havebeen published over the last five years and reporting on their use, or misuse, of methodology. Further, thispaper proposes a solution to the problem by adopting a CAQDA technique which utilises a generic style ofmethodology. A tool used by Quantitative researchers, known as ‘R’, is available which is a free, open sourcestatistical programming language. Within the last five years R has become the lingua franca for statisticians andapplied workers to publish reference implementations for novel quantitative techniques. No such tool withsufficient flexibility exists for qualitative researchers. We describe the initial development of a transparent fileformat and research process which keeps the researcher close to the data and provides strong safeguardsagainst accidental data alteration. This has two main effects. The transparency of the file format keeps theresearcher close to the data, and ensures that the researcher keeps in mind the process used to analyse the datarather than the tool in use. The second effect, also related to the open source, transparent plain text basis of thetool, means that an environment for fostering innovation in qualitative data analysis can be easily providedand freely distributed among workers in the field.