Response to “Calling for a Shared Understanding of Sampling Terminology in Qualitative Research

What Is Already Known?

Grounded theory is a widely used method of qualitative research, with a particular approach to sampling and analysis.

What Does This Paper Adds?

Application of grounded theory in nursing research is often flawed, detracting from the value of this distinct method.

In their critique of our recent paper on theoretical sampling in the Journal of Advanced Nursing (McCrae & Purssell, 2016), Gentles and Vilches (2017) show the importance of terminological clarity in qualitative research. As keen exponents of grounded theory, Gentles and Vilches display a sophisticated and nuanced understanding of the principles and practice of theoretical sampling. Our paper, however, was written from a different standpoint to theirs. We strive to improve the quality of nursing research, and our scrutiny of grounded theory studies was within a theme that has also included systematic reviewing in nursing journals (McCrae, Blackstock, & Purssell, 2015). The same methodological standards should apply to research by nurses as to any other researchers, but as nursing scholars we are conscious of the particular needs of students and researchers in our discipline.

Nursing, indeed, has a strong qualitative tradition. Nurses consider patients holistically, being more orientated to care than cure. While medicine typically measures outcomes, nurses tend to be more concerned with the patient’s experience: How it feels to be the human being receiving treatment in a clinical environment rather than merely assessing symptoms. Interestingly, the original work on grounded theory by Glaser and Strauss (1965) was in palliative care, and Strauss had previously conducted notable research in mental hospitals (Strauss & Shabshin, 1961).

Nonetheless, qualitative research by nurses is not always of good quality. Rigor is as important in qualitative studies as in any other form of scientific investigation. As well as lecturers who teach, supervise, and examine academic work by nursing students, we have both had years of experience on research ethics committees, and our shared observation is of the principles and procedures of qualitative research frequently being applied loosely or haphazardly. Often, we have found researchers claiming to use grounded theory when core components of that method are not applied. Undoubtedly, the quality of nursing research has improved, judging by academic journals, but flaws persist in published studies, particularly in qualitative design and analysis.

Our paper, described by Gentles and Vilches as a “methods overview,” was actually a systematic investigation of the specific activity of theoretical sampling. In doing so, we made some assumptions that are challenged by Gentles and Vilches. Our critics rightly state that sampling in grounded theory is not necessarily of participants: The sampling unit is ideally theoretical or conceptual material rather than people. However, the prevailing approach in the studies included in our review was to sample by participants. Perhaps Gentles and Vilches could have examined a few of the studies to see this tendency. It is another example, we believe, of researchers following a priori convention instead of the inductive approach promoted by the grounded theory model.

Our figure of 50% of studies failing to apply theoretical sampling was likely to be an overestimate, according to Gentles and Vilches, because our definition of theoretical sampling was too narrow. We considered purposive and theoretical sampling as distinct. While it may be argued that the latter is a type of the former, there was no doubt that in papers we examined, the approach was purposive to the exclusion of theoretical. Several studies overtly sampled on people, with the number apparently decided in advance or by the limits of recruitment. Again, Gentles and Vilches could have looked at the studies in our review (our eligibility criteria were transparent and straightforward). Without doing this, their verdict on our numbers is speculative.

The meaning of purposive sampling is contentious, as indicated by Gentles and Vilches. There are technically correct definitions, and those that apply (rightly or wrongly) in practice. In the nursing literature (and possibly other fields too), purposive sampling is taken to mean sampling based on known attributes of participants (rather than emergent theoretical elements). This is the reality of qualitative research, and numerous methodological texts used by nursing students support such comprehension (e.g., Cresswell, 2003). Purposive sampling (as we understand the term) may be appropriate for the early stage of fieldwork, but for grounded theory to be applied properly, theoretical sampling is necessary for the development of analysis and to achieve theoretical saturation.

It is important to challenge misunderstandings and misuses of qualitative research technique. Dilution of the core concepts of grounded theory, if perpetuated in academe and in journals, dilutes the model. This debate illustrates inconsistent and nebulous definitions in this enterprise. While terminological latitude may be acceptable to researchers and scholars, study participants, patients, and other consumers of research have a right to expect consensus on fundamental concepts: who was selected for interview, and why? Gentles and Vilches make a useful contribution to the pursuit of clarity, and we wish them well in this valuable work.