Theoretical sampling

It is important to understand that in qualitative research, 'size' does not mean 'significance'. Here is a long-standing issue that many qualitative researchers face, especially those whose research environments are strongly quantitative: how to deal with qualitative sampling? Achieving significant data leads researchers to sample participants according to their knowledge and experience regarding a specific phenomenon: commonly described as 'purposive'. Purposive or purposeful sampling can be defined a priori, that is to say, before the beginning of data collection process (Charmaz 1990). Purposive sampling and its underlying decision-making process are, usually, driven by 'tradition' and 'best practices'. For example, Mason (2010) gathered sampling indications from qualitative researchers: Morse (1994) suggests conducting 30–50 interviews for ethnography, ethnoscience and grounded theory; for a phenomenological study, she indicates at least six interviews/participants. Creswell (1998) conveys 20–30 interviews for grounded theory and from five up to 25 interviews for phenomenology. At any rate, the issue is still there, as Mason (2010) stresses: 'While these numbers are offered as guidance, the authors do not tend to present empirical arguments as to why these numbers and not others for example. Also the issue of why some authors feel that certain methodological approaches call for more participants compared to others, is also not explored in any detail' (Mason 2010, p. 11). Saturation is the regulatory idea for sampling size in qualitative research and has to do with the significance of collected data. But again, Mason (2010) quotes Guest et al. (2006, p. 59) to suggest that 'although the idea of saturation is helpful at the conceptual level, it provides little practical guidance for estimating sample sizes for robust research prior to data collection'. Even if qualitative researchers can indicate a possible number of participants to be involved in the written research proposal, the overall process of data collection remains uncertain because of saturation. What is really unpredictable is another 'qualitative' sampling method, known as 'theoretical sampling', which is a unique feature of grounded theory. 'With grounded theory strategies, theoretical development turns on theoretical sampling. Here, the researcher collects new data to check, fill out, and extend theoretical categories […] Hence, theoretical sampling fits into the research and analytic process much later than initial sampling of sites, people or documents. By the time theoretical sampling is planned, a researcher would have some hunches or even hypotheses which he or she wishes to check. Thus, theoretical sampling shapes further data collection as the researcher pursues developing conceptual ideas rather than amassing general information' (Charmaz 1990, p. 1163). So, if 'participants in different situations are chosen as they are needed to help the researcher clarify understanding' (McCallin 2003, p. 204), how can we negotiate the research process with funding agencies? Our experience involves a study conducted in a large hospital in Northern Italy on the validation and application of standards on the involvement of children and their families in the healthcare process. Together with a panel of experts, we chose to work on the therapeutic self-care outcome, which induced us to develop within the scope of this project with a quantitative design, a secondary study on the psychosocial processes of therapeutic self-care. The aim of the secondary study was to build a theory emerging from the data on the psychosocial processes of children and their families related to therapeutic self-management in terms of self-care, in relation to the emotional and psychological dimension. In particular, the qualitative part of the research was included in a quantitative study, enabling us to define a benchmark theoretical model regarding the styles of disease management during and after hospitalisation. The qualitative study proposed at this stage is generating a cultural shift in the way all the members of the healthcare team listen to the history and perspective of children, adolescents and their families. This is triggering a process that enables qualitative research to create a space of its own, which is increasingly appreciated by all the members of the healthcare team, who are discovering the usefulness of this method. In particular, the physicians and nurses of the healthcare team have started to highlight the importance of integrating quantitative studies on therapeutic treatments with qualitative research. According to Tarozzi (2013), qualitative researchers should see funding agencies as a research constraint to be taken into account while designing research, as they 'have an impact […] on the methodological level. In this sense, they affect in particular the qualitative research process. From this perspective, qualitative research seems to be more fragile and exposed to the danger of inopportune intrusions than quantitative research. Being based on somewhat less rigid and controllable procedures, qualitative research is exposed to a major risk of intrusion or to be subdued to the funder's expectations' (Tarozzi 2013, p. 5). The difficulty of 'getting funded' that qualitative researchers often face due to the unpredictability of sampling, must become an incentive to negotiate, convey, and – why not? – educate funding agencies. In fact, not being able to predict sample size and participants' characteristics strictly does not mean lack of rigour and method, or of clear and thorough explanations. So, as funding agencies have an impact on the methodological level and are one of the research constraints, mutual (researchers' and funding agencies') expectations must be shared and examined to negotiate effectively: a shared idea of qualitative research; of theoretical and epistemological underpinnings; and credit to the rigour of qualitative research. This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. No conflict of interest has been declared by the authors.