Chapter 4. Methods of Discovery: Data, coding and theory generation
4.4 Theoretical Sampling
Beyond gathering first-hand accounts from all three therapy profession groups included in the study there was limited theoretical motivation for the inclusion of the first four participants. Following the analysis of the first four interviews, theoretically motivated questions began to arise which guided the next phase of recruitment. One theoretical consideration at this stage related to gender. There is a well-established literature,
96 extending beyond the health and care professions, which has explored the relationship between gender and aspects of professional practice such as professional identity (Hatmaker, 2012; Ten Hoeve, Jansen and Roodbol, 2013). As the four initial participants were all female, an early theoretical sampling consideration was that male therapists should be sought so that the researcher could explore the extent to which tentative concepts identified in the data might be gendered. A second consideration was prompted by sensitising concepts to be found in published literature regarding the relationships between professional identity and status and expertise, knowledge and skills (Apker, Propp and Zabava Ford, 2005; Nancarrow and Borthwick, 2009; King et al, 2015). Therefore, a theoretical sampling question which arose was whether the extent to which a therapist regards herself as highly specialist or experienced might influence the concepts being identified.
At this stage, the practical benefits of being a researcher with a network of professional connections in the substantive area of inquiry were immediately apparent; something which reflects the co-production feature of constructivist grounded theory (Charmaz 2006, 2014). The researcher was able to put out feelers and spread the word with contacts to indicate that in the next phase of her research she was looking to talk with male therapists and, or those in highly specialist roles. The inevitable slowing of the rate of recruitment necessary when theoretical sampling can be a challenge in a research study which is required to complete within a period of academic registration, however benefiting from the professional connections described, it was possible to achieve saturation of concepts in this research from nineteen interviews gathered over a period of 22 months between June 2015 and April 2017; something which is discussed further in section 4.6.3 of this chapter.
Further theoretical sampling considerations arose as new data were added to the study, constant comparison between existing and new data undertaken, and as the researcher began to explore tentative theoretical possibilities through the process of memo-ing.
New theoretical considerations were not necessarily apparent after each subsequent interview but more often after clusters of interviews were gathered and analysed. For example, after interviews five to seven, the researcher reflected that all participants to
97 this point had been experienced practitioners who had been qualified and practising for at least six years. This led the researcher to wonder about how a more novice practitioner might talk about supervision and whether indicators of tentative concepts identified in the first seven accounts would also be found in newly qualified practitioners’ accounts. This prompted a sampling focus on recruiting newly registered therapists for the research. In addition, this theoretical consideration guided the researcher to return to earlier data to look for instances which might reveal something about early career supervision and which may, until now, have gone undiscovered.
It is only once engaged in the concurrent collection and analysis of data that the value of theoretical sampling, constant comparison and memo-ing become apparent. Table 3, overleaf, summarises the major theoretical considerations which influenced the criteria for further sampling with the considerations and associated sampling characteristics or criteria set out in the order in which they arose during the research.
Theoretical Consideration Theoretical Sampling Participant Characteristics and Criteria
Gendered Male Therapists working in a new area or practice Therapists with emerging, non-traditional roles
Related to the therapist’s actual or perceived isolation from colleagues
Therapists actual or perceived visibility to others
Therapists who are physically isolated from colleagues; lone working in community services, large hospital sites, satellite services, the only member of their profession working in the setting or employer, working in rural settings
Therapists whose role is unlike others in the locality; high degree of specialism, emerging, non-traditional roles
Therapists who have different contractual arrangements to colleagues; secondment to a different provider e.g. health worker seconded to social services, temporary locum or bank contract, fixed term contract
Non-statutory services; independent sector, not-for-profit social enterprise or charity sector
Acute, community, speciality; specialist Urban, metropolitan, rural
Subject to cultural or ethnic influence
Declared ethnicity, trained outside UK, worked outside the UK Table 4: Theoretical sampling considerations and participant characteristics or criteria
98 Theoretical sampling was then enacted in three interrelated ways:
• framing the characteristics of subsequent interview participants
• prompting the researcher to review collected data afresh for previously undiscovered theoretical indicators
• shaping interviewer prompts to probe for instances which may support further theory building
Taking, for example, the theoretical consideration about the extent to which actual or perceived isolation from colleagues may influence emergent tentative concepts, purposive recruitment of participants who were working in more isolated settings was undertaken. This led the researcher to talk to therapists who worked in both urban and rural community settings where the practice involves regular lone-working with patients in their homes. As with the theoretical question around novice experiences, the researcher again returned to earlier data to look for possible concerns about practitioner isolation which had not previously been noticed.
The constant comparison of new and existing data for tentative concepts supports elaboration by identifying different dimensions of each concept. The memo excerpts presented in Appendix I, illustrate this theoretical development between interviews eight and eleven in relation to a concept of ‘isolation’; extending from a concrete, physical isolation to the recognition of a virtual or perceived isolation linked to practice status. The reference to other interview participants in each memo excerpt reflects the to and fro between new and earlier transcripts as the researcher seeks to saturate emergent categories.
The third way in which heightened theoretical awareness influenced data gathering was in sensitising the researcher to incidents and instances such that, during subsequent interviews, the researcher may include prompts which might not have seemed relevant before. For example, when interviewing Leanne, a therapist working in a rural community setting, the researcher asked about other settings where Leanne had worked as a therapist. It was then possible to explore with Leanne whether there were
99 similarities of differences in her experiences in the different settings and what might or might not account for these experiences.
The combination of constant comparison and theoretical sampling also ensures that the researcher remains alert to the possibility that some participant characteristics are aligned with multiple theoretical considerations; a newly qualified therapist working in a rural community setting might provide insights into factors including level of experience, isolation and practice setting. Accounts from Pauline and Bella, from two different therapy backgrounds, illustrate this effectively. While Bella had been qualified for slightly longer and Pauline was a novice practitioner, both practised in community settings, one in an inner city and the other in a rural context. However, both offered similar insights arising from the isolation of lone working; risk, visibility, access to role models and so on.
Through the three aspects of theoretical sampling described, the researcher can work efficiently to saturate the emerging concepts from which a theoretical perspective is ultimately constructed. Early tentative concepts serve to steer the researcher to sample and to guide the interviews to move towards the saturation of tentative selective concepts or the elimination of those which are not central to the resolution of the participants’ main concern. Crucially, it is what is discovered in the data that guides this and not the researcher’s interests or assumptions.
Having engaged in both constant comparison and theoretical sampling it is now much easier to grasp what Glaser is so passionate about in his ongoing defence and promotion of the methods. It is now possible to recognise how the method was instrumental in developing a theoretical perspective about supervision for AHPs and how the influence of theoretical sampling moved the interviews beyond a rich description of AHP supervision practices. With the benefit of the experience of doing grounded theory research, it is now apparent that had all nineteen interviews been collected before analysis was commenced it would have remained difficult to move beyond this rich description to the more conceptual exploration of supervision experiences presented in this thesis, not least because the interviewer may well have stuck more rigidly to and
100 consistently applied the same topic guide. Combining constant comparison and theoretical sampling, as experienced by this researcher, produced an evolution of interviewing which supported progress in the research from descriptive to conceptual while moving back and forward between the transcripts in the analysis.
Recognising the highly descriptive nature of some earlier accounts, (see excerpt 3 from illustrative memos in Appendix I) the researcher was prompted to ask later participants for concrete examples and illustrations from supervision to support the elaboration of tentative concepts; Can you recall what you took to your last supervision session?; Why did you take it?; What happened?; And then?; And now?; And in the future? This need for a grounded theory researcher to remain flexible as she generates data from interviews is recognised by Birks and Mills (2015) who refer to the potential evolution of any interview topic guide as a study progresses; citing as an example Fletcher and Sarkar (2012).
Benefitting from the increased theoretical awareness which develops through constant comparison of incidents in the data, the researcher becomes sensitised to look for instances which elaborate emergent concepts and in turn this builds the researcher’s confidence in subsequent claims of saturation of central concepts; something that will be illustrated further in section 4.6.3 below.