A variety of non-probability sampling methods were used during recruitment including convenience, purposive and snowball sampling (Rubin and Babbie, 2009). Purposive sampling was employed throughout as this is recognised
as being appropriate where a limited number of knowledgeable ‘experts’ in a field exist (Tongco, 2007). It has been used with success in similar areas in the past as a means of seeking out “groups, settings and individuals
where…the processes being studied are most likely to occur” (Denzin and Lincoln, 2011). Academic staff and Teacher Practitioners (TPs) and pre- registration employer/trainer participants were invited to participate based on their roles and area/s of practice, while recent graduates were invited based on their gender, geographical location and area of current practice. This allowed the “gaps” in the data to be filled. Key stakeholders were invited first, and only after analysis of these interviews were the next interviewees invited to take part. A variety of perspectives were sought in order to provide a “more stable view of reality based on a wide spectrum of observations from a wide base of points in time-space” (Dervin, 1983). See individual empirical chapters for detail about the sample.
As a means of triangulating the findings a range of participants from different stakeholder groups were interviewed. Individual viewpoints were verified against others (Shenton, 2004) and provided the opportunity to “check out bits of information across informants” (Van Maanen, 1983).
Traditional quantitative social science research may be viewed as a fairly linear process, beginning with a review of the literature, sample selection, data collection then analysis. However, qualitative research is more of a cyclical, iterative process. This was applied during this programme of research. The researcher continually reflected on the research as a whole asking “how far do the methods… and theories that are used do justice to the subject and the data” (Flick, 2009). The linear process of collecting data then analysing at a later date was replaced with the notion of an “interwoven procedure” (Flick, 2009) such that new participants were constantly invited throughout the study as opposed to “all in one go” at the start. Use of a cyclical process was also beneficial in particular for a novice researcher whose skills were developed early in the process.
Convenience sampling was also used to enable the researcher to access interviewees with limited disruption. Those in a convenient geographical area (easily reached by public transport from Cardiff) were approached and
invited to participate. Where face-to-face interviews were not possible then telephone interviews were conducted, as described later in the chapter.
Another mechanism of recruitment employed was snowball sampling. Interviewees were asked at the end of the interview whether they could identify anyone else who may wish to participate in the study thus affording the technique the alternative name “respondent-driven sampling” (Salganik and Heckathorn, 2004). Interviewees were asked either to provide the
researcher with the individual’s contact e-mail address (with permission) or to ask the individual themselves to get in touch with the researcher if they were interested in participating. The technique has often been employed in
situations where it was desirable to access “hard to reach” populations and individuals for example in studies with unemployed men (Atkinson and Flint, 2001) and individuals with HIV virus (Sifaneck and Neaigus, 2001). Its use in this study allowed participants to utilise their social networks to suggest potential interviewees.
Glaser and Strauss (1967) state that interviews should continue until
saturation is achieved, meaning that no new codes or themes are generated during analysis. However, it is possible to argue that saturation may never fully be achieved in a study, especially in a study such as this which includes a self-selection sample. Indeed, O’Reilly and Parker (2012) have an
opposing view and challenge the acceptance that ‘adopting saturation as a generic quality marker is inappropriate’, instead focusing on issues of transparency and epistemology (O’Reilly and Parker, 2012).
In this study interviews were conducted until the rate of new code and theme generation significantly slowed (i.e. near saturation), all invited participants
who had agreed to be involved had been interviewed and all reasonable efforts to recruit new participants had been made (within the practical, time- restricted constraints of a programme of doctoral research). However, it is not possible to say that saturation was achieved.
2.6.1 Informed consent
Informed consent is grounded in the principles of “individual autonomy and secondarily that of beneficience” (Marzano, 2012) thus respecting people’s choice to participate and avoiding potential harm to participants. Gaining informed consent is vitally important in research involving human participants (Mason, 2002). A number of codes of ethics were consulted when
considering this and in drawing up participant information sheets and consent forms (The British Sociological Association, 2002; The British Psychological Society, 2010). In line with the recommendations made by such societies procedures were put in place to ensure consent obtained was informed.
Participants were informed about the overall purpose of the research, the design and any possible risks and benefits from participation (Brinkmann and Kvale, 2015). They were also informed about confidentiality, who would have access to the data and how information would be disseminated (Brinkmann and Kvale, 2015).
Participants were e-mailed the information sheet and consent form beforehand (at least two weeks in advance of the interview). Before conducting the interview the researcher checked that they had the
opportunity to read the information sheet and that it had been understood. Participants were given the opportunity to ask questions before completing a hard copy of the consent form. For telephone interviews the same process was undertaken via e-mail.
Consent was obtained in a written format through participant consent forms as this is the recognised preferred mechanism (Brinkmann and Kvale, 2015).
Where interviews were conducted by telephone signed copies of the consent form were scanned and e-mailed between participant and researcher. As there were no child participants or vulnerable adults it was deemed that all participants were capable of signing to acknowledge informed consent had been achieved.
Interviewees were given the opportunity to refuse to participate in the study and were encouraged to be frank and open in their responses (Shenton, 2004) by stating:
“There are no right or wrong answers, it is just your opinion that is
important”
at the outset of each interview. They were also given the opportunity to withdraw at any time, without needing to give an explanation.
2.6.2 Confidentiality
Confidentiality in research refers to an agreement with the participant/s about what is to be done with the data after it has been generated, for example who it may be shared with (Kaiser, 2012).Participants in this study were reassured that data would be kept confidential within the research team (supervisors Professor Dai John, Dr Louise Hughes and Dr Sion Coulman) at all times. Transcripts were anonymised such that no personal identifiers were left in any quotes used in dissemination of the study findings. Audio and transcript data were stored on the researcher’s password protected laptop and the recording device (Philips Voice Tracer DVT5500) was locked in a cabinet within CSPPS. Only the researcher and supervisory team could access this.