Qualitative research is concerned with meaning in context and involves a subjective interpretation of the data, (Willig, 2001). The conundrum that attempts to undermine qualitative methods is that any sample size sufficiently large enough to be statistically representative cannot be analysed in depth, (Yardley, 2000). Equally, small sample sizes cannot be considered generalisable. However, findings from qualitative research must be considered genuine. Yardley (2000) offers various characteristics of good qualitative research:
- sensitivity to context meaning a grounding in the approach,
- commitment suggests prolonged engagement with the topic, development of competence and skill in the methods used, and immersion in the relevant data,
- rigour refers to the resulting completeness of the data collection to enable comprehensive and complete analysis and interpretation,
- transparency relates to clarity and sufficient detail to provide depth of appreciation and understanding of the process,
- coherence relates to clarity and cogency [simplicity, logic and reason].
Several authors provide guidelines for good practice in qualitative research to ensure rigour, (Elliott et al., 1999; Henwood and Pidgeon, 1992) but there appears to be no consensus considering the breadth of research from differing ontological and epistemological perspectives, (Madill et al., 2000). Instead, I relied upon Lincoln and Guba’s key text of 1985 that related to the trustworthiness of qualitative research and its findings. It posed four questions that relate to truth value, applicability, consistency and neutrality even though some argue that the quantitative terms of reliability and validity are just as applicable as they symbolise good rigorous research, (Robson 2002).
- Truth value: ‘how can one establish confidence in the ‘truth’ of the findings of a particular inquiry for the subjects (respondents) with which and the context in which the inquiry was carried out?’, (Lincoln and Guba, 1985, p290). Whilst an understandable question, ‘truth’ remains a nebulous concept as reality is ‘a multiple set of mental constructions’, (Lincoln
86
and Guba, 1985, p295). Instead I as the researcher must have shown that the constructions were sufficiently well represented and credible. This introduces the concept of ‘credibility’ which relates to internal validity in quantitative research, (Lincoln and Guba, 1985, p295). Within this research it was important that clinicians fulfilled the inclusion and exclusion criteria and were accessed from as wide a geographical range as practicable to demonstrate the purposive sampling. By doing so this increased the range of experiences of clinicians regarding graduates from differing universities, different courses and different cohorts. It was also important to accurately reflect these findings from subsequent data analysis within the presentation of this thesis e.g. with accurate reference system as an audit trail.
- Applicability: ‘How can one determine the extent to which the findings of a particular inquiry have applicability in other contexts or with other subjects (respondents)?’ (Lincoln and Guba, 1985, p290). This relates to the likelihood of possible transferability to similar circumstances or situations, noting that the similarity is the important issue. Whilst compared to the concept of external validity in quantitative research it is thought to differ from external validity with its statistical confidence levels because, qualitatively, contexts do differ, (Lincoln and Guba, 1985, p298). As will be discussed further on p95 this study utilised a purposive sample to explore physiotherapy clinicians’ views. Given the number of institutions that provide pre-registration physiotherapy training nationally (see Appendix 2) it can be argued that there is applicability beyond this purposive sample towards other members of the physiotherapy profession.and resonance with other healthcare professionals.
- Consistency: ‘How can one determine whether the findings of an inquiry would be repeated if the inquiry were replicated with the same (or similar) subjects (respondents) in the same (or similar) context?’, (Lincoln and Guba, 1985, p290). The ability for future replication of a study requires reliability, stability, consistency and predictability but assumes that the subject matter to be evaluated is also dependable, stable and replicable. This introduces the concept of ‘dependability’ that relates to reliability in quantitative research, (Lincoln and Guba, 1985, p299). As the emphasis of the data collection required perspectives from clinicians based upon their experiences the
87
outcomes could never be predicted nor standardised. However the process of data collection was regulated using an agreed format of semi-structured interview and questionnaire design. These were both verified and agreed by the National Research and Ethics Service (NRES), (see Appendix 5 and 6).
- Neutrality: ‘How can one establish the degree to which the findings of an inquiry are determined by the subjects (respondents) and conditions of the inquiry and not by the biases, motivations, interests, or perspectives of the inquirer?’ (Lincoln and Guba, 1985, p290). This asks for assurances that the study findings are independent and hence introduces the concept of ‘confirmability’ that relates to the concept of objectivity in quantitative research, (Lincoln and Guba, 1985). This will be discussed in the next section.
Reflexivity and bias
Reflexivity is the term given for a researchers’ conscious reflection of how the study’s design and execution affected the study itself, (Langdridge, 2007). This is important because Interpretive Phenomenological Analysis places an emphasis on making sense of participants’ experiences (Howitt and Cramer, 2008; Smith and Osborn, 2003). It accepts that:
such an exploration must necessarily implicate the researcher’s own view of the world as well as the nature of the interaction between researcher and participant, (Willig, 2001, p53).
Hence one must be sure that due consideration has been made to ensure that a reasonable interpretation has been made and presented in this thesis. Willig (2001) differentiated between personal reflexivity and epistemological reflexivity. The former involves reflecting upon ways in which ‘our own values, experiences, interests, beliefs, political commitments, wider aims in life and social identities have shaped the research’, (Willig, 2001, p10). In contrast epistemological reflexivity considers the study design and ‘encourages us to reflect upon the assumptions (about the world, about knowledge) that we have made in the course of the research, and it helps us to think about the implications of such assumptions for the research and its findings’, (Willig, 2001, p10).
88
Regardless of structure of the data collection, whether qualitative or quantitative, bias can exist. Regardless of sincere assurance of anonymity and confidentiality participants may wish to protect themselves and not speak openly or honestly. Alternatively, they could be swayed in their responses by the researcher’s personal characteristic or by the way the data collection is conducted.
Despite one’s natural reflexive nature bias is possible within any data collection. Hence, in this data analysis it was important to recognise the possibilities of the varied types of bias; gender bias (Fontana, 2003); halo effect (i.e. that which occurs when behaviour is evaluated well only because the participant exhibits likeable attributes, (Mateo and Kirchhoff, 1991); acquiescence responses (i.e. those who agree with statements regardless of content, (Polit and Hungler, 1999); or prestige bias (i.e. that which occurs when an opinion is held only because a prestigious person holds that same view, (de Vaus, 1996) ).
As a researcher, I was aware that some participants knew me having worked with, or close to, me professionally or knew me to hold an academic post in their professional field. It was possible that such biases could occur and this required careful listening and careful reiteration. At times it proved useful to ask the participant to explain something more fully that might have been thought obvious to us both ‘for the benefit of the tape’ [i.e. the readers of the thesis] as a vehicle of unpicking their perceptions. Alternatively contrasting points of view were offered to open discussion further. Hence, at times, and when thought prudent I openly played Devil’s advocate to develop conversations.
Langdridge (2007) offers a list of useful questions that the researcher should ask oneself to consider how thoughts and actions determine the outcome. Within the planning, data collection and subsequent analysis of my research for this thesis the decisions and decision- making process were recorded in a diary. The supervisory meetings that oversaw my doctoral studies proved invaluable.
89