«LEARNING TO TEACH PHYSICAL EDUCATION IN PRIMARY SCHOOLS: THE INFLUENCE OF DISPOSITIONS AND EXTERNAL STRUCTURES ON PRACTICE by Ian Pickup, BA (QTS) A ...»
Incorporating biographical approaches As discussed in chapter 3, personal experiences in school PE and sport as a pupil are thought to be relevant frames of reference for trainee teachers who import life history data into their accounts of classroom events (Goodson, 1983). It was therefore important for an element of biographical methods to be included in the data collection and analysis process so that experiences preceding and out-with the immediate context of ITT could be considered as part of the process of becoming a primary school teacher of PE. Goodson and Sikes (2001)
There are likely to be many influences, experiences and relationships within any teacher’s life which have led to their developing a particular philosophy of education and taking on a specific professional identity which informs their work. Then there are the various contexts and conditions within which teachers have to work which have an effect upon what they do and how they do it (pp. 22Relevant approaches to biographical research include highly detailed studies of an individual life built up over a number of years (Plummer, 1983) or a combination of methods in which life stories provide a further perspective on a topic. Described broadly as biographical research, appropriate methods include biography, autobiography, story, discourse, narrative writing, personal history, oral history, case history, life history, personal experience and case study. Connelly and Clandinin (1999) list several other approaches to what they call narrative enquiry, including the use of annals and chronicles, photographs, memory boxes, interviews, journals, letters, conversations and documents. In light of this, it was deemed important for this research to present the possibility of discussing personal experience across the life cycle, and afford this relevance as an influencing factor in the development of knowledge and dispositions and hence on ensuing practice. Therefore, the design of interview schedules supported probing in this regard and provided the participants with the opportunity to reflect and comment upon experiences retrospectively. The intention was to present this as interpreted and edited life history data (Hitchcock & Hughes, 1989) rather than as a firstperson life history. Although this latter approach would undoubtedly provide a valuable and in-depth account of the processes at large, research design favoured the integration of life story approaches into a broader study. In designing interview schedules, a list of prompts was created to ensure that participants were steered towards biographical recollections and to stimulate probing regarding historical events (see example interview schedules, appendix 7).
This was particularly the case with respect to the origins of dispositions and knowledge of school PE curricula based on personal experiences. The interest in developing a level of biographical understanding through interviewing the respondents also had ethical implications (discussed below) and encouraged the adoption of techniques to foster a supportive and informal relationship with each informant.
Validity and Reliability Questions of validity and reliability have been central to research design, completed in full awareness of criticisms of qualitative research in this regard. Whilst concurring with Cohen, Manion and Morrison’s (2003) assertion that it is ‘unwise to think that threats to validity and reliability can ever be erased completely’ (p. 105), the aim has been to reduce such threats through careful design. In short, validity is explained in terms of what constitutes a credible claim to truth (Silverman, 2000) and as such is highly important; if this research was to be thought invalid then it would be worthless to the community of practice. In qualitative
research, validity is often couched in terms of:
honesty, depth, richness and scope of the data achieved, the participants approached, the extent of triangulation and the objectivity of the researcher (Cohen, Manion & Morrison, 2003, p.105).
In this regard, the intention was to produce an honest, deep and rich account of the phenomena at large. Denzin and Lincoln (1994) discuss these issues in terms of internal and external validity, reliability and objectivity. Internal validity reflects the degree to which findings correctly map the phenomena in question; external validity reflects the degree to which findings can be generalised to other similar settings; reliability reflects the extent to which findings can be replicated, or reproduced, by another enquirer; objectivity reflects the extent to which findings are free from bias. Validity in qualitative research is generally seen differently to validity in positivist research where goals may be controllability, predictability, the derivation of laws and universal rules. Some qualitative researchers have suggested that ‘understanding’, ‘fidelity’ and ‘authenticity’ are more appropriate terms to use in qualitative contexts (Guba & Lincoln 1989; Mishler, 1990; Maxwell, 1992; Blumenfeld-Jones, 1995).
Despite such claims, it remains that in-depth inquiry does not automatically secure a sufficient level of validity and reliability (Silverman, 2001; Hammersley, 1992). Qualitative research has been criticised for a tendency to ignore less clear or contradictory data. This problem referred to by some as ‘anecdotalism through limited quantification’ (Bryman and Bell, 2007, p. 638) stems from a tendency to use data in relation to conclusions or explanations and to provide selective evidence to support a particular contention.
Attempts to make qualitative research more valid typically include triangulation and / or respondent validation. The former represents the use of two or more methods of data collection to explain more fully the richness and complexity of human behaviour (Cohen, Manion & Morrison, 2003, p. 112) and stems from a view that reliance on one method may lead to a biased analysis and interpretation. In this research, where the emphasis was on individual experiences, triangulation was used to glean fresh insights into the phenomena under investigation, rather than to ‘settle’ validity questions. More pertinently, respondent validation was utilised by presenting emerging results to participants in order to refine findings (Reason & Rowan, 1981). Advocates of this approach suggest that this increases validity and objectivity through a process of checking and confirmation of findings. This does, however, assume that the participants themselves are expert commentators on their actions (Fielding & Fielding, 1986) and, although regular responses from participants and others were sought to help confirm findings, such feedback could not necessarily be construed as direct validation of findings. Instead, in order to improve the validity of the research, Silverman’s (2000, p. 177-178) ‘five interrelated ways of thinking critically about qualitative data analysis’ were adopted as a broad framework for research design and analysis. This stems from a standpoint of attempting to refute initial assumptions about the data in order to achieve objectivity (what Silverman, p. 178, terms the ‘refutability principle’). By taking such an approach, the aim was to secure both internal and external validity in this research through the adoption of rigorous and transparent approaches. This was achieved through the investigation of numerous cases over the timescale of the study in order to test out emerging findings and compare results. Whilst this approach did not look outside the specific institutional case until the final stage of the investigation, by commencing analysis of the data whilst new cases were being developed the results were subjected to a constant checking and comparison process.
This involved the on-going development and cross referencing of data to parts of the dataset generated through different methods. This is ‘comprehensive data treatment’ (Mehan, 1979;
ten Have, 1998) whereby anecdotalism is avoided by repeated inspection and analysis of data ‘until generalisation is able to apply to every single gobbet of relevant data you have collected’ (Silverman, 2000, p. 180). Allied to this, cases that did not conform to emerging findings were sought out and addressed. This deviant-case analysis ensured that the ‘provisional analytic scheme was constantly confronted by negative or discrepant cases’ (Mehan, 1979, p. 21) until all the data within the analysis framework had been satisfactorily captured. Silverman’s final recommendation for ensuring validity is to utilise appropriate numeric data which could ‘offer a means to survey the whole corpus of data lost in intensive, qualitative research’ (Silverman, 2000, p. 185). By using the PSPP and PIP in Stage 1 research, I was able to incorporate the use of descriptive statistics to generate a sense of the data as a whole and to use this as a checkpoint for cross referencing throughout subsequent stages.
Reliability is seen as a ‘synonym for consistency and replicability over time, over instruments and over groups of respondents’ and ‘is concerned with precision and accuracy’ (Cohen,
Manion & Morrison, 2003, p. 117). As Kirk and Miller (1986) suggest:
qualitative researchers can no longer beg the issue of reliability. While the forte of field research will always lie in its capability to sort out the validity of propositions, its results will (reasonably) go ignored minus attention to reliability.
For reliability to be calculated, it is incumbent on the scientific investigator to document his or her procedure (p. 72).
Rather than apply the ‘canons of reliability used in quantitative research, which may be unworkable in qualitative forms of inquiry’ (LeCompte & Preissle, 1993, p.332) reliability has been addressed in this study by aiming to produce comprehensive data which is an accurate representation of what actually occurs in the field (Bogdan & Biklen, 1992, p.48). In qualitative research, reliability generally refers to the degree of consistency with which instances are assigned to the same category by different observers or by the same observer on different occasions (Hammersley, 1992, p. 67). Other researchers were not enlisted to work in this investigation; instead, the full dataset was revisited on each occasion when a new case was introduced. The constant revision of the data and resultant updating of findings provided confidence that at the end of the research process, accuracy and reliability had been achieved.
Reliability is a problematic concept in naturalistic studies where the uniqueness and individuality of the context and the actors within the context are regarded as a key feature. It is also a particular challenge whilst using interviews as a research method; it has been suggested that there may be as many different interpretations of qualitative data produced through this method as there are researchers (Kvale, 1996, p. 181). A key part of analysing data derived in interviews is the categorisation and coding of statements made by participants and it is desirable for this process to be a reliable one; the results of the categorisation should not be
influenced by the individual carrying out the categorisation. As Kvale (1996) argues:
A strict requirement of intersubjective reliability in all forms of interview
analysis may, however, lead to a tyranny by the lowest possible denominator:
that an interpretation is only reliable when it can be followed by everyone, a criterion that could lead to a trivialisation of the interpretations…this may again involve a consensualist conception of truth: that an observation or an interpretation is only considered valid if it can be repeated by everyone, irrespective of the quality of the observation and the argumentation (p. 181).
There is also potential for disregarding the expertise and knowledge that the researcher may bring to the analysis process. This has led to a consideration of my own role in the research
process throughout the study. As Kvale (ibid., p. 181) goes on to say:
The question may involve an externalization of the interpretation of the meaning to fixed rules and criteria, rather than going beyond method and drawing upon the craftsmanship of the researcher, on his or her knowledge and interpretive skills.
Concurring with Smith (1988), this research design accepted that methods for data collection could not be viewed as interpretation-free. From the outset, the study design concurred with Ball’s view that ‘data are a product of the skills and imagination of the researcher and of the interface between the researcher and the researched’ (1990, p.169). The participants were viewed as partners in knowledge development and an ‘egalitarian’ researcher-researched relationship was fostered (Devis-Devis, 2006, p. 41). It was also acknowledged that my own life experiences would have a degree of influence on how data were interpreted and, although wishing to report credible findings, it was accepted that being entirely objective would not be possible. My own positioning as an active professional in PEITTT was acknowledged throughout the research process, with this in mind.
Field notes were used throughout each stage of the investigation (samples included in appendix 7, alongside annotated interview transcripts) in which reflections were noted regarding possible researcher influence on the interpretive process in keeping with a reflexive approach (Mays & Pope, 2000). Notes were checked against transcriptions of interviews, helping to interrogate any initial assumptions and informing the development of subsequent interview schedules. Throughout the study, field notes were used to support the confirmation or rejection of initial impressions and to support an objective stance. The wider potential for reciprocity between me and the researched which enables corroboration of data and provides participants with insights that could lead to the changing of their own practices and circumstances was also acknowledged. The participants themselves helped to frame questions and interpret data which may have had some impact on the practices of trainee teachers themselves. This is what Friere (1973) called ‘conscientization’, through which participants become energised and better placed to change practice.