Agencia Gubernamental de Control
RESOLUCIÓN N.° 730/AGC/15
Thomas (2016) writes that ‘reliability and validity are not your principal concern when doing a case study’ (ibid: 63) arguing that since there is no expectation that the results will be repeatable, they are irrelevant. He adds that they are borrowed from other disciplines and methods but make no sense in case study research. However, this view is not supported by Yin (2003), who feels that researchers should conduct their research as if being closely watched by an auditor (though it is not clear who the auditor is). It seems to me that all research needs to be seen to be adopting appropriate measures to ensure the conclusion is robust, and, to this end, Yin (2003) develops the auditor idea in suggesting that procedures should enable any auditor to arrive at the same conclusions.
A case study that is positivist, and using quantitative data to draw conclusions, can adopt repeatable procedures – thus ensuring reliability, in Yin’s (2003) terms. But if one is taking an interpretative approach, using qualitative data, then Thomas’s (2016) argument is stronger.
Golafshani (2003) proposes that reliability could be redefined for qualitative researchers, so that it reflects how trustworthy the results are. Trustworthiness
makes sense in my case, and, although I accept Thomas’s (2016) view that the data from my study is unlikely to be repeatable in another setting, I do concur with Yin’s (2003) view that procedures need to be replicable, for the purposes of showing trustworthiness, rather than producing the same result. Yin (2003) and Cohen et al. (2011) identify replicable procedures as ‘construct validity’. For me, it means carrying out a case study that might be useful as an example, or illustration, of what can happen in a similar setting when there are similar dilemmas.
Validity, like reliability, is from the positivist paradigm (Golafshani, 2003) and needs to be adapted for interpretative researchers using qualitative data unlike Thomas (2016), where it is not a principal consideration. Indeed, Macklin and Whiteford (2012) argue that qualitative and interpretative research and researchers need not adhere to practices ‘underpinned by positivist reasoning processes but by practical rationality’ (ibid: 87). Macklin and Whiteford’s (2012) argument for practical rationality is one for doing what works to answer the question one wants to answer ethically. Thomas (2016) and Macklin and Whiteford (2012) argue that researchers do not need to consider processes such as validity from the positivist paradigm, which goes against the view of Yin (2003) who views validity as important in case study research. Gorard (2013) supports Yin (2003) where ‘The ‘validity’ of any findings refers to their real-life applicability and to their robustness when examined sceptically’ (ibid:159). For Gorard (2013), real-life applicability relies upon due care and attention from the researcher. The due care and attention of the researcher indicates a development of ‘tacit knowledge’ (Thomas, 2016). However, the
above discussion does not answer how my thesis addresses validity. Essentially, validity will be ensured through construct and internal validity (Cohen et al. 2011 and Yin, 2003), where data is analysed transparently. Reliability and validity are the basis for establishing generalisable research, but generalisability is more than just reliability and validity. One must also consider what, if anything, arising from this study of particular leaders dealing with particular dilemmas, that could usefully be extrapolated for use in other settings It is not necessarily the purpose of interpretative case study research to provide generalisations, but, identifying methods for studying schools, as well as gaining an understanding of how some leaders might deal with dilemmas, are both important. This case study can support other studies or provide an exception, through the creation of an illustrative example, that queries generalisation. Yin (2003) argues that researchers using case study should aim for ‘analytical generalisations’ and not concern themselves, as he sees it, in Level One Inference ‘statistical generalisability’ where inferences are made from a sample of the population. For, Yin (2003) case study is not a sampling unit and therefore cannot be related or extrapolated to a whole population. However, it does indicate the ‘case’ being studied and can be useful for ‘analytical generalisations’ where one or more cases support or refute a theory. This can be achieved through the generation of examples based upon small-scale studies of particular cases.
The generation of examples, for Flyvbjerg (2001 and 2006), is a good thing, as the examples can be used to illustrate and make clear points. Flyvbjerg (2001) sees the creation of examples as historical, because cases have
always been used to illustrate and make arguments clearer. It could also be said that case studies facilitate the creation of exemplars, as Flyvbjerg (2006) argues in ‘Five Misunderstandings about Case-Study Research’. It is, therefore, possible to use case studies to illustrate, support or falsify theories. And this, for Flyvbjerg (2006), is the strength of case study research. He argues that unique cases have moved forward our understanding of the world. He relates that even Galileo had observed that ‘Aristotle’s law of gravity was not based upon observations across a wide range’ (ibid: 74). Flyvbjerg (2006) also describes case study as a ‘black swan’, which acts to contradict a theory, and therefore, generalisation is not always necessary and can be overrated. The view of Macklin and Whiteford (2012) is similar to Flyvbjerg’s (2001) and Symmonds and Gorard (2008) regarding ‘doing what works’ within ethical parameters. Generalisation, for Flyvbjerg (2001) is one of many practical skills researchers need and can use. This fits with Symmonds and Gorard’s (2008) argument that all researchers need concern themselves with is whether they can answer the important educational questions.
The stance taken in this current study is that answering the research questions is central, and that, as an interpretative and qualitative study, it does not need to adhere to positivist thinking and processes. But, a consideration of how research might be made robust and applicable to other settings is worthwhile. As Thomas (2011) writes, conducting research should also develop my Technical skills (techne), theoretical understanding (episteme), my tacit knowledge (phronesis) and my actions (praxis) as a researcher. The identification of the paradigm and methodology in this thesis which has
developed my learning needs to be considered alongside a suitable research tool that collects good data.