• No results found

Traditionally, the benefit of a piece of research depends on its validity and reliability (Devellis, 2012), as these reflect the quality of the study. Validity ―refers to the

accuracy of a result. Does it capture the real state of affairs?‖ (Robson and McCartan,

2016: p.105). Validity also refers to an assessment of whether a piece of research and its design accurately depicts the theory or theme that it is intended to illustrate (Briggs and Coleman, 2007).

Reliability traditionally ―relates to the probability that repeating a research procedure

or method would produce identical or similar results‖ (Bush, 2007: p.92). Moreover, an

estimate of reliability is defined as the evaluation of the stability of measures managed at diverse times with respect to the same participants or implementing the same standard, or ―the equivalence of sets of items from the same test (internal consistency)‖ (Kimberlin and Winterstein, 2008: p.2277).

However, the processes of assuring validity and reliability can differ according to the research approach or how the data are collected. Writers such as Lincoln and Guba (1985), Sandberg (2005) and Atkinson and Hammersley (2007) argue that the traditional understandings of these concepts are not consistent with the ontological and epistemological stance of qualitative research and they must be reconceptualised or substituted with alternative criteria. In qualitative studies, therefore, reliability and validity are replaced with the ideas of trustworthiness, comprising credibility, dependability, transferability and confirmability (Lincoln and Guba, 1985), and relatability (Bassey, 2001).

Since this study was intended to yield qualitative data by involving people who could change their opinions or beliefs due to social or political influence, it was difficult to achieve a high level of reliability in the traditional sense. It is not feasible to accurately

159

duplicate an interpretative study, owing to the differences and individuality inherent in the research context (Cohen et al., 2011). Moreover, Briggs and Coleman (2007) claim that when using semi-structured interviews, it might be difficult to guarantee replicability, since participants are deliberately treated as potentially unique respondents. In research that recognises participants‘ subjectivity, and the impact of context and experience, consistency of outcomes over time is not expected. Therefore, instead of conventional reliability, writers propose the criteria of dependability, understood in terms of integrity on conducting and reporting the research, a reflexive stance, and openness about one‘s own values and positionality (Miles and Huberman, 1994). Dependability was enhanced by ensuring the willingness to participate of the participants and the schools, by means of initial telephone contact and careful explanation of the research issue and study aims, stating the need for willing participants and to be able to observe their classroom practices. Moreover, respondents‘ convenience was considered a high priority when scheduling interviews, in order to ensure that they would be comfortable and not feel rushed. As advised by Hair et al. (2003), I carefully read the transcripts of the interviews when inputting the data to verify that there were no logical flaws and that the answers provided by individual participants were not contradictory; also, I used the same strategy when conducting observations along with informal conversational interviews, when I chose only five schools for purposive sampling in order to obtain a great deal of information within the limited amount of time granted by the Saudi Embassy, and this led to no logical flaws.

Moreover, as advised by Miles and Huberman (1994), I have also retained all materials used in retrievals from (documents, transcripts and recordings). As for reflexivity, since researcher bias can be an issue, I have been open about my positionality (chapter one, section 1.8) and my approach to dealing with issues encountered in the research (section 3.11.3 in this chapter).

Another problematic notion in qualitative research is internal validity, traditionally understood in terms of how well the research outcomes match reality. The reason this is problematic is that research rooted in interpretivism rejects a realist ontology and accepts the possibility of multiple (and sometimes conflicting) ―realities‖. An alternative

160

is credibility, which concerns the presentation, not of a single absolute ―truth‖ but of a reflection and meaningful interpretation of participants‘ experiences and perceptions. Atkinson and Hammersley (2007) claim that qualitative research is capable of achieving credibility by using certain means, such as member checks and triangulation. In this study, I used appropriate means, consistent with the qualitative approach, to ensure that my research was credible. For example, I used three data collection methods – interviews, observations (and informal conversational interviews) and documents –, which contributed to the credibility of my study. All of these methods provided in-depth data that enriched my understanding of how inclusion works in Saudi Arabia and, in particular, the way it is subjectively perceived and interpreted by principals in schools. Since I also used observation (and informal conversational interviews) in this study, for example, I was able to compare the interviewees‘ responses with the results of the observations (and informal conversational interviews) in order to achieve credibility. Triangulation approaches in the social science sphere enable researchers to outline or describe in greater depth the complexity and richness of people‘s behaviours through examining a phenomenon from various perspectives (Cohen et al., 2011). Thus, it can be seen that combining methods enhanced the accuracy of the data obtained (Robson and McCartan, 2016).

In order to support the trustworthiness of the data gathered in this study, I also discussed the interview questions and observation strategy with my supervisor, other experts, such as the Director of the Centre of Special Educational Needs in Saudi Arabia, and school principals and professionals involved in the special educational needs field, to seek their advice in terms of improving the data collection. Furthermore, since the interview schedule was developed in English, an English-language specialist assessed the appropriateness of the wording of the questions and their face value. These questions were then translated into Arabic. The semantic correspondence of the Arabic and English versions was verified by experts and an independent translator, as I mentioned previous who holds a certificate in translation. Subsequently, I showed participants their interview script before using it (member checks).

161

There is a need to recognise that qualitative approaches have been questioned with regard to generalisability or external validity (Yin, 2009); it is commonly suggested that, regardless of the depth of understanding reached in a certain case, this does not mean generalisability as traditionally conceptualised can be achieved. The problems associated with generalising results have been touched upon with regard to utilising only a limited number of cases (Punch, 2005). However, generalisation is not the concern for my research; the emphasis is instead on understanding the cases under examination in their specific context (Punch, 2005).

In consideration of the above discussion, the present research does not intend to generalise the findings to other schools in other areas of Saudi Arabia; rather, the objective is to deliver insight into a very specific context. The example of the Eastern Province was not viewed as a representative sampling unit, owing to the fact that there are several large ARAMCO schools in the area, meaning this particular locality is unique. Accordingly, this choice was made bearing in mind the provision of a real-life context with the purpose of analysing and gaining insight into the inclusion of children with Down‘s syndrome in Saudi Arabia, facilitating the fulfilment of the research‘s aims and objectives.

This does not, however, preclude the possibility of wider application, subject to the principle of transferability, a judgement by the reader of the appropriateness of transfer, based on similarity of context. The researcher‘s role is to inform such judgement by providing detailed contextual information; this is done in chapter two, part two, and through the rich verbal and pictorial information in chapter 4, 5 and 6.

A similar concept to transferability is relatability, which was introduced by Bassey (2001) specifically in the education context (Blaikie, 2010). It is defined as the extent to which the research findings are sufficient and appropriate for others working in the same field (such as school principals) to relate their own experience and decision–making to the case(s) described (Bassey, 1984, 2001). In other words, it concerns whether the situations described are sufficiently in harmony with readers‘ experience for them to benefit from the findings (LeCompte and Preissle, 1993), or the extent to which they find relevance in the researcher‘s interpretations (Pound, 2000). Relatability combines

162

the notions of resonance (meaning recognisability) and transferability (Sanger, 1994), and also bears comparison with Stake‘s (1995) idea of ―naturalistic generalisations‘‘, which are not predictable regularities, but guiding inferences drawn by readers, depending on the fit between what Lincoln and Guba (1985) call the sending and relieving contexts.

Relatability is a useful concept in studies where it is not possible, because of the small number of cases and/or the particularly of the context, to claim generalisability. Indeed, in Bassey‘s (1984, 2001) view, it is more important than generalisability. It allows readers to evaluate the research findings, with full awareness of the methods and sample used (Bassey, 2001), and use them for their needs (Corbin and Strauss, 2008). Examples of relatability could include schools of the same type (e.g. primary, urban or rural, public or private), similar size and with similar ethnic, cultural and socio-economic characteristics (Bell, 2005).

There is no guarantee that the findings can be applied, but by providing relatable empathetic insights (Bassey, 2001), a study may stimulate worthwhile thinking (Bassey, 1999), allowing readers to adapt the research findings and implications to their own contexts (Davey, 2015), opening up possibilities for action (Andrew, 2003).

As with transferability, therefore, the first step in facilitating relatability is to provide a rich account of the study and its context (Blaikie, 2010), including the physical, organisation and social factors influencing policies and practices (Andrew, 2003). Such particularisation can provide readers with a vicarious experience that can contribute to the social construction of knowledge (Stake, 2000). In this thesis, for example, I provided a detailed account of the research background in chapter 1 and in chapter 2 , I described the Saudi context of the education of children with Down‘s syndrome and the move towards inclusion, including legislation, the structure of the education system, social and religious attitudes towards disability, the role played by ARAMCO, and the cultural phenomenon of ‗wasta‘. In the three findings chapters, I make extensive use of quotation to illustrate practices and conditions in the selected schools, and in chapter 6 in particular, I provide rich details on the physical environment.

163

Relatability is also promoted by the use of triangulation of methods and data sources as a heuristic tool (Denzin and Lincoln, 2005), providing different and complementary information (Gray and Malins, 2016), in order better to capture the richness and complexity of the research setting. This approach reflected in the ways in which issues of language, culture and personal experience influenced my data collection and interpretations (Blaikie, 2010). In all these ways, I have afforded opportunities for readers to evaluate the relatability of this study.

Having discussed trustworthiness and relatability, let me now conclude this chapter by discussing ethical considerations, specifically informed consent, anonymity and confidentiality and issues regarding my own position as a researcher.