• No results found

3.8 Data analysis

3.8.4 Validity, reliability and generalisability

As a qualitative researcher, I was concerned about using the terms ‘reliability’ and ‘validity’ in my study, which Wellington (2015, p.43) sees as ‘difficult to understand and define’. Guba and Lincoln (1994) argue that validity and reliability are issues that are a concern for positivist or quantitative research, rather than for qualitative research; the reason being that, in quantitative research, the methods used can be replicated to gain the same results, which is not possible in qualitative research. Kirk and Miller (1986, p.20) define ‘reliability’ as the degree to which research findings are ‘independent of accidental circumstances of the research’, and ‘validity’ as the degree to which the findings are ‘interpreted in a correct way.’ This suggests that reliability is about how a piece of research can be relied on for its accuracy or honesty and how the findings can be replicated in other circumstances. Lincoln and Guba (1985) suggest the use of qualitative research could mean greater dependability and credibility or trustworthiness, transferability, and conformability, rather than reliability and validity, which will make it possible for other researchers to replicate their work. Lincoln and Guba (1985) prefer the term ‘trustworthiness’ in qualitative research and stress the need for it to be credible. In addition, Creswell and Miller (2000) argue that qualitative researchers need to show how valid and credible their data are by using choosing different methods to validate their studies and through their chosen paradigm. Triangulation of evidence from different sources has been recommended by Hughes (2001) as one of the means by which researchers can add credibility to their study. I used four key steps to improve the credibility and dependability of the study. First, I critically reviewed existing literature on the postcolonial lens and early language and Ghana government policy documents, to ensure that my research questions and objectives relate to my main research themes (changes in language and literacy policies, impact of policy changes on teachers’ practices within lower grades), as shown in the literature review. Second, I provided justification for my choice of an interpretivist paradigm and the use of in-depth interviews for data collection, in line with a life history study. The data used in this study are the true voices of my participants, which were elicited from them during the interview sessions. As Hughes (2001) suggests, participants’ voices constitute valuable knowledge. As a researcher with an interpretivist view, I was open to the different interpretations that my

92 participants attached to the understanding of language and literacy policy and the effects of the use of policy in their classroom experiences. Indeed, Creswell and Miller (2000) argue that researchers with constructivist or interpretivist views believe people have pluralist or different interpretations or views on social reality. Thus, the understanding and interpretation to a particular issue may differ from participant to participant and these need to be presented as such. Using Wellington’s (2015) views on how to make one’s interview credible, I validated the transcription of participants’ interviews by calling and talking to them on the phone to seek further clarification on things that were not very clear or audible on the voice recorder. I also met participants during my follow-up visit to Ghana, where they had the opportunity to check the typed scripts to confirm whether or not I had accurately captured what they had said during the interviews.

For the expert interviews, the communication and validation of the transcribed scripts was through emails. This was quite easy and fast because the experts had the opportunity to correct the script and send it back to me. By asking all my participants to validate their responses, I ensured I had represented participants’ views accurately, as suggested by Creswell and Miller, who argue that validity in qualitative research depends on ‘how accurately the accounts represent participants’ realities of the social phenomena and is credible to them’ (ibid 2000, p.125). Goodson and Sikes (2001, p.36) further suggest that researchers can ‘ask informants what they think about any analysis or written accounts’ and also consider what action to take if some informants disagree with the information attributed to them. This will help give credibility to participant voice, because Cohen et al. (2011, p.179) are of the opinion that ‘if a piece of research is invalid then it is worthless’. Third, my interviews with experts provided another source of evidence for triangulation. The fact that the data from the experts interviewed were in line with the data provided by participants in the study and sometimes came up with different perspectives on some of the issues raised, validates the findings (Creswell and Miller, 2000).

It is appropriate at this point to give thought to the issue of generalisation. Larsson (2009) suggests generalisation in qualitative research is a topical issue because of the different approaches taken and interpretations on the matter. In qualitative research, generalisations ‘are explicit and constitute the explanation/generalization schema that is the basis for scientific reasoning’ (Payne and Williams 2005, p.295). This reveals the difficulty in how qualitative results, which are not based on numbers and scientific testing, can be generalised. The aim of qualitative research is generally not to provide scientific reasoning, which can be generalised,

93 but to provide a deeper understanding of phenomenon or participants’ experiences, as this study has done. As a result, the term ‘transferability’ used by Lincoln and Guba (1985) comes to mind. In addition, Wellington (2015, p.43) argues that qualitative researchers can consider the ‘extent to which a piece of research can be copied or replicated in order to give the same results in a different context with different researchers’. I believe the adoption of life history and the use of in-depth interviews can be used to study other aspects of teachers’ lives in Ghana. This is because any ‘similarity’ between my research context and ‘other research contexts’ can be considered a form of generalisation (Larsson, 2009, p.32). Although the experiences of individual teachers may be difficult to generalise, as these are personal to their individual situations, there are other areas of the study which are transferable to other similar studies on teachers.

3.9 Researchers’ professional life and positionality

The research design and the choice of life history was influenced by my own life history as a professional teacher. As a teacher researching the lives of other teachers, I believe I can understand the life stories of my participants better by reflecting on my own professional life first. This is in line with Berger (2015, p.220), who describes reflexivity as ‘self- appraisal’ or ‘self-evaluation of the researcher’s positionality’ and considering how these may affect the results of the study. I started my teaching career as a religious education teacher in higher education. After teaching for a number of years, I decided to do a postgraduate diploma in education and was appointed to head a large school in Accra. This was the beginning of my interest in early childhood and lower grades. With time, I became more involved in children’s language and literacy learning, where I began to realise the difficulties some teachers faced with regard to the policy in lower grades.

The choice of life history fits into my cultural background and personal interest. I love storytelling and I remember how I used to sit and listen to my grandmothers’ stories whenever we visited the village on school holidays. By combining my professional job, interest in language and literacy development and my Ghanaian cultural values, I saw the opportunity of using life history as a method of researching teachers’ lives and children’s language and literacy development. I came into this study fully aware that the values I hold, as a Ghanaian and a teacher, would influence each stage of the research process. As May (2001) argues, researchers add their personal value to their research from the beginning to the end by the design of the project, the way data is collected and interpreted and how the results are used.