• Tidak ada hasil yang ditemukan

3.10 VALIDATION PROCEDURES

3.10.2 Qualitative validation procedures

Creswell (2009: 190) states that validity has different connotations in qualitative studies compared to quantitative studies, and is not the same as reliability or generalisability.

While qualitative validity refers to the procedures employed by the researcher to check for the accuracy of findings, qualitative reliability is an indication of the consistency of the researcher’s approach when applied by other researchers across different projects. Gibbs (in Creswell, 2009: 190-191) suggests several reliability procedures that researchers can follow to ensure that their research is consistent and reliable. Some of these include:

• Checking transcripts for mistakes

• Ensuring that the coding of data is consistent

• Cross checking codes developed by different researchers in a team by comparing results

• Communicating with members of the research team through regular meetings, and sharing analysis

The first two procedures were followed, whilst the third and fourth were achieved through discussion with my supervisor.

152

Validity in qualitative research is based on determining whether the findings are accurate from the point of view of the researcher, participant, or the readers of an account.

According to Marshall and Rossman (1995: 143) all research must meet criteria against which the trustworthiness of the project can be evaluated. Krefting (1991: 217-221) suggests twenty-three strategies to establish the trustworthiness of qualitative research, and eleven different methods of determining the trustworthiness of qualitative findings are presented by Tashakkori and Teddlie (1998: 90-93), while Creswell (2009: 191) also recommends the use of multiple validity strategies to ensure the accuracy of findings.

Some of these strategies (a few of which were used in this study,) include:

• Triangulation of different sources of information and converging sources to develop a sound justification for themes

• Member checking to determine accuracy of qualitative findings through taking the final report, or themes revealed to participants, to ascertain whether they think that these are accurate

• Spending prolonged time in the field to develop in-depth understanding of the phenomenon being explored

• Presenting negative or discrepant information that goes against the flow of the themes since real life comprises different perspectives

• Peer debriefing, which involves interpretation by another person who reviews the account and asks questions about it

• External auditing by someone not familiar with the researcher to provide an objective assessment of the project

• The use of rich, thick descriptions of the setting or themes to convey findings so as to

“transport readers to the setting” and give them a sense of “shared experiences”

(Creswell, 2009: 192).

The validation procedures with regard to this study involved triangulating quantitative and qualitative datasets, presenting negative or discrepant information that went against the flow of the themes, peer debriefing, external auditing by someone not familiar with the

153

researcher, as well as the use of rich, thick descriptions of the setting or themes to convey findings.

While Cohen et al. (2000: 105) state that the terms validity and reliability apply to quantitative and qualitative research, Lincoln and Guba (in Krefting, 1991: 215) suggest that these terms are relevant for quantitative inquiry, but are inappropriate for qualitative research. Alternate terms such as accuracy of representation, credibility, and authority of the writer are suggested by Agar (in Krefting, 1991: 215).

The global qualitative concept of trustworthiness was introduced by Lincoln and Guba (in Tashakkori & Teddlie, 1998: 90; Marshall & Rossman, 1995: 143-145; Krefting, 1991:

214-216) as an alternative for many of the issues associated with quantitative design and measurement. Four general criteria are suggested, which, when combined, can determine the trustworthiness of an enquiry. These include:

• The credibility of findings of the study, and the criteria by which it can be judged

• The transferability and applicability of the findings in another context

• The consistency of the findings in the event that the study is replicated

• The neutrality of findings, i.e.: freedom from bias or prejudice.

Lincoln and Guba (in De Vos, 2005: 346) advocate the use of alternate terms credibility, transferability, dependability and conformability to evaluate qualitative research, rather than the terms internal validity, external validity, reliability and objectivity which are terms appropriate to describe quantitative inquiry.

Each of these alternative terms is described as follows:

1. Credibility is the alternative to internal validity, in which the goal is to show that the study was conducted in a manner that ensures that the participants were appropriately identified, and that the pattern of interactions, setting and research process were accurately described. “The strength of the qualitative study that aims to explore a

154

problem or describe a setting, a process, a social group or a pattern of interaction will be its validity” (De Vos, 2005: 346). Based on this statement, my research can be considered valid.

Linked to credibility is the establishing of truth value of the study. According to Lincoln and Guba (in Krefting, 1991: 215), truth value establishes whether the researcher has demonstrated confidence in the truth of the findings for the participants, and the context in which the research was undertaken. The truth value is usually obtained from the study of human experiences as they are lived and the meanings attached to these experiences. Sandelowski (in Krefting, 1991: 216) states that truth value is perhaps the most important criterion for evaluating qualitative research, and suggests that when the researcher presents accurate descriptions or interpretations of human experience, which can be immediately recognised by the people who share that experience, then the qualitative research is deemed credible.

In this study credibility was ensured by the use of the non-probability sampling strategy, involving the purposive sampling technique, to select the participants for the one-to-one interviews. The interviews were conducted in contexts where the participants were comfortable, that is, at the schools where their deaf children were attending, or where they themselves were employed. The participants were apprised of the purpose of the interviews and were asked for permission to audio-tape the conversations, after they were ensured of confidentiality. The transcripts are an accurate representation of the interviews. The researcher is confident that the descriptions and interpretations of the participants’ lived experiences have been accurately presented.

2. Transferability is the alternative term suggested by Lincoln and Guba (in De Vos, 2005: 346) for external validity or generalisability which entails “demonstrating the applicability of one set of findings to another context”, and this may be problematic in a qualitative study. Lincoln and Guba’s (in Krefting, 1991: 216) perspective on the applicability of findings is determined by the “degree of similarity or goodness of fit

155

between two contexts”. Sandelowski (in Krefting, 1991: 216) argues that generalisation is not relevant in qualitative research and it is an illusion since each research situation is unique in terms of the researcher, participants and interactions.

Tashakkori and Teddlie (1998: 65) also maintain that for most qualitative studies, generalisability or “transferability of results” to other individuals, situations and times is considered irrelevant.

Nonetheless, as long as the original investigator presents sufficient descriptive data, shows how the data collection and analysis has been guided by the original theoretical framework, and states the theoretical parameters of the research, the problem of applicability or transferability can be addressed (De Vos, 2005: 346; Lincoln & Guba, in Krefting, 1991: 216). Another way of enhancing the generalisability of a study is through the triangulation of multiple data sources, “to corroborate, elaborate or illuminate” the research project (De Vos, 2005: 346). Research design that involves multiple participants, using more than one data collection method, can strengthen the transferability of the findings of a study.

In this study, transferability was ensured through sufficient descriptive data. The parameters of the research problem, the setting, the population and pattern of interactions, as well as the theoretical framework have been clearly identified.

3. Dependability or consistency is the alternative to reliability. Consistency, according to Guba (in Krefting, 1991: 216), is defined in terms of dependability, and this criterion considers whether the findings would be consistent if the study were to be replicated with the same participants or in a similar setting. While quantitative inquiry is based on the assumption of a single unchanging reality, qualitative research on the other hand may be complicated by external and unexpected variables. The positivist concept of an unchanging social context is in direct contrast to the qualitative research, where the social context is constantly being constructed and transformed. Therefore, the concept of replication is itself problematic, and variability is expected in qualitative research (De Vos, 2005: 346).

156

Field and Morse (in Krefting, 1991: 216) assert that the emphasis in qualitative research is the uniqueness of human experience, so rather than evaluating for identical repetition, variation in human experience is sought. Duffy (in Krefting, 1991: 216) maintains that the main aim of qualitative research is to learn from the participants rather than to control them. Thus, in my mixed methods study the main purpose of conducting in-depth, one-to-one semi-structured interviews with hearing parents of deaf children was to investigate their lived experiences and to attach meaning to these experiences, so as to gain a deeper insight into the way they manage their parenting role, and to learn from their unique experiences.

4. Confirmability, suggested by Lincoln and Guba (in De Vos, 2005: 346) is the alternative to objectivity. Sandelowski (in Krefting, 1991: 216) proposed the term

“neutrality” which like conformability, refers to freedom from bias in the research procedure. This criterion is appropriate to evaluate the confirmability aspect of trustworthiness in qualitative research, since it seeks to assess whether the findings of the study could be confirmed by another, and whether the data help to confirm the general findings that culminate in the presentation of implications. In this study, the validity of the research project rests on the use of the mixed method research design, the triangulation of data sources, and the accurate descriptions and interpretations of participants’ experiences. The interview data was coded following the guidelines suggested by Tesch (in Creswell, 2003: 192). Furthermore, the interview data were submitted for scrutiny to an independent coder, who confirmed that there was saturation in the data after conducting an audit of the transcripts of the interviews. The researcher is confident that in the presentation of the findings, a true reflection of the phenomenon of hearing parents raising deaf children has been captured.

157