REFERENTES CONCEPTUALES DE LA SECUENCIA DIDÁCTICA

The use of mixed methods gives the confidence to believe the results and findings. Keeves (1997) and Denzin & Lincoln (2008) note that methodological triangulation especially when used in mixed methods approach has proven to be a very effective strategy that improves research credibility and trustworthiness by overcoming researcher biasness, and providing rigour, breadth, complexity, richness and depth. In qualitative research, issues of credibility, dependability and conformability help to determine trustworthiness of research findings. This study attained credibility and trustworthiness in the many ways.

4.11.1. Credibility

Chilisa & Preece (2005), Korstjen & Moser (2018), equate credibility to internal validity in quantitative research saying research evidence is therefore credible if it represents as adequately as possible the multiple realities revealed by the participants. Chilisa & Preece (2005), Anney (2014) contend that credibility of research findings can be ascertained by how long the researcher takes in the field. Thus, prolonged stay and engagement with participants in the field enhances the credibility of a study. Anney (2014: 276) says, “…..prolonged engagement in the fieldwork helps the researcher to understand the core issues that might affect the quality of the data because it helps to develop trust with study participants.” Pitney, & Parker (2009) call this participant checks. Although some researchers have proposed a period of engagement with participants in field, this depends on the design adopted for a study. Ethnographic studies for instance can take as long as 6 months or more to understand the context being studied. This study used the convergent parallel mixed methods design, which in itself houses several advantages in ascertaining the credibility of the results.

Several other steps were taken to ensure credibility in this study. First, the researcher is a professional within the field of education. Using this advantage, before collecting data, the researcher created rapport with participants to introduce himself as a member of the teaching profession and assured them on the significance of the study. This was to develop trust with participants. Further, after observation of lessons and well in data analysis, the researcher sent text messages to observed teachers requesting them to provide any other terms that were difficult to sign and any other data that they could have forgotten telling the researcher. The

researcher received additional information from some participants. On 7th July 2017, all the transcribed transcripts of the interviews were emailed to respective research participants so that they could read through and state whether some information could be removed from the interview or some information could be added. Phone calls followed up to ask the participants to work on the sent emails. Below was the email:

Dear participant,

I appreciate your wonderful contributions to the study interview that you participated in early this year in February. As per requirement we would like to ensure that you as a participant is satisfied with the information you provided on curriculum implementation for learners with special education needs

Attached is a transcribed transcript of the conversation we had. Kindly go through and make subtractions or additions. Indicate which parts you feel may not be necessary or were wrongly said and you wish to say them rightly. This will help us present a true reflection of curriculum implementation for our learners. This does not mean however that we should change the picture as it truly was initially but to simply correct or adjust the presentation of facts.

Kindly note that this conversation is still bound by ethics and so is shared between you and I as a researcher and no one else. As promised earlier during the interview, this is an academic study with respect for ethics and so will not reflect identities in the final report,

I will appreciate if you have time to go through the document and make the necessary notes! You can use track changes or comments so that I am able to see areas which need to be attended to.

One of the participants among ESOs replied to the email saying everything was okay except spellings and grammar. Phone calls were made to participants to ask them to verify some information they provided. This was part of the member check strategies that was to allow participants go through their own work and make comments on the data they provided. It was also an act of assuring them the genuiness of the study. Chilisa & Preece (2005) say member check can be formal or informal. Pitney & Parker (2009), acknowledge that participant checks can be conducted in two ways; by transcript verification to see the accuracy on their interviews or by asking the participants to verify the findings i.e. (debriefing).

As part of peer debriefing, the preliminary study findings were presented on 7th _{July 2017 at} the Masters and Doctorates seminar week at UNISA and the researcher benefited from the comments that were given as part of debriefing. Further, as part of debriefing process, a paper titled “Special Education teachers involvement in curriculum development in Zambia”, an extract from the results of this thesis was presented at Southern African Society for Education (SASE) conference on 4th October 2017 in Botswana. Feedback and comments were given to the researcher. For instance, one comment that the researcher benefited from peer debriefing was;

“Your findings should not be negative things only, show the reader also the positives. What are the strengths and the weaknesses of the study?”

The comment made the researcher to get back to the scripts to re-analyse the data and identify the positive elements of the study, thus a positive check on researcher subjectivity. Negative case analysis is part of credibility. Verbatims with contrary views were noted and recorded in the verbatims. However, since they were fewer, they did not influence the final conclusions of the results and findings. As a researcher, it was only necessary to highlight views in a balanced manner.

The analysis of qualitative data by use NVIVO qualitative software played a significant role in ascertaining the credibility of this study (see process in 4.9.2). Journals were created as part of memos in NVIVO. Besides, every activity during data collection process was manually recorded in a diary on a daily basis. For instance, extra data of SETs observed, whether male or female, disabled or not were recorded in a diary. In NVIVO, reminders on key points related to data collected each day were noted down as key issues for analysis and reflection.

Further analysis by use of NVIVO qualitative analysis software enhanced the credibility of the findings for this study. For instance, a Pearson correlation coefficient of qualitative responses in the three provinces showed no differences among the provinces (see table 5.15). Coding density of inapproapriate and unclear responses was run in NVIVO and compared between provinces (see figure 5.3). When the qualitative analysed data was compared with the quantititave data, similarities were observed. These processes helped the researcher to be focused and ensured that the data collected was credible enough.

4.11.2. Dependability

Dependability (trustworthiness) Ary etal (2010), is equal to reliability in quantitative research refers to the consistency and replicability of findings of similar findings. According to Williamson, Radford & Bennetts (2003: 131), dependability requires that the argument is complete, allowing the reader (or reflective designer) to follow and understand it without unexplained leaps from argument to conclusion.” In this study, the interviews and observations were conducted in different provinces. The questionnaires had both qualitative

and quantitative questions which collected responses from different schools, districts and provinces. Triangulation was the strength that determined the dependability of qualitative data in this study. The different instruments used to collect data complemented each other. The responses collected provided a similar picture about curriculum implementation and its challenges. Ary et al (2010: 503) explain the replication logic in determining the dependability measure in qualitative research saying “………the more times a finding is found true with different sets of people or in different settings and time periods, the more confident the researcher can be in the conclusions.” SETs, ESOs and CSs from different settings and involved at different times but giving similar findings as demonstrated in chapter 5 shows how such findings are dependable.

One other strategy used in qualitative research to ensure dependability is documentation, Ary et al (2010) call it audit trail. In this study, data documentation was critical from the onset of data collection through analysis to reporting. Data was collected and stored according to the type collected. Quantitative data was on questionnaires and entered in SPSS. Qualitative data was coded in NVIVO in which memos or diaries were created for reference from time to time during analysis. The study collected detailed interview data which was transcribed into more than 243 pages. These were coded in NVIVO qualitative software providing a high coding density on certain themes that came out (see appendix ‘E’). This further satisfies the qualitative measure of dependability.

4.11.3. Confirmability

Confirmability is another measure for trustworthiness in qualitative research. It is a measure of biasness on the researcher. With this in mind, the researcher should be able to present findings that are not skewed to his or her interests. Ary et al (2010) equate confirmability to objectivity in quantitative research. Ary et al (2010: 504) say confirmability and objectivity “both deal with the idea of neutrality or the extent to which the research is free of bias in the procedures and the interpretation of results” Confirmability concerns the aspect of neutrality (Korstjen & Moser 2018). “You need to secure the inter-subjectivity of the data. The interpretation should not be based on your own particular preferences and viewpoints but needs to be grounded in the data” (Korstjen & Moser 2018: 122). Audit trail like in dependability is used as a strategy to ensure confirmability. Data analysis followed steps as

described in 4.9.1 and 4.9.2. Bias in mixed methods is overcome by triangulation. In this study, the qualitative findings from questionnaires confirm the results from interviews and observation and vice versa. Further, coding density in NVIVO qualitative analysis for instance cannot allow a researcher to skew the findings. Responses demonstrating understanding of the concept of curriculum adaptation were written by respondents through the open and close ended questionnaire. These were picked as they were and used as examples. These strategies ensured that the results of this study can be confirmed by other researchers.