• No results found

ITERATIVE-SIMULTANEOUS-PROCESS

3.6.2 Data Verification

3.6.2.1 Validity

The term validity indicated whether the research items measured a description of what it is supposed to measure (Bell, 2011). The researcher attempted to measure some content, constructs, predictions, etc about the effects of orphanhood on the academic performance of double orphans in CHHs. In this study, the OVC semi-structured questionnaire, for instance, measured the effects of orphanhood on learning and cognition of OVC in CHHs. Whilst the quality of a study is usually measured by its validity and reliability; some qualitative researchers assume that other criteria could be more appropriate. Four alternative constructs to assess qualitative enquiries, namely credibility, transferability, dependability and conformability (Vos et al, 2011) were proposed to check on validity of this study.

In this study,‟ credibility‟ was taken to be the most important of the four because it is linked to internal validity. For the purposes of credibility, this enquiry clearly stipulated the parameters of the study throughout the research process covering issues of triangulation and prolonged engagement in the field. In addition, Campbell and Fiske in Cohen et al (2011) contend that it is through triangulation that concurrent validity in qualitative research is demonstrated.

117

On the issue of „transferability‟ or generalisability the researcher was aware that qualitative research findings may be problematic in as far as other settings, populations, and treatment arrangements (external validity) is concerned. To counter such drawbacks, the theoretical and geographical boundaries mentioned in chapter one clearly stated how the theoretical framework and the issue of triangulation on data-collection greatly strengthened the vitality of the study.

Validating the content of the interview schedules was necessary to check whether each question was meant to answer the stipulated research questions. Gay (2010) asserts that content validity is usually determined by expert judgment. Therefore, the research supervisor, Dr. MC. Maphalala at UNISA and other specialists in Psychology of Education assisted in validating the instruments. Cronbasch in Nachmias and Nachmias (2007) assert that construct validity involves relating the instrument to an overall theoretical framework in order to determine whether the instrument is tied to concepts and theoretical assumptions within the study. The main construct in the study was orphanhood. Construct validity was also ensured by asking simpler questions at the beginning and moving to more difficult ones towards the end. The language used in the interviews, however, remained simple for more accessibility for Zimbabwean second language speakers. In fact the participants were allowed to respond in the language each preferred to use, be it English or ChiShona. Transcriptions were then made by the researcher herself as someone who is competent and fluent in both languages.

Being present during the exercise was quite beneficial as it was possible to respond to any queries the respondents presented. To some extent, the interview schedule ensured proper construct validity by explaining in simple terms like what orphanhood, learning and cognition meant. From the results, predictions were made possible on how orphanhood status affected learning and thinking processes. Any replication of this study may enhance future generalisations which Gay (2010) call predictive validity.

Validity for the qualitative aspects of the enquiry was also enhanced by adopting a social phenomenological research design which led to the establishment of some truth since details came from the concerned OVC experiences. The audio-recordings and the transcriptions also enhanced the truthfulness of the enquiry. Triangulation on instrumentation also assisted in

118

ensuring that the resultant findings were the truth about the OVC learning in CHH of Chipinge district. Data sources were also varied (see data matrix- table 3), in order to enhance validity of the study.

3.6.2.2 Reliability

Reliability in research refers to the degree to which procedures give results of the same kind under constant conditions on all occasions (Bell, 2011). In concurrence, Nachmias and Nachmias (2007) contend that reliability is an indication of the extent to which a measure contains variable errors, i.e. errors that differ from observation to observation during the measuring stance.

Similar to reliability is Lincoln and Guba‟s (1999 in de Vos, 2011) concept of dependability. Replication of the study may not be problematic as such, but what might differ are the manners in which participants may present their lived experiences since anyone‟s social world is always being constructed. The study can also be assessed on the basis of its conformability or objectivity (de Vos et al).

Being more qualitative than quantitative, the enquiry did not require statistical reliability computations, though the aspect of using reliable instruments remained a pre-requisite. The test– retest reliability procedures for data collection instruments was carried out at pilot level where pilot participants assisted in identifying sections of the instruments that sounded vague or ambiguous. Corrections were made and instruments were retested until they became more user- friendly. A few printings that showed threats to reliability were corrected for readability purposes as well as clarity. The idea of working with a slightly larger number of participants (Denscombe, 2007) and sorting several experts‟ scrutiny on the open-ended questionnaire, observation guide, etc (Gay, 2010), was one other strategy planned to ensure reliability of instruments before and after the pilot testing.

As noted by many qualitative researchers, it can be difficult to accomplish reliability in naturalistic events (de Vos et al, 2011) such as the CHH situation involving different household

119

sizes and many other variables affecting OVC. Most of these aspects are not static. The qualitative process can also be subjective for no investigator studies a phenomenon exactly like the other. The idea of pilot testing the instruments with participants of an almost similar background in Chimanimani district was a noble idea in order to ensure use of more accurate instruments. In addition, instrument triangulation as well as allowing more time in field work helped to enhance validity and reliability of instruments (Creswell, 2009; Strong and Hensley, 2009; Morgan, 2010).

A visiting schedule was prepared for each school and CHHs so that respondents were not rushed into the procedures taken for data collection. Afternoons were noted to be more ideal times for working with the OVC at each school just to make sure that respondents were exposed to similar situations and that the participants‟ responses to the individual and focus group interviews were not directly influenced by too many external variables.