QUALITY ASSURANCE SYSTEMS: HISTORICAL PERSPECTIVE OF SOUTH AFRICA AND INTERNATIONALLY
4.4 RESEARCH DESIGN
4.4.2 Sample size and distribution
The researcher purposefully selected three academic programmes and one support unit that were reviewed during 2008, 2009 (two programmes) and 2012 and which were still functioning. An important consideration was to ensure that the Heads of the respective departments/units which were assessed in the original review, were still currently holding these positions. The
University of Kwa-Zulu Natal (UKZN) has recently been structured along the following four College systems:
1) College of Humanities. 2) College of Health Sciences. 3) College of Law and Management.
4) College of Science, Agriculture and Engineering.
The following programmes were purposefully sampled to obtain their experiences of the review process:
1) College of Science, Agriculture and Engineering: School of Mathematics (Academic unit); 2) College of Law and Management: School of Accounting, Economics and Finance
(Academic unit);
3) College of Health Sciences: School of Pharmacy (Academic unit);
4) College of Law and Management: School of Management Studies Educational unit (Support unit).
4.4.3 Validity
Validity and trustworthiness are significant issues in effective research and essential requirements for both quantitative and qualitative research. In the quantitative phase of this research project the following positivist principles were observed (Cohen et al., 2011:180):
Table 4.1: Principles observed in research project
PRINCIPLE APPLICATION
Controllability In controllability variables are controlled that can threaten the internal validity. This research project utilised the natural setting of HE being the principal source of data which is valid and relevant.
Replicability Replication by analysing multiple cases confers confidence about the research because the resulting evidence is much stronger. Uniqueness of responses is thus validated in this research project
Predictability Prediction is to determine whether the primary objective of a particular research study could forecast some event in the future. The emergence of relevant criteria from this study could be validated in the future by their application.
Context- applicability
Context applicability validates that the setting, situation or environment is applicable to the research study. Consequently the HE context in this research project is validated in the quality assurance context.
Content validity This principle ensures that the research instrument represents the domain that it purports to cover. Consequently the data from questionnaires and data systems utilised in this study were validated in the construction of the instrument. Neutral
objectivity
Neutralism refers to an objective account of the situation uncontaminated by personal bias, political goals, and judgement. The unbiased value of responses and data from this study was validated by quantifying them.
Table 4.2. Action taken to control aspects of validity
PRINCIPLE ACTIONS TAKEN
Controllability Heads of programmes were requested to provide feedback via questionnaires. Both open-ended and closed questions were utilised in the questionnaire.
Replicability Same questionnaires were utilised amongst all the respondents. Similar data sets were extracted for all the programmes reviewed.
Context-applicability The context of discovery is part of the research process that asks questions and formulates hypotheses based on a particular context. Responses from Heads of Departments in their natural settings (context-driven) were obtained before and after review processes.
Content validity A mixed method design was followed. Numerical data obtained quantitatively and descriptive data obtained qualitatively were fragmentally collated in the discussion section of this research project.
Neutrality objectivity
I attempted to approach this research project neutrally. Care was taken to be unbiased in analysis of data.
The following validity tests were ensured in this research project (Cohen, Manion & Morrison, 2011:213):
Content validity was ensured by selecting a variety of different academic and support programmes that were representative of the different structures within the higher education institution. Three out of four colleges constituting UKZN were used as case study material.
Criterion-related validity was achieved by comparing responses to criteria from one programme to responses to criteria on other programmes.
Construct validity was confirmed by obtaining underlying responses by asking open- ended questions in the questionnaire.
The validation in mixed methods research utilised in this study is enshrined by Onwuegbuzie & Johnson (2006:57) as follows:
The sample was integrated by drawing from three diverse Colleges constituted by a small to a large number of enrolments - thus a combination of diverse sections.
The Heads of Units were regarded as the inside participant, while the researcher was regarded an outside participant. Views were balanced in the sense that I did not coerce the inside participants. Responses obtained were not biased to support any ideology. The potential weaknesses of a single approach, quantitatively by utilising HEMIS data,
were compensated for by the strength of the qualitative approach - by using questionnaires.
Sequential effects between quantitative and qualitative data collection and analysis were minimised as there was no linkage between these two processes. The independency of each source was maintained as the quantitative approach drew from the HEMIS system and the qualitative approach analysed responses from the questionnaires.
As the qualitative approach yielded ratings on a scale of one to five, conversion of qualifying numerical data or quantifying qualitative data was assisted in yielding inferences in a statistically neutral approach. Paradigmatic mixing and theoretical beliefs and practices yielded useful results as personal perspectives were obtained.