New clusters of universities have emerged in South Africa in the past decade, with fewer truly research-focused institutions following mergers, according to a study by the Centre for Higher Education Transformation, CHET. Three clusters revealed by an analysis of universities using nine input and output variables look very different to the formal categories used to differentiate the tertiary sector - with potentially major implications for policy.
"We need evidence-based rather than ideologically-driven policy-making," CHET director Professor Nico Cloete told the Stakeholder Summit on Higher Education held in Cape Town last month, sparking heated debate in the commission on institutional differentiation. An empiracle set of data had painted a picture of the higher education system as it actually was.
Historically there have been different forms of differentiation in South Africa. Apartheid differentiated institutions by race and there was a binary divide between universities and technikons (polytechnics).
Post-apartheid, differentiation was driven by the funding formula and then by restructuring and mergers from 2000 that slashed the number of institutions from 36 to 23 including 11 research universities, six universities of technology and six 'comprehensive' universities that combine formative and vocational higher education.
But a very different picture emerged from Cloete's 'empirical clustering' analysis of differentiation, which used information from government's Higher Education Management Information System, research publication data and university financial statements for 2008.
The CHET study combined six input and three output variables* and clustered institutions in relation to how they showed up against the variables. The distance University of South Africa was excluded because its huge student numbers and low success rates messed up the statistics, said Cloete.
The object of the study was to investigate universities by purpose, not to rank them. The three clusters of universities that emerged performed vastly different purposes in terms of student intake and success, staff qualifications, research outputs and income, among other things.
In the 'red' cluster were five research-intensive universities - Cape Town, Pretoria, Rhodes, Stellenbosch and the Witwatersrand. They produced the bulk of postgraduates and future academics and had high student success and graduation rates, high proportions of academic staff with PhDs, high research outputs, high income and low staff-student ratios.
In the 'green' cluster were nine universities - Free State, KwaZulu-Natal, North-West, Fort Hare, Limpopo, Western Cape, Johannesburg, Nelson Mandela Metropolitan and Zululand.
This group scored in the middle on the variables and included former research-intensive institutions whose performance in terms of research, success rates, postgraduates and staff qualifications declined following mergers with historically disadvantaged institutions, as well as a three formerly disadvantaged universities and three 'comprehensive' universities.
In the 'blue' cluster were eight institutions including two rural historically disadvantaged universities and six universities of technology. They had relatively lower postgraduate enrolments, success and graduation rates, qualified staff, research outputs and income but high enrolments in science, engineering and technology and high staff-student ratios.
The clusters fulfilled very different but equally important functions, Cloete stressed.
Interestingly, there was a sharper class than race attachment to the three constellations. For example, the five universities in the red cluster - all formerly 'white' - produced 45% of black African and mixed-race PhDs. "If we want to raise the number of black postgraduates and academics, we need to understand which part of the sector is producing them."
Surprising findings were a low negative correlation between staff-student ratios and student success, a low positive correlation between the percentage of academics with doctorates and student success - and a high correlation between lecturer qualifications and masters and doctoral enrolments, and research outputs.
A major implication, said Cloete, was that it was time to move away from a 'one policy fits all' approach to differentiated policy with differentiated funding.
"For example, there is an efficiency question for the blue group because their student success rates are low. The efficiency factor would be to improve student success. For universities with high success rates, the efficiency factor would be to produce more PhDs and more black graduates."
The green group was interesting because with certain policy interventions these universities could move into either the red or the blue group. "But it would be very difficult, unless there was collapse of the whole system, to move from the red to the blue group or visa versa."
South Africa, Cloete concluded, had a three-cluster system different to the current formal classifications and it was important to look at how this fitted into the state's development policy. "The problem is that there is no government policy on regional development."
A shortcoming of the data was the absence of direct quality and development indicators. The government did not have the capacity to collect the full range of information, said Cloete, and he called for a specialised agency to be created to do this work.
Reporting back on the commission Professor Derrick Swartz, Vice-chancellor of Nelson Mandela Metropolitan University, said policy ambiguity meant the higher education funding system only partially supported differentiation.
South Africa, he said, had 'strategic' differentiation in which institutions were evolving within and across formal categories in complex and dynamic ways - 'self-differentiation'. "What is missing is a coherent policy framework within which this evolution takes place."
It was important to make progress on differentiation in the interests of institutional self-confidence and stability of planning, and because scarce resources made it necessary to strategically concentrate investment to enhance optimal growth paths for all institutions.
Also, with a new Department of Higher Education and Training and with human resource development planning and reforms underway, there was an opportunity for universities to position themselves within a wider post-school system.
The commission, Swartz said, felt the CHET framework revealed an 'as is' picture but did not explain the legacy of inequality that created the differentiated outcomes. Any prospective differentiated 'model' should transcend rather than reinforce unacceptable institutional and social inequalities. "There was a very passionate argument over this," he said.
The commission felt a differentiation framework should provide some steering but also be sufficiently flexible to allow individual institutional trajectories to be negotiated in a realistic and responsive way, taking into account local and regional dynamics.
Differentiation should not be focused on rankings but on purposes, enabling the system to pursue multiple purposes more effectively and coherently.
Any model of differentiation should enable the sector to enhance student access and success, should be linked to the wider post-secondary system and to human resource development planning, and should not stunt research-intensive institutions while also strengthening under-developed institutions and supporting others to find optimal growth paths.
The commission suggested empirical clustering could offer a basis for comparing institutions within the system and enable deep questions to be asked about underlying correlations and intra-systemic shifts. It offered a heuristic tool to describe the system and to isolate the most fundamental axes of differentiation - but it must be further nuanced to take into account other variables and to identify and define key drives of transition within and across the clusters.
While the empirical clustering model should not be seen as a classification or a planning tool, with further development it could offer starting points to work towards an appropriate framework for higher education differentiation that allowed evolutionary transition across categories and should be linked to adequate funding - especially for under-capitalised institutions - as well to regional and local economic networks.
"Whatever we do, the system must allow portability of students, academics and knowledge across the sector. An academic must be able to have currency," said Swartz. It should also support key transformation goals such as equity of access and success.
"After a stormy start, we made more progress than in the last five years," said Theuns Eloff, Vice-chancellor of North-West University.
"We agreed on the need to fund differentiation. We agreed that differentiation needs to be developmental in character, that probably purpose is the best indicator, that we can't compromise efficiency and quality, and on self-differentiation within a funding framework."
Higher Education Minister Blade Nzimande concluded that differentiation should "be driven by a transformation imperative premised by leveling of the playing field" - institutionally by tackling the challenges of formerly black universities, and also by transforming the sociology of higher education spaces, which implied creating inclusive, caring universities.
* The input variables were: percentage of headcount enrolment in science, engineering and technology; masters and doctoral enrolments; student to staff ratios; permanent staff with doctoral degrees; private and government income; and student fee income. The output variables were student success rates, graduation rates and weighted research output units per permanent staff member.
There were many factors in favour of differentiation, he continued. Important for South Africa was that "high student participation and high differentiation go together, and that supports development," said Cloete. "In a high participation system, there should be a large number of different institutions catering for many kinds of students.
Receive UWN's free weekly e-newsletters