Predicting institutional research performance from published indicators: A test of a classification of Australian university types

Abstract

In 1998, the Australian federal government published comprehensive performance indicator data on universities. This article makes use of several indicators to examine correlates of research productivity in Australian universities, and to test the explanatory power of Marginson's (1997) typology of universities. Research productivity is measured using the quantity known as the 'research quantum' (RQ), which is a performance-related component of operating grant funding. Results of the analysis reveal that a large proportion of the variability in research performance can be attributed to two latent constructs. One relates together favourable student-staff ratios, a high proportion of graduating students continuing into further study, and a high proportion of research qualified staff. The other connects graduates' experiences of low quality courses, good student progress rates and low attrition with high graduate employability. These factors also predict membership of the four Marginson types. The results confirm that Australian universities can be separated into four main reputational groups based chiefly on research performance. Some implications for the use of performance indicators in a more market-driven system are discussed.