Background and aim Higher education reform in South Africa gave rise to new legislative and policy developments which pose interesting challenges to educators. One is to achieve generic outcomes for all programs, irrespective of fields of knowledge. These outcomes include the development of critical, problem-solving and communication skills, responsible citizenship and progressive abilities to think epistemologically, as defined by level descriptors. One aim of these outcomes is to ultimately develop a metaconceptual awareness1 in learners. We found that existing conceptual frameworks of students differ considerable, which affects the achievement of these desired outcomes. In this paper we report on an empirical approach to identify the conceptual frameworks of students and evaluate if it relates to prior knowledge and motivation. Experimental designStudents from the Faculties of Natural Science, Health Sciences and Engineering (N=441) who enrolled in the prescribed module in Philosophy of Science, participated in the study. They were requested at the start of the module to respond voluntary to a questionnaire consisting of: 1. A question which ask students to make a clear sketch of their image of the universe2 2. A self-developed multiple choice test to determine prior knowledge, consisting of questions on general knowledge in the field of science, technology and its history, and on views of the universe2.3. The Motivated Strategies for Learning Questionnaire (MSLQ)3 to measure motivational attitudes.The sketches on the universe were interpreted and ordered into a taxonomy, based on standard rules for classification. Research data were analyzed by appropriate statistical methods. Results and discussion 1. The sketches of the students could unequivocally be classified into five basic clusters or taxa: Primeval, Classic, Contemporary, Phenomenological and Improvised. Each of these taxa refers to a particular kind of mental representation which differs with respect to the underlying conceptual framework. The taxonomy could be developed further into two higher level categories, leading to more than 40 taxa. A two-way frequency table of the five basic taxa and students grouped into their various faculties or sub-faculties, indicated a statistically significant relationship (p<0.01). An effect size of 0,406 furthermore indicates a potential of it being important in practice4. These observations, together with the conceptual content of the representations, show that the taxonomic dendrogram provides an instrument which might assist an empirical approach to distinguish between conceptual frameworks of higher education students.2. Both the questionnaire on general knowledge and on views of the universe proved to be usable, based on reliability and item analysis (Cronbach's alpha coefficients of 0.616 and 0.507 respectively). A two-way frequency table of the basic taxa and high, medium or low performance in the questions on general knowledge as well as in questions on the universe, show statistically significant relationships (Chi-square p<0.01) of these kinds of prior knowledge with representations of the universe. 3. An ANOVA indicated that students who's representations fall into the contemporary taxon showed a statistically significant stronger motivation in the MSLQ evaluation than students with representations in the primeval (p<0.05) and classical (p<0.02) taxa.Preliminary results show that this empirical approach has potential to evaluate conceptual change in students.References1 Vosniadou, S., Ioannides, C., Dimitrakopoulou, A., & Papademetriou, E. (2001). Designing Learning Environments to Promote Conceptual Change in Science. Learning and Instruction, 11, 381-419.2 Lemmer, M.M., Lemmer, T.N. & Smit, J.J.A. (2003). South African students' views of the universe. International Journal of Science Education, 25, 563-582.3 Pintrich, P. R., and De Groot, E. (1990). Motivational and Self-Regulated Learning Components Of Classroom Academic Performance. Journal of Educational Psychology, 82, 33-40.4 Ellis, S.M. and Steyn H.S. (2003), Practical significance (effect size) versus or in combination with statistical significance (p-values), Management Dynamics, 12, 51-53.