The statistical assessment of learning outcome: Knowledge structure

Abstract

Knowledge structure represents understanding of the world; studying it constitutes grounds for understanding learning. Multidimensional scaling has been used in previous research studies on revealing the characteristics of knowledge structure. This study examined how multidimensional scaling and property vector fitting can be used to reveal longitudinal qualitative changes in knowledge structures. Learners' knowledge structures appear to converge to a common structure as their learning experience grows. This study provides methods for statistically describing qualitative information about the evolution of knowledge structures during learning - something that has been elusive in previous research.