Network Representations of Knowledge about Chemical Equilibrium: Variations with Achievement.

This study examined variation in the organization of domain-specific knowledge by 50 Year-12 chemistry students and 4 chemistry teachers. The study used nonmetric multidimensional scaling (MDS) and the Pathfinder network-generating algorithm to investigate individual and group differences in student concepts maps about chemical equilibrium. MDS was used to represent the individual maps in two-dimensional space, based on the presence or absence of paired propositional links. The resulting separation between maps reflected degree of hierarchical structure, but also reflected independent measures of student achievement. Pathfinder was then used to produce semantic networks from pooled data from high and low achievement groups using proximity matrices derived from the frequencies of paired concepts. The network constructed from maps of higher achievers (coherence measure = 0.18, linked pairs = 294, and number of subjects = 32) showed greater coherence, more concordance in specific paired links, more important specific conceptual relationships, and greater hierarchical organization than did the network constructed from maps of lower achievers (coherence measure = 0.12, linked pairs = 552, and number of subjects = 22). These differences are interpreted in terms of qualitative variation in knowledge organization by two groups of individuals with different levels of relative expertise (as reflected in achievement scores) concerning the topic of chemical equilibrium. The results suggest that the technique of transforming paired links in concept maps into proximity matrices for input to multivariate analyses provides a suitable methodology for comparing and documenting changes in the organization and structure of conceptual knowledge within and between individual students.