The coherent versus fragmented knowledge hypotheses for the structure of matter: an investigation with a robust statistical methodology

Abstract

Research in cognitive psychology and science education has acknowledged two antagonistic theoretical perspectives concerning students' knowledge. One perspective considers students' knowledge as coherent or theory-like, explaining the learning process in terms of framework theories and stable mental models, whereas the other considers it fragmented and based on the existence of knowledge elements that can be activated, interconnected and organized accordingly. Although both perspectives continue to demonstrate empirical support, a third point of view focuses on the role of the particular topic and suggests that the stability of the in situ formed conceptions varies under different circumstances being also affected by the researchers' interpretation of data. Since methodological issues are by far fundamental and crucial for the above theoretical premises, in the present research an attempt is made to test the above hypotheses by applying a robust statistical tool, the Latent Class Analysis (LCA). Data were taken from students' responses (N = 329, aged 14–15) using an instrument designed to access students' understanding of the structure of matter and its changes of states that had previously been used to identify students' progression through a series of mental models reported in the literature. However, the results did not support the hypothesis that participants would hold these as coherent mental models. The present research adds to the ongoing dialogue on such a crucial issue for its pedagogical implications by contributing to the methodological pathways for an advancement of science education research through theory building.