Profiles of inconsistent knowledge in children’s pathways of conceptual change.

Abstract

Conceptual change requires learners to restructure parts of their conceptual knowledge base. Prior research has identified the fragmentation and the integration of knowledge as 2 important component processes of knowledge restructuring but remains unclear as to their relative importance and the time of their occurrence during development. Previous studies mostly were based on the categorization of answers in interview studies and led to mixed empirical results, suggesting that methodological improvements might be helpful. We assessed 161 third-graders' knowledge about floating and sinking of objects in liquids at 3 measurement points by means of multiple-choice tests. The tests assessed how strongly the children agreed with commonly found but mutually incompatible statements about floating and sinking. A latent profile transition analysis of the test scores revealed 5 profiles, some of which indicated the coexistence of inconsistent pieces of knowledge in learners. The majority of students (63%) were on 1 of 7 developmental pathways between these profiles. Thus, a child's knowledge profile at a point in time can be used to predict further development. The degree of knowledge integration decreased on some individual developmental paths, increased on others, and remained stable on still others. The study demonstrates the usefulness of explicit quantitative models of conceptual change. The results support a constructivist perspective on conceptual development, in which developmental changes of a learner's knowledge base result from idiosyncratic, yet systematic knowledge-construction processes.

Exploring the Shift in Children’s Incline Motion Predictions: Fragmentation and Integration of Knowledge as Possible Contributors

“…This would appear to fit in well with ideas concerning knowledge integration (Linn, 2006;Schneider, 2012;Schneider & Hardy, 2013;Schneider & Stern, 2009). Fragmented ideas are interlinked and coordinated within larger configurations of knowledge, and reinforcing such connections over time should result in more systematic forms of reasoning (Thaden-Koch, Dufresne, & Mestre, 2006;Wagner, 2010).…”

Informative tools for characterizing individual differences in learning: Latent class, latent profile, and latent transition analysis

“…The application of latent transition analyses is illustrated by a study on conceptual change in the domain of children's development of the understanding of floating and sinking of objects in water (Schneider & Hardy, 2013). Using three continuous indicator variables, each assessed at three measurement occasions, a latent profile transition analysis identified five different knowledge profiles (Figure 4a).…”

“…• these models contribute to our understanding of learning and individual differences Kleickmann, Hardy, Pollmeier, & Möller, 2011;Schneider & Hardy, 2013;Smith, Carey, & Wiser, 1985;Vosniadou & Brewer, 1992). To fully characterize learning processes research therefore needs to account for both quantitative and qualitative individual differences at a specific measurement point as well as in change over time.…”

“…Classes are typically labeled based on the patterns the class-conditional parameters convey, and typically tied to theory informing the research, as was done in the empirical examples from the previous section (Abenovali at el., 2017; Hickendorff et al, 2010;Schneider & Hardy, 2013). In latent profile analyses, testing which class means differ from each other or from the overall sample mean can aid the interpretation of the profiles.…”

“…If the class-conditional parameters are constrained to be equal over time, it is advisable to test the number of latent classes simultaneously for all measurement points (e.g., Schneider & Hardy, 2013), where it is possible that a class is 'empty' at a specific measurement point (for instance if learners develop new solution strategies between measurement points). If the class-specific parameters are unconstrained, the number of latent classes is usually determined separately for each measurement occasion.…”