Middle-School Science Students’ Scientific Modelling Performances Across Content Areas and Within a Learning Progression

Abstract

This paper focuses on students' ability to transfer modelling performances across content areas, taking into consideration their improvement of content knowledge as a result of a model-based instruction. Sixty-five sixth grade students of one science teacher in an urban public school in the Midwestern USA engaged in scientific modelling practices that were incorporated into a curriculum focused on the nature of matter. Concept-process models were embedded in the curriculum, as well as emphasis on meta-modelling knowledge and modelling practices. Pre–post test items that required drawing scientific models of smell, evaporation, and friction were analysed. The level of content understanding was coded and scored, as were the following elements of modelling performance: explanation, comparativeness, abstraction, and labelling. Paired t-tests were conducted to analyse differences in students' pre–post tests scores on content knowledge and on each element of the modelling performances. These are described in terms of the amount of transfer. Students significantly improved in their content knowledge for the smell and the evaporation models, but not for the friction model, which was expected as that topic was not taught during the instruction. However, students significantly improved in some of their modelling performances for all the three models. This improvement serves as evidence that the model-based instruction can help students acquire modelling practices that they can apply in a new content area.

Keywords:

• Transfer
• Model-based learning
• Science education

Acknowledgements

This research was funded by the National Science Foundation under Grant ESI-0628199 to the MoDeLS project at Northwestern University. The opinions expressed herein are those of the authors and not necessarily those of the NSF. We appreciate the support of the entire MoDeLS research group, across multiple institutions. We particularly appreciate the input of Brian Reiser, the project PI, and of James Hagerty.