Network analysis of misconceptions in FCI data

In this paper we discuss the misconceptions detectable in Force Concept Inventory responses by analysing student selection of distractor items. We discuss the appearance of distinct, coherent, clusters of distractor items which appear in this data. Certain distractor items are structurally central in these groupings and sub-groups within these clusters may be interlinked by “connector” items. Our analysis suggests that these clusters represents a coherent, robust, non-Newtonian worldview. We further suggest that this worldview could be destabilised by correcting a small number of conceptual mistakes.