Where Are the Readings Behind Your Concept Maps? Annotation-driven Concept Mapping

Concept maps are widely used in education and business. They are reckoned to stimulate the generation of ideas, facilitate requirement elicitation, and serve as the first step in ontology building, and as such, they are used by a wide variety of practitioners, including designers, engineers, instructors, and others, to organize and structure knowledge. However, the rationales that drive the creation of the map's concepts and relations are often implicit in the practitioners' heads. These rationales should be often sought in the reading material, provided by instructors or stakeholders, and 'processed' by the learners or designers to deliver the concept map. This poses a traceability issue where concepts can not always be traced back to the document excerpts that originated the concepts in the first place. However, the existence of such traces is envisioned to bring two major gains. Concept-to-excerpt traceability would facilitate third-party observers (e.g., stakeholders, instructors) checking out the source of possible misunderstandings in the concept map. Excerpt-to-concept traceability would enable document reading to be strategic, i.e., framed by the concept map. This work pursues the synergy between Concept Mapping and Strategic Reading by means of highlighted annotations. This results in annotation-driven concept mapping, a process where concept maps unfold in tandem with the annotations derived from the readings. Through the interplay of annotation and mapping, learners get a headstart on both activities. As mappers, learners no longer have to resort to their reminders but text annotations provide the raw material to build up the concept and relations. As readers, learners no longer wander around the readings but concepts and relations might serve as focus drivers. This work proves the feasibility of this vision by making a popular tool for concept mapping, CmapTools, annotation-driven. A focus group (n=5) is used to anticipate its utility. The results identified two moderating variables: the level of elaboration of the reading material and the degree of abstraction of the assignment topic.