Advance in Detecting Key Concepts as an Expert Model: Using Student Mental Model Analyzer for Research and Teaching (SMART)

Improving summary writing through formative feedback in a technology‐enhanced learning environment

“…SMART analytics use an expert summary to automatically generate an expert concept map alongside key concepts and their propositional relations embedded in that expert map. In a previous study (M. Kim et al, 2019), the SMART algorithm was able to identify key ideas in the expert summary with 79% accuracy. The same algorithm is applied to elicit a student model from a student summary.…”

“…SMART elicits students' concept maps from their summaries as re‐represented their internal mental models. Thus, one can evaluate learner comprehension by examining multi‐dimensional aspects of concept maps (Clariana & Taricani, 2010; Gijbels, Dochy, den Bossche, & Segers, 2005; M. Kim et al, 2019; Zimmerman et al, 2018). A central assumption of the SMART design is that students' summary revisions facilitated by SMART feedback can indicate their evolving mental models towards the structure of the author's mind (Anzai & Yokoyama, 1984; Collins & Gentner, 1987; Johnson‐Laird, 2005; Pirnay‐Dummer & Ifenthaler, 2011; Seel, 2004; Smith et al, 1993).…”

“…Using these similarity measures (for more details, M. Kim et al, 2019), SMART provides formative feedback information. Students receive personalized feedback that encourages them to develop a more cohesive summary focused on the key concepts and relations in the text.…”

“…The distinction between memory for text and comprehension of the content is an important one in theories of discourse comprehension. These theories posit that deep comprehension emerges from the construction of a coherent and elaborated mental model, or mental representation, of the text (Graesser, Singer, & Trabasso, 1994; K. Kim & Clariana, 2017; M. Kim, 2015; M. Kim, Gaul, Kim, & Madathany, 2019; Kintsch, 1988). Mental models include information that is explicit in the text as well as necessary inferences that connect ideas from different parts of the text and connect information from the text to information in prior knowledge.…”

“…A model‐based approach uses a concept map—a network of interrelated concepts (Axelrod, 1976; Schvaneveldt & Cohen, 2010) elicited from a summary—to generate indices in multiple dimensions of knowledge structure (Ifenthaler, 2014; M. Kim, 2012; Pirnay‐Dummer & Ifenthaler, 2011; Spector & Koszalka, 2004). Model‐based ASEs such as AKOVIA (Ifenthaler, 2014), GISK (Kim, 2018), HIMATT (Pirnay‐Dummer & Ifenthaler, 2011) and SMART (M. Kim et al, 2019) evaluate the quality of students' knowledge representation and use concept maps for feedback to help them consider appropriate content and connect important ideas together. Scholars have shared that students made meaningful changes in their summaries through concept map feedback (Clariana, 2010; Ifenthaler, 2010; K. Kim, 2018; M. Kim, 2015; Pirnay‐Dummer & Ifenthaler, 2011).…”