Text Summaries or Concept Maps: Which Better Represent Reading Text Conceptualization?
Tristan E. Johnson, Pablo N. Pirnay-Dummer, Dirk Ifenthaler, Anne Mendenhall, Selçuk Karaman and Gershon Tenenbaum
Capturing students’ mental models has been proposed as a viable means to measure students’ understanding and conceptualization of given learning materials. Mental models are usually represented by either short text summary or a graphical map (i.e., concept map). This study aimed attesting which learner’s mental representation associates higher with three criteria: original text, expert concept map, and expert text summary. HIMATT, a mathematical framework proven to share a sound reliability of mental model in both semantic and graphical formats, was used to elicit the association between students’ mental models and the three criteria reference models following studying two book chapters. The findings indicate that across all association indices, students’ text summary elicitations were stronger than the students’ concept map elicitations over the three criterion-reference models. Moreover, stronger similarities emerged for the four structure indices (surface, graphic, structure, and gamma) than for the three semantic indices (concept, proposition, and balance) within the text summary and concept map formats. The results are attributed to students’ strong familiarity with written representation of the learning materials rather than creating concept maps. Furthermore, the results indicate that reading a text stored in long-term memory and retrieving and representing it into a concept map is harder than retrieving and representing it in a written text. Further research must clarify to what extent the practice of transforming written materials into graphical maps will improve the validity of using concept maps for mental conceptualization.
Keywords: mental models, elicitations, concept maps, text summaries, semantic measures, graphical measures, text conceptualization, reference models, HIMATT, convergent validity