Visualizing Domain-specific and Generic Critical Online Reasoning Related Structures of Online Texts: A Hybrid Approach

Besides “traditional” educational media, young professionals in higher education use the Internet to obtain information. To utilize their online research in professional contexts, they critically evaluate the information they access and its sources. One dimension of this evaluation is an assessment of the linguistic state of the online sources, either implicitly or explicitly. This computational educational linguistic study applies methods from computational linguistics to online sources visited by young professionals from three fields (law students, teacher trainees, and medicine student) and develops partly novel visualizations that allow to quickly discover similarities as well as differences between multi-heterogeneous Internet sources, that is, sources that exhibit various topics, genres, and textual structure, among others. The visualizations also allow a comparison of search behaviour between different professional fields. In this way, we found that (1) genre classification has a significant impact on reliability scores, (2) young professionals’ search approaches vary by their professional field, and, (3) the best predictor of reliability is indeed the linguistic profile of an online source.