ConceptGrid provides a template-style approach to check natural language responses by students using a model-tracing style intelligent tutoring system. The tutor-author creates, using a web-based authoring system, a latticestyle structure that contains the set of required concepts that need to be in a student response. The author can also create just-in-time feedback based on the concepts present or absent in the student's response. ConceptGrid is integrated within the xPST authoring tool and was tested in two experiments, both of which show the efficacy of the technique to check student answers. The first study tested the tutor's effectiveness overall in the domain of statistics. The second study investigated ConceptGrid's use by non-programmers and non-cognitive scientists. ConceptGrid extends existing capabilities for authoring of intelligent tutors by using this template-based approach for checking sentence-length natural language input.
Available at: http://works.bepress.com/stephen_b_gilbert/42/
This is a manuscript of an article published as Blessing, Stephen B., Shrenik Devasani, Stephen B. Gilbert, and Jivko Sinapov. "Using ConceptGrid as an easy authoring technique to check natural language responses." International Journal of Learning Technology 10, no. 1 (2015): 50-70. DOI:10.1504/IJLT.2015.069449. Posted with permission.