home - login - register

PERC 2011 Abstract Detail Page

Previous Page  |  New Search  |  Browse All

Abstract Title: Finding Meaningful Search Features for Automated Analysis of Students’ Short Responses to Conceptual Questions
Abstract: The Pathway Active Learning Environment synthetic tutoring system is capable of collecting large numbers of students' short responses to open-ended questions.  The analysis of these responses may provide insight into the utility of the system, as well as information about student understanding of physics.  The free-response nature of our data lends itself to qualitative analysis, however large data sets benefit from automated analysis.  Natural language processing and data mining approaches, such as clustering, have been of interest across a variety of fields for automating the analysis of qualitative data.  However, content-specific vocabulary, an abundance of search features, some of which are irrelevant, and inherent limitations on computers abilities to match meaning are challenges that must be overcome.  In this poster we discuss the extraction of features from human-grouped responses that may be meaningful and also useful for training automated systems for the analysis of future data.
Abstract Type: Contributed Poster

Author/Organizer Information

Primary Contact: Christopher M. Nakamura
Kansas State University Physics Department
116 Carwell Hall
Kansas State University
Manhattan, KS 66506
Phone: 785-532-7167
Fax: 785-532-6806
Co-Author(s)
and Co-Presenter(s)
Sytil K. Murphy
Kansas State University Physics Department

Dean A. Zollman
Kansas State University Physics Department

Michael Christel
Carnegie Mellon University Entertainment Technology Center

Scott M. Stevens
Carnegie Mellon University Entertainment Technology Center