Using lexical analysis software to understand student knowledge transfer between chemistry and biology

TitleUsing lexical analysis software to understand student knowledge transfer between chemistry and biology
Publication TypeConference Proceedings
Year of Conference2009
AuthorsHaudek, KC, Moscarella RA, Urban-Lurain M, Merrill J, Sweeder R, Richmond G
Conference NameNational Association of Research in Science Teaching Annual Conference
Date PublishedApril 17-21
PublisherNational Assocation for Research in Science Teaching
Conference LocationGarden Grove, CA
KeywordsAACR, AACR-pub, Biology, chemistry, Lexical analysis, Transfer
AbstractTime and resources often prevent using constructed response assessments in large undergraduate science. We investigate the utility of using lexical analysis software to categorize student responses and uncover undergraduate student misconceptions in chemistry and biology. Students were randomly assigned a question set consisting of two questions relating to a single topic. Student responses were analyzed using SPSS Text Analysis for Surveys, using a custom library of science-related terms. The resulting analyses of student responses suggest potential barriers and connections between students understanding of these topics. Only 38 out of 160 students linked reaction spontaneity with thermodynamics. Student responses for one question set were rated using a scoring rubric by two independent scorers. Analysis of an acid/base question set showed student deficiencies in predicting pH behavior of functional groups in biology. After this scoring was complete, discriminant analysis was used to create classification functions that could predict human expert scores with 65.4% and 82.4% accuracy (p < .000). This study suggests that computerized lexical analysis may be useful for automatically categorizing large numbers of student open-ended responses.
URLNARST presentation: http://aacr.crcstl.msu.edu/system/files/NARST_talk09.pdf
0
Your rating: None