TY - ABST

T1 - Investigating scientific literacy documents with linguistic network analysis

AU - Bruun, Jesper

AU - Evans, Robert Harry

AU - Dolin, Jens

PY - 2009

Y1 - 2009

N2 - International discussions of scientific literacy (SL) are extensive and numerous sizeable documents on SL exist.Thus, comparing different conceptions of SL is methodologically challenging. We developed an analytical toolwhich couples the theory of complex networks with text analysis in order to obtain clear visual images of whatis meant by SL expressed in written text. The raw text was first parsed into one statement sentences. Then, alinguistic type network was created with nodes being the words used in SL texts, and a link between two wordsestablished if they were adjacent to each other in the one statement sentences. Using the program Pajek, wedrew a map of the text showing the number of times a concept appeared in the one statement sentences, andthe strength of links between words. Different SL texts was analysed in this way. The network description allowedfor different calculations on the network data. For example a minimal description length approach partitionedthe network in to groups of words, which was then seen to represent different visions appearing in thediscussion of SL. In short, the networks allow for quantitative analyses as well as a quick visual overview of SLdocuments.

AB - International discussions of scientific literacy (SL) are extensive and numerous sizeable documents on SL exist.Thus, comparing different conceptions of SL is methodologically challenging. We developed an analytical toolwhich couples the theory of complex networks with text analysis in order to obtain clear visual images of whatis meant by SL expressed in written text. The raw text was first parsed into one statement sentences. Then, alinguistic type network was created with nodes being the words used in SL texts, and a link between two wordsestablished if they were adjacent to each other in the one statement sentences. Using the program Pajek, wedrew a map of the text showing the number of times a concept appeared in the one statement sentences, andthe strength of links between words. Different SL texts was analysed in this way. The network description allowedfor different calculations on the network data. For example a minimal description length approach partitionedthe network in to groups of words, which was then seen to represent different visions appearing in thediscussion of SL. In short, the networks allow for quantitative analyses as well as a quick visual overview of SLdocuments.

M3 - Conference abstract in proceedings

SP - 239

BT - ESERA; European Science Education Research Association

PB - European Science Education Research Association (ESERA)

Y2 - 31 August 2009 through 4 September 2009

ER -