Lexical Networks Constructed to Correspond to Students' Short Written Responses: A Quantum Semantic Approach

A simple method to construct lexical networks (lexicons) of how students use scientific terms in written texts is introduced. The method is based on a recently introduced quantum semantics generalization of a word-pair co-occurrence. Quantum semantics allows entangled co-occurrence, thus allowing to model the effect of subjective bias on weighting the importance of word co-occurrence. Using such a generalized word-pair co-occurrence counting, we construct students' lexicons of scientific (life-science) terms they use in their written responses to questions concerning food chains in life-science contexts. The method allows us to construct ensembles of lexicons that probabilistically simulate the variability of individual lexicons. The re-analyses of the written reports show that while sets of top-ranking terms contain nearly the same terms irrespective of details of the method used to count co-occurrences, the relative rankings of some key-terms may be different in quantum semantic analysis.