Content Analysis for Recognizing Threads in Conceptual Change Research

Since its origins - the publication of the milestone paper of Posner, Strike, Hewson and Gertzog in 1982 - conceptual change has been one of the main research issues in science education. The literature on conceptual change is wide and some researchers have already tried to identify the main threads in this research strand. In some review papers the main articles have been classified according to different criteria, like “fragmentation and coherence”. Our work stems from these overview papers and aims at contributing to the literature review on conceptual change research. In particular, we will present and discuss a new method for identifying research threads or currents of thought. The analysis of the collection uses both a quantitative linguistic software and an algorithm inspired by the classical information retrieval procedure LSI (Latent Semantic Indexing). The selection of the terms to build the reference dictionary needed for the algorithm is semi-automatic. In the paper, we discuss the extent to which such software and algorithm are useful for (i) aggregating and classifying similar papers in terms of topics; and (ii) checking the coherence between the results of the automatic classification procedure and the effective differences among the papers, in order to visualize more deeply the trends in conceptual change research.