Impact of Educational Escape Games on Students’ Semantic Lexicon in Environmental Science: A Word Association and AI-Based Analysis

Abstract

Gamification, including escape games, has proven to be an effective tool for changing environmental behaviour. However, little research has been done on the impact of this game type on students' conceptual networks, particularly in relation to environmental science and chemistry. Therefore, this study examines the semantic lexicon of primary and secondary school students using word association tests with selected call words from these subjects. Students were asked to provide answers immediately before the game and exactly one week later. The pre- and post-tests were analysed using the rather robust but complicated Garskof-Houston formula and the results were compared to the analysis of a smaller generative AI running on a laptop to find a more user-friendly way to evaluate this type of survey. The Garskof-Houston formula showed that the relations of the words "mixture", "precipitation" and "rain" changed the most, but the smaller, and therefore presumably weaker, LLM found the most changes for the relations of the words "compound", "precipitation" and "rolling". Since the meaning of a concept can be defined as a list of words with which it is associated, these results suggest that learners' concepts have changed, which is quite important from the perspective of constructivist learning theory. However, further research is needed on the use of LLMs for this type of evaluation.