The effects of transformative and non-transformative discourse on individual performance in collaborative-inquiry learning

The effectiveness of computer-supported collaborative inquiry learning in STEM education is well-documented in the literature. At the same time, research indicates that some students struggle to articulate relevant concepts, to make their reasoning explicit, and to regulate their learning-all of which are necessary for effective collaboration. In this study, 106 college students completed tasks related to Ohm's Law in a simulation-based, collaborative-inquiry learning environment. Using qualitative analysis, multilevel modelling, and data-mining techniques, we investigated the relationship between student engagement in transformative and non-transformative learning processes and learning outcomes. The results revealed that by using the appropriate feature engineering and algorithms, we could build accurate machine-learning models that could automatically identify transformative and non-transformative discussions on a large scale. Additional qualitative and quantitative analyses indicated that when groups engaged in additional interpretation and sustained mutual understanding, their members tended to have statistically better individual-learning outcomes. These analyses also indicated that when groups engaged in additional orientation and proposition generation, their learning outcomes were statistically lower. Approximately two-thirds of the students considered their group work helpful in completing inquiry tasks. Explanations of these results and research recommendations are provided.