Abstract

This paper describes a technique for locating indicators of success within the data collected from complex learning environments, proposing an application of e-research to access learner processes and measure and track group progress. The technique combines automated extraction of tense and modality via parts-of-speech tagging with a visualisation of the timing and speaker for each utterance developed to code and analyse learner discourse, exploiting the results of previous, non-automated analyses for validation. The work is developed using a dataset of interactions within a multi-user virtual environment and extended to a more complex dataset of synchronous chat texts during a collaborative design task. This methodology extends natural language processing into computer-based collaboration contexts, discovering the linguistic micro-events that construct the larger phases of successful design-based learning.

Citation

Thompson, K., Kennedy‐Clark, S., Wheeler, P. & Kelly, N. (2014). Discovering indicators of successful collaboration using tense: Automated extraction of patterns in discourse. British Journal of Educational Technology, 45(3), 461-470. Wiley. Retrieved August 10, 2024 from https://www.learntechlib.org/p/148351/.