Corpus analysis of problem-based learning transcripts: an exploratory study

Problem-based learning (PBL) is one of the most researched areas of medical education; numerous studies and meta-analyses have examined its overall effectiveness compared with that of traditional methods. Fewer studies have looked at the processes occurring within the PBL group, considering that a key difference between PBL and traditional teaching is the amount of talking in which students engage: a typical PBL week might involve 3-5 hours of talk. Little research has attempted to capture the vocabulary used or to analyse the cognitive processes that occur during PBL discussions and that might contribute to the advantages of this method of learning.

We professionally transcribed a recording of a facilitator and seven students engaging in a PBL case consisting of three sessions. The text was analysed by using Wmatrix2 (UCREL, Lancaster University, UK), a powerful linguistic program used for corpus analysis. This allowed the inter-sessional comparative quantification of the use of technical vocabulary, as well as of reasoning, questioning and explaining episodes.

We have compared the frequency of technical words and of reasoning, explaining and questioning episodes between sessions and related these differences to the focus of each session. We have been able to build concordances of important words and phrases within sentences which can be used to identify and monitor examples of questioning, explaining and reasoning.

In this exploratory study we have shown that corpus analysis can be a useful tool with which to analyse PBL transcriptions. This technique can be used to monitor the development of a technical vocabulary and skills in scientific and clinical reasoning as students progress through a PBL curriculum. We propose this methodology will become a powerful tool to help explore the much wider cognitive and linguistic developments of students and facilitators as they engage in PBL discourse.