Most materials developed for EFL learners use corpora derived from L1 texts to help identify what vocabulary and expressions learners need to know (Lee et al., 2019; Verspoor et al., 2012). However, this is problematic because the vast lexicon of the L1 speaker is nearly unattainable for most L2 learners (Hou et al., 2016; Sinclair, 2004). The presentation demonstrates how L2 corpora can be used to help bridge this gap and develop materials for L2 learners based on learners’ backgrounds and needs.

To do this, the presenters will describe the development and use of a large-scale longitudinal corpus consisting of texts collected from first-year Japanese university students (n = 500) at multiple points of time and across different modes of production: academic writing, academic presentations, and academic discussions. They will show how the data from this corpus was correlated to demographic and linguistic information gathered from each participant to help identify the lexical needs of specific groups and, how this information was used to develop materials for different proficiency levels and for different tasks. Finally, they will give examples of how these materials, and the process used to create them, could be adapted to other contexts.