Data-Driven Learning in Enhancing Learners' Language Idiomaticity

As data-driven learning has been advancing across new frontiers in recent years, there is still a paucity of studies on data-driven vocabulary learning model that brings about effectiveness in teaching and learning practices. Idiomaticity, which serves as an important indicator for language proficiency, needs abundant contextualized language input for the acquisition of target words. In this regard, the present study explores whether computer-assisted language learning is effective in vocabulary acquisition, and in the differentiation of synonymous words on the part of learners. Pre-/posttests and questionnaires were administered among an experimental group (N=26) and a control group (N=26). Results showed that the experimental group, who was instructed under the data-driven learning model, got a higher mean score than the control group, who received traditional dictionary-consulting instruction. The former also finished the posttest within a much shorter period of time. A significant relationship was found between the pretest scores and posttest scores among the experimental group whereas the scores in the control group did not reach statistical significance. Therefore, there was a significant improvement in learners' performance on collocation production under the data-driven learning model, whilst dictionaries did not prove to have such a contributing effect. This study provides some suggestions for how to enhance learners' idiomaticity by improving collocation performance under the data-driven model.