Home > Research > Activities > Corpus Linguistics Conference 2021
View graph of relations

Corpus Linguistics Conference 2021

Activity: Talk or presentation typesOral presentation


Title: Evaluating the effectiveness of indirect data-driven learning in second language writing error correction: A pilot study report
Abstract: First mentioned by Johns (1990), data-driven learning (DDL), which uses corpus data as the learning material, becomes a new inductive and learner-oriented learning. Research on relevant topic reveals that DDL is effective on lexico-grammatical errors (e.g. Crosthwaite, 2017; O’Sullivan & Chambers, 2016) and the indirect DDL can lead to higher error correction rate (Yoon & Jo, 2014). However, in China, DDL is new to teachers and students due to lack of evidence to show its effectiveness. The current research aims to answer to what extent the indirect DDL can be effectively applied to correct the four most frequent lexico-grammatical errors in EFL writing (errors of articles, prepositions, verbs and word choices), and how the feedback, the concordances and other consulting resources (e.g. online dictionaries, textbooks) interact with students in error correction activities, in a Chinese university. This study adopts the mixed classroom intervention design to evaluate students’ outcome of error correction, under three conditions (without DDL, DDL only and the combination of common consulting resources and DDL). This study evaluates the actual process of error correction, where students are free to choose material for error correction, with error correction spreadsheet and stimulated recall. Lastly, the questionnaire and interview will be conducted for both students and their teacher to know about their perception and difficulties of DDL application in EFL writing class. The pilot study shows that students used DDL material as a tool for reactivating or supporting their prior knowledge to correct errors. Concordances were frequently considered as the most useful referencing material. It resulted in higher error type recognition rates and repair rates. Indirect feedback was important to help students decide which consultation resources should be used to recognise error types and correct errors. Both students and their teacher held the positive attitudes towards DDL-mediated error correction, although they still experienced some difficulties.

Electronic data

Event (Conference)

TitleCorpus Linguistics Conference 2021
Abbreviated titleCL2021
Degree of recognitionInternational event