Using corpora to aid in learning collocations

The potential of corpora, language databases comprised of authentic language materials from a variety of sources, has gradually trickled down to ESL and EFL classrooms (McCarthy, O'Keeffe, & Walsh, 2010) and has been associated with data-driven learning (DDL) where learners observe language patterns and improve language accuracy. This study examined whether the familiarity with corpora would improve learners' preference to use corpora and the ability to solve V-N collocation problems, learners' knowledge of V-N collocations and learners' perception of the usefulness of corpus tools for solving collocation problems. Three groups of participants were recruited for three treatments: a one-time workshop combined with long-term in-class practices (TG), a one-time workshop only (CG), and no treatments (BG), and a pre- and post-test comparison was made to measure collocation knowledge gain of the three groups. Besides the post-test, the post-test with reference tools were given immediately after the post-test where the use of reference tools was allowed. The reference tools used on the post-test with reference tools and the performance of those who used corpora and those who did not was analyzed between the post- and post-test with reference tools. The results showed that those receiving treatments (TG and CG) that also chose corpora as reference tools were more successful than those who did not. Also, although collocation knowledge gain was not reflected in the score improvement from the pre- and post-test in all three groups, a correlation between the frequency of out-of-class use of corpora and pre- and post-test score improvement was found in TG, but not in CG nor BG. More positive perception of corpora as reference tools was found in TG than CG, but both TG and CG rated favorably for corpora's utility to solve lexical collocation problems. This study contributes to the current literature by applying DDL to lexical collocations, specifically V-N collocations and by examining the factors influencing learners' preference of reference tools, the ability to solve V-N collocation problems, potential knowledge gain of collocations and the perception of learners.

Name thesis:degree_name

M.A.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Teaching of English Sec Lang

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2016

Author dc:creator

• Wen, Fei Chen

Contributors dc:contributor

• Ionin, Tania
• Sadler, Randall

• Collocations
• Corpora
• Vocabulary
• Data-driven learning
• Inductive learning

Statement dc:rights

• Copyright 2016 Fei Chen Wen

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/90952

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/90952