開發學術英文學習網站及寫作輔助工具並調查其效能

Collocation knowledge is essential for English academic writing. However, many L2 academic writers still have difficulties in collocations. Some researchers proposed that the development of collocation tools for academic purposes (EAP) writers might be a good solution. In this paper, we introduced a newly developed corpus-based academic collocation tool- ACE (Academic Collocation Engine). Supported by AntCorGen, an efficient corpus crawling tool created by Laurence Anthony, we collected about 10,000 articles from the PLOS One journal database and compiled a multi-disciplinary 500-million-word academic corpus (ACE corpus). Our new academic collocation tool allows learners to search not only bi-grams but also 3 to 5 grams. Learners can further combine the search with parts-of-speech of the collocates such as adj., v., n. adv, to narrow down the search. ACE shows the patterns that meet the queries and list the results in the order of frequency. To further validate if the new tool can assist EFL learners in finding appropriate academic collocations, we compare this tool with one well-known, sate-of-the art corpus search tool COCA (Corpus of Contemporary American English). Thirty-five students participated in the study and they were randomly assigned into two group. One group used the COCA academic corpus, and the other used the ACE tool. They were asked to answer 25 gap-filling collocation questions. The results indicated that the ACE group has higher raw scores but there was no statistically significant different between the posttest scores of the two groups. The two tool succeeded in helping to find more academic collocates. According to the user survey, the results showed that students found that the interface of ACE is much more friendly. The preliminary results indicate that this new tool is user-friendly and shows promises in helping EFL learners find suitable academic collocation.