Corpus-Based Data-Driven Learning in Reading English for Academic Purposes
[Thesis]
Rasikawati, Ira
Beers, Scott
Seattle Pacific University
2020
148
Ph.D.(Educ.)
Seattle Pacific University
2020
One of the second language (L2) instructional approaches perceived as promising is data-driven learning (DDL) - the use of computer-generated concordances for language teaching and learning. DDL provides students with the opportunity to analyze language forms across contexts found in the concordance output. Notwithstanding the growing research support, DDL has not been widely adopted in L2 classrooms. Studies examining the use of DDL have mostly been carried out in a case study context with a small sample size. To bridge the gap between research and instructional practices, more empirical studies are needed to corroborate the claim that exposing students to the authentic language samples in corpora can effectively promote the development of various language learning areas. This study was an exploration on the use of corpus-based DDL in Reading English for Academic Purposes (EAP) classrooms in an urban university in Indonesia. The study investigated whether different instructional methods result in a difference in vocabulary knowledge, syntax knowledge, listening comprehension, and reading comprehension in different levels of English proficiency. There were 153 freshmen from seven study programs who participated in the study. Randomly assigned to the experimental and control groups, students in the experimental group had corpus-based DDL, while those in the control group received regular vocabulary and grammar instruction. The study found mixed results suggesting the effects of instruction in different directions. While the effect of corpus-based DDL was statistically significantly different for the syntax knowledge measure, the statistically significant differences for vocabulary measures were in favor of the regular vocabulary and grammar instruction. The interaction between the instructional method and English proficiency showed a statistically significant difference for the syntax knowledge measure in support of the higher proficiency level of the corpus-based DDL group. These findings illuminate the understanding of how corpus-based DDL instruction may work in EAP courses targeting undergraduate EFL students. The recommendations include the instructional implications of the study findings and directions for future research.