Doctor of Philosophy (Ph.D.)
Degree Granting Department
Curriculum and Instruction
John I. Liontas, Ph.D.
Yiping Lou, Ph.D.
Sanghoon Park, Ph.D.
Michael Sherry, Ph.D.
Jennifer Wolgemuth, Ph.D.
COCA, ELL, instructional technology, TESOL
Because of their prevalence in both spoken and written English, collocations—words that go together—are important for the English language learner. Collocations that contain prepositions have been shown to pose particular difficulty (Hong, Rahim, Hua, & Salehuddin, 2011). Collocational errors are common, even among advanced learners (Laufer & Waldman, 2011). The purpose of this pilot study was three-fold: (1) to determine whether the use of a digital corpus by English language learners is effective in learning prepositional collocations, (2) to determine if proficiency level exerts any influence on effectiveness of using a digital corpus to learn prepositional collocations, and (3) to examine learner perception regarding the use of a digital corpus to learn prepositional collocations.
Forty-four international undergraduate students participated in this mixed-method study, and were randomly assigned to either a treatment or control group. Study materials were delivered via Canvas, a learning management system. All participants completed a pretest (gap-fill format) consisting of 15 target prepositional collocations to establish a baseline. Target collocations were presented to treatment group participants via an instructional module utilizing the Corpus of Contemporary American English (COCA), while control group participants used an instructional module without COCA. Following instruction, participants completed an immediate posttest and a delayed posttest two weeks later. Treatment group participants completed a survey to gauge their satisfaction with COCA and perception of its usefulness.
Immediate and delayed posttest gains were compared using statistical analysis, but results did not show a statistically significant difference, suggesting that the impact of COCA on collocational learning is inconclusive. Utilizing factorial analysis of variance for pretest to immediate posttest gains, there was a statistically insignificant effect of group (treatment vs. control) after controlling for proficiency. For pretest to delayed posttest gains, a statistically significant effect of group was indicated, although the effect size was small. A statistically significant effect of proficiency on test score gains (for both pretest-immediate posttest and pretest-delayed posttest) was shown after controlling for group. There was no statistically significant interaction effect for either pretest-immediate posttest gains or pretest-delayed posttest gains. Results seem to indicate an impact of proficiency on gains, although higher baseline pretest scores may have played a role. Thematic analysis of feedback from the post-study survey revealed several areas that participants emphasized in their responses: user interaction/interface, usefulness, context/examples, functionality, and layout/design. Responses were positive overall, suggesting that participants viewed COCA favorably in terms of satisfaction and usefulness. There were mixed responses regarding user friendliness and ease of use, highlighting the importance of effective training on the use of COCA prior to instructional integration. The most frequently cited positive attribute of COCA was the use of authentic examples of collocations in context. Overall, despite some inconclusive quantitative results, qualitative findings suggest positive perceptions and benefits of using COCA for learning collocations, and the extension of this pilot study to a main study seems feasible and promising.
Scholar Commons Citation
Siegel, Matthew A., "Digital Corpus Use in Learning L2 Prepositional Collocations: A Pilot Study" (2019). USF Tampa Graduate Theses and Dissertations.