Analyzing Student Writing: A Corpus Linguistics Study

Abstract

The aim of this study was to examine the written compositions of 40 fifth-semester students at Universitas Negeri Padang. The objective was to determine their strengths and areas that need improvement. The study utilised a manual analysis technique, investigating factors such as word frequency, sentence length, grammatical structures, contextual usage, and collocations. This thorough method yielded an in-depth comprehension of the pupils' writing proficiencies. The findings unveiled notable disparities in writing habits among various academic fields. Scientific writing typically employs a straightforward and factual style, characterised by shorter phrases. In contrast, humanities assignments sometimes involve longer sentences that include comprehensive explanations and critical assessments. Students with a higher level of proficiency showed a wider range of vocabulary and better correctness in grammar. In contrast, students with a lower level of proficiency used simpler words and made more frequent grammatical mistakes, such as inconsistent tenses and problems with subject-verb agreement. The majority of students demonstrated accurate usage of common collocations, while there were some instances of improper usage, suggesting a need for a better grasp of these terms. The results also highlighted a dependency on simple conjunctions and a deficiency of more sophisticated transitional words, which affected the writing's coherence and refinement. Based on these findings, the study recommends implementing focused teaching tactics to target specific areas of difficulty. To increase students' writing skills, it is crucial to provide them with targeted grammar teaching, engaging vocabulary tasks, and many opportunities to practice collocations. Moreover, providing discipline-specific writing education can assist students in modifying their writing styles to suit various academic circumstances. Future studies should explore the integration of automated analytic methods and broaden the scope to encompass a wider range of student samples and multimodal data in order to gain a more thorough insight into students'.