Parmadi, Alifia Rimas (2023) ERROR ANALYSIS OF ICNALE MODULE FROM INDONESIAN SUBSETS ENHANCED BY LOUVAIN ERROR TAG SET VERSION 2.0: A CORPUS BASED STUDY. Masters thesis, Faculty of Humanity.

Text Tesis_Final.pdf Restricted to Repository staff only Download (3MB)

Abstract

There is a lack of error analysis studies on annotated corpora, particularly those utilising
metadata from Asian students and, more significantly, from the International Corpus Network
of Asian Learners of English (ICNALE) (Ishikawa, 2013). This study aims to analyse the errors
based on Indonesian subsets’ ICNALE edited essays module using the Louvain error tagging
manual version 2.0 (Granger, 2022) and investigate the errors’ consistency made by the
students based on ICNALE’s proficiency levels. The results show that lexico-grammar errors
were the least frequently discovered, making up just 0.10% (2 data), while grammar errors
were the most frequent, accounting for 40% (1039 data). Moreover, the data findings show that
there is a lack of consistency in the occurrence of errors across different proficiency levels. In
terms of grammar errors, the B1_1 intermediate lower level demonstrates a higher frequency
of committed errors compared to the A2 waystage level. Meanwhile, in the field of word errors,
the A2 waystage exhibited the highest level of errors. Following that, the punctuation error
B1_2 at the intermediate upper level had the most number of errors in the essay, with B1_1 at
the intermediate level coming in second place. The inconsistency of the error in the students’
essay is also influenced by the corrector's preferences for writing styles. Having fulfilled the
said aims, I have provided the error annotation of the Indonesian subset from the ICNALE
edited essays module and made it accessible for everyone on CQPWeb.

Actions (login required)

View Item