Computational text analysis of intermediate and high intermediate reading passages for ESL learners / Anealka Aziz Hussin
The main concerns of the present study are on issues of matching reading materials to intended learners and standardizing reading materials difficulty level for learners with similar reading ability. This study intends to identify ways to improve the present practice of selecting and adapting readin...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2010
|
| Subjects: | |
| Online Access: | http://ir.uitm.edu.my/id/eprint/14903/ http://ir.uitm.edu.my/id/eprint/14903/1/TP_ANEALKA%20AZIZ%20HUSSIN%20ED%2010_5.PDF |
| Summary: | The main concerns of the present study are on issues of matching reading materials to intended learners and standardizing reading materials difficulty level for learners with similar reading ability. This study intends to identify ways to improve the present practice of selecting and adapting reading materials for ESL learners so that the process can be done in a more objective, consistent and comprehensive manner. To do that, the study proceeds to determine text characteristics that can significantly differentiate the difficulty level of intermediate and high-intermediate reading passages (referred as IR and HIR passages respectively), determine additional predictors that can enhance the efficiency of the Flesch Reading Ease formula to estimate passage difficulty at sentence and word level and develop a set of instruments that can estimate the difficulty level of these passages in a more precise manner. The Descriptive Correlational approach is used as the research design for the study to achieve the above objectives. The samples of the study come from IR and HIR passages. Three computational tools, the Flesch Reading Ease formula, Writer's Workbench 8.18 and Word Smith Tools 4.0, are used to extract information related to passage difficulty at text, sentence and word level respectively. The study also replicates and improves on Vogel and Washbume's (1928) process of developing readability formula. Descriptive statistics namely mean, range and standard deviation, and inferential statistics such as t-test, correlational analysis and multiple regression analysis are used to analyze the data collected. It is found that all the text characteristics used in the study can significantly differentiate the difficulty level of reading passages between the two levels. However, not all these characteristics have significant relationships with passage difficulty level. Two of the significant ones namely the Flesch Reading Ease score (FRE) and the coverage of high frequency words (HFW), are finally selected as predictors to develop a more refined difficulty level formula. Together with the established recommended difficulty range for IR and HIR passages, a set of instruments to estimate passage difficulty level on the refined Intermediate scale are constructed. This study is of great significance to language instructors, material writers, language learners, test-setters, publishers, researchers and self-access learners. With a more precise readability formula and its accompanying instruments, selection and adaptation of reading passages can be done in a more objective, consistent and comprehensive manner. |
|---|