Abstract
This study investigated whether the use of sentiment analysis methods would improve the reading comprehension skills of students learning English as a Foreign Language. A quasi-experimental pretest-posttest control group design was used in this research, which enrolled 60 intermediate level students enrolled as EFL students. The students were randomly assigned to an experimental group (n = 30) or a control group (n = 30). The experimental group was taught using reading instruction based on sentiment analysis techniques, while the control group received reading instruction through traditional teaching methods (e.g., teacher lectures, classroom discussions).
There were no significant differences found between the two groups in the pretest (Experimental: M = 62.40, SD = 6.85; Control: M = 61.90, SD = 7.10; p > 0.05). However, significant differences were found in the posttests, with the experimental group (M = 78.60, SD = 6.20) significantly outperforming the control group (M = 68.10, SD = 6.75). There was a large effect size (Cohen's d = 1.60) and a large proportion of variance explained by the type of curriculum used for the posttest (40%). Additionally, positive correlations were demonstrated between the posttest scores for reading comprehension and the following questionnaire dimensions: 1) Learner engagement (r = 0.69); 2) Perceived improvement in comprehension (r = 0.74); 3) Motivation and ease of learning (r = 0.71); and 4) Overall satisfaction (r = 0.77) at p < 0.01. Thus, these results demonstrate that utilizing sentiment analysis-based instruction enhances both the cognitive and affective aspects of reading comprehension. Furthermore, this study demonstrates that using this type of instruction will benefit EFL students' ability to learn.