IELTS and TOEFL Applicants’ Self-assessment and Actual Test Performance: A Mixed-Methods Study of Relationship and Variation

Abstract

Recent research on self-assessment (SA) has primarily dealt with its relationship with students' scores. However, few studies have systematically explored SA in language proficiency tests. This mixed-methods study examined differences between IELTS and TOEFL applicants' self-assessments and their actual test scores. It also explored the sources of variations between these two assessments. The study sample included 81 IELTS (n= 51) and TOEFL (n=30) participants. Data collection involved the applicants' self-assessments, their test scores, and semi-structured interviews. Quantitative data were analysed using correlation and regression analyses, while qualitative data were examined through thematic content analysis. The statistical analyses revealed moderate to moderately high correlations between self-assessed and actual test scores. Self-assessments in speaking, reading and writing for both IELTS and TOEFL showed strong correlations with test scores. However, self-assessments in listening were only moderately correlated with actual test scores of both IELTS and TOEFL. In addition, regression analyses indicated that self-assessments in reading, speaking and writing for IELTS, as well as reading and speaking for TOEFL, were significant predictors of test scores. On the other hand, self-assessments in listening for both IELTS and TOEFL, as well as TOEFL writing, were poor predictors of actual test scores. Furthermore, qualitative data analyses highlighted the influence of factors such as experience, psychological aspects, linguistic abilities, background knowledge and feedback in explaining the variations between self-assessments and actual test performance. In conclusion, the paper discusses the findings and implications of the study in the context of language proficiency testing