Slim Khemakhem email@example.com
Investigating the predictive validity of IELTS for a teacher education program in UAE
This research study investigates the predictive validity of IELTS test scores as a graduation requirement for a Teacher Education Program in the UAE. The argument is that the use of the IELTS cut-off score (band 6) as a predictor of post-graduation performance in schools may not be justifiable based on differences in context-related interaction and their impact on language proficiency. Subjects in the present study were given a mock IELTS speaking test, and were recorded while giving classes on teaching practicum to examine the correspondence between the two performances and the impact of each context on their lexical diversity scores.
A comparison between lexical diversity scores in each context using index D, revealed that there was no correlation between the two sets of scores as lexical diversity scores in the classroom were clearly lower than IELTS scores. Moreover, a comparison between the subjects’ band-scores on the IELTS speaking test and their mentors’ grades on teaching practicum revealed that there was no correlation. Corpus linguistic tools were also used in this study to compare content-word frequency lists produced in the IELTS speaking test with those produced in the classroom. Results showed that there were clear discrepancies, which indicated that context was a determining factor for the subjects’ lexical choices. A further qualitative analysis took an emic perspective following a Conversation Analysis approach in analysing the subjects’ talk in the classroom. Seedhouse’s (2004) framework for the analysis of the architecture of classroom interaction and Walsh’s (2011) combined Corpus Linguistics and Conversation Analysis (CLCA) approach were used as frames of reference. The analysis revealed that teacher repeats of lexical items in the classroom played a major role in reducing the subjects’ lexical diversity scores due to their frequency and variety. A total of 16 types of teacher repeats was identified. Those types were classified into 6 main categories related to different features of classroom discourse such as control of interaction, asking questions, giving feedback, and so on. Consequently, the predictive validity of IELTS scores was found to be weak for the context of a teacher education program.
A two-step argument-based validation process is suggested at the end of this research study such that IELTS scores can be included as part of the requirement for the graduation of the Bachelor of Education program students. The first step is to adopt a complementary assessment based on a rubric created for the purpose of assessing classroom-based English language proficiency (CBELP). It is suggested that this assessment is administered by a trained mentor during classroom-based observations conducted on teaching practicum. The second step is to formulate a composite score based on the minimum required band-scores for both the IELTS and the CBELP assessments. The interpretation of that score should be supported by a validating argument leading from score interpretation to related decisions regarding the graduation of Bachelor of Education students.
The thesis makes a contribution to theory in the fields of testing, lexical diversity and language proficiency, and furthers our understanding of classroom discourse.
Khemakhem, S. (in press). Investigating the predictive validity of IELTS for a teacher education program in UAE. (Thesis). University of the West of England
|Acceptance Date||Jan 31, 2016|
|Keywords||English language proficiency, lexical diversity, language testing, IELTS, predictive validity, argument-based validity, classroom discourse/ interaction, teacher repeats, corpus linguistics|
Slim Khemakhem_Thesis_Final Draft.pdf