Word recognition skill and academic success across disciplines in an ELF university setting
Michael Harrington, University of Queensland, Australia
Thomas Roche, Southern Cross University College, Australia
https://doi/10.0000/0000
Volume 3, Issue 2, 2014
Abstract: Previous research (Harrington & Roche, 2014) showed that the Timed Yes/No Test (a measure of vocabulary size and response speed) is an effective tool for screening undergraduate students at risk of failure in English-as-a-Lingua-Franca (ELF) university settings. This study examines how well performance on the test predicts grade point averages across different academic disciplines in one of those contexts, an ELF university in Oman. First year students (N= 280) from four academic disciplines (Humanities, IT, Business and Engineering) completed Basic and Advanced versions of the Timed Yes/No Test. The predictive validity of word recognition accuracy (a proxy for size) and response time measures on GPA outcomes were examined independently and in combination. Two patterns emerged. Word accuracy was a better predictor of academic performance than response time for three of the groups, Engineering the exception, accounting for as much as 25% of variance in GPA. Response time accounted for no additional unique variance in the three groups after accuracy scores were accounted for. In contrast, accuracy was not a significant predictor of GPA for the Engineering group but response time was, accounting for 40% of the variance in academic performance. The findings are related to the use of the Timed Yes/No Test as a reliable and cost-effective screening tool in Post Enrolment Language Assessment (PELA) applications in ELF settings. 
Keywords: vocabulary size, recognition speed, academic performance, screening, academic English