Abstract/Summary

The main aim of this study was to investigate how oral fluency is assessed across different levels of proficiency in the Test of English for Educational Purposes (TEEP). Working with data from 56 test-takers performing a monologic task at a range of proficiency levels (equivalent to approximately levels 5.0, 5.5, 6.5, and 7.5 in the IELTS scoring system), we used PRAAT analysis to measure speed, breakdown, and repair fluency. A MANOVA and a series of ANOVAs were used to examine the differences between fluency measures at these different levels of proficiency. The results largely replicate previous research in this area (Tavakoli, Nakatsuhara & Hunter, 2020) suggesting that a) speed measures distinguish between lower levels (5.0 and 5.5) and higher levels of proficiency (6.5 and 7.5), b) breakdown measures of silent pauses distinguish between 5.0 and higher levels of 6.5 or 7.5, and c) repair measures and filled pauses do not distinguish between any of the proficiency levels. Using the results, we have proposed changes that can help refine the fluency rating descriptors and rater training materials in the TEEP.