Investigating the Relationship between Fluency Measures and Second Placement Test Decisions

Abstract

This study investigates the relationship between measures of fluency and placement decisions in 90 L2 writing samples written for a placement test at a university academic English program. The purpose of this study is threefold: (a) to find out what patterns can be described in the difference between measures of fluency and the placements; (b) to examine the relationship between six measures used in previous literature as a gauge of fluency, and (c) to determine the extent to which fluency can predict students’ placement test results. The results showed a significant difference between the two placement levels for most of the fluency measures. Also, the factor analysis contradicted claims by Wolfe-Quintero et al. (1998) that words per T-unit and words per clause should be considered fluency rather than complexity measures. Discriminant analyses showed that all six measures together showed the highest predictability overall. However, for the intermediate level, raw fluency measures were the best predictors, while the best prediction of placement into the advanced level was a combination of raw fluency and complexity measure.