Assessment of authorial voice strength in L2 argumentative written task performances: Contributions of voice components to text quality

Abstract

The purpose of this study was twofold: (a) to examine the level of authorial voice strength among Iranian second language (L2) writers; and (b) to investigate the relationship between L2 learners’ authorial voice strength and the quality of their argumentative written task performances. Argumentative writing samples were elicited from 129 upper-intermediate L2 learners in writing courses. To quantify learners’ voice strength, these samples were scored by two raters using an analytic voice rubric. Raters also provided a holistic rating of the overall authorial voice strength in written argumentations. The quality of argumentations was measured using the TOEFL scoring rubric. While descriptive results indicated that learners demonstrated a low level of voice strength in their argumentations, results from Multiple Regression Analysis (MRA) suggested positive associations between voice strength along with two of its dimensions and the quality of writings. Moreover, results from Multiple Correspondence Analysis (MCA) pointed to the association of low and mid-level of writing quality and low voice strength, and the prevalence of high and mid voice strength in learners with high proficiency in writing. Finally, while an Item Response Theory (IRT) analysis revealed that the ‘presence’ dimension of authorial voice was the most difficult one for L2 learners, a Differential Item Functioning (DIF) analysis showed that the difficulty of the three voice dimensions did not differ significantly across genders. The findings were discussed regarding English L2 writing within the Iranian context.