Abstract
The authorial voice is an essential part of success in academic writing. For the argumentative essay genre, developing a solid authorial voice contributes to increased writing quality and authorship beliefs in students who are non-native speakers of English (NNSE). However, these students face challenges in expressing their authorial voice due to the complexity of authorial voice as a pedagogical construct. Moreover, learning tools developed to address authorial voice challenges do not combine learning linguistic features with self-reflection in a student-facing manner. This study is thus a proof-of-concept paper that tests the usability and effectiveness of AVATAR (Authorial Voice Toolkit for Authorship Reflection), a prototype authorial voice learning technology. The study first reviewed the literature to operationalize authorial voice as performance and self-perception. Performance aspects of authorial voice encompass the linguistic choices a student-writer makes, the factors influencing these choices, and a reader’s interpretation of the written text. Self-perception incorporates how students feel about themselves as writers, including self-efficacy and authorial confidence. Next, AVATAR was developed as a theoretically grounded intelligent tutoring system that provides integrated affordances for the performance and self-perception of the authorial voice of undergraduate argumentative essay writing. The study then adopted a pre-use vs. post-use one-group research strategy to test the usability and effectiveness of AVATAR’s affordances for both performance and self-performance aspects of authorial voice. This research was undertaken at a prominent Fijian university in an NNSE higher education setting. The participants (n = 23) were NNSE first-year students enrolled in an English academic writing course.
Participants utilized an authentic argumentative essay assignment to test AVATAR’s affordances. Data were collected using a combination of quantitative and qualitative methods. A post-use usability survey and semi-structured interviews determined the usability and learnability experiences of the participants. For the testing of AVATAR’s self-perception affordances, pre- and post-use surveys and interviews were utilized to detect changes in participants’ authorship beliefs on authorial confidence and self-identification as authors. This was further informed by participants’ self-scores for their authorial voice in the initial and edited versions of their essays. AVATAR’s affordances for the performance aspects of authorial voice were evaluated through tracking changes in the authorial voice linguistic features present in the initial and edited version of participants’ essays. The initial and edited essays were also scored for authorial voice strength using a voice analytic rubric. Post-use interviews further informed these findings. Analyses revealed positive feedback for usability and learnability’s interface and automated reports, although a lack of familiarity with authorial voice and a perceived need for teacher input affected participants’ use of AVATAR. Findings also revealed evidence of AVATAR’s self-perception aspects having a positive impact with a statistically significant change in participants’ post-use authorial confidence and the self-scores participants gave themselves for authorial voice. Participants’ interview responses also indicated an increased awareness and appreciation of authorial voice. AVATAR’s visualization and integration of self-assessment reflection with practicing inserting authorial voice linguistic features were highlighted as the factors that prompted changes in participants’ self-perception of authorial voice. Findings for the performance affordances showed that most participants utilized AVATAR to edit their essays for authorial voice. There were also statistically significant changes in most dimensions of the voice analytic rubric used to score the initial end edited essays. The prevalent reasons for prompting participants’ engagement with affordances for authorial voice performance were AVATAR’s automated reports visualization of authorial voice in participants’ essays and the reflective self-assessment task in AVATAR. However, for both self-perception and performance aspects, some participants expressed the need for teacher feedback to validate their authorial voice development and explicate the authorial voice’s importance. The study ends with a discussion of the implications for authorial voice pedagogy and the development and implementation of authorial voice learning technologies in higher education settings.