The effects of written corrective feedback processing conditions on L2 learners' written texts and levels of depth of processing

Abstract

The present PhD attempted to contribute to two relevant SLA-oriented lines of research: (i) effects of composing medium on written texts; and effects of feedback processing conditions on writing processes and products. The intended contribution was empirical, including a central methodological aim. The motivation for these global aims derives from the following considerations. On the one hand, in response to the mass shift seen in language classrooms to more online, digital learning environments, L2 writing scholars have advocated for research to explore and compare the effects of traditional pen-and-paper versus digital composing environments on the cognitive processes involved in writing and feedback processing, as well as on the characteristics of the resulting written texts (e.g., Vasylets & Marín, 2022; Vasylets et al., 2022; Zhi & Huang, 2021). On the other hand, SLA-oriented research on written corrective feedback (WCF) has received ample attention throughout the years (as recently reviewed by Bitchener, 2021; Roca de Larios & Coyle, 2021; Hyland & Hyland, 2018; Kang & Han, 2015, 2021), with studies that focus specifically on the processing of WCF gaining increased attention more recently. The construct of depth of processing (DoP, Leow, 2015, 2020) of WCF has become a key concern in theoretical and empirical feedback research. Studies in this domain have employed diverse methodological procedures -including think-aloud protocols (e.g., Bowles & Gastañaga, 2022: Caras, 2019; Kim & Bowles, 2019; Leow et al. 2022; Sachs & Polio, 2007) and written languaging (e.g., Cerezo et al, 2019; Manchón et al, 2020; Suzuki, 2012, 2017) to obtain data on potential (i) effects of DoP on how deeply L2 users engage with the feedback provided on their writing, and (ii) correlations between DoP and language acquisition (usually operationalized in terms of improvements in text revisions). Importantly, scholarly debates have more recently focused on research methodological considerations regarding data elicitation procedures in this research. In this regard, some critics (e.g., Leow & Manchón, 2021; Manchón, 2023a) have called for more controlled, methodologically oriented studies in which the validity of the data collection instruments is tested, whilst also advocating for investigations in more diverse writing environments. In response to these calls, the main aims of the present doctoral thesis were to contribute empirically to previous research by exploring writing and feedback processing in both pen-and-paper and digital environments, and to shed light on the affordances of diverse introspective measures (individually and combined) commonly used for WCF processing. To achieve these global aims, the following research questions guided our study: RQ.1 How does writing in a traditional pen-and-paper environment versus writing in a computer-mediated environment affect L2 written production in terms of CAF measures? RQ. 2 How does the experimental manipulation during WCF processing affect L2 written production (in terms of CAF measures) in pen-and-paper versus computer-mediated writing environments? RQ.3. How does the experimental manipulation during WCF processing affect L2 learners’ levels of depth of processing of the feedback received in pen-and-paper versus computer-mediated writing environments? To answer RQ2 and RQ3, the study explored the methodological affordances of three WCF processing conditions: (i) think-aloud protocols, (ii) written languaging, and (iii) a combination of think-aloud protocols and written languaging in two writing environments (computer-mediated and pen-and-paper writing conditions). The study followed a pre-test/treatment/post-test design in which 36 English undergraduate students participated. Participants were invited to write an initial text (pre-test) in time-constrained conditions. The writing task was the problem-solving, picture-based “Fire Chief” task (Gilabert, 2007), which was completed by half of the participants (18) online, via GoogleDocs, and by the remaining 18 participants on pen-and-paper. Regardless of the writing and processing conditions, all participants received unfocused, direct WCF on their initial written texts. The participants were then invited back to process the feedback received, according to the treatment group to which they were assigned: (i) think-aloud only, (ii) written languaging only, and (iii) simultaneous think-aloud and written languaging. The final task (post-test) invited participants back to rewrite their original text under the same conditions as in the pre-test. Once the processing data had been collected, the think-aloud protocols were transcribed and coded following Leow’s (2015) definition of DoP and the written languaging data was coded according to the levels of engagement and noticing, guided by the coding scheme elaborated in Cerezo et al. (2019). The written products were analysed in terms of a range of CAF measures. Results show that, as regards composing medium, computer-mediated written texts were initially found to be more accurate and more fluent when compared to more traditional pen-and-paper written texts. Additionally, composing medium played a role on how the participants engaged with feedback: the pen-and-paper condition was more successful in engaging students in metalinguistic languaging, which, as a result, led to higher levels of accuracy in subsequent revised texts. In terms of feedback processing conditions, results show that the combination of think aloud and written languaging whilst processing WCF was not only the processing condition that provided the most insights into WCF processing, but it also constituted the most favourable treatment condition for promoting higher levels of DoP. These deeper levels of processing also led to higher L2 accuracy in subsequent text revisions. These results represent relevant, novel insights into writing and feedback processing in diverse writing environments, while at the same time the insights obtained point to equally relevant and novel methodological implications for future research by shedding light on the affordances of WCF processing instruments and conditions. In addition, and from a pedagogical perspective, the results also present a series of potential implications relevant to pedagogical decision-making in the second language classroom with regards to writing task implementation and feedback processing conditions.