Source Text Use In Integrated Writing

Over the past two decades,integrated writing tasks have become increasingly popular in large-scale English language proficiency tests such as TOEFL iBT,IELTS,and TEM.Integrated writing performance has been the focus of much investigation into the underlying constructs of integrated writing.Unlike traditional independent writing tasks,integrated writing tasks are text-responsible and require test-takers to use the source text while composing their essays.Most studies on source text use in integrated writing have probed source text use by comparing test-takers’ writing products and the source text via textual analysis.Fewer studies have investigated source text use from a process perspective and mainly relied on verbal protocols as the core source of data.Focusing on integrated reading-to-write tasks,this study adopted a mixedmethods approach to investigate the relationship between source text use,task scores,English language proficiency,and source text features.Fifty-three Chinese EFL university students performed two summarization tasks,and the researcher recorded their eye movements,keystrokes,and on-screen activities across the whole writing process.After the writing tasks,fifteen participants were invited to stimulated recall interviews.Innovatively,this study defined a series of transition-based eye movement measures to capture the interaction between target text production and source text processing and operationalize the dynamic process of source text use.Supplemented by stimulated recall interviews,this study further explored the cognitive processes underlying source text use.Results revealed that participants’ source text use behaviors were not significantly different across English language proficiency levels or source texts.Still,the interaction between English language proficiency and source text feature did explain a certain amount of the variation of source text use.Hierarchical regression analyses showed that the effects of source text use on summary scores mainly depended on participants’ English language proficiency.This finding suggests that English language proficiency might moderate the relationship between source text use and summary score.Based on the findings,this study provided theoretical and practical implications for language teachers,test designers,L2 writing researchers,and language learners and offered suggestions for future eye-tracking-based integrated writing process research.