Writing from sources is an important academic literacy skill essential for university students to succeed in academia. Nonetheless, because it involves a set of complex cognitive, metacognitive, and self-regulatory processes and strategies, it is extremely challenging. 

Existing research primarily focused on the cognitive processes of source-based writing, adopting qualitative and case-study based methods. While the research generated a nuanced understanding of the intricate mental struggles and issues during the reading-to-write process, it did not investigate the contextual and behavioural aspects of the process, such as the regulation of time, environment and motivation. There is also a paucity of research adopting quantitative means to connect important antecedent, process and outcome variables to generate a comprehensive picture with sufficient clarify to guide practice and further research.

The proposed study will attempt to address the above gaps in the literature. It will adopt Structural Equation Modeling (SEM), an advanced modeling technique appropriate for investigating relationships among multiple variables simultaneously, to examine the effects of key antecedent and process factors on academic papers written for disciplinary courses. Six measures will be adopted to assess these important variables identified from existing research, as follows: two antecedent factors (the reading-to-write integrated skill (Plakans & Gebril, 2012) and knowledge and competence of source use and reference (Hu & Lei (2014)); three process variables (task representation (Wolfersberger, 2013), control strategy (Tsung & Dornyei, 2006), and discourse synthesis strategy (Spivey, 1997)), and one outcome variable (intertextuality practices). The outcome variable will be measured in terms of source integration (Keck, 2006), citation density (Cumming et al., 2018) and rhetorical functions (Petric & Harwood, 2013)).

The study will gather data from a sizeable sample (up to 306) of undergraduates majoring in humanities and social science subjects in Hong Kong universities. Participants will submit a recently completed source-based argumentative essay, respond to measures in a questionnaire and take an integrated reading-to-write test. Accordingly, data will consist of written texts, questionnaire data, and test scores. Data analysis will identify and compare profiles of the intertextuality practices of successful and unsuccessful writers and associate each profile with the antecedent and process variables. It will then model the three sets of variables and verify their relationships projected based on existing research and self-regulated learning theory.

The research will have pedagogical, theoretical and policy implications. Pedagogically, the research can generate valuable diagnostic information for university-level instruction and assessment of source-based academic writing in English. Theoretically, the research will make important contributions to two lines of scholarly enquiry, namely, research into diagnostic language assessment and research into source-based writing. At the policy level, the research will provide important implications for undergraduate curriculum design, university admission requirements, and have far-reaching influence on English writing education at the secondary level.