Learning Self-regulated L2 Writing Under a Cognitive Model

This research study has been conducted to experiment an innovative teaching approach believing that when less use of self-regulated learning strategy leads to lower metacognitive learning attitude which results in low prociency. This study has investigated the consequences of instructionally aroused cognitive involvement load for self-monitoring and self-assessment through facilitation of metacognitive learning strategy use for improving L2 writing skills of Thai undergraduate students. This innovative instructional model for teaching self-regulated L2 writing has been named as Strategic Self-Regulated Metacognitive Activities or S 2 RMCA. The approach of this model has been developed to promote self-regulated learning management. For stimulating the use of learning strategies, a set of self-monitoring and self-evaluating assessment rubric named Strategy Inventories for Learning L2 Writing (SILL2W) has also been devised. A set of questionnaire, pre-post-tests, checklist, and interviews were employed for collecting and analyzing the data. Results of data analyses have shown effectiveness and feasibility of the S 2 RMCA model for teaching self-regulated L2 writing. Satisfactory results have also been shown by participants in their L2 writing skills. So far, research studies conducted on cognitive involvement load, a continuing challenge has always been there regarding the accurate measurement of load via self-reporting and this study has also faced that challenge.


Introduction
Several research studies have shown solid correlation among self-regulation, language learning strategies, metacognition and cognitive involvement load from different angles and viewpoints. Cubukcu (2009) has portrayed the relationship between self-regulation, metacognition, and autonomy and demonstrated that low autonomy is closely related to low self-regulated behaviors in learning. Chamot and Harris (2019) have asserted that promising evidence is still anticipated relating to how to design and implement the language learning strategy mediations. Evans, Kirby, and Fabrigar (2003) have claimed that there is synchronous interplay among the processes of writing, learning approaches and selfregulation. Farrington and colleagues (2012) have directed at researcher speci ed learning strategies as the limitations of existing studies that are not essentially selected by teachers or learners. Raya (2011) has discussed about the important role played by teachers in creating learning opportunities that can support self-regulation and re ectivity of learners. In the domain of language learning strategies and selfregulated learning, research has rarely been meant to have an in uence on classroom practices (Conley, 2014). As asserted by Conley, it is even hard to design such research. There are not enough models to train teachers on instructional models that can foster self-regulated learning of L2 writing. This study has come forward with an innovative model of teaching and learning L2 writing -Strategic Self-Regulated Metacognitive Activities (S 2 RMCA) for fostering self-regulated and metacognitive learning strategies through instructionally stimulated cognitive involvement load on learning processes. S 2 RMCA is designed to engage learners in self-controlling their participation in planning, self-monitoring, and self-evaluation.

Literature Review And Conceptual Framework
Literature engaging learners to achieve their goals through self -regulated learning, explaining, modelling, and scaffolding of the learning strategies to students in regular classrooms are needed (Pressley, Lindsay, Fingeret, Re tt, & Raphael-Bogaert 2007). Issues like relationship of theoretical underpinnings with strategy training and successful learning need attention (Gkonou & Inceçay, 2016). Focus of some recent studies have been the relationships among the language learning strategies and self-regulation (Rose, 2012;Chamot, 2018;Gri ths, 2013;Oxford, 2018). Requirements of active engagement and involvement of the learners in the process of learning were discussed (Cotterall, 2000;Zimmerman & Schunk, 2008). Very little is explored about associating learners' thinking with the production in texts. Sub processes included in the Hayes and Flower's (1980) model were translating thoughts into texts and examining in relation to a learner's long-term memory and a task environment. Hayes (2012) has restructured his previous model and formed a control over level of motivation, setting goals, planning, and writing schemata. Studies have also examined sub processes of L2 writing in terms of revising, uency of text generating, and restructuring of texts (Hall, 1990;Chenoweth & Hayes, 2001;Larios, Murphy & Manchόn, 1999). Some other studies that have compared cognitive strategy use between L1 and L2 writing (Arndt, 1987;Whalen & Menard, 1995;Cummimg, 1989;Sasaki, 2000). This study has looked at self-regulated learning as learners having control over their own thinking, using metamemory, choosing learning strategies and learning behavior to solve problems in developing L2 writing. Pour-Mohammadi, Zainol and Cheong Lai, (2012) have claimed that useful and strategic mediations of language learning strategies can support to improve L2 writing. In the S 2 RMCA model, concentrating, processing information, extracting meaning, organizing ideas, composing texts, or developing sentences and paragraphs, monitoring own compositions, engaging with study aids, and understanding the tasks are listed as metacognitive strategies. Tseng, DÖrnyei, and Schmitt (2006) have not included the learning strategies in their effort to evaluate "strategic-learning". R. Oxford and Amerstorfer (2018) have claimed that there has been still a void particularly in the instructional modelling of teaching self-regulated L2 writing. Lack of resources as well as guidance for quantifying use of learning strategies by learners in classroom environments is a signi cant challenge at present (Gunning, 2011). For facilitating selfregulated and metacognitive learning strategy use, this study has developed a self-questioning checklist -Strategy Inventories for Learning L2 Writing (SILL2W) with 40 inventory items. SILL2W has been designed especially in self-questioning form to help facilitate self-monitoring and self-evaluation of L2 writing. Oxford (2017) has claimed that there are less information sources of cultivating self-regulated and metacognitive learning strategies for L2 writing skills. The S 2 RMCA model is inspired by S-W-SR (Weinstein and Palmer, 2002) and ER-GO-CA (Olejnik and Nist, 1992). The study of Phoocharoensil, S., Moore, B., Gampper, C., Geerson, E., Chaturongakul, P., Sutharoj, S., and Carlon, W. (2016) has revealed some problems that Thai EFL undergraduate students often encounter in L2 writing are due to not only L1 in uence, but also due to confusion over the target language and its complex grammatical system. As cited in Nopmanotham (2016, p. 37), the study conducted by Pimsan (2003) has revealed that mainly metacognitive, cognitive, social, and affective strategies in L2 writing have been the lower strategies used by Thai graduate students. Table 1 below shows the components and the self-regulated and metacognitive learning strategies of the S 2 RMCA model.  DÖrnyei and Ryan (2015) have argued that research have not achieved any answer to the primary concerns about what splits strategic learning activities from regular learning activities. Macaro (2006) has stated that learning strategies are located in the working memory and Oxford (2017) has argued that application of learning strategies requires working memory. On the other hand, Sweller (1988) has hinted that working memory has limited ability to keep information. Recommendations of Cognitive Load Theory have been bene cial for designing teaching and learning materials to handle learners' working memory for successful learning. Involvement Load Hypothesis (ILH) of Laufer and Hulstijn (2001) has stated that when there is higher involvement load on the learning processes, there is more effective learning. This study has developed instructional materials to stimulate cognitive involvement load on learners to manage self-regulated learning of L2 writing. Conley (2014) has advised that students should be trained to monitor the use of learning strategies through self-evaluation and peer feedback. Oxford and Amerstorfer (2018) have claimed that the relationships of multiple factors of self-regulated language learning strategies, contexts, and individual differences can be brought together in strategy instructions to meet up the requirements of learners with distinct pro ciencies. Studies have tried to meet individual learners' goals and characteristics (Oxford, 2018;Chamot & Harris, 2019). Cao, Y. (2012) has reported that training of metacognitive strategy could show positive results in web-based English learning. Grounded on the thoughts as described above, this study has made an effort to response to the following research questions.
1. Can the S 2 RMCA model help L2 writing learners to be self-regulated in a regular classroom environment and improve L2 writing skills? 2. Whether instructionally induced cognitive involvement load can promote learners' self-regulated and metacognitive learning strategies?

Method
This study was carried out as an experiment to promote use of self-regulated and metacognitive learning strategies by 26 Thai undergraduate engineering students divided into three groups. Duration of the research was an academic semester of 17 weeks. Groups A and B were taught under the instructional model of this study. Group C students did not receive the treatment of task demands for stimulating cognitive involvement load. Because of the nature of participation, Group C can be marked as the control group. In the starting of the semester, students were informed about self-regulated learning approach and use of self-regulated and metacognitive language learning strategies. A survey was conducted to assess the self-regulated and metacognitive learning strategy use by the students. Research tools of this study were of two types -Instructional and Testing. Data analyses were done to gather information regarding the in uence of instructionally induced cognitive load on the learners thinking processes and to notice the impacts of using self-regulated and metacognitive learning strategies in L2 writing. Tools of data collection were self-reports, a semi-structured questionnaire, observation, a survey questionnaire, pre-post test scores, and two online tools.

Data Analyses And Findings
The analyses started with examining the reliability and validity of the instructional instrument SILL2W through multiple statistical tests for each group of participants and for each category of learning strategies independently. Total 40 learning strategies have been included in the SILL2W construct covering the areas of motivation, self-correction, self-regulation, and metacognition.    indicates that the metacognitive items [2,3,4,5,6,8,9,10,11,13,14,15,16,17]   In table 5 below, the Kruskal-Wallis test mean ranks for three writing tasks of Group B students (pre, mid and post) are presented. The scores in the Mean Rank column can be used to compare the differences in writing performances of week 2 and week 17. Students were involved in self-monitoring and selfevaluation to fend for and treat errors by using the SILL2W and the writing assessment rubrics. The differences in the mean scores can be interpreted as showing better performances because of increasing scores. The results of Group B show statistically signi cant differences in scores of pre, mid and post tests and number of words. However, Group B had 7 students and the sample size was relatively small for statistical analysis. Therefore, the scores of mean ranks (with arrow markers) can be considered for interpretation. Table 5 Kruskal -Wallis test mean ranks for Group B students (pre, mid and post)

Page 9/16
A Wilcoxon Signed-Rank non-parametric statistical test was also run to evaluate the pre and post test scores in order to determine whether the group's population mean ranks varied through a paired difference test. Table 6 below shows the mean rank and sum of rank along with the negative and positive ranks in pre and post test scores of Group A, B and C. Table 6Negative and positive ranks of pre and post test scores of Groups A, B and C convincing indication that the participants of the experiment groups have shown improvement in the test scores. Experiment groups' performance scores attained through grading with the assessment rubrics showed signi cant differences whereas the control group's performance did not show any signi cant difference between pre and post test scores. Table 7Wilcoxon Signed-Rank test statistic based on negative ranks of Groups A, B and C The self-reporting checklist has been designed to elicit students' opinions on experiences of cognitive pressure for self-regulated learning. Table 8 below shows the results of Chi-Square statistical difference in use of learning strategies based on the self-reports. A Chi-Square with low value means high correlation between two sets of data.   Figure 1 displays Group A's self-report on cognitive load experience. 38.5% students rated their experience at 4 out of [1 to 6]. 23.1% commented optionally [more than 6 or 6+] that they were exhausted because of cognitive involvement load, but they could learn. 30.8% students rated their experience at 3 and rest of the students rated their cognitive involvement load at 5. Figure 1 Group A's self-report on cognitive involvement load Figure 2 displays self-report of Group A's using metacognitive strategies at the end of the semester. The question was whether they learned about planning, organizing, monitoring, and evaluating own writing to improve. 84.6% students reported as "YES", and 15.4% students reported as "LEARNED SOME" of the metacognitive strategies. Figure 2 Group A's self-report on metacognitive strategy use Figure 3 demonstrates Group B student's self-report on using metacognitive strategies. 42.9% students con rmed trying to understand the task requirements before writing the draft 1, 14.3% students chose "STRONGLY YES", and 14.3% students chose "SOMEWHAT NOT". 18.6% students were neutral about using the metacognitive strategies before starting a writing task. Figure 3 Group B's self-report on application of a metacognitive strategy after pre test Figure 4 reveals Group B's self-report on application of the metacognitive strategies during the midsemester week. 57.1% students con rmed "SOMEWHAT YES" about trying to understand the task requirements before starting a task and 42% students con rmed being "Neutral". But no student chose "NOT" for using strategies, which con rms students' being aware of using metacognitive strategies after receiving cognitive involvement load.

Conclusion
The self-reports in the pie charts above present evidence of the participants developing self-regulated and metacognitive learning strategies. From the self-reports, it can be interpreted that the SILL2W could generate some understanding and willingness among learners for self-regulated learning and employ metacognitive learning strategies. Based on the results, the answer to the rst research question would be that the S 2 RMCA model can be helpful for L2 writing learners to be self-regulated and improve L2 writing skills. The comparisons of test scores between the beginning and the end of the semester show statistically signi cant differences in the quality of writing skills and use of the learning strategies. While answering the second research question, it can be suggested that instructionally induced cognitive involvement load can facilitate use of self-regulated and metacognitive learning strategies for improving L2 writing. Students of both A and B groups have reported experiencing high cognitive load and learning.
The development and improvement in the writing qualities of the learners have con rmed that selfregulated and metacognitive learning strategy training if integrated in pedagogy can facilitate learner's self-regulated learning of L2 writing. Learners of this study have executed their learning management as planned in the S 2 RMCA model through instructionally demanded and controlled self-regulated metacognitive actions. The instructional model S 2 RMCA and the instructional instrument SILL2W have been useful for teaching and learning self-regulated L2 writing in a regular classroom.

Recommendations
Based on the results and experience of this study, few suggestions can be offered. Teacher education programs should provide training on developing teaching approaches for enhancing self-regulation and metacognition. Pedagogy should provide guidance for learning management from early stages of learning which is necessary for self-regulated learning at higher level of education. Pedagogical frameworks should be constructed for educators and material developers to enhance self-regulation and metacognition in both formal and informal learning settings. Cognitive aspects of learning should not be