This thesis presents an investigation into the Malaysian secondary school students’ syntactic construction and their metalinguistic understanding of writing. Through a mixed-method design, it also examines the relationship between these two aspects and their teachers’ metalinguistic understanding of syntactic complexity and how they make judgments about writing quality. The study was conducted in two phases. The first phase involved the collection of 120 essays from student participants with different levels of proficiency. Several one-hour writing sessions were administered in three different schools. To examine their syntactic constructions, the essays were manually analysed and coded based on the coding frame, which was divided into several sections consisting specific syntactic measures. The second phase involved several writing conversations comprising two sections: 1) open-ended questions aimed to elicit perceptions and believes on writing and grading essays; 2) elicitation tasks aimed to elicit metalinguistic understanding among teacher and student participants. Students’ essays were also used in the writing conversations to elicit students and teachers’ comments on the writing. The 12 student representatives were chosen based on their proficiency levels and the syntactic structures used in their essays. All six English teachers teaching the students were involved in the writing conversations. There are two key findings derived from this study. Firstly, the detailed analysis of the essays suggests that syntactic complexity of writing based solely on the presence of syntactic features may not determine essay quality and effectiveness. Secondly, while teachers and students are still not confident of their metalinguistic knowledge and understanding, it was evident that students’ perceptions of good writing mirrored their teachers’, which are very much focused on accuracy. There was no evidence of students or teachers discussing their linguistic choices in writing, especially in achieving the rhetorical goals. While explicit grammar knowledge may help in improving accuracy, the study also suggests that it should also be used to foster discussion about writing that goes beyond language accuracy.