Instructor modeling of metacognitive processes has been shown to support student learning. Teacher educators have leveraged this research to support teacher candidates in framing, decomposing, and explaining disciplinary thinking and decision-making, particularly in reading development. However, little work has focused on supporting teacher candidates' use of similar pedagogical moves when teaching writing. This project investigated teacher candidates modeling of writing before and after instruction aimed to support their ability to model. What we noticed was a trajectory of learning to frame, decompose, explain, and highlight writing content using visual and metacognitive representations to make the cognitive processes of writing more visible. However, we noted places for growth in teacher candidates' knowledge of content and students, so that candidates might more aptly decompose and highlight their disciplinary thinking in ways that are appropriate for the teaching context.