Can Multimodal Data Channels Guide the Scaffolding of Self-Regulated Writing?

• In Event: Using Multimodal Trace Data to Support Learners Across Tasks and Contexts

Abstract

Objectives and Theoretical Framework. Writing is an important skill for college students to develop for their college and future careers. Experienced writers demonstrate using self-regulatory skills during writing, including planning, metacognitive monitoring, and self-reflection (Zimmerman & Risemberg, 1997). However, developing these skills is challenging for students and for their instructors because every writer needs to develop an awareness of which strategies work best for them and when to use them, emphasizing the importance for developing scaffolding for writing instruction.

Graham’s (2018) Writer(s)-within-Communities Model argues that emotions and other self-regulatory behaviors can be used to manage writing processes. For example, a writer may take a break from writing if they are frustrated. As such, we used multimodal data channels to examine self-regulatory processes during learning (Azevedo & Gašević, 2019) by investigating student writing processes and facial expressions of emotions in a first-year writing course.

Methods, Data Sources and Results. 62 undergraduate students enrolled in a course that uses a Writing-about-Writing curriculum (Downs & Wardle. 2007) engaged in a 30-minute writing session. They were prompted to: describe in as much detail the step-by-step process of how you complete a formal writing task.

For our independent variable, we coded students’ written responses to the writing prompt as planning, performance, or self-reflection (Winne, 2018). We created a proportion score for each SRL process and conducted a cluster analysis that yielded a 4-cluster solution.

For our dependent variables, we used keystroke-logging behaviors collected from InputLog (Leijten & Van Waes, 2013) and facial expressions using iMotion’s Affectiva Module (iMotions, 2018), both collected during writing. Keystroke behaviors were time spent producing, inserting, and deleting text, as well as pausing during those behaviors. Videos of facial expressions included proportion scores (0-100 range) for expressing joy, anger, fear, contempt, surprise, disgust, and sadness.

A one-way MANOVA revealed no significant differences in keystroke-logging behaviors between clusters (p > .05). A second one-way MANOVA also did not reveal significant differences in facial expressions of basic emotions between clusters (p > .05).

Scientific Significance. Although not significant, results have implications towards implementing self-regulated writing scaffolds in the classroom. Teachers should consider the written product may not necessarily reflect the self-regulatory processes students engaged in while completing the given task. This emphasizes the importance of collecting multimodal data channels to better understand self-regulated writing processes during writing.

Additionally, it is important to consider when to provide the support, how or how often to do so, and what that support will look like. For example, if students struggle with planning, it will be helpful to provide support early on in a writing session to ensure they implement effective planning strategies. In contrast, if students demonstrate a lack of monitoring or self-reflection, scaffolded support should occur later on. It will also be helpful to consider applying individualized scaffolded support that cater to each student’s particular writing needs. This will ensure they are successful as they write while continuing through college, and ultimately throughout their careers.

Authors

• Michelle Taub, University of Central Florida
  Presenting Author

• Joel Schneier, University of Central Florida
  Presenting Author

• Allison Macey Banzon, University of Central Florida
  Non-Presenting Author

• Lindsey Olivera, University of Central Florida
  Non-Presenting Author

• Sierra Outerbridge, University of Central Florida
  Non-Presenting Author

• Marissa Salas, University of Central Florida
  Non-Presenting Author

• LaVonda R. Walker, University of Central Florida
  Non-Presenting Author