Does Automated Feedback Improve Writing Quality?

• In Event: Roundtable Session 15
  In Roundtable Session: Implications of Assessment Systems and Tools for Formative Classroom Practice and Improved Learning

Abstract

The current study examines data from students in grades 4-8 who participated in a statewide computer-based benchmark writing assessment that featured automated essay scoring and automated feedback. We examined whether the use of automated feedback was associated with gains in writing quality and with improved first-draft performance or accelerated growth on a subsequent writing prompt. Three-level hierarchical linear modeling reveled that writing quality improved across revisions, though growth decelerated over time. Females and students with higher prior writing achievement scored higher for first-draft performance, but grew at similar rates as their peers. No significant transfer effects were observed, but the effect of SES on growth in writing quality was no longer significant in models describing performance on a follow-up prompt.

Authors

• Joshua Wilson, University of Connecticut

• Natalie Olinghouse, University of Connecticut

• Gilbert N. Andrada, Connecticut State Department of Education