Identifying active ingredients and uptake patterns in the implementation of an AI-based writing support tool: Insights from a randomized controlled trial
Huang, Y., Palermo, C., & Wilson, J. (2025). Identifying active ingredients and uptake patterns in the implementation of an AI-based writing support tool: Insights from a randomized controlled trial. Computers and Education: Artificial Intelligence, 9, Article 100479. https://doi.org/10.1016/j.caeai.2025.100479
This secondary analysis used treatment-group data from a randomized controlled trial of MI Write, an automated writing evaluation system used by 902 middle school students across three U.S. districts, to examine which system components and usage patterns were associated with writing outcomes.
Key takeaways
- The amount of writing completed in MI Write and the number of revisions were associated with gains in argumentative writing performance.
- Uptake of these components varied across districts, and stronger engagement was linked to better outcomes in one district.
- Student factors such as baseline writing performance, teacher caring, and demographics predicted writing-task completion and revisions.
- Teacher-level factors explained much of the variation in writing-task completion, highlighting the role of implementation context.
Why it matters
Applies implementation science to identify the active ingredients of an AI-based writing support tool and to show how contextual factors shape uptake and effectiveness.
Coauthors
Yue Huang, Corey Palermo, Joshua Wilson
Topic tags
External links