|Center on English Learning and Achievement|
|How does writing affect learning? A review of the
Robert L. Bangert-Drowns, Marlene M. Hurley, & Barbara Wilkinson
Three decades ago, a team of British educators championed the idea that writing could enhance academic learning. The process of writing, they argued, closely resembles processes of speaking, thinking, and learning. Writing also leaves a "residue," a document that can serve as a tool for reflection, discussion, and revision. Writing might facilitate learning if its processes and products were thoughtful, expressive, and integrated in classroom discourse.
Over the last twenty years, researchers have tested the efficacy of writing-to-learn in numerous studies, but the results of these studies have been ambiguous. Though many showed improved academic achievement from writing, others reported detrimental effects.
Our meta-analysis* of school-based writing-to-learn programs shows that writing can have a positive impact on achievement. There are several reasons why the effect of writing on learning may vary. The frequency, nature, and social context of the writing tasks might influence writing’s effects. Writing also takes time from other learning activities that are more or less productive. Students with low confidence, interest, or skill in writing might find additional writing tasks distracting and burdensome. And writing tasks might be well or poorly aligned with classroom assessments.
Using statistical procedures to analyze the previous research on writing-to-learn, we sought to identify conditions that might best enhance the learning effects of writing. We identified 45 studies that compared normal classroom instruction to writing-intensive instruction on the same content. These studies had been conducted in elementary grades through college and in all sorts of subjects. We coded each study on fifteen variables representing its publication history, the quality of its research design, the context of the learning activities, the intensity of the writing treatment, and the types of writing tasks required of students. We translated the study outcomes into a common metric - the effect size (the standardized difference between the mean achievement scores of writing-to-learn and conventionally taught students). We looked for relations between study features and study outcomes measured as effect sizes.
In three-fourths of the studies, writers outperformed conventional students, but the typical improvement was a small one. In twenty-four of the studies, students completed writing assignments in class, so researchers could record the time spent on the writing tasks. What appears to matter more than the amount of time given to an assignment is the nature of the writing task, the kind of thinking that gets done. One factor reliably enhanced the effect of writing-to-learn: When writing prompts urged students to reflect on their learning processes - the challenges they faced and the strategies they employed - the educative effects of writing were substantially improved.
In general, these studies and other research suggest that writing can benefit learning, not so much because it allows personal expression about subject matter as because it scaffolds metacognitive reflection on learning processes. And the cost need not be great: even relatively brief tasks can boost learning. Additional research and classroom investigation should further clarify how writing benefits learning.
*A meta-analysis is a systematic review of previous studies on a topic. It involves four phases: identification of studies to include in the review, coding of study features, calculation of effect sizes, and statistical analysis of effect sizes to examine the relations between study features and study effects.
For more information about this study, contact Robert Bangert-Drowns at