by Kelly N. Tracy, PhD
Western Carolina University
October 21, 2013

Graham, S., and Hebert, M. A. (2010). Writing to read: Evidence for how writing can improve reading. A Carnegie Corporation Time to Act Report. Washington, DC: Alliance for Excellent Education.

“The evidence is clear: writing can be a vehicle for improving reading” (p. 6). 

Ten years ago The National Commission on Writing in America’s Schools and Colleges deemed writing the “neglected ‘R’” and called for a “writing revolution” that included doubling the amount of time students spend writing. In the years following, extensive reports such as Reading Next (Biancarosa & Snow, 2006) and Writing Next (Graham & Perin, 2007) supported the idea that writing is a powerful tool for improving reading, thinking, and learning.  Now as much of the country implements the Common Core State Standards, there is a renewed push for more and better writing. As educators try to determine how to improve student learning and include more writing within the same time limits, it is important to revisit Steve Graham and Michael Hebert’s (2010) Writing to Read, which gives strong evidence that writing, an essential skill itself, also improves reading comprehension. 

For decades researchers have emphasized the strong connection between reading and writing, both in theory and in practice. Multiple studies have demonstrated that writing can improve comprehension. What has been less clear is what particular writing practices research supports as being effective at improving students’ reading. To determine those practices, Graham and Hebert (2010) undertook an in-depth meta-analysis of experimental and quasi-experimental studies that examined the effectiveness of writing practices on improving students’ reading in grades 1 -12. They acknowledge the limitations of excluding other forms of research and recognize the significant contributions of that research; at the same time, they share that completing this sort of meta-analysis allowed them to focus on studies where cause-and-effect could be inferred and effect sizes calculated.  Their meta-analysis generated three recommendations:

1. Have students write about the texts they read. “Writing about a text proved to be better than just reading it, reading and rereading it, reading and studying it, reading and discussing it, and receiving reading instruction” (p. 14). Specific types of writing about reading that had statistically significant effect sizes included responding to a text through writing personal reactions or analyses/interpretations of the text, writing summaries of a text, taking notes on a text, and creating and/or answering questions about a text in writing. The benefits of these types of writing were stronger, particularly for lower-achieving students, when they were tied with explicit instruction on how to write.
2. Teach students the writing skills and processes that go into creating texts. Teaching students about writing process, text structures, paragraph or sentence construction, and other writing skills improves reading comprehension; teaching spelling and sentence construction skills improve fluency; and teaching spelling skills improves word reading skills.
3. Increase how much students write. An increase in how often students write improves students’ reading comprehension. Graham and Hebert recommend more writing across the curriculum, as well as at home to achieve more time spent writing.

What may be most important in all of Graham and Hebert’s findings is that infrequent writing and lack of explicit writing instruction minimize any sort of effect on reading from the writing practices they recommend. Their report also supports earlier calls for emphasizing writing in the classroom and across content areas. Writing is a critical skill, important in its own right; given the evidence that consistent writing time and instruction not only improves writing but also reading, gives us an even more compelling case for finding time in our school day for more writing.

Reader response is welcomed. Email your comments to LRP@/