Written composition has long been regarded as a cognitively challenging task. It simultaneously makes significant demands on language, memory, and thinking. For lengthy compositions, the writing process entails multiple work sessions extending over days, weeks, months, and even years. The diligence required can be challenging emotionally as well as cognitively. At the heart of meeting these demands is the capacity for self-regulation. Not surprisingly, then, the psychological research on executive functions has much to say about writing skill-both its successes and failures.

A model of how working memory, as conceived by Baddeley (1986), supports the planning of ideas, translating ideas into written sentences, and reviewing the ideas and text already produced was proposed by Kellogg (1996). A progress report based on research from the past 17 years shows strong support for the core assumption that planning, translating, and reviewing are all dependent on the central executive. Similarly, the translation of ideas into a sentence does in fact require also verbal working memory, but the claim that editing makes no demands on the phonological loop is tenuous. As predicted by the model, planning also engages the visuo-spatial sketchpad. However, it turns out to do so only in planning with concrete concepts that elicit mental imagery. Abstract concepts do not require visuo-spatial resources, a point not anticipated by the original model. Moreover, it is unclear the extent to which planning involves spatial as opposed to visual working memory. Contrary to Baddeley’s original model, these are now known to be independent stores of working memory; the specific role of the spatial store in writing is uncertain based on the existing literature. The implications of this body of research for the instruction of writing are considered in the final section of the paper.

Preparing a written outline during prewriting and composing a rough rather than a polished first draft are cognitive strategies that may lessen a writer's work load. The present laboratory and field research examined whether these strategies enhance writing performance. In an experiment, I manipulated the use of these strategies by college students in a letter writing task. The students' writing process, efficiency, and quality were examined. The results showed that preparing a written outline, compared with not doing so, increased the time spent translating ideas into text, improved the quality of letters, and failed to enhance overall efficiency. The use of rough versus polished drafts affected when the students reviewed their work, as expected, but had no influence on quality or efficiency. A survey of science and engineering faculty revealed that the frequency of using written outlines correlated positively with writing productivity, whereas use of polished drafts was uncorrelated with productivity.