Effects of a period of planning prior to writing (''pre-writing'') and of the writing medium on the content and quality of writing were examined in two between-subjects experiments. In Experiment 1, subjects wrote short narratives either immediately after instruction or following a 5-minute pre-writing period; half of each group used pen and paper; the other half used a word processor. Both pre-writing and word processing resulted in more changes to text content, resulting in compositions that were judged as more sophisticated and creative. Experiment 2 demonstrated that pre-writing using a written outline decreases content revisions compared to non-written pre-writing. Verbal protocols indicated that non-written pre-writing was focused on content generation, whereas outlines were associated with more conceptual planning. To account for these data we propose several additions to the Hayes and Flower (1980) cognitive model of the writing process. Both pre-writing and word processing are viewed as expanding the alternatives to be considered, i.e. ''option expansion''. During writing the options are resolved during pauses and through changes to text content, i.e. ''option resolution''.