A study examined the extent to which computers affect the writing process. In order to test several hypotheses about writing with a computer 15 experienced writers wrote persuasive letters under four conditions--using pen and paper once and a computer the other three times. All of the subjects were experienced with computers having worked daily or almost daily with the machines on which they were tested. In two of the computer conditions, writers used "Andrew," an advanced computing system and related software. The other advanced work station differed only in the size of its screen. The quantitative measures employed were (1) time spent composing the letter, (2) total words produced, and (3) number of words produced per minute. The results indicated that time to compose was longest in the advanced work station and shortest in the pen and paper condition, while the highest number of words produced was in the advanced work station, with the lowest in the pen and paper. There was no significant difference between conditions on the words per minute measure. The letters also received scores for content and mechanics from two readers and from a third reader when scores were more than one quartile different. Analysis showed that the texts produced in the advanced work station were significantly better. Analysis of revision protocols for 8 of the 15 subjects indicated more planning for the pen and paper copies. The study suggests that if computers are to be used effectively in the writing classroom, students must learn not only word processing, but awareness and adaptation of their writing processes. (Tables and five references are included.) (AEW)