To investigate differences in the writing processes of novice and expert writers, a test of writing skills was developed that required subjects to write a business letter in defense of a particular system of public transportation in response to a letter objecting to that system. In the first of three experiments, significant differences were found between the way ten expert writers (college graduates with extensive writing experience) and ten novices (first-semester freshmen with low verbal Scholastic Aptitude Test scores) generated ideas for their letters, generated outlines from given lists of ideas, and translated given outlines into text. In the second experiment, novices were asked to translate given outlines into letters, and the investigator studied the effects on their writing of completing a questionnaire that focused attention on the concerns of the reader of the letter. Results suggested that only 4 of 30 novice subjects addressed their reader's concerns in their letters and that the questionnaire had no effect on performance. In the third experiment it was found that two manipulations--removing the given outlines and increasing the specificity of the complaint letter--dramatically increased response to the audience. The results have implications for the training of writers and for the management of document generation within large organizations. (The paper includes materials used in the experiments and sample letters written by subjects.) (Author/GT)