This paper describes a study designed to investigate the effectiveness of narrative versus step-by-step instructions for a computer task. The participants in this study were 31 undergraduate education students enrolled in a computer literacy class at the University of Memphis during the Summer 1996 semester; none of the participants had prior knowledge of e-mail. Participants were randomly assigned to one of the two treatment groups and were individually observed as they completed the steps for retrieving, replying, spell checking, and sending e-mail using software on a mainframe computer. After completion of their respective treatments, participants completed a survey designed to assess their attitudes toward the instruction. There was no performance time difference between the two groups, but the step-by-step treatment made fewer errors during the more complex tasks. There was a lack of difference in attitude between the two groups. A figure illustrating the performance rating form used by the researchers is included, as well as tables containing data on errors on task and errors by treatment. (DLS)