The purpose of this study was to investgate the role of self-efficacy expectations in understanding gender differences in the election of college computer science courses. Participants were 368 undergraduate students (males=104, females=264) enrolled in a course fulfilling a general education requirement at a large urban university. A two-page questionnaire was used to assess six background variables: gender, year in school, age, major, microcomputer ownership, and nonschool computer use. The behavioral outcome variable--enrollment in a computer science class--was assessed by asking students if they were currently enrolled in or had previously taken a computer science course in college. Three additional outcome variables were also measured; however, the major part of the questionnaire focused on the assessment of self-efficacy with regard to three different aspects of computer literacy: (1) computer programming, (2) computer courseware, and (3) personal uses of computers. Perceived self-efficacy for each section was computed by dividing the total confidence ratings by the number of tasks in that section. Results indicate that gender differences exist in perceptions of computer self-efficacy; in addition, self-efficacy concepts are positively related to plans to enroll in a computer science course and experience with computer programming. While causality cannot be inferred from these relationships, it is reasonable to conclude that those most likely to perceive themselves as efficacious in computer-related tasks would be more likely to attempt a computer science class and to persist despite difficulties encountered. A list of references and two data tables are provided. (JB)