The impetus for this study was the need to gain a better understanding of what interaction activities in the virtual classroom affect student outcomes. The purpose was to determine which perceptions of interactions contributed to predicting student outcomes of satisfaction and future enrollment in Web-based courses, while controlling for student characteristics. The problem is that the interaction that occurs in the Web-based classroom is markedly different than what occurs in the traditional classroom setting. The study was a secondary analysis using data from 388 student evaluations of Web-based courses. Using Astin's Input-Environment-Outcome (I-E-O) conceptual framework, influences of student characteristics [inputs] and virtual classroom interactions [environment] on student outcomes were examined. Student input predictors were perceptions of computer skills; knowledge of electronic communications; number of Web-based courses taken; distance living from campus; and age. Environmental predictors included interactions with the instructor, students, technology, and perceptions of presence. Hierarchical, multiple regression analyses were performed to answer two research questions: 1.Do students' self-reported ratings of interaction help predict their satisfaction in a Web-based course, while controlling for student characteristics? 2.Do students' self-reported ratings of interaction help explain their willingness to take another Web-based course, while controlling for student characteristics? The most significant predictor of both student outcomes was students' perceptions regarding their interaction with their instructors. Second, satisfaction and enrollment were affected by students' perception of the technology as contributing to wasted time. Third, students who did not miss the face-to-face interactions as much tended to be more satisfied and were willing to take other online courses. Finally, information on distance living from campus helped in predict