This research was an exploratory study of which teachers, students, and teacher-student combinations benefit most from a particular aspect of computer-assisted instruction. The study is based on teacher and student evaluations of different teaching styles while using instructional software. Three teaching styles (monitoring, coordinating, and mediating) were used with each of three software packages in three fifth grade classes. The classroom teachers and 18 students, selected to represent differing preferred learning styles, were interviewed about the advantages and disadvantages of each style/software dyad. Contrary to expectations, the effectiveness of the three styles did not seem to depend on the primary instructional style of the teacher or on the preferred learning styles of the students. Rather, the results suggested that, for optimal effectiveness, all three styles should be used with every instructional software program. The mediating style provides a demonstration of the software, allows the teacher to highlight the important concepts, and shows the students what achievement is possible. Coordinated activities provide additional time and contexts for students to learn the concepts. The monitoring style allows students to work with the concepts at their own pace. Demonstrating these styles to teachers in their classes seemed very useful for encouraging them to use the styles themselves. (Contains 52 references.) (Author/KRN)