In 1997 we first offered American government classes online as well as face-to-face classes. We administered pre- and posttests to our students to measure their general knowledge of American government, political attitudes, demographics, and some behaviors. Following an initial report in 2001, we continued to gather data for 10 more years; this current study covers nearly 3,200 students during 13 years. We examine the sample as a whole and changes in audiences and outcomes, over time, for the two teaching formats. Although the kinds of students taking online classes have become more similar, a few differences persist. Learning outcome differences continue to be insignificant. Neither format has a clear advantage in students' changes in attitudes, but the online classes increased students' newspaper reading. Class dropout rate and faculty workload both favor face-to-face classes, but flexibility in scheduling and student demand clearly favor online classes.