Universities and even high schools are moving more and more to online instruction as a cost-effective way to reach more students with fewer resources. This naturally raises the question: Can online learning be effective? (The question is not "Is online learning effective?" because just like face-to-face instruction, online instruction includes a diverse array of techniques.) In this paper I compare online and flipped face-to-face versions of an introductory astronomy course. Both versions were designed around student-centered learning principles, but the specific implementation of these principles varied according to the strengths of each type of instruction. Normalized Hake gains on the Star Properties Concept Inventory (SPCI) were quite similar for both classes: 56% and 58% for the online and flipped face-to-face versions, respectively. The gains obtained by students with low pre-test scores were as good as the ones achieved by students with high pre-test scores.