Comparing Online and In-Person: Guidelines for Creating Better Outcomes Online

Concurrent Session 6

Brief Abstract

In Spring 2017, Georgia Tech launched an online for-credit version of its CS1 class. In Spring 2018, students in this online version achieved better learning outcomes than students in the traditional version. In this session, we present the experiment and the design of the course that led to these outcomes.


David Joyner is the Associate Director for Student Experience in Georgia Tech's College of Computing, overseeing the administration of the college's online Master of Science in Computer Science program as well as its new online undergraduate offerings. He has developed and teaches CS6460: Educational Technology, CS6750: Human-Computer Interaction, and CS1301: Introduction to Computing, all online.

Additional Authors

Extended Abstract

Socially, online education is often viewed as an inferior product to traditional alternatives. The assumption is that the advantages of online learning may be in the access, flexibility, or cost, but that the trade-off is for an inferior learning experience. Significant research supports these observations: learning outcomes from online classes have often, though not always, been observed to be inferior to traditional classes.

In Spring 2017, Georgia Tech launched an online version of its CS1 class for on-campus students. The course was designed to target a course whose size could often not meet demand, as well as the demand from students for more flexible class options that would allow them to enroll during internships, athletic commitments, or already busy schedules. Knowing the common conception of online learning, though, the university invested significant effort into designing a course from the ground up that specifically takes advantage of the unique opportunities of the online interface. Additionally, knowing the prior outcomes from other studies, the university compares students in the online section to students in the traditional section every semester using a mixture of surveys, pre-test measures, and post-test measures to ensure that online students are not receiving an inferior product.

After equal outcomes were observed the first two semesters, in Spring 2018 a difference between the sections was observed: with statistical significance, students in the online section outperformed students in the traditional section. Although students in both sections started the semester with comparable pre-test scores, students in the online section ended the semester with higher scores on the shared post-test. In addition, as with previous semesters, students rated the online experience as exceptional. This new semester also delved more deeply into certain aspects of the student experience: among other things, it found that students in the online section reported spending less time on the course (~7 hours per week, compared to ~9 hours per week in the traditional section), but they spent a greater percentage of their time on active learning activities. While students in the traditional version of the course reported spending 40% of their time on active problems and exercises, students in the only course reported spending 75% of their time on such activities.

In this presentation, we will present the design of the online course, and specifically how it manages to encourage students to spend a greater portion of their time on active learning activities. A common perception of online learning has been that it is less active, and yet with the design guidelines we have developed, the online experience can absolutely be more active.

Of particular note is that this online course is also offered as a MOOC (massive open online course), which means that it also scales to a massive online audience. Over 100,000 students have registered for the same course experience on the MOOC provider edX, so not only can this active learning approach be used to create superior outcomes for for-credit students, but it can also be scaled to a massive audience.