Association for Public Policy Analysis & Management

MOOCs and online degree programs have been heralded as having the potential to transform higher education, particularly in STEM. They have promise of increased access to disadvantaged students as well as lifelong learning, among other benefits. To date, most MOOCs have yielded disappointing results because of extremely high attrition rates, enrollment by those who already have degrees, and highly visible failures of attempts to replace on-campus courses. Online degree programs have also not proven to have substantial labor market returns. 

Many of the courses and programs studied in previous research have been free (or low-cost) and/or of unclear quality. We study the first fully online version of a prestigious STEM degree. In particular, we look at Georgia Tech's new Online M.S. in Computer Science (OMSCS). GA Tech’s computer science department, one of the top 10 in the nation, announced in 2013 that it would offer a fully online version of its prestigious master’s degree for one-seventh of cost that out-of-state students pay for its on-campus degree ($7,000 vs. $45,000). The program is open to all students who hold a relevant B.A. and have an undergraduate GPA of 3.0 or higher. The program has garnered extensive media attention because it’s the first offered by a high quality institution at low cost and resulting in a degree that won’t say “online”. As such, the department is staking its reputation on making the educational experience a serious one. 

We answer two questions with this research. First, where and from whom does demand for high quality online STEM education come from? To do so, we collect data on and compare the characteristics of all applicants to OMSCS to all applicants to GA Tech's traditional on-campus master's degree program. We are particularly interested in whether the online program increases access to any particular groups of students. Second, we ask what the educational impacts of admission to OMSCS are? In other words, does this online program simply substitute for other educational experiences or is it sufficiently unique that failure to be admitted implies students will not pursue alternative formal educational options? To answer this question, we apply a regression discontinuity design exploiting a threshold in the admissions process to data from the National Student Clearinghouse that observes whether applicants ultimately enroll anywhere in the U.S.

Our data suggest that, in its first year of operation, OMSCS enrolls three times as many students as the largest on-campus computer science master's degree program in the country. Initial demand for this product has thus been quite high. Only 10% of applicants to the on-campus program are American, with nearly all of the remainder from India or China. In contrast, nearly 80% of the applicants to OMSCS are Americans. On-campus applicants are nearly all in their early to mid 20s. OMSCS applicants range from their 20s to their 50s, with an average age of 35.