| Last Updated:
Creating DOI. Please wait...
Nicole Barbaro, WGU Labs
Meredith Riggs, Carnegie Mellon University
Jason Levin, WGU Labs
Carolyn Rosé, Carnegie Mellon University
The COVID19 pandemic has caused unprecedented disruption to higher education and student learning. However, disentangling the impact of the pandemic on academic outcomes at traditional higher education institutions is made difficult by the fact that academic outcomes were affected by both the pandemic itself and by the abrupt, often rocky, online transition. Unique features of WGU’s education model -- fully online, program mentors (i.e., a faculty member that supports a student through the entirety of their time as a student), and new terms beginning each month -- afford the opportunity to evaluate the impact of the pandemic on student academic outcomes independent of the transition to online learning.
Leveraging WGU’s 100,000+ student population and detailed institutional data, including behavioral click-stream data in the learning portal and qualitative meeting notes from program mentors, we use topic modeling and statistical analysis to (1) evaluate the impact of the pandemic across WGU’s diverse student population, and (2) evaluate the important role of social support via student’s program mentors on student outcomes during this crisis. Through analysis of a 20-topic topic model applied to notes written by mentors about their interactions with students, we were able to separate students with evidence of COVID-19 related issues (e.g., having COVID-19, working on the front lines, increased child care issues) from those without (with 13% of students fitting into the HighCOVID category). We assessed the impact of evidence of COVID issues on number of course units earned for the Spring 2020 semester. We also separated students with above median number of interactions with a course mentor from those with lower than median number of interactions with a course mentor (HighSupport vs LowSupport). HighSupport students earned significantly more course units than LowSupport students (effect size = .17 s.d.). HighCOVID students also received more support on average than LowCOVID students. An interaction analysis showed that for HighSupport students there was no effect of COVID issues, and the same was true for LowSupport students, thus we have evidence for an important effect of social support regardless of COVID issues. These findings highlight the importance of social support for higher education institutions that are striving to deliver rigorous online education at scale during a crisis.