A look at student performance during the COVID-19 pandemic.

Purpose: This study aims to use self-reported publicly available student assessment data from the time period when there was an abrupt change in instructional method at the start of COVID-19 to assess potential for differences as a result of course delivery mode. Design/methodology/approach: A general linear model using 837 student evaluations from 191 US public higher education institutions investigates the impact on student performance and how performance was related to a number of covariates, namely, online experience of instructor prior to shut downs, discipline of study and size of institution. Findings: The analysis finds an overall grade point average (GPA) increase of 0.10 (out of 4.0) associated with the shift away from face-to-face instruction. In exploring potential covariates, only institutional size was significant in explaining this increase in GPA. This supports the notion that despite hardships inherent with the abrupt switch to online education across the country that student grades as a whole did not suffer. Research limitations/implications: The source of data was self-reported. In addition, GPA is an imperfect measure of student learning. Despite this, because GPA is highly correlated with student satisfaction, retention and matriculation, it is relevant. Practical implications: This study suggests that the rapid transition to online instruction did not negatively impact student performance and may have marginally increased these marks. These findings were cross-disciplinary and not influenced by the instructor's prior online teaching experience. These findings support the idea that institutions and instructors should be more willing to use a variety of delivery modes going forward. Originality/value: The data set used is uniquely large and varied in the number of institutions, professors, students and discipline. The COVID-19-induced transition from largely in-class instruction to mostly online or remote instruction allowed for a natural experiment that eliminates the sample selection problem associated with most other instructional method comparison studies. [ABSTRACT FROM AUTHOR]