Topic modeling of the student emails sent before and during the birth of COVID-19 in physics and math classes

Abstract

The COVID-19 pandemic caused physical classes to suddenly transition to online learning all over the globe two years ago, resulting in students becoming more active in online email communication. The emails sent by the students were observed to contain students' concerns and needs for teacher support during the early stages of worldwide online classes due to COVID-19. As such, this study was interested in those email contents that were explored and analyzed through topic modeling, network analysis, and grounded theory. Six hundred twenty-three emails sent by seventy students in physics and math classes were analyzed using InfraNodus. This tool can perform topic modeling and visualize network graphs of verbal text data such as emails. By topic modeling and network graphical analysis, the findings revealed that the main topic clusters of the student emails corpus pertain to class assessments – questions and tests. Moreover, the influential keywords in the network graphs were coded, and the emails representing those keywords were further categorized using grounded theory. Doing so led to the finding that students needed teacher support on the content and supportive pedagogy. Supportive pedagogy needs may include test goals, schedule, content, and procedures, reviewing the test solutions and answers, and providing necessary test accommodations. Further study on teacher support in the online physics class and the effect of delivering teacher support on the student’s performance can be a topic of future research.