Finding and Categorizing COVID-19 Papers in CS Education

Abstract

In the first 2 years following the outbreak of COVID-19, many papers have been published regarding the impacts and adaptations of the pandemic on computer science education. As a first step towards a systematic literature mapping, this study attempts to develop a process for searching and a categorization schema for papers. The goal of this project is to produce a literature map which will be used to provide an initial assessment of the state of research, as well as a framework for future research directions. Limiting our search to papers published in the ACM Digital Library in the publications sponsored by SIGCSE, we first create and validate a query and inclusion/exclusion criteria for papers. Using a double evaluator model, we find high agreement with a Cohen's Kappa of 0.93, resulting in 42 papers across 6 conference proceedings. We further validate these findings by independent checking against all papers from SIGCSE2021 TS. We then develop categories across three dimensions: In activity: we find remote teaching, remote assessment, remote work, virtual events and general impact of pandemic. In measurement: we find grades, non-grade assessment, attendance/retention, affect/perception, and mental health. In population: we find K-12 students, university/college students, Educators, and the sub-categories of introductory/CS0/CS1 students, gender, and race. Double rater assessments initially produced a relatively low Kappa score of 0.58, but after protocol revision, and the production of additional categories, the kappa score was raised to a very high 0.94.