Exploring Dunbar's Number Through COVID-19: Analyzing Daily Testing and Case Dynamics in Dunbar-Sized Groups and Their Impact on Twitter Sentiment via Granger Causality

The COVID-19 pandemic enables the exploration of traditional psycho-sociological concepts, such as Dunbar's number, which suggests a limit of about 150 people with whom an individual can maintain stable social relationships, although estimates vary. In this study, we provide a new estimate of Dunbar's number by analyzing COVID-19 datasets from 2020 - before vaccines were available - focusing on the dynamics of daily testing and positive cases. We investigated sentiment changes in time using a large corpus of English-written, COVID-19-related tweets from Twitter. Our approach involved calculating the sizes of virtual population groups that theoretically include just one individual who was tested or tested positive. In the first year of COVID-19, many individuals did not fully believe in the existence or serious impacts of the disease - thus, knowing at least one positive case could influence them to accept the severe reality and, potentially, express their sentiment on Twitter. The knowledge of at least one such individual within one's Dunbar circle could lead to more negative sentiment in tweets. The larger Dunbar's number is, i.e., the more people an individual could know well, the easier it is to know at least one new tested or positive case. Employing Granger causality, we compared the sizes of groups with at least one tested individual against various hypothetical values of Dunbar's number to estimate the minimum number of connections necessary within such groups. Our findings suggest that Dunbar's number could be about 175, which aligns closely with the traditionally cited figure of 150.