A deeper look at the 2020 Facebook boycott and related themes of misinformation: A text mining analysis of topics, emotion and sentiment.

Facebook has repeatedly come under fire from consumers, companies and government agencies in recent years owing to the prevalence of misinformation on its platform as well as the way it handles information related to social justice and public health, among other things. Using a large data set of 604,269 social media mentions sourced from popular social platforms (eg Twitter and Reddit), this study set out to discover themes associated with misinformation and the Facebook advertising boycott that occurred in July 2020. To understand the discourse, a linguistic analysis approach called theme extraction was used. This method employs machine learning and natural language processing to reveal relationships in the data that may otherwise be buried in the mass of social media mentions. The most prominent theme identified among social media users was the desire that Facebook and other social media platforms actively stop the spread of misinformation. Other trending topics included #StopHateForProfit, lockdown protests, hate speech policy, news as spam, right-wing politics and suspended ads. Managerial implications for advertisers are discussed as they relate to social media management and how misinformation impacts brand engagement on social platforms. [ABSTRACT FROM AUTHOR]