Examining Engagement with Disinformation Accounts on Instagram Using Web Archives

Abstract

Disinformation poses a serious threat to society. Researchers have shown that many people spread disinformation, especially in relation to the COVID-19 pandemic, on social media platforms like Instagram. However, performing disinformation analysis is difficult if those users have been banned. Previous research has shown that web archives are useful for performing disinformation analysis on pages that are no longer live, but this approach is limited by web page capture availability. To better understand how disinformation is spreading on Instagram, we examine how disinformation actors are utilizing hashtags and account mentions to boost engagement by extracting engagement metrics from archived Instagram account pages. We use the data gathered from these archived webpages, or mementos, to perform network analysis showing how engagement connects users. We then perform clustering based on hashtag frequency to examine how people searching for reputable content are being exposed to disinformation by identifying groupings that contain both health authority accounts and anti-vax accounts. Our findings indicate that roughly one-fifth of intra-group hashtags are also distributed inter-group. Limited memento availability remains an obstacle to comparing reputable accounts to disinformation accounts, but a higher percentage of Instagram account page mementos from the Archive.today web archive are able to be scraped than those from the Internet Archive’s Wayback Machine.