“Don’t research us”—How Mastodon instance rules connect to research ethics

Abstract Mastodon is a decentralized social network that has recently gained in popularity due to the platform changes of Twitter (now X). When it comes to collecting and analyzing data, the technical structure of such a decentralized network not only has methodological but also ethical implications. Mastodon consists of a large number of individual instances (around 17k), and each of these instances typically have their own set of rules, which may also address the use of data. Against this backdrop, we investigated whether and how Mastodon instances address the scientific use of data. Our analyses focused on active instances with English-language rules. Based on a combination of quantitative and qualitative content analysis, our results show that only a small portion of instances mention the scientific use of their data. Instead, the majority of rules rather focuses on user behavior and interactions. Based on the results, we formulate recommendations for researchers who want to work with Mastodon data. The recommendations are informed by the results of our empirical study and guided by general ethical principles for the examination of data from social media.