### Explanation Snowball sampling is a strategic methodology employed in network analysis, particularly effective for investigating populations that are otherwise difficult to observe directly. This method is useful in the context of social media platforms where channels and chats serve as nodes or websites with cross-references. Nodes could be interconnected through forwards, mentions, and internal hyperlinks, which function as the network edges. This technique is useful because it helps identify narratives and map out the structure of a network of connected accounts, even if such capability is not initially available technically. Process begins with a selected initial sample (or 'seed') and expands through multiple steps, identifying relevant actors within this network through links, post forwards, mentions and so on. The seed is crucial as it sets the direction and scope of the sampling. However, the choice of the seed can introduce biases, influencing the resulting sample and network representation. ### Input - posts or accounts ### Output - corpus (graph) of related posts and accounts ### Examples - [Reverse-engineering the work of Telegram ‘political technologists’: Mapping entities involved in building narratives surrounding Russian President on Polish Telegram](https://wiki.digitalmethods.net/Dmi/SummerSchool2023ReverseEngineeringTelegramNarratives) ### Tools - [Telegram Snowball Sampling](https://github.com/thomasjjj/Telegram-Snowball-Sampling) - [Issuecrawler](http://www.govcom.org/Issuecrawler_instructions.htm) ### Types - behavioural ### Weakness - [[SOWEL-7. Copying Content]] ### Functionality - {{related functionality}}