Recommendation networks among educational video channels on YouTube
Cynthia Pasquel-López, Lucía Rodríguez-Aceves, Gabriel Valerio-Ureña
Frontiers in Education
Network design
Nodes
412 YouTube channels classified within the educational-video or EduTube ecosystem
Ties
directed channel-to-channel recommendations visible on platform pages during the October 2021 collection
Methods
The study used snowball discovery from 37 entry channels, described degree and reciprocity, coded channel attributes, and applied QAP correlations and MRQAP models to account for relational dependence.
Source: Frontiers in Education
Reviewed summary
Researchers snowballed from 37 educational YouTube channels to a directed recommendation network of 412 channels and 1,303 links collected in October 2021. Degree, reciprocity, QAP, and MRQAP analyses linked outgoing recommendations with digital engagement and found homophily in reciprocal channel ties. The article extends educational SNA beyond classrooms and helps researchers examine how platform recommendations organize discoverability and potential collaboration among learning-content producers.
Channels with more outgoing recommendations tended to show greater measured digital engagement, and similar channels were more likely to reciprocate, although recommendation direction and platform mechanisms complicate interpretation. The network depends on 37 seeds, one collection month, platform recommendations, channel classification, and third-party engagement measures. Algorithmic personalization and deleted or hidden links are not fully observed, so results are cross-sectional associations rather than creator influence or causal platform effects.
SNA
How SNA was used
The study defined 412 YouTube channels classified within the educational-video or EduTube ecosystem as nodes and directed channel-to-channel recommendations visible on platform pages during the October 2021 collection as ties. The study used snowball discovery from 37 entry channels, described degree and reciprocity, coded channel attributes, and applied QAP correlations and MRQAP models to account for relational dependence.
Source
Frontiers in Education, 7, Article 845647
DOI: 10.3389/feduc.2022.845647
