Social media companies make important decisions about what counts as 'problematic' content and how they will remove it. Some choose to moderate hashtags, blocking the results for certain tag searches and issuing public service announcements (PSAs) when users search for troubling terms. The hashtag has thus become an indicator of where problematic content can be found, but this has produced limited understandings of how such content actually circulates.Using pro-eating disorder (pro-ED) communities as a case study, this paper explores the practices of circumventing hashtag moderation in online pro-ED communities. It shows how: (1) untagged pro-ED content can be found without using the hashtag as a search mechanism, (2) users are evading hashtag and other forms of platform policing, devising signals to identify themselves as 'pro-ED', and (3) platforms' recommendation systems recirculate pro-ED content, revealing the limitations of hashtag logics in social media content moderation.

Expanding the debate about content moderation: scholarly research agendas for the coming policy debates. Internet Policy Review, 9(4).

This article proposes ‘sexist assemblages’ as a way of understanding how the human and mechanical elements that make up social media content moderation assemble to perpetuate normative gender roles, particularly white femininities, and to police content related to women and their bodies. It investigates sexist assemblages through three of many potential elements: (1) the normatively gendered content presented to users through in-platform keyword and hashtag searches; (2) social media platforms’ community guidelines, which lay out platforms’ codes of conduct and reveal biases and subjectivities and (3) the over-simplification of gender identities that is necessary to algorithmically recommend content to users as they move through platforms. By the time the reader finds this article, the elements of the assemblages we identify might have shifted, but we hope the framework remains useful for those aiming to understand the relationship between content moderation and long-standing forms of inequality.

No abstract

Heavily used hashtags on Instagram and other platforms can indicate extensive public engagement with issues, events or collective experiences. This article extends existing research methods to paint a fuller picture of how people engage collectively with public issues online. Focussing on Instagram content often deemed ‘problematic’, we develop and test what we call a ‘hashtag practice’ approach. This approach targets the hashtag #depressed, and also moves beyond it to (a) incorporate the posts immediately preceding and following a root post, (b) more inclusively sample content associated with the hashtag to combat filtering bias, (c) consider collocated hashtags and (d) draw on contextual cues in the interplay between posts’ visual content, captions and profile management. The method shows the prevalence and significance of aesthetic and memetic practices, and caution in embodiment in mental health posts, revealing more diverse forms of engagement with mental health on Instagram than previous research suggests.