The final speaker in this morning session at the ECREA 2026 conference in Brno is Eva Mayerhöffer, whose focus is on the role of digital trace data in understanding exposure to alternative information. One point of departure for this is often to start with a list of alternative and problematic news outlets which are then identified in survey responses, data donations, or digital trace data; the focus here is then also often only on the top domains, i.e. the most frequently accessed sources.
But such a reduction to the most notorious sites substantially underestimates exposure, since there is a very long tail of nice outlets, actors, influencers, and other sources that users also encounter; these may be ephemeral and little known, but cumulatively can still account for a great deal of exposure. There’s a need to further examine and understand this long tail. Here, too, though, it’s still researchers rather than users who define what is considered to be ‘alternative’.
This particular project worked with browser history data donations from some 130 Danish adults (out of some 2,600 invitees who were invited to donate their data). These datasets amounted to some 2.6 million visits to nearly 17,000 domains; the content of those pages was then scraped and classified paragraph by paragraph by a zero-shot classifier to identify whether the text positioned itself against the societal or political mainstream.
In any of this, though, news is a very small part of the overall set of sites visited; commercial sites, search engines, social media, Webmail services, and others dominate. Mainstream news media account for 8.4%, alternative information for 0.6%, and known hyperpartisan and alternative content for 0.4%. Across all site types, the classifier found a substantially greater likelihood of encountering alternative information in the long tail than in the most mainstream domains; this also means that focussing only on the most prominent sites misses out on the majority of alternative information encountered by users.
Further, alternative content encounters were very unevenly distributed: 105 of the 134 data donors encountered no alternative information at all, while a small subset of 11 donors reported virtually all (97%) of the encounters with alternative information. We must therefore also not overestimate the use of alternative news: a small number of users are highly active, and for them this long tail is certainly relevant, but for many others it is not at all.
