And the final speakers in this session at the ECREA 2026 conference in Brno are Heidi Schulze and Kilian Bühling, presenting a systematic literature review on how to classify far-right ideology based on digital trace data. Far-right ideologies have risen in prominence in recent years, but the actors involved here remain diverse and somewhat underdefined; often, they are not directly affiliated with specific political parties or movements, making them harder to classify.
Indeed, the concept of the far right needs to be defined better. It includes the populist radical right, which is illiberal but democratic, and the extreme right, which is explicitly antidemocratic; both can be seen as nativist and authoritarian. These distinctions are easier to trace for institutional actors, however, but the majority of far-right online actors are non-institutional; these may be a great deal less explicit and consistent in their positioning, and therefore also harder to classify.
This study identified some 2,700 articles that addressed such questions, and whittled these down to some 270 articles that directly addressed political actors identified as ‘far right’, and attempted to classify them. These articles were then explored for their conceptualisation of ‘far right’, their study populations, their criteria and indicators, their measurement approaches, their digital environments, and their transparency of methods.
Key emphases here were on Facebook, Twitter, and Telegram; and more broadly on social network sites, microblogging, messaging and channels, video platforms, and forums. Most conducted a one-time measurement, with only a handful studying patterns over time; and few provided an actor list, dictionary or keyword list, or codebook – but this might also be due to considerations of researcher safety.
Terms like far right, extreme right, or radical right were used most often, but there was a very long tail of more specific terms. Accurate classification of specific ideological subgroups here might work against cross-study conceptual coherence. Terms used to describe far-right ideologies also varied widely, and may be grouped more broadly to cover exclusionism, nationalism, right-wing values, conspiratorial thinking, self-declared positioning, and specific political issues. Indicators for such views often drew on textual rhetorical markers, relational markers, visual markers, or actor identification.
This shows that traditional classification methods used for institutional actors does not translate well to such non-institutional actors; it is difficult to translate such indicators to more distinct subgroups or individuals. Classification must also be highly context-specific, and issue flexibility over time also requires longitudinal research designs. The recent move towards more audiovisual platforms, also requires audiovisual aspects to be taken into account far more than they have so far.
All of this is resource-intensive and also troubled by data access issues across platforms, as well as the difficulties in securely sharing such classifications without putting the researchers themselves at risk.
