The two-step flow theory, formulated by Katz and Lazarsfeld, described a communication system in which the media disseminated information and opinion leaders filtered, interpreted, and transmitted it to the public. Today, despite digital transformation, this logic still holds – not only through influencers but also in the comment sections of political news, where a new functional division between more active commentators and more passive readers can be observed. This is the subject of a paper now published: Han, J. (2026). Commenters and lurkers: Navigating the two-step flow of communication in online news discourse. New Media & Society, 28(9), 4566-4588. https://doi.org/10.1177/14614448251363954.
In digital environments, opinion-forming processes affect these two groups differently. Active users, who seek to influence others and interact with journalists and political actors, function somewhat like modern opinion leaders. Motivated to express their positions and influence debates, they are particularly sensitive to news bias that signals their political orientation. They consistently internalise these cues (partisan cue-taking) and remain largely indifferent to the content of other users’ comments. For them, reading and writing comments are interconnected activities that deepen engagement with news content but do not alter their positions.
In contrast, the lurkers – passive readers – mainly rely on the bias present in the comments. They infer the ‘climate’ of public opinion from the dominant political orientation in the comment sections and align their positions with what they perceive as the majority view. This process is explained by the bandwagon heuristic, which leads lurkers to interpret comments as indicators of broader public sentiment.
The results reveal a cascade process: news bias influences commentators, who reinterpret and disseminate these partisan cues; subsequently, lurkers shape their perceptions based on the comments these commentators produce. Thus, public opinion flows from a small group of active participants to a larger, passive audience, reinforcing the relevance of the two-step flow model.
This dynamic has ambivalent democratic implications. On the one hand, the cascade effect can promote civic engagement by exposing more users to political debate and encouraging participation. On the other hand, it reveals how vulnerable online discussions are to coordinated manipulation or visibility moderation, which functions as a new form of censorship. Real cases illustrate this risk: South Korean pro-government teams strategically posted favourable comments to influence the political environment; official Chinese propaganda flooded platforms with pro-government content to redirect attention and suppress dissent; and, for example, Russian troll campaigns during the 2016 US elections spread polarising narratives on decentralised platforms.
In light of these results, news portals and journalistic media should review their comment-curation algorithms. Rather than highlighting the most recent or popular comments, they should prioritise a diversity of political perspectives, mitigating polarising effects and offering lurkers a more balanced view of public debate. The New York Times model, which rigorously filters content and malicious messages, highlighting only high-quality comments in the NYT Pick section, is a promising example – although intensive moderation may limit the spontaneous expression of diverse ‘voices’.
In summary, this study highlights the multidimensional nature of public opinion formation in digitalised environments. Commentators and lurkers play distinct yet complementary roles: the former reinforce partisan cues and act as opinion leaders, while the latter rely on the perceived majority sentiment in the comments. This dynamic reaffirms the relevance of the two-step flow in the algorithmic era and emphasises both the potential of comment sections and their vulnerability to manipulation. Platforms that adopt balanced and inclusive moderation systems could transform these spaces into more resilient arenas for civic participation.