Answers are grounded in this paper and cite their sources.
The topic of influence has existed for millennia, as people have sought to convince, persuade, or coerce others into adopting a perspective or taking an action they would not have taken on their own. The internet has allowed influence practitioners to refine their craft in various ways, including through online social networks, online communities, and blogs. Typical studies of influence aim to identify the most influential community member but do not attempt to pinpoint those most susceptible to influence. In this work, we advocate exploring the distribution of both the probability of being influential and the probability of being susceptible to influence among members of text-based online communities. We employed an agent-based model of such communities to conduct our studies. We compared the distributions of the probability of being influential generated by our Influence Score algorithm with those produced by traditional Centrality and Link Analysis algorithms. Our findings revealed that our Influence Score algorithm produced a distribution that, at the 0.05 significance level, was like that produced by the Hubs Link Analysis algorithm. The other algorithms were shown, also at the 0.05 significance level, to be different. The distribution of the probability of being susceptible to influence showed very few members with probabilities matching those of the most susceptible, but indicated a greater likelihood that multiple members could be considered highly susceptible to influence. These results suggest areas for future research and present ideas for leaders or organizers of text-based online communities to consider as they contemplate how to protect their communities from influence schemes.
No takes yet. Share an insight, caveat, or question.
Pinto-Coelho et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: