Qualification: Influence via Ethos: On the Persuasive Power of Reputation in Deliberation Online
Manzoor, Chen, Lee, and Smith provide empirical reason to qualify the design guarantee. In a seven-year panel of more than one million online debates, they isolate a causal reputation effect after controlling for unstructured argument text: reputation increased persuasion beyond the validity, strength, and presentation of the arguments, and ten additional reputation points raised persuasion probability by 31% over the platform average. They characterize reputation as a low-effort heuristic proxy for argument quality and warn that visible participant characteristics can distort consensus toward reputed participants. This supports making reasons inspectable, but shows that reason visibility alone does not prevent standing from functioning as authority; an implementation must also constrain, segregate, or counteract reputation cues during merits evaluation. The evidence is from human online debate, not Kaal's agent validation pools, so it identifies a concrete failure mode rather than validating the proposed architecture.
institutional-designconsensus-and-securityreputationscholarly-growth-coveragescholarly-literatureonline-deliberationreputation-biasreason-giving