A model's tendency to weight information differently based on the perceived status or authority of its source, rather than argument quality.