Statistical or anecdotal? How digital human doctors' message evidence influences the effectiveness of online health science popularization

Purpose This study aims to examine how statistical (vs. anecdotal) evidence used by digital human doctors influences consumers' information adoption in online health science popularization. It further explores the mediating role of perceived deceptiveness and the moderating role of digital human type. Design/methodology/approach Three experiments were conducted in digital health communication contexts to test the effect of message evidence on consumers' information adoption. The studies also examined whether perceived deceptiveness mediates this effect and whether it varies across different types of digital humans. Findings The results show that statistical (vs. anecdotal) evidence used by digital human doctors enhances consumers' information adoption by reducing perceived deceptiveness. However, when health information is delivered by mirrored rather than virtual digital humans, this mediating effect becomes non-significant. Originality/value This study extends message evidence research to the context of online health science popularization and identifies perceived deceptiveness as an important psychological mechanism underlying consumers' responses to digital human doctors. It also reveals the boundary role of digital human type in shaping the persuasive effects of message evidence in AI-driven health communication.