How Traceability Information and Live-Stream Credibility Shape Agricultural Product Purchase Intention: An Online Experiment and SEM Analysis

Agricultural products sold through live-streaming channels are difficult to inspect before purchase, which makes buyers dependent on product and source cues. Drawing on the Stimulus-Organism-Response framework and signaling theory, this study tests whether traceability information and streamer credibility jointly predict online purchase intention through product trust and perceived risk. Study 1 was a 2 × 2 between-subjects online experiment (N = 163). Study 2 surveyed experienced agricultural live-streaming consumers (N = 363) and used partial least squares structural equation modeling. In Study 1, both manipulated signals had significant main effects and a positive interaction on purchase intention, F(1, 159) = 7.575, p = 0.007, partial η² = 0.045. Conditional-process analysis showed that the indirect effects through product trust and perceived risk were present when the other signal was high but not when it was low. Study 2 produced significant main-effect paths and a smaller positive traceability-information × streamer-credibility path (β = 0.064, bootstrap 95% CI [0.003, 0.127]). Both mediators carried significant indirect associations. The trust-minus-risk indirect-effect contrast was significant in Study 2 but not Study 1, so the results do not establish universal dominance of either mediator. The full Study 2 model explained 49.2% of the variance in purchase intention. Ten-fold PLSpredict assessment showed positive indicator-level predictive relevance and slightly lower prediction errors than the linear-model benchmark for all four purchase-intention indicators. These findings support a context-bound signal-consistency account: verifiable product information and a credible source can reinforce one another, while trust and risk remain empirically distinct pathways.