Tao Liu, Xin Wang
Assignment to the bundle causally raised hypothetical charitable allocation, but the design cannot show music was necessary or uniquely effective, concurrent measurement cannot establish mediator ordering, and the beneficiary shift is a descriptive composition change, not generalized prosociality.
BACKGROUND: Music listening and collective musical activity are linked to prosocial outcomes, yet everyday social-media music sharing is little studied. Using the social grooming model, we tested the overall effect of a multicomponent relational-responsive music-sharing bundle on hypothetical charitable allocation; the design did not isolate music or any single component.
METHODS: In a 7-day, three-arm randomized field experiment on WeChat Moments (583 randomized; mITT N = 568), the focal contrast compared relational-responsive music sharing with informational posting under the same daily posting target (n = 377); a third arm used minimal music. The primary outcome was tokens allocated to two charities from a fixed 10-token endowment. Perceived social presence and social identity were measured concurrently at endpoint; perceived social support was an exploratory moderator.
RESULTS: The bundle increased allocation over informational posting [M = 4.08 vs. 3.28; b =0.758, 95% CI (0.377, 1.138), d = 0.425]; informational posting did not differ from minimal use. The serial presence → identity association was positive [0.053 (0.021, 0.099)]; presence-only and identity-only associations included zero. Unpaid helping did not differ [OR = 1.20 (0.79, 1.81)]. The bundle raised total allocation and shifted composition toward the music-education beneficiary.
CONCLUSIONS: Assignment to the bundle causally raised hypothetical charitable allocation, but the design cannot show music was necessary or uniquely effective, concurrent measurement cannot establish mediator ordering, and the beneficiary shift is a descriptive composition change, not generalized prosociality.