The Impacts of Live Chat Expressions on Service Refunds in Online Crowdsourcing Platforms
Fangze Dai, Lingfeng Dong, Liqiang Huang, and Yu Tu
International Journal of Electronic Commerce,
Volume 29, Number 1, 2025, pp. 99-130.
Abstract:
Successful service refund processes in online crowdsourcing platforms require ongoing communication and negotiation between workers and requesters, mostly facilitated by live chat. This study explores how workers’ live chat expressions affect the likelihood of a service refund request in online crowdsourcing platforms. We rely on a unique archival dataset, which includes 35,178 orders with 1,597,305 live chat recordings of 33,140 requesters and 1,351 workers on a major Asian crowdsourcing platform, to test the research model. The findings show that, overall, requesters’ likelihood of seeking a refund is negatively impacted by workers’ affective expressions, such as politeness intensity and sentiment intensity. It is generally positively impacted by workers’ cognitive expressions manifested by the amount of information and response time. Delving more deeply, we further observe that the impact of workers’ amount of information varies across different decision stages. Specifically, in the prepurchase stage, the amount of information provided by workers exhibits a U-shaped effect on refund requests, while in the postpurchase stage, it has a consistently positive linear effect. This suggests that information has varying roles at different stages in shaping refunds. Moreover, we identify that business familiarity weakens the relationship between the amount of information and refund requests, but strengthens the impact of sentiment intensity on refund requests. However, it typically does not significantly moderate the effects of response time and politeness intensity. This study contributes to the literature by empirically demonstrating how different live chat expressions, that is, cognitive and affective, and business familiarity influence refund requests across transaction stages on online crowdsourcing platforms, offering new insights into service refund dynamics and communication strategies.