Dynamic interaction effects between online reviews and spoilers: a PVAR approach

Spoilers, as content elements, have received increasing attention from academics, especially for narrative products, such as films and books, where spoilers tremendously influence product sales and consumer experience. However, from the perspective of user-generated content, the relationship between online reviews and spoilers remains unclear. The study addresses this research gap by collecting 279,433 reviews of 465 movies on Douban.com and constructing a panel vector autoregression (PVAR) model to dynamically analyze the interaction between spoilers and three essential indicators of online reviews (volumes, valences, and variances). Our research indicates that spoilers can stimulate the generation of online reviews and foster a convergence of opinions. In addition, we further verified that the status of previous online reviews could influence spoiler behavior. Specifically, in the initial phase, individuals are more prone to engaging in spoiler behavior during subsequent stages when there is a decreased volume, higher valence, and greater variance in reviews. These findings contribute to an enhanced understanding of the spoiler effect by review platforms and film marketers, helping them develop more effective and flexible word-of-mouth operational strategies in the future.