More Than a Button: How Creator-Initiated Interaction Influence Online Tipping on UGC Video Platforms

As user traffic becomes a key source of competitive advantage on UGC video platforms, creators have begun actively initiating interactions in their videos to boost users' positive engagement and ultimately promote content monetization. Based on dual-processing theory, this study investigates how creator-initiated affective (low-cognitive-effort) and cognitive (high-cognitive-effort) interactions influence users' online tipping. Specifically, we employed large-scale data collection via the Bilibili API using Python-based crawling scripts, followed by computational preprocessing to construct a dataset of 353,501 videos from 25,277 creators. We applied regression algorithms and interaction effect estimation to empirically test our hypotheses. Furthermore, content type plays an important moderating role. This study not only enriches theoretical understanding of interactive media effects but also demonstrates how computational approaches - such as automated data mining, algorithmic modeling, and large-scale empirical analysis - can advance the application of computer technology in digital media interaction research.