Online Evaluation Information Cascade and Its Impact on Consumer Decision Making: Analyzing Movie Reviews Using Sentiment Corpus

Abstract

User-generated content on self-media platforms significantly influences the market. In the era of Web 2.0, consumers make purchasing decisions based on electronic word-of-mouth (eWOM) from these platforms. This research illustrates how sentiment value of eWOM content guides consumers’ behavior by empirical study based on the causality approach. The research calculates the sentiment value of 160,000 textual film reviews using sentiment analysis program which is based on the sentiment corpus and addressing sarcasm. It also measures the complexity of information by calculating entropy to capture the information cascade process of eWOM converging into a reputation signal. The findings demonstrate the causal relationship between positive eWOM content and consumers’ decision-making processes, as well as the extent to which consumers rely on eWOM textual information when making consumption decisions. When the eWOM information conveys a clear reputation signal, it will have a lasting impact on future box office revenue. This article is the initial empirical literature that quantifies the commercial value of online text sentiment information through a causal and dynamic perspective. It is also the first literature to empirically illustrate the formation process of sentiment information cascade during the diffusion of eWOM among the netizens and capture the required time lag for the formation of online consumer reputation signal.

Keywords

• Sentiment analysis
• information cascade
• entropy
• user-generated content
• Chinese text
• movie reviews