A machine learning method for measuring information disclosure in sharing economy platforms

Research on e-commerce, social technologies and privacy has overwhelmingly treated information disclosure as a survey-based, subjective, and unidimensional construct. A few studies employing semantic analysis on objective textual data, on the other hand, are constrained by the manual coding method with limited number of categories and thus prone to bias. Building upon the social penetration theory, we introduce an innovative method of measuring information disclosure using machine learning algorithms in the context of sharing economy platforms. We propose that information disclosure should be examined from two dimensions, i.e., breadth and depth, and machine learning techniques could effectively compute the high-volume factual data of information disclosure. Using 1,200 hosts' self-description data in Airbnb as an example, we report the computational and evaluation processes of operationalizing information disclosure. The research thus provides new theoretical lens and empirical support through which information disclosure in digital age could be better understood and efficiently assessed.