Data for a Dynamical Study of Feed–Search Coupling and Information Cocoon Breakthrough
Description
This dataset contains the empirical data and analytical materials associated with the study “Does the Information Cocoon Exist and How Can It Be Broken? An Empirically Calibrated Dynamical Study of Feed-Search Coupling and Individual Breakthrough.” The study examines whether the information cocoon is a measurable and stable phenomenon by analyzing the dynamic coupling between algorithmically pushed information feeds and users’ active search behavior. Using Baidu Index time-series data for 60 films released in China between 2018 and 2024, the dataset provides the empirical basis for calibrating and evaluating dynamical models of feed-search interaction, search suppression, and individual breakthrough. Dataset Contents: - Raw Data: Film-level Baidu Index time-series data. - Baseline Regression Results: Static regression results for each film, including estimated coefficients, p-values, model fit statistics, and pass/fail indicators for theory-constrained model structures. - Dynamic Regression Results: Rolling-window dynamic regression outputs used to evaluate the stability and robustness of feed-search coupling relationships over time. - Equation Structure Summaries: Summary tables ranking candidate equation structures by fit, pass rate, coefficient direction, statistical significance, and consistency ratio. - Summary Files: Aggregated results across all films, including baseline regression summaries, dynamic regression summaries, and overall structure-discovery outcomes. Together, these materials support the estimation of feed-search coupling relationships, the identification of robust dynamical structures, and the reproduction of the main empirical analyses reported in the paper. The dataset is provided as supplementary research material for transparency, verification, and academic reuse.