Information Cocoons in Economic Statistics Education: A Study of Teaching Reform Pathways

Abstract

In the era of digital economy, intelligent algorithms—such as platform recommendations, short video pushes, and search ranking mechanisms—have become primary channels for college students to access information and data. The resulting "information cocoon" effect is increasingly influencing how economic statistics students perceive data and approach statistical reasoning. Specifically, students’ data sources are becoming more platform-dependent, visually oriented, and preference-driven, which undermines their ability to collect diverse data and weakens their awareness of sampling principles. Moreover, the filtering mechanisms embedded in recommendation algorithms impair students’ capacity to identify key statistical biases—including sampling bias, selection bias, and availability bias—leading to an overreliance on procedural tools, reinforced learning pathways, and a neglect of statistical ethics. To address these challenges, this study proposes a three-stage teaching reform model centered on “cognitive awakening – bias detection – anti-cocoon practice.” It introduces a data bias measurement formula and an operational flowchart to guide implementation. The reform is designed to enhance statistical literacy and strengthen data governance competence, ultimately aiming to shift the focus of statistics education from merely using data toward understanding, evaluating, and governing data. By doing so, it seeks to cultivate a new generation of ethically responsible data professionals capable of thriving in an increasingly digital society.