Measuring the Psychological Effects of Algorithmic News Consumption on Political Polarization

Abstract

The rise of algorithm-driven news recommendation systems on social media platforms has reshaped information consumption patterns, raising concerns about political polarization. Algorithmic curation, designed to maximize engagement by tailoring content to user preferences, may reinforce ideological biases, create echo chambers, and amplify selective exposure. This study investigates the psychological effects of algorithmic news consumption on political polarization, examining cognitive, emotional, and behavioral responses to personalized news feeds. A quantitative survey was conducted with 400 social media users, spanning diverse age groups, education levels, and political affiliations. Constructs measured include algorithmic exposure, selective perception, cognitive bias, emotional response, and political polarization. Structural Equation Modeling using SmartPLS was employed to test the relationships between algorithmic news consumption and polarization outcomes. Results indicate that higher exposure to algorithmically tailored content significantly increases selective perception and cognitive bias, which mediate the relationship between algorithmic exposure and political polarization. Emotional responses, including anger and moral outrage, further exacerbate polarization effects, suggesting that algorithmic curation not only shapes information exposure but also amplifies affective drivers of political division. The study contributes to media psychology and political communication literature by providing empirical evidence linking algorithmic news curation with psychological mechanisms that drive polarization. Findings highlight the role of cognitive and emotional mediators, emphasizing the need for platforms, policymakers, and users to recognize and mitigate the effects of algorithmic filtering. Limitations include reliance on self-reported survey data and cross-sectional design. Future studies should incorporate longitudinal and behavioral tracking methods to capture real-time interaction patterns and causal mechanisms.