Wan, J., Araujo, T., Helberger, N., & de Vreese, C. (2026). How does political efficacy condition clicks on politics? Understanding information-selection behavior in algorithmic feeds over time. Journal of Computer-Mediated Communication, 31. https://doi.org/10.1093/jcmc/zmag022
Summary
This paper examines how political efficacy — an individual’s sense that they can understand and influence politics — shapes the way people select political information within algorithmically curated feeds. Moving beyond the common framing that algorithms alone determine exposure, the authors argue that user-side psychological dispositions meaningfully condition information-selection behavior even when the feed is fully algorithmic. Using a combination of survey data and actual behavioral usage histories from 186 Dutch users of Google Discover, they distinguish within-person change over time from between-person differences, focusing on outcomes such as the proportion of political content selected and the diversity of exposure.
Key Contributions
- Links the psychological construct of political efficacy to observed, real-world information-selection behavior in algorithmic feeds.
- Combines self-report survey measures with genuine behavioral trace data from a fully algorithmic feed (Google Discover), rather than relying on self-reported or simulated exposure.
- Analytically separates within-person temporal dynamics from between-person variation in political information selection.
- Brings a user-agency perspective to debates typically centered on algorithmic curation and selective exposure.
Methods
- Pairs survey data on political efficacy with logged usage histories from Google Discover.
- Sample of 186 Dutch individuals.
- Models both within-person change over time and between-person differences.
- Examines at least three facets of information selection: proportion of political content selected and exposure diversity (the full set is truncated in the available abstract).
Findings
- Political efficacy helps explain individual-level differences in how people select political information from algorithmic feeds.
- Information-selection behavior varies both within individuals over time and between individuals.
- Between-person differences in political information selection are conditioned by users’ political efficacy.
- Specific effect sizes and directions are not recoverable from the truncated abstract.
Connections
This work sits in the broader research program scrutinizing whether algorithmic feeds drive exposure or whether user choices dominate, connecting to the large-scale platform studies of Guess2023-ai, Guess2023-ur, Gonzalez-Bailon2023-uy, Nyhan2023-gb, and Bakshy2015-rn, all of which weigh algorithmic versus individual contributions to political exposure. Its emphasis on user-side selective exposure and information diversity resonates with the filter-bubble and echo-chamber debates in Flaxman2016-lm and Del-Vicario2016-uj, while its reliance on behavioral trace data aligns it methodologically with Ulloa2024-jm.