Transparent and interactive news feed design

What is the boost?

The news feed plays an important role in how people consume and share information. The algorithms that curate news feeds guide users through the information landscape; if designed transparently and allow the user to interact with the sorting algorithm, news feed algorithms can increase users’ autonomy.

How does the boost work?

The criteria for determining how information is sorted in a news feed are displayed to users and users can change the importance of the different criteria (see figure below). People can thereby not only observe how the ranking of articles and posts changes under different settings, but also set their personal preferences clearly, without having to rely on inscrutable algorithms.

Which competences does the boost foster?

Understanding the logic of content ranking and curation in an online news feed, as well as setting one’s own preferences (e.g., towards more news-oriented or more personal information).

Which challenges does the boost tackle?

Blind reliance on inscrutable algorithms (which tend to prioritize popular and emotional content) for curating and sorting the news people consume.

What is the evidence behind the boost?

This boost has not yet been directly tested. However, experiments have shown that click-based curation can promote emotional, moral, or low quality information; a rule-based algorithm with other factors than popularity and recency for ranking could mitigate these effects.

How is the boost implemented?

Implementation would require the cooperation of platforms, which would have to replace their news feed algorithms with rule-based versions, as well as change their interface.

Example of a transparently organized news feed on social media. Types of content are clearly distinguished, sorting criteria and their values are shown with every post, and users can adjust weightings. Based on Figure 3 in [Lorenz-Spreen et al. (2020)](https://doi.org/10.1038/s41562-020-0889-7).
Example of a transparently organized news feed on social media. Types of content are clearly distinguished, sorting criteria and their values are shown with every post, and users can adjust weightings. Based on Figure 3 in Lorenz-Spreen et al. (2020).

Key reference

Lorenz-Spreen, P., Lewandowsky, S., Sunstein, C.R., & Hertwig, R. (2020). How behavioural sciences can promote truth, autonomy and democratic discourse online. Nature Human Behaviour, 4, 1102–1109. https://doi.org/10.1038/s41562-020-0889-7