Can changing social media algorithms alter user beliefs?

Question

Accepted Answer

Research Indicates Facebook's Algorithm Does Not Significantly Alter User Beliefs

A comprehensive research project investigating Facebook's operations around the 2020 US presidential election has yielded findings that challenge the notion that the platform's content-ranking algorithm fundamentally shapes users' beliefs. The study, a significant collaboration between Meta and academics from several US universities, provided researchers with extensive access to internal company data. Through experiments involving tens of thousands of users over a three-month period, the project aimed to understand the algorithm's impact on political attitudes and behaviours, with initial results suggesting a limited direct effect on belief systems.
Experiments Showed Limited Impact on Political Attitudes

The core of the research involved altering how users received content on Facebook and Instagram for three months. One group experienced a chronological feed, displaying posts from newest to oldest, while another group continued to use the standard algorithm-driven feed. Despite the chronological group spending approximately half the time on the platforms and encountering more ideologically mixed content, measured political attitudes did not change significantly. This suggests that even substantial alterations to content delivery mechanisms do not necessarily translate into shifts in users' deeply held political views.
Reshared Content and Like-Minded Content Had Minimal Effect on Beliefs

Further experiments explored the impact of suppressing specific types of content. When reshares were suppressed, the proportion of political content users saw decreased, leading to a reduction in political knowledge. However, this intervention did not affect downstream political attitudes or behaviours. Similarly, an experiment that suppressed content from like-minded users, pages, and groups, which constitute a significant portion of what users see, also failed to alter ideological extremity or belief in false claims.
Algorithm Influences On-Platform Experience but Not Core Beliefs

While the algorithm may not directly change users' beliefs, it demonstrably influences their on-platform experiences. Researchers found that the algorithm is "extremely influential in people’s on-platform experiences," impacting what users see and how much they engage with the platforms. This indicates that the algorithm plays a crucial role in shaping the immediate user interface and interaction patterns, even if it does not penetrate to the level of altering fundamental political attitudes or beliefs.
Study Confirms Ideological Segregation on the Platform

Despite the findings regarding the algorithm's limited impact on beliefs, the research did confirm a significant issue of "ideological segregation" on Facebook. The study found that politically conservative users tend to be more siloed in their news sources compared to liberal users. Furthermore, a substantial majority of political news URLs rated as false by Meta's third-party fact-checking program were disproportionately viewed by conservatives more than liberals, highlighting existing societal divisions rather than algorithmic causation of those divisions.

Reading history

Hide →

This history is cleared when the page is closed.

Can changing social media algorithms alter user beliefs?

Loading

Was this article helpful?

Helpful

Not helpful

Report an issue

1

0

0

Share with friends