The next speaker in this session at the 2026 International Communication Association conference in Cape Town is Jin Wan, whose interest is in how political efficacy conditions clicks on political content in algorithmic feeds. Political efficacy here means people’s belief in themselves within the political world: this includes internal efficacy (confidence to participate in politics) as well as external efficacy (confidence in the responsiveness of the political system).
How do people with different levels of such efficacy differ in their information selection approaches in algorithmic environments, then? Do they seek a different proportion of political content; do they seek different source exposure diversity; do they attract more negative political information?
This project worked with the Google Discover dataset, which is a popular, fully algorithmically curated content feed; it contains multiple content cards representing different thematic interests, and users’ selection of such content cards points to a distinct personal preference and selection.
The project approached this via a combination of survey and data donations, resulting in some 190 data packages containing 70,000 URLs that were visited via Google Discover. Some 8.7% of these URLs can be understood as political, across 304 unique sites. 66 of the participating users never cliched on any political content; the rest visited at least one such URL.
Users with higher internal and external efficacy were more likely to click on political information; users who cliched at least once there was a slight decrease in the proportion of political information over time, but this is in line with overall click-rate decreases over time. Higher internal efficacy also resulted in higher exposure diversity; this led to more diverse information diets over time. This trend was reversed for people with less internal efficacy. Negative political information had no significant impact; this may be because most political coverage is negative by default.
Overall, then, more internally efficacious users are more resilient to concerns about algorithmic content feeds, but there may also be a growing divide between more and less internally efficacious users here, which is a concern.
