TikTok’s ‘For You’ Algorithm Delivered Asymmetric Partisan Content to Both Democratic and Republican Accounts in 2024 Study

A research paper titled 'Systematic partisan content skews in TikTok during the 2024 US elections' was published in Nature. The study documented systematic, asymmetric differences in partisan exposure on the platform's 'For You' feed, which delivers content almost entirely through algorithmic recommendation.

Researchers conducted 323 audit experiments using controlled sock puppet accounts on TikTok.

The sock puppet accounts were seeded with Democratic or Republican content across three US states, and data collection occurred over 27 weeks during the 2024 US presidential election campaign. More than 280,000 recommendations were collected during the experiments. 5 percent more cross-partisan content than Republican-seeded accounts.

The cross-partisan content received by Democratic-seeded accounts was largely anti-Democratic material. These partisan asymmetries persisted even after adjusting for engagement metrics. The asymmetries are concentrated among high-reach Republican channels.

Asymmetries for Democratic-seeded accounts were concentrated in policy domains including immigration, crime and foreign policy, while asymmetries for Republican-seeded accounts were concentrated in the policy domain of abortion. @Nature reported that the 'For You' feed of TikTok offers a setting in which user agency is sharply constrained compared with platforms where users retain substantial control over their feeds.

The findings show partisan imbalances in political information exposure on a platform dominated by algorithmic recommendations.