“In six out of seven countries — all but Germany — Tweets posted by accounts from the political right receive more algorithmic amplification than the political left when studied as a group,” Twitter said in its blog post detailing the findings of its analysis of whether recommendation algorithms amplify political content. The study examined tweets from elected officials of seven countries— Canada, France, Germany, Japan, Spain, the United Kingdom, and the United States. Tweets from news outlets were also studied in this exercise.
The social media platform compared the political content on the Home timeline and the latest tweets section.
Twitter’s study revealed that its algorithms amplify political content from elected politicians regardless of which party they belong to at the time. It also found that “right-leaning news outlets see greater algorithmic amplification on Twitter compared to left-leaning news outlets”. However, as highlighted in the paper, these third-party ratings make their own, independent classifications and as such the results of the analysis may vary depending on which source is used.
Twitter’s research confirms the dangers of social media including death in several countries and may shape the conversation around political content on social media.
How did Twitter conduct this study?
Researchers Luca Belli and Rumman Chowdhury started with the following questions:
- How much algorithmic amplification does political content from elected officials receive in Twitter’s algorithmically ranked Home timeline versus in the reverse chronological timeline?
- Are some types of political groups algorithmically amplified more than others?
- Does this amplification vary across political parties or within a political party? Are these trends consistent across countries?
- Are some news outlets amplified more by algorithms than others? Does news media algorithmic amplification favor one side of the political spectrum more than the other?
They analysed millions of tweets from April 1 to August 15, 2020, from Twitter accounts in seven countries to find answers to these questions. Belli and Chowdhury said they used public, third-party sources to identify political affiliation of the elected leaders. They also clarified that the content of the tweets was not used to “infer political views”.
Moreover, analysis was performed on millions of tweets containing links to articles shared by people on Twitter. The media bias ratings of the news outlets were ascertained based on data from two independent organisations – AllSides and Ad Fontes Media.
How will Twitter address issues emerging out of these findings?
Twitter did not announce any steps to mitigate the inequities put forth by the research.
“Establishing why these observed patterns occur is a significantly more difficult question to answer as it is a product of the interactions between people and the platform,” read the post by the platform. Simply put, Twitter does not know why the algorithm favors a particular political ideology.
Twitter, however, did assert that “algorithmic amplification is not problematic by default – all algorithms amplify”. The company said it was concerned if certain tweets received preferential treatment not as a result of the way in which users interacted, but because of the inherent code of the algorithm.
Root cause analysis is required in order to determine what, if any, changes are required to reduce adverse impacts by our Home timeline algorithm — Twitter in its blog post
Will Twitter make this data available to other researchers?
The platform said that it is making aggregated datasets available for third-party researchers on request but added that privacy concerns make it difficult to provide access to the raw data.
The ML Ethics, Transparency and Accountability (META) is conceiving ways, including a partnership, to leverage privacy-preserving technology to provide this data responsibly.
You can read the entire paper here.
- Twitter’s ‘soft block’ feature is for users who need some space from annoying followers
- Twitter releases principles for regulators to counter the ‘age of techno-nationalism’
- How will Twitter determine bad vibes in a conversation?
- Twitter tests new feature wherein users can report tweets as misleading to tackle misinformation
Have something to add? Post your comment and gift someone a MediaNama subscription.