Twitter is not reality.
The company's algorithm favors right-leaning news outlets, right-leaning political parties, according to a Twitter study and a blog post that was published with it. This means that long-disputed claims about anti-conservative bias in social media are untrue.
What is the right-wing content amplification? According to the study, users only need to look at the Home timeline, which is the Twitter timeline that is not presented in reverse chronological order. This timeline uses an algorithm that helps users understand what they see. That is the problem.
The study revealed that content on Twitter's homepage timeline is ordered and selected by personalization algorithms. These algorithms can be used to increase visibility by ranking certain messages higher and amplify others by consistently ranking them higher.
Amplified messages tends to trend in a specific direction.
The study that examined tweets from seven countries (Canada France Germany Japan Spain Spain the U.S.) found that the main political right has higher algorithmic amplification than those of the mainstream political left in six out of seven countries.
Rumman Chowdhury (director of Twitter's META) team, which is responsible for machine learning ethics, transparency and accountability, felt this was a critical point and made it clear in a lengthy Twitter thread.
Twitter researchers note a Trump-sized caveat, notably.
They cautioned that "when comparing the amplifications of individual politicians' accounts rather than party in aggregate, we found not an association between amplifications and party membership."
The study authors found that right-wing amplification on Twitter doesn't just apply to political parties. It also applies to news media outlets.
"In line with this overall trend our second set findings from the U.S media landscape showed that algorithmic amplifying favours right-leaning news outlets."
This is why amplification is happening.
SEE ALSO: People fight algorithms for a better and more equitable future. You can, too.
Chowdhury tweeted, "Here's the complex: The team did phenomenal job identifying *what* was happening." "Establishing the reason these patterns are occurring is a much more difficult question to answer, and something META (Meta-Explorer) will investigate."