Twitter admits it multiplies the French more than the left

Twitter has released the longest and rarest survey to reach messages posted by political figures across Europe, according to their party. This observation has improved in all educated countries: the right is amplified more than the left. Not sure why social networking.

Tweets from right-wing political parties are, on average, more frequently posted on user news feeds than those on the left.

This is a consequence Long study 27 pages, Posted by Twitter On October 21, 2021, ” Algorithmic multiplication of politics on Twitter ”, In which six researchers analyzed two components:

• Publications of elected politicians in seven states (Canada, France, Germany, Spain, UK and USA),
• Publications related to information sites are more left or right oriented.

How your Twitter feed works

When you open Twitter, your homepage will display tweets from the accounts you choose to follow, and the algorithm from the accounts will be considered “appropriate” for you. There are two ways to order them: Chronological thread (You see when they post tweets), and Algorithmic classification, It claims to present “popular tweets”.

This order was proposed By 2016 by default via Twitter (Although it can be changed manually, the timeline can be returned to the display). It is determined by a process in the Twitter algorithm Mechanical learning (The software is constantly learning from your actions), it will show up at the top of your home feed. These choices were accurately analyzed by the authors of the October 21 study.

They compared the magnitude of the proliferation of tweets classified as “right” politically between “left”, chronological texts and “weighed” texts by algorithm. It received millions of tweets over a four-month period. The results are clear: The parties on the right are bigger than those on the left on TwitterSimilar to the media displayed on the “right” compared to the left.

The Republican Party is much stronger than the rebellious France

Of the seven countries observed, Germany was not the only one to experience this bias. To six other states, including France Tweets posted by political right-wing accounts received more algorithmic multiplication than those on the left », Can we read In a Twitter blog post.

For example, the chart below shows how the tweets of some political groups are more prominent than others.

In the case of France, the study took into account tweets from delegates and delegates by 7 groups: Rebel France, PS, Environmental Democratic Solidarity Group (EDS on the map), Modem, LREM, Republicans and unregistered.

We have enlarged France on this map so that the results can be analyzed very neatly, which gives the image below.

We can clearly see that the tweets of rebel France were “multiplied” by at least 110% in all parties, and that those from the Republicans were multiplied by more than 150%. Modem and Republic are considered government parties in the study, with environmental activists taking second and third place.

⁇ A large percentage of the promotion is that the ranking algorithm gives this tweet group more ‘match scores’, so it appears more often in news feeds than chronological chronology. 2, explains the study.

In short, Twitter acknowledges that tweets from Les Républicains representatives appear more frequently in users’ timelines (not timelines) than in rebellious France, PS or non-subscribers.

Another map of the study supports this finding: the proliferation of right-wing parties is about 150% and the tweets of left-wing parties in France are about 125%, i.e. 25% difference.

This imbalance is more important than we see in the United States and Japan worldwide, but less obvious than in Germany (where the tax is 160%, versus … 50% on the left).

Why does the right know more than the left?

Twitter, despite the good intentions of transparency, is very embarrassing: the platform can’t say exactly why its algorithm supports right-wing content more than left-wing content.

⁇ It is very difficult to succeed in identifying the causes of this evolving system because it is the result of interactions between humans and the site. , We can read. ” Naturally algorithmic multiplication is not a problem – all algorithms multiply. Algorithmic multiplication is problematic if there is a priority treatment caused by the way the algorithm is designed. ⁇

The social network thinks it should lead. ” Further analysis of the source of the problem “To determine if he should change his method” Minimize negative impacts சையில் Chronology of users.

However, this first study is already crucial in showing that large technology companies are finally beginning to realize their enormous responsibility in conducting public debate. Many years later they showed themselves to be “neutral.”. This awareness is so important during the election period that, as in France, the Assembly elections and the presidential election take place in May 2022.

There are more political tweets than any other tweets

As for news sites, Twitter also noticed this Media content that is tilted significantly to the right is slightly more amplified than sources that are tilted to the left », But the gap between the two is very small.

On the other hand, ” Politically significant tweets from elected officials, regardless of their party or if their party is in government, have more methodical access Popular order, chronologically compared to political tweets, Twitter points out in its official statement.

This finding is not surprising given that it exists for elected officials Usually a certificate badge on TwitterTherefore, it has more reliability than other users.

How to rewrite your Twitter news feed chronologically

From 2018 onwards, you have the option to select the chronological order (or rather chronologically, precisely) if you want to make Twitter’s amplification algorithm’s choices less or more revealing.

To do this:

• Go to the home page
• Click The small star is shown in the upper right

• ” Check out the latest tweets instead ⁇
• Twitter is telling you now, ” You see when the latest tweets are released .

You are now (slightly) less exposed to the site’s algorithms.

Continuation in the video