Propagande et politique: les nouvelles règles du game
Organisation size: Small
Publication date: 2/12/2021
Credit: Jehanne Berge, Guillaume Derclaye, Eric Walravens
Biography: The Ad-Lens collective, which we are a part of, was created in order to gather data on online political party spending. We have been working on monthly reporting, freely published each month. For this specific project, Jehanne, Guillaume and Eric have teamed up to bring a story to investigative magazine Médor. We are all independant journalists working with several media in Belgium.
Social networks are fundamentally changing public debate. Polarizing content is favored, micro-targeting makes it possible to send messages to categories of the population, and party spending has exploded, particularly so in our country Belgium.
Our project is based on data from the Facebook ads library and on crowdsourced data from users. We have treated the data in a way that allows the public to understand the issues at stake, and to tell a story based on numerous sources, including top politicians. Additionally, a newsgame allows the reader to put him/herself in the shoes of a political community manager.
When the size of political spending on Facebook was revealed, compared to other EU countries, thanks to the work of the Ad-Lens collective which we are a part of, it created a bit of a shock, and an important debate has taken place in Belgium, mostly on social media. The political parties involved are avoiding the debate, because they have an interest in the status quo. However, a hearing was organized in the national parliament on the issue.
Most and foremost, with the help of the Ad-Lens collective, we have worked on data from the Facebook Ads Library
* analysis of the amounts spent by politicians on political ads on Facebook, both at Belgian and European level from August 2019 until the end of October 2021.
* Collection and sorting of thousands of spending data
* Qualitative analysis of the biggest campaigns run by Belgian parties (at party level but also at the level of political figures)
* Highlighting of gender targeting and gender bias of the algorithms
Since the library does not allow to obtain information on targeting, we have launched an open survey, which resulted in nearly 80 detailed responses from the public. Those allowed us to understand some of the targeting used by parties, especially those that they would relunctantly admit they were using.
We have also published the full details of a standard questionnaire that was send to political parties, so as to allow the public to read the answers in full, as well as the written answers from Facebook.
An important part of this work was the creation of a newsgame : on the Medor website, the reader was offered a scenario in which he/she could work as the community manager of a political party, and explore, from the inside, the reasons that pushed political parties to spend so much in Facebook ads.
What was the hardest part of this project?
The amount of data made available on the Facebook ads library was a bit daunting, especially as we had to compare political spending all across Europe to confirm our intuition that Belgium was leading the way. So there was a lot of time invested in gathering, sorting, checking the data, and also to publish it in a way that would make most sense for the public debate. Admittedly, datavisualization in our article is mainly just charts, but we believe they tell an important story.
What can others learn from this project?
What we are most proud to have achieved is to get a story based on a large amount of data that is also highly readable. Our intention was never to show off with the most beautiful charts (altough we are applying here now), but to expose a problematic trend for the public debate, in a way that is understandable and enjoyable to read.