Читать книгу Rethinking Digital Native Comunicators Training - Cristina Pulido Rodriguez - Страница 11

The first three field experiments we are going to discuss consisted in the creation of ad hoc Facebook (2017 and 2018) or Instagram (2019) profiles managed by the students during a four to six weeks period.

In every case, the predefined profiles created by the students had to follow a list of 50 news, buzz or conspiracy inclined media and adopt a conduct coherent with a specific ideal-type (a political tendency, an interest in conspiracy theories or a passion for sports news, for example). At one point, the students had to add the other profiles taking part to the experiment as “friends”.

As the exercise intended to make the students go beyond “common sense” knowledge by implementing rigorous methods, the students were encouraged to take daily actions (posts, shares, likes, comments, interactions,...) in order to test both general hypothesis and their own hypothesis on how the Facebook algorithm works in relation with news media content distribution.1 Does the algorithm really prioritize contents posted by friends over contents shared by pages ? What kind of news content do you see when you show a strong interest for a subject ? Which actions have the strongest impact on news personalization ? The students recorded their observations, progress and screenshots of their newsfeeds in a digital notebook.

The first experiment was carried out in 2017 by the AJM (Université de Neuchâtel, Switzerland) alone. It was dedicated to the study of Facebook’s algorithm and political news in the context of the 2017 French presidential election campaign. What kind of news do you see if you show your support for a political candidate ? And is the customization of a news feed reversible ? And if so, how quickly ? In order to answer these questions, eight “supporters” profiles were created, two for each of the main presidential candidates: François Fillon, Benoît Hamon, Marine Le Pen and Emmanuel Macron. It resulted in “politically polarized” media contents on newsfeeds combined with the weight of recent news events and the importance of the content shared by friends. For the students, the key finding of this first edition was the extent of the impact of their actions on the content they see and the rapidity of this process.

We wanted the second experiment to focus even more on the challenges these algorithms cause to newsmedia. The objective was also to add a comparative and international dimension, with two other schools from the same journalism schools network (ARPPEJ, Alliance internationale de recherche sur les pratiques et la pédagogie en journalisme), the MJMN (Université de Lorraine, France) and the EJL (Université catholique de Louvain, Belgium) joining the experiment.

The second experiment was conducted right after the announcement by Mark Zuckerberg in January 2018 that Facebook would modify the algorithm of its newsfeed and prioritize “meaningful social interactions” and content posted by friends.2 The students had to test both the prevalence of friends content over media content and the influence of showing an interest for a specific theme (conspiracies, buzz, local news, Russian presidential election) on the content one sees on their Facebook’s newsfeed. The experiment led the student to experiment first hand the weight of the contents posted by friends on a Facebook News Feed. During the last phase of the experiment, they also tested the reversibility of a show of interest and found out that some degree of reversibility was possible, especially from a more “serious” behavior to an interest in buzz content. The reverse appeared more difficult to achieve.

Due to difficulties linked to Facebook tightening up security, tracking down fake accounts and shutting them down but also because of the growing importance of Instagram in media strategies and audience practices, it is on this social network that we conducted the 3rd experiment. The mobile friendly character of Instagram and the fact that accounts accessed from multiple devices had been deleted by Facebook during the second experiment led us to build up an individual experiment with a group perspective. The protocol was adapted: each account was managed by a single student while the focus on results sharing and group discussions had been reinforced.

The high number of influencers brand pages on Instagram as well as the popularity of the Stories format led to renewed questioning about news content visibility and the dissemination of information on social networks. However, the main hypothesis was unchanged: a correlation between our actions on social media and the news visible on our news feeds and the weight of algorithms in these arbitrations.