How collaborative journalism worked with Comprova
Days of training, lots of sticky notes and karaoke were some of the building blocks of the Brazilian election project
By Sérgio Lüdtke (Versão em português brasileiro)
Competition can be a great motivator, but a shared purpose for all is better, which is what I discovered after 12 weeks of working collaboratively on Comprova. During this period, journalists from 24 competing media companies worked together to investigate the veracity of dubious claims and content that circulated on social networks during the Brazilian presidential campaign of 2018.
The fluidity of collaborative work depends on engagement of the organizations taking part and the clarity of the project’s goals. With Comprova, having the editorial processes defined and a project scope was instrumental in creating a solid foundation for our work together.
Before the start of the project, journalists gathered for a 2-and-a-half-day bootcamp where they received training and met their new collaborators. Working together like this is a fundamental step for collaborative work, especially if that collaboration relies on teams that will work remotely. Personal contact helps to create relationship networks, but above all it is an opportunity to reveal affinities and reinforce the bonds of trust that proved to be necessary for Comprova.
At the bootcamp, Comprova journalists learned about network-monitoring methodologies and were trained to use a range of tools like CrowdTangle, NewsWhip’s Spike and Torabit, that allowed them to search for questionable content during the campaign. Face-to-face training with hands-on activities and group-work is necessary to strengthen relationships and to promote an exchange of experiences. Verifiers complemented their training with several hours of online courses, but it was the time when we were together that most contributed to partner engagement during the project.
Back in the newsrooms, communication between the group was vital to the project. The initial intent was to use Slack, however, Brazilians newsrooms have not heavily adopted the platform and having other tools to learn — the CMS, CrowdTangle, NewsWhip, etc. — the group opted for the familiar platform WhatsApp to communicate with each other about debunks and what we might work on together that day or week. WhatsApp does not allow for threaded messages so conversations can be lost and search can be clunky. Even so, the group generated 18,500 messages during the project, so while WhatsApp was not ideal for partner communications, it did work for the purposes of our collaborative project.
Obviously, at times, when there was more than one subject under discussion and many people involved in the conversation, the communication became a bit confusing. During the period that we investigated a video that showed “Bolsonaro 2018” written in the fields of a plantation, we exchanged 545 messages in the group in a single day.
In collaborative work, many conversations are generated. In the case of Comprova, we chose the management model of consensus, which encouraged partner participation. In this horizontal model, it was important that everyone’s opinion was respected and that space for disagreements was provided. In the project, verifiers were encouraged to comment, intervene, and propose solutions even for verifications in which they were not directly involved. The decision process adopted does not necessarily require more time since consensus is often not reached by a vote, but rather can be simply observed. A consensus model does require mature decisions, which is most welcome in investigative work that aims to be rigorous.
Traditional newsrooms have more rigid hierarchies with more vertical decision-making processes. A consensus model tends to give more value to the knowledge, opinions and participation of all. A survey of the verifiers at the end of the project showed that almost 90 percent of them did not have prior collaborative networking experience. All said they were satisfied with the consensus management model and with collaborative work, and almost 80 percent said they were fully satisfied with the level of collaboration from colleagues.
The openness to participation of all implies more commitment, distributes the responsibilities and creates a greater zeal for the final result. With Comprova, engagement among partners helped to ensure the accuracy of our work and to more easily achieve the zero-error goal together.
The scope and objectives for Comprova were clear from the beginning, and questions about any doubts were quickly solved. With transparency and openness to participate, the verifiers, in addition to collaborating with colleagues, became auditors for the objectives of Comprova.