How Zimbabwe’s fact checkers monitored the 2018 election using Check

During Zimbabwe’s 2018 election, civil society and fact checkers collaborated to debunk fabricated media and local-language misinformation spread via WhatsApp.

Last month, protests erupted in Zimbabwe due to the massive rise of fuel prices and living costs. Clashes between protesters and security forces led to the killing of several people, and hundreds were arrested as security forces in Zimbabwe fired live ammunition and teargas.

During the fuel protests, many Zimbabweans woke up to a complete internet blackout, according to OONI. Facebook, WhatsApp and Twitter were reportedly blocked.

The clashes were the worst in the southern African country since elections last year, when six civilians were shot dead by police, reported The Guardian.

Campaign ads at the streets of KweKwe (Courtesy: Wafaa Heikal)

Zimbabwe’s 2018 presidential elections
The harmonized elections came at a crucial moment in the country’s history. Robert Mugabe, the former president, retired in November 2017, opening the path for a new era of politics in the Southern African nation. For the first time in 16 years, Zimbabwe invited international election observers to monitor the vote on 30 July 2018. As with all elections in the digital media age, concerns about the impact of mis- and disinformation were high. It was in this fascinating context that in July 2018 I traveled to Zimbabwe to join The Carter Center mission, in an effort to support civil society and media efforts in tracking and addressing misinformation that could limit citizen access to credible information during this crucial moment in Zimbabwe’s democratic process.

“The Center’s expert mission was limited in scope, and did not conduct a comprehensive assessment of the electoral process. The mission focused on several key issues, including the legal and electoral framework, election administration, political and electoral environment, campaign period, women’s participation, civil society engagement, and electoral dispute resolution in the post-election period. The team did not observe election day proceedings in a systematic manner, and only visited a limited number of polling stations and tally centers in and around Harare. The team closely monitored post-election developments, including the dispute resolution process and the legal challenge of the presidential election results to its conclusion. Throughout its deployment the team met with various of stakeholders including election administration authorities, political parties, candidates, citizen observers, human rights organizations, media organizations, government officials, and other international observers and stakeholders.

Additionally, the Carter Center experts provided limited technical assistance to the Zimbabwe Election Support Network (ZESN) to support its data collection efforts for long-term observers through the ELMO open-source data collection tool. The Center also provided training to civil society organizations using a collaborative fact-checking tool named CHECK, including ZimFACT, a fact-checking platform in Harare and to the Center for Innovation and Technology (CITE) in Bulawayo.”

Check: A Tool to Enable Collaboration in fact-checking project

The goal of Meedan’s partnership with The Carter Center was to introduce civil society organizations in Zimbabwe to a culture of fact-checking and verification by leading a pilot collaboration to fact-check mis/disinformation shared online during the election day using Check, Meedan’s collaborative verification tool.

The collaboration was built between ZimFact, Zimbabwe’s first fact-checking initiative, based in Harare (northern Zimbabwe), and the Centre For Innovation & Technology (CITE), a vibrant civic technology incubator that supports initiatives led by young people, journalists, and designers. CITE is based in Bulawayo (South Zimbabwe).

In Harare, I was lucky to meet the skilled journalists and correspondents of ZimFact, launched only 4 months prior to the election, and to train them on how to use Check, open-source investigation techniques, and fact-checking processes. Over the course of the elections, ZimFact produced a series of high-quality fact checks, and Check was the very first fact-checking tool to be embedded in their newsroom workflow.

In Bulawayo, I had the pleasure to introduce the CITE team to fact-checking tools, resources, and debunking techniques. We fact-checked live examples during the training. We also worked in collaboration to customize a verification list to fact-check claims during the voting day.

When I arrived in Zimbabwe, election-related propaganda was all over the media. As an Egyptian, I am all too familiar with this kind of information ecosystem: Mainstream media is owned by the state, and its goal is simply to amplify and spread one side of every story — that of those in power. Wandering the streets of Harare, it was impossible not to notice the dominance of Zanu-PF posters and billboards. Zimbabwe’s digital streets — Facebook, Twitter, YouTube, were similarly flooded with pro-Zanu-PF content: memes, articles, and videos all designed and spread to support the long-time ruling party.

Zimbabwe’s media coverage of the run-up to general elections from May 30 to July 22, 2018 was largely biased and partisan; below are the results according to a study by local organisation, Media Monitors.

Top 10 political parties covered by the media (Source: Media Monitor)

Types of misinformation

The team of 20 journalists, reporters, and citizen journalists of ZimFact and CITE worked over the days leading up to the election, and on the day of the election itself, to debunk online mis- and disinformation.

On voting day we agreed to use a six-step process of verification that established the source of the reports, the place of occurrence, the time when the incident happened, type of violation, corroborating content from other sources, and the authenticity of any media shared with the report.

“We identified nearly 80 claims posted on social media during and after election day. We found that one-third of the claims we tracked were false, and another one-third were verified as true. The rest we found to be either misleading or inconclusive, meaning that there wasn’t enough information to determine the accuracy of the claims.” — Combatting misinformation online during Zimbabwe’s election. SEAN NDLOVU, Co founder at Center for Innovation & Technology (CITE)

Here are a couple of examples we observed during the voting period.

  • Doctored front pages of print newspapers

A manipulated copy of a billboard purportedly published by Zimbabwe’s Herald newspaper circulated widely on social media, saying Vice President Constantino Chiwenga had sabotaged President Emmerson Mnangagwa ahead of the elections.

  • False Context
    A picture of a crying soldier was circulating on Twitter the day after elections, using the reverse image feature on Check, fact-checkers found it was taken on October 27, 2017, in Nairobi, Kenya.

The day before the elections, pictures of ballot papers was heavily shared on FB, Twitter, and WhatsApp. By following basic visual verification steps, we found a ‘SAMPLE’ stamp on the ballot papers shared.

  • Staged Video
https://checkmedia.org/zimbabwe-election-2018-fact-checking/project/1507/media/16364

A video claiming to show Nelson Chamisa accepting money from Grace Mugabe at an undisclosed hotel in South Africa was circulated on FB pages and groups heavily during the elections (20k views), and screenshots of it were circulated through WhatsApp groups.

The video wouldn’t be verified by the team during the elections There’s no silver bullet for verification, the video status stayed inconclusive as we couldn’t verify the video during the voting day, here’s the reasons we think that the video might be staged; the video is in poor quality (and has no color), we can’t see any faces in the video, there’s no sound, and there is no video time code to indicate the time and date. Most hotel CCTV cameras have this feature.

Audio messages on WhatsApp: local language misinformation

During our collaboration, we observed a high volume of misinformation circulating via WhatsApp, which is a particularly popular platform in Zimbabwe due to its relatively low cost of use. Particularly interesting was that misinformation was circulating in Shona. One topic of misinformation was about the way of the ballot papers were being folded.

Recommendation and conclusion

We observed a spike in mis- and disinformation in the tense days that followed the contested elections. The lesson that we will carry forward into future initiatives is to extend the monitoring and fact-checking efforts to cover the days and weeks following elections, particularly when the result is in clear dispute and social tensions are running high.

We are grateful to the ZimFact, CITE and Carter Center teams for their efforts on this pilot initiative and look forward to continued collaboration in future.

*Special thanks to @Tom_El_Rumi for editing this blog.

To request access to Check, get in touch here.