Misinformation Amplification in the Nigerian Election

By Integrity Institute chief research officer and co-founder Jeff Allen

On February 25th, Nigeria is holding elections for their president and vice president, as well as elections for their Senate and House of Representatives. And of course, where there is a democratic election, there will be misinformation spread on online platforms to accompany it. At the Integrity Institute, we are tracking how platforms amplify misinformation in advance of democratic elections around the world, in preparation for the coming wave of democratic elections in 2024.
Monitoring how platforms amplify misinformation outside the United States is critical, as companies typically do not invest the same level of resources into protecting international elections, and it is essential that international elections be actively monitored to ensure platforms are not having an outsized negative impact on the information ecosystem. Elections organizations in Nigeria have already made calls on Facebook to collaborate more on fighting online misinformation.
We should not necessarily expect to see the same Misinformation Amplification Factors that we saw in our U.S. analysis for the 2022 midterms. Social media companies generally invest fewer resources towards integrity issues outside the U.S., which will impact the extent to which misinformation spreads and the platforms can be exploited by bad actors. Countries use platforms differently and have distinct cultural norms around them. And fact checking organizations will have varying levels of resources at their disposal to counteract misinfo. For all these reasons, it is crucial to monitor and study the amplification of online misinformation internationally, so that voters, civil society organizations, governments, and platforms can respond accordingly.
Our methodology for analyzing misinformation amplification in Nigeria is the same as what we applied around the U.S. midterms. We collect fact checks from the Nigerian based fact checkers that are part of the International Fact-Checking Network, extract all social media posts referenced in the fact checks, and label them as misinformation or accurate information, according to the fact check. We then compare the engagement (views, likes, replies, shares) that the misinformation posts got relative to the baseline engagement the uploader of the misinformation got on posts within two weeks prior to the misinfo post.
You can find all of our results on our Misinformation Amplification Factor Tracking Dashboard for Nigeria.

Findings

Because there are fewer fact checking organizations in Nigeria, which produce an overall lower volume of fact checks, we can only report misinformation amplification factors for Facebook and Twitter. Overall, we do see comparable results to what we found in the U.S. We see Facebook has a misinformation amplification factor that is significantly less than that of Twitter.
 
The average MAF for each platform, with 90% confidence intervals, is:
  • Facebook: 4.3 (2.1 - 8.8)
  • Twitter: 15 (8.9 - 24)
These numbers are fairly compatible with the numbers we saw in the U.S. midterms, although Twitter is lower. We still see examples of misinfo tweets being amplified by factors of thousands, 10,000 for the most amplified tweet. But the lower average value could be a reflection of the lower popularity of Twitter in Nigeria relative to the U.S. and the role it plays in the news ecosystem.

WhatsApp

One key difference between the U.S. and Nigeria is the importance of WhatsApp. By some measures, WhatsApp is the most popular platform in Nigeria. And this is reflected in the fact checkers work. About 16% of fact checks in Nigeria are debunking misinformation that is spreading on WhatsApp. However, because of the lack of data available on WhatsApp, it is impossible for us to estimate a MAF. It is clear that WhatsApp is a significant vector for misinformation in Nigeria, and it would be extremely helpful for WhatsApp to release data that allows external organizations to estimate the extent to which WhatsApp design is amplifying misinformation.

PageRank Still Works

We continue to use our social media aware PageRank calculation to see how well it performs in separating spreaders of misinformation from sources of accurate information in Nigeria. As in the U.S., PageRank does an excellent job. Low PageRank scores are dominated by sources of misinformation, which high PageRank scores go exclusively to accurate sources.
 
 

Next Steps

We will continue to monitor the spread of misinformation through the election in Nigeria, as well as continue to track more elections around the world throughout 2023 and 2024.
Jeff Allen

Jeff Allen is the co-founder and chief research officer of the Integrity Institute. He was a data scientist at Facebook from 2016 to 2019. While at Facebook, he worked on tackling systemic issues in the public content ecosystems of Facebook and Instagram, developing strategies to ensure that the incentive structure that the platforms created for publishers was in alignment with Facebooks company mission statement.

Previous
Previous

Let Section 230 Stay

Next
Next

Widely Viewed Content Report, 2022 Q3