WhatsApp posts shared by political parties in India do not reveal a significant amount of misinformation or hate speech, according to a Rutgers Study.

The study, “What circulates on Partisan WhatsApp in India? Insights From an Unusual Dataset,” by SC&I Assistant Professor of Library and Information Science Kiran Garimella and Simon Chauchard, associate professor of Political Science and Distinguished Researcher at the University Carlos III (Madrid), was published in the Journal of Quantitative Description: Digital Media. 

JQD:DM named Garimella and Chauchard inaugural winners of the “Best Paper Using Data from the Global South” for the paper, noting the awards were drawn from the collection of papers they published through the end of 2023.

Their findings surprised them, Garimella said, because they expected to find more discussions about local politics as well as political misinformation while analyzing images posted by Whatsapp groups managed by political parties in India (they examined images only because images made up the majority of their data).

“What we found, Garimella said, “is that most of the content is not political. It’s about entertainment and local events the political parties were organizing, such as events in the village, as well as random selfies of people gathering together. So even though these are WhatsApp groups created and managed by political parties, we found only 20% of the content related to politics. That was very surprising to us because it seemed to defeat the purpose of the groups. However, our hypothesis is that these groups are not functioning as a tool for spreading propaganda but as a tool for organizing and creating a sense of community among the local party workers.”

JQD:DM named Garimella and Chauchard inaugural winners of the “Best Paper Using Data from the Global South” for the paper, noting the awards were drawn from the collection of papers they published through the end of 2023.

They were also surprised, Garimella said, that they did not find more misinformation and hate speech. Garimella said they determined that 2% of the content shared misinformation and 3% contained hate speech.

There are caveats, Garimella noted, explaining that if he and Chauchard had examined overall WhatsApp content, not just images, instances of misinformation and hate speech would have been much higher, particularly regarding content shared about Muslims.

It’s important to note, Garimella said, that even though the levels of misinformation and hate speech they found were rare, these posts are still harmful.

It’s also important to note, Garimella said, that even though the levels of misinformation and hate speech they found were rare, these posts are still harmful. “While we did not find much hate speech and misinformation in these WhatsApp groups, this does not mean this is not a problem on Whatsapp. It is especially problematic in this context because the misinformation is being sent from a source the recipients trust, so they are more likely to believe and act upon it. So even though it’s rare, it’s impactful.”

Garimella said he and Chauchard utilized two methodological innovations to gather the data. The first is the novel on-the-ground survey they conducted. They visited political party offices in India and asked party officials if they would add them to their WhatsApp groups.

“In India whatsapp is huge in politics,” Garimella said. “We were added to 500 groups, and we pulled the data from there. The methodology we used by combining online data collection with offline surveys is a huge novelty and really gave us new insights.”

The other methodological innovation they employed was collecting the type of personal data they did – chat data – on WhatsApp.  

Garimella and Chauchard utilized two methodological innovations to gather the data. The first is the novel on-the-ground survey they conducted. The other methodological innovation they employed was collecting the type of personal data they did – chat data – on WhatsApp.  

“The study provides a basis for which others can study a platform like WhatsApp in an ethical manner by getting permission from the participants,” Garimella said.

The broader context for undertaking this research, Garimella said, was that he had conducted some earlier work looking at the WhatsApp group ecosystem of Indian political parties and found that these political parties use WhatsApp very extensively. But at that time, he could not gain internal access to such groups.

“Our findings may not generalize beyond this population,” Garimella said, “because even though it’s a large sample it’s a biased sample. This is because we depended upon the participants to add us to their groups. There might be bias in that – such as perhaps they did not add us to the really spicy groups where all the hate speech is. We don’t know that. But our findings still give us a window into how Whatsapp groups are managed by political parties in India, as well as some hypotheses on the role of WhatsApp and social media in elections and politics, not just in India but across the world.”

In the future, Garimella said, he and Chauchard plan to improve upon the sampling they did for this study to continue to analyze how WhatsApp is used in elections in India.

Learn more about the Library and Information Science Department on the Rutgers School of Communication and Information website.

Image source: “What circulates on Partisan WhatsApp in India? Insights From an Unusual Dataset,” Used with permission from Kiran Garimella.