What CSMaP Experts Are Watching Ahead of the 2024 Election: Part Two

This is the second article in a series on the 2024 election. Click here to read part one.

In this year of elections, many worry that a variety of factors could undermine the integrity of the electoral process, including advances in generative AI, a surge of disinformation and foreign influence campaigns, threats to election officials, and declining trust in media and other democratic institutions.

While it’s too early to rigorously research and understand the impact of these concerns, our scholars and experts — past, present, and future — are closely watching several issues in the 2024 elections. 

This article, the second in a series, highlights several themes that emerged from our team.

Foreign Influence Campaigns

With nearly half the world’s population eligible to vote this year, politicians, electoral monitors, and civil society groups have raised concerns that foreign interference campaigns will again try to influence public opinion and spread disinformation on social media. Sarah Graham, CSMaP’s Research Operations Manager, worries some of these fears have gone too far.

Before 2016, when Russia tried to interfere with our elections, politicians and other elites often downplayed the risks of foreign influence campaigns. But now they’ve swung the other way, often exaggerating and overhyping the effects of online interference and disinformation without solid rigorous empirical evidence or research to support their claims.

While the threat of online foreign interference is indeed real, as demonstrated by research examining the impact of Russia’s campaign on Twitter in 2016, the extent and consequences of these threats need to be rigorously examined through evidence-based research. In the coming months, I’m watching how elites discuss and amplify suspicions of interference and disinformation in electoral processes. In doing so, they risk inflating perceptions of online interference beyond actual levels, potentially fostering more discord and distrust in our public discourse.

To counteract this hype, it is crucial to have access to large-scale datasets held by social media and messaging platforms, so I’ll also be watching how these platforms share information with the public, researchers, policymakers, and the media. Their datasets can enable researchers to conduct thorough investigations that accurately measure the impact of online interference, thereby avoiding both overestimation and underestimation of the threat.

Messaging Apps & the Global South

Given that this is a huge election year in the Global South (India, South Africa, Mexico, and local elections in Brazil, to name a few), CSMaP Faculty Research Affiliate Tiago Ventura is interested in misinformation circulating through messaging apps, most notably WhatsApp, and its effects on political attitudes and behaviors. 

WhatsApp is not like a traditional social media platform, such as Twitter and Facebook, which serves users content in a feed based on a mix of algorithms and who they follow. But even though WhatsApp doesn’t have a news feed, most people in the Global South report getting their daily news through the app. 

Political content circulates on WhatsApp mostly through users sharing it across groups, which gives a condition of quasi-anonymity to most viral political content (including misinformation). This makes the kind of networks people are embedded in very important and gives researchers many interesting things to look at.

I currently have in the field a set of deactivation experiments with WhatsApp users in India and South Africa, and will do the same in Brazil in the Fall, to understand the effects of exposure to misinformation on the app before elections on polarization, misinformation beliefs, and some other political outcomes. I am also developing a data donation pipeline to collect group-based data in the app.

Social Media Research & Data Access

Finally, in order to understand the impact of digital media on elections, researchers must have access to social media platform data. But this has become much more difficult since the last election, with several platforms reducing data access. Twitter/X, for example, previously allowed researchers to collect up to 10 million tweets per month using the Academic API. But Elon Musk shut down this free program soon after acquiring Twitter, and now charges academics thousands of dollars a month. 

Access to the API is not the only limitation to observing what’s happening on the platform, however, according to Christopher Barrie, who will join CSMaP this fall as Core Faculty.

On June 12, X made “likes” anonymous. According to Musk, this led to a large increase in recorded likes. One interpretation of this change is that users are liking content they otherwise would not have liked publicly. There are two potential consequences of this. 

First, we know that liked content receives an algorithmic boost and is consequently recommended to more users. If content that otherwise would not have been liked receives this boost, this has the potential to turn up the volume on more contentious and disagreeable content. Second, the invisibility of likes makes it more difficult to detect bot activity. This is because bot farms conventionally like particular types of post in short bursty periods of time — a giveaway that the accounts engaging this way are inauthentic or automated. 

All of this is, of course, very salient to monitoring elections, understanding the kinds of information users are consuming, and fostering a reliable and (to the extent possible) civil information environment. For all of X’s talk about open-sourcing code and the importance of transparency, it is arguably now harder than ever to conduct research with data from this platform. This should be of considerable concern as we enter a likely contentious election period in the U.S.