Reports & Analysis

The Digital Information Environment & Global Elections

National survey data from the 2016, 2020, and 2024 elections shed light on how candidates' campaign strategies impact voter policy recall.

Deactivating multimedia on WhatsApp in Brazil consistently reduced exposure to online misinformation during the pre-election weeks in 2022, but did not impact whether false news was believed, or reduce polarization.

Results from the American Institutional Confidence poll's second wave show that the public's confidence in technology, and tech companies, has markedly decreased over the past five years.

That could affect their votes — and their safety from covid-19.

We find that YouTube’s recommendation algorithm does not lead the vast majority of users down extremist rabbit holes, although it does push users into increasingly narrow ideological ranges of content in what we might call evidence of a (very) mild ideological echo chamber.

Republican candidates have dramatically increased how much they share from unreliable sources in just two years.

Our research finds that after a prominent male media personality targets a female journalist, the prevalence of hateful speech targeting those journalists increases in the immediate aftermath, often taking days to decrease. 

She’s gaining followers and ‘likes’ on other social media platforms, our research finds.

Twitter’s decision to deactivate trending topics in Ethiopia did not reduce the volume or toxicity of tweets about the civil war.

Our research suggests conservative politicians are ‘ratioed’ more often, which may explain why they’re in your timeline.

Those who dislike Democrats and don’t mind white nationalists. That includes plenty of Republicans with college educations.

This explains how social media can both weaken — and strengthen — democracy. Groups opposed to fundamental tenets of liberal democracy also have found their megaphone.

Taking Twitter data, we want to better understand which issues matter most to voters in the Georgia Senate election. We find that voters pick up on topics mentioned in attack ads and that voters tend to not view this as a national election.

Continuing previous work, we find that politically motivated popular users have the most influence in online discussions around Black Lives Matter and Common Core State Standards.

Twitter put warning labels on hundreds of thousands of tweets. Without a hard block, tweets continue to spread — especially tweets by President Trump.

We examined three recent tweets from the president and found that harmful speech on Twitter spiked in their immediate aftermath, but soon returned to baseline.

Now, there are ways to find out. New data shows that machine learning can identify content created by online political influence operations.

To better understand how an issue becomes politicized over time, we examine Black Lives Matter and Common Core State Standards and track how they evolved over time.

A new study shows how states are adapting classic propaganda tactics to social media, and why policymakers must consider how information spreads across platforms to protect voters from these covert campaigns.

In 2020, even small amounts of fake news about the coronavirus can have dire consequences. Unfortunately, it seems quite difficult for people to identify false or misleading news.

As the largest Democratic primary field in modern history continues to narrow, we wanted to examine how the debates drove the conversation online, and specifically how voters changed their opinions over time.

In the aftermath of a violent terrorist attack in Sri Lanka, the government shut down access to social media sites, with widespread implications.

In general, people don't tend to share a lot of links to fake news websites, but those that do are more likely to be older and more politically conservative.

We find that IRA-operated Twitter accounts shared less junk news than one might have expected — relying instead on local news sources.