Marvelous AI Teams Up with Scholar Sarah Oates to Track How Twitter Commentary Disadvantages Female Candidates in the 2020 U.S. Primaries 

October 22, 2019 — Researchers found that female candidates in the 2020 U.S presidential primary election faced significantly more character and identity attacks online than their male counterparts. Sarah Oates, Olya Gurevich, Chris Walker, and Lucina Di Meco spent the last several months tracking discourse on social media. Over this period, they studied trends in how different Democratic candidates in the 2020 U.S. Presidential primary election were being talked about on Twitter. Using data from Marvelous AI StoryArc, the team identified and categorized different social media narratives based on candidates’ policy, ideology, character, identity, and electability. 

The team believes understanding trends in social media narratives in the 2020 U.S. primaries can help better inform voters. “Social media narratives drive a lot of how people think about candidates. And how people think and feel about candidates is important in the voting process,” said Oates.

“What it means to be a democrat and republican now is much more fluid and changeable than it was 20 years ago. That means media are more important, and social media are particularly important because they engage us on a much more emotional level.” Oates continued. 

The researchers believe machine learning is extremely helpful in discovering and tracking tracing online narratives. Marvelous AI is committed to a combination of human content analysis and artificial intelligence for the identification and trending of narratives.

“Innovation in terms of tagging and tracking narratives is so critical…We should try to make important aspects of online conversation visible,” Said Oates. 

Marvelous AI StoryArc can track familiar narratives of misogyny online. 

“We know that women are asymmetrically attacked and denigrated on the online hemisphere. This research makes that clear and transparent.” Said Oates

Download the official American Political Science Association paper