A program developed by Syracuse University’s School of Information Studies and the S.I. Newhouse School of Public Communications is examining a discrepancy in political opinions shared on social media.

The Share of Voice Project questions the fairness of social media pie charts in accurately projecting the popularity of presidential candidates among U.S. citizens. Jennifer Grygiel, an assistant professor and social media professional in Newhouse, and Jennifer Stromer-Galley, a professor in the iSchool, are the two main contributors to the project.

The combined efforts of educated professors from both schools allows for a more comprehensive undertaking of the Share of Voice project, the two professors said.

“I am always looking for opportunities to join forces with colleagues from other schools at SU, and (Grygiel’s) idea was an important one,” said Stromer-Galley, who is also director for the Center for Computational and Data Sciences.

Grygiel said the main goal of the project for her is to help educate journalists on new media and specifically social media data and how it works.

As Election Day draws closer, voters have taken to social media and filled it with their political opinions. People have, for example, been using the hashtags of election slogans such as #ImWithHer and #MAGA, short for “Make America Great Again”, to stand with Democratic presidential nominee Hillary Clinton and Republican presidential nominee Donald Trump, respectively.

But the quantitative data content of these postings aren’t always the most accurate or representative of how the nation as a whole feels about the presidential election, Grygiel pointed out.

Grygiel and Stromer-Galley are striving to reveal to voters what the gathering of social media data really means by exposing its potential bias with the Share of Voice project.

“Twitter produces its own Share of Voice metric, but it’s a black box, meaning we have no idea what goes into their measure,” Stromer-Galley said. “Professor Grygiel wanted to create a metric that was transparent so that consumers of news and this campaign could better judge and evaluate it.”

A big aspect of the project, Grygiel said, is the fact that viewers of the presidential debates are misinterpreting this data to be an accurate representation of U.S. citizens’ political views.

The project illustrates the common theme of how statistical and social media analysts can easily manipulate U.S. citizens just by the use of quantitative descriptions of social media activity without mentioning potential bias, Grygiel said.

“I was noticing around the debates that social media was being presented in a way that was misleading and people were seeing it more as poll-level data instead of the raw social media that it is,” Grygiel said.

She added that the Share of Voice data is generated around key words.

“I wanted to make sure people understood how it was created so that they could see that there is potential for bias and the need for transparency in this,” Grygiel said.

Published on October 17, 2016 at 10:03 pm

Contact Kayla: kseiberl@syr.edu