From left, Arnout van de Rijt, Vivek Kulkarni, and Steven Skiena form an interdisciplinary faculty research team from Stony Brook University to assess that women are significantly underrepresented in media coverage.
From left, Arnout van de Rijt, Vivek Kulkarni and Steven Skiena from Stony Brook University are part of an interdisciplinary research team that has assessed the gender disparity in the media. PHOTO CREDIT: STONY BROOK UNIVERSITY.

Every few weeks, Ruchi Shah, a senior biology major, will take a look at Stony Brook-related science and research news.

A collaborative project between the computer science and sociology departments at Stony Brook University has helped quantify and define the gender disparity in the media called the “paper ceiling.”

After analyzing thousands of newspaper articles, the scientists found that male names outnumber female names five-to-one in the media.

When the scientists further categorized the names as either those of famous people whose names appear multiple times or those of obscure people whose names appear once or twice, the gender disparity was predominantly seen in famous people.

Advertisement

Overall, there was about an equal number of obscure mentions of male and female names, which is representative of the proportion of men and women in society.

The gender disparity in the media comes from the fact that most news is about famous people—those at the very top of their fields—and men still hold most of these positions in our society today.

This paper ceiling, a lack of women in the media, is representative of the glass ceiling, a lack of women in top positions that exists today.

“We live in a world in which women are catching up with men, and even outperform them, such as in college graduation rates,” Arnout van de Rijt, associate professor of sociology at Stony Brook University, said. “Yet the news is all about men, and we worry that this heavily skewed rate of coverage normalizes in the minds of audiences the notion that newsworthiness is a quality that only men possess.”

Advertisement

Another major finding of the study is regardless of whether a newspaper was liberal or conservative, or whether it is predominantly run by men or women, the gender disparity in media coverage persists.

In order to conduct the study, computer scientists created an analysis program that scoured through over 3,000 newspapers from 1983 to 2009.

The program scanned each article, extracted the text and analyzed the article, looking for any names that were mentioned.

Once names were extracted, they went through a process known as anaphora resolution, which allowed for the identification of the name’s gender.

“We have a large list of male and female names collected from the U.S. Census and the system, given a name, will look up how many times the name occurred as a male in the U.S. Census,” Vivek Kulkarni, a graduate student and co-author of the study, said. “If it almost always occurred as a male name, then the system will tag it as a male.”

Advertisement

Scientists then aggregated the counts and looked for differences between the genders by using statistical tests.

“Our large-scale computational news analysis provides a way to measure changing gender disparity over time, providing a way to look backward through history to see how things were different,” said Steven Skiena, professor of computer science at Stony Brook University.

This project is also representative of the power of interdisciplinary research.

“Past social science researchers lacking an interdisciplinary tie to computer science had had to resort to tedious human hand-coding of news sources, so this database was a huge leap forward,” van de Rijt said.

In the future, the research team said it hopes to expand the study to determine if there are differences in how men and women are talked about in the media through analyzing sentiments and language.

Advertisement
Tagged:

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.