Blog post by: Juliana Viola, Intern at Affectiva
Today in the US and around the world, women are undeniably underrepresented in politics. American women make up just 19.4% of Congress and 24.9% of state legislators. Globally, just ten women serve as head of state and nine as head of government. This lack of diversity brings huge consequences; time and time again, studies have documented how diversity can spur workplace innovation and boost productivity. Therefore, increasing the representation of women, specifically women of color, in government offices would likely lead to a more effective government.
Along the same vein, as an avid news junkie, I have often noticed homogeneity in the panel discussions I watch on TV. Political panels in particular are often comprised of mostly men. I wondered how I could capture metrics about how panel members emote and participate in the discussion, and how these metrics might vary by gender. For example, how is airtime split between men and women? Since the American public relies on political talk shows for perspective, these panels would ideally represent a diversity of voices to interpret objective information.
I also wondered about the emotional climate of these debates. Which emotions are most prevalent in news debates, and what might this data suggest about the state of American politics?
In order to answer the above questions, I downloaded a few thousand publicly available videos of news panel debates, using search terms like "Fox News panel discussion," "George Stephanopoulos Powerhouse Roundtable," and "Meet the Press panel" to generate the results.
In order to obtain emotion metrics about each video, I uploaded all of the videos to Affectiva's Emotion as a Service platform. These emotion metrics are estimated by automatically analyzing the facial expressions of the panelists in the video.
Next, in order to analyze how verbally assertive men versus women are, I examined how quickly male versus female panelists grabbed the floor following a pause in the conversation. To do this, I removed the audio from the video files, and processed the audio using Affectiva's speech-based gender estimation model, which is able to predict the gender of the person speaking in an audio file. I performed this analysis on a subset of my entire data set: a collection of CNN debates and a collection of Fox News debates.
For the full article and relevant charts, go to: