2011 – 2013
Do major US newspapers cover international news less than they used to?
What if we could empirically answer questions about our media in realtime?
These are the driving questions behind Ethan Zuckerman’s Media Meter research agenda. We’re working to quantify media diets so that individuals can be more reflective and news companies themselves can be known for what they give the world — or don’t.
As part of my Research Assistant work with Ethan at MIT, I helped design and implement a distributed media research project to crowdsource tens of thousands of judgments with Amazon Mechanical Turk. I scaled up our teams’ initial media coding with crowdsourced labor to analyze 40 years of articles from four national newspapers for international coverage trends. I worked closely with Rahul Bhargava, Nathan Matias, Ethan Zuckerman, and Dan Schultz to scrape archived data and develop coding procedures.
My primary role was to design, run, and quality assess large jobs on Crowdflower, where I learned the best practices to efficiently administer huge crowdsourced jobs while maintaining quality responses. I also studied the history of nutritional labels (our metaphor of choice) and designed a video and series of concept interfaces. With Nate Matias, we presented the project at conferences like the Media Lab’s Understanding Networks and the Mozilla Festival.
Rahul, Nathan and I also built a hardware display to house this research by retrofitting a newspaper vending box from The Berkshire Eagle with an LCD screen, Arduino, and a custom-designed serial dial (right). Passerby turn the dial to explore “front pages” composed of tree graphs visualizing decades of New York Times data. (Photo via Nate Matias)
Another output is this poster Rahul designed using The New York Times data. It clearly visualizes both general trends in international and domestic coverage, as well as the disproportional news impact of specific events like invasions of Iraq, the Olympics, and March Madness: