There Is A Media Slant, And Readers Might Be Responsible Professor and economist Matthew Gentzkow, the recent winner of the John Bates Clark Medal, discusses how to predict media slant and use big data in economics.

There Is A Media Slant, And Readers Might Be Responsible

There Is A Media Slant, And Readers Might Be Responsible

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript

Professor and economist Matthew Gentzkow, the recent winner of the John Bates Clark Medal, discusses how to predict media slant and use big data in economics.


From NPR News, this is ALL THINGS CONSIDERED. I'm Audie Cornish. Your favorite newspaper has a political slant, thanks to you, the reader. That's the finding of University of Chicago economist Matthew Gentzkow. His work measuring media slant has, in part, won him this year's John Bates Clark Medal. The prestigious honor from the American Economic Association is given to the country's most promising economist under the age of 40.

In his work Gentzkow uses what's called big data, huge volumes of information he sifts through for answers. And we're joined by Matthew Gentzkow himself. Welcome to the program, and congratulations.

MATTHEW GENTZKOW: Thank you so much.

CORNISH: So let's start with this media slant research. Basically what you and your colleague at the University of Chicago found is that newspapers are swayed by market forces, right. They purposely cater to the political slant of their readers.

GENTZKOW: Yeah, so Jesse Shapiro, my colleague here, and I looked at several hundred newspapers in the U.S., tried to use some automated techniques to measure their political slant and then looked at what factors seem to drive that. One factor is catering to the readers. Others we looked at were things like the preferences of owners or influence from politicians.

And basically what we found was, you know, newspapers in Texas are more conservative because the people in Texas are more conservative. Newspapers in Massachusetts are more liberal because readers in Massachusetts are more liberal.

CORNISH: Now on its surface this might seem pretty obvious, right, like newspapers are a business, and they cater to their customers. Tell us what the economic value is.

GENTZKOW: We're trained as economists, and so we're used to looking at firms that are trying to maximize profits and thinking about their decisions. So there's a big literature in economics on how do firms choose, you know, what flavor of ice cream to sell or what kind of car to make, the characteristics of their product.

The question we started with was is that a good description of what media firms do. Of course they're trying to maximize profits to some degree, but I think there are a lot of other ideas out there that maybe a big part of the media is owners, say, trying to persuade consumers of their views, that really this economic model would not do a good job of explaining what's going on in the media.

So our initial question was if we just think of media firms like ice cream firms, how far can we go in explaining what we see? And the answer, which was somewhat surprising to us, was pretty far.

CORNISH: Now this award you've gotten, it's given to economists under 40, and you're part of a generation of economists that have really taken advantage of big data collection that your University of Chicago colleague Austan Goolsbee has described the large datasets you work with as almost unfathomable. I mean, how has this changed things for economists and the work you're able to do?

GENTZKOW: Almost unfathomable might be a slight exaggeration on Austan's part. But I think it is right that I was lucky enough to go to graduate school, to wander into this field at a time when the availability of data, the availability of technology was just exploding.

People have looked at this kind of thing in the past, and what it meant in the past was I want to study the content of newspapers, so I get some poor graduate students to sit down with a stack of dozens and dozens of issues of some newspapers and read through them, checking off every time somebody uses a certain word.

That's a time-consuming, laborious process that's really limited in how much you can do. The scale of newspapers you could look, the number of issues you could look at was just really, really limited. So having all of that text means you can look at all of the same questions on a much bigger scale and get traction on them in a way you couldn't before.

CORNISH: Is there some other dataset out there that you want to get at, some other big question you want to tackle?

GENTZKOW: There are a lot of big questions I'd love to tackle. In this area of media, I think one thing that's incredibly exciting right now is the ability to take these same kinds of questions and look at them across many different countries all over the world. So there's text data available, just as for the U.S., for China and Russia and Middle East and lots of other places where, I think to be honest, these sorts of questions matter to a degree that dwarfs how much they matter in the U.S.

Whether the U.S. media is skewed to the right or skewed to the left, we could debate about, but the role of media in what's happening in places like China and Russia I think is just enormous. So one really exciting frontier for all of us that a lot of other people, not me, have been making really exciting progress on is looking at these same issues in other countries.

CORNISH: University of Chicago economist Matthew Gentzkow, winner of this year's John Bates Clark Medal for Young Economists. Thank you so much for speaking with us.

GENTZKOW: Thank you so much, Audie.

Copyright © 2014 NPR. All rights reserved. Visit our website terms of use and permissions pages at for further information.

NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.