Skip to content Skip to navigation

Matthew Gentzkow on the media, political polarization and the power of data

July 14, 2016

“Illegal alien” or “undocumented worker?”

Mere words, yes. But their connotations are so ideologically loaded that economist Matthew Gentzkow sees a window to understanding: when combined with today’s data-analysis technologies, the use of one phrase over the other reveals a great deal about right- or left-leaning biases.

Developing meaningful methods to quantify weighty words is one of the innovative specialties that Gentzkow has honed to explore the economics of the media industry – the extent of political slant, the impact of media on politics and society, and vice versa.

Gentzkow, a senior fellow at Stanford’s Institute for Economic Policy Research, is known for capitalizing on text mining and automated content analysis to expose the interplay between media and politics. He has probed prickly questions that few economists of prior generations have ventured to deeply investigate, partly because of the once cumbersome task of manually sifting through paper records — work that can now be done by computers.

As people wondered if Americans were becoming more ideologically separated because of the growing amount of diverse Internet sites, Gentzkow found an answer: No.

Turns out, fears that Americans would become highly segregated media consumers by staying in ideological “echo chambers” have been overstated, according to the widely cited study Gentzkow authored with Jesse Shapiro, an economist he frequently collaborated with while both were at the University of Chicago.

“The bulk of my work is motivated by trying to understand media markets more broadly,” Gentzkow says. “There's an intersection of economic forces and economic incentives with some really important political and social effects.”

Gentzkow’s ambitious, data-driven research in the media arena has cast him into the upper echelons of economics. He was awarded the John Bates Clark medal in 2014 for being the top American economist under the age of 40. The honor is considered second only to the Nobel Prize in the field.

“Having more data is very powerful and very useful and makes everything easier, but there’s still the hard work of identifying either randomized experiments or good natural experiments,” he says. “And that is typically the limiting factor, the bottleneck.”

His findings have landed at perfect times, providing valuable insights just as the media industry was struggling amid the new world of the Internet and online advertising. The economic dynamics of the evolving media industry, he has found, shares similarities with the birth of the print-dominated media landscape from a century ago.

And in analyzing daily newspapers dating back to 1869, Gentzkow has found that while the existence of newspapers increases voter participation, the political slant of newspapers surprisingly does not affect the share of votes for a particular political party. Newspapers, in other words, do not seem to lead to polarization.

There is, however, an expanding usage of partisan language in media and political speeches, Gentzkow says. It’s a modern phenomenon he plans to explore further.

“If you tune into Fox News or CNN, you’ll immediately notice how people on the right or left talk very differently. Democrats will talk about undocumented workers and Republicans talk about illegal aliens,” he says. “It’s very clear those partisan differences are big today and it’s also clear that those differences are deliberate.”