Professor demonstrates that mainstream media has a liberal bias

Everyone believes that mainstream media outlets have a liberal bias, right? Well, a UCLA professor says he has actually proven it.

Furthermore, while some conservative outlets may lean to the right, their bias is less than the liberal bias of most mainstream outlets.

These are the results determined by Dr. Tim Groseclose, a professor of political science and economics at UCLA who has spent years constructing precise, quantitative measures of the slants of media outlets. His methods and conclusions are detailed in his book, “Left Turn: How Liberal Media Bias Distorts the American Mind.”

He will be speaking as the guest of The Maine Heritage Policy Center on Thursday in Portland. To register, go to http://www.mainepolicy.org.

Groseclose measures the political content of news and converts that content into an SQ, or “slant quotient,” of the outlet. To determine bias, he then compares SQs of news outlets to the PQs, or “political quotients,” of voters and politicians.

He concludes that all mainstream media outlets have a liberal bias, and that while some supposedly conservative outlets—such as the Washington Times or Fox News Special Report—do lean right, their conservative bias is still less than the liberal bias of most mainstream outlets.

Groseclose contends that the general leftward bias of the media has shifted the “political quotient” (PQ) of the average American by about 20 points on a scale of 100.

With “Left Turn,” readers can easily calculate their own PQ to decide for themselves if the bias exists. This timely, much‑needed study brings fact to this often over‑heated debate. Sound interesting? Take the quiz to determine your own PQ.

To find out what PQs are and how they reveal bias, consider this excerpt from Tim Groseclose in “Left Turn”:

“Come on. Political science isn’t really a science,” said my friend Dawson Engler one day, trying to goad me.

Engler, one of the country’s premier computer scientists, is currently a professor at Stanford, where his specialty is operating systems. He has constructed his own operating system—twice.

He is the type of person who succeeds at nearly anything he tries. Born in Yuma, Arizona, during high school he placed second in the “Teenage Mr. Arizona” bodybuilding contest. After graduating from Arizona State University, he enrolled in the highly prestigious computer‑science PhD program at MIT.

It is unusual for a PhD student to publish a paper in a peer‑reviewed scientific journal. Yet Engler published eight while a doctoral student. Shortly after Stanford hired him, for a brief period he dated one of the actresses from “Baywatch.”

When Engler goaded me, both of us held positions at MIT, and he knew that my position was in the political‑science department. At MIT, which is filled with “real” scientists and engineers, you often hear quips like Engler’s. So when he made it, I was prepared.

“Look,” I said. “We can both agree that if you can graph something, then you can describe it mathematically.”

“Yeah,” said Engler.

“And people, all the time, talk about politicians being left wing or right wing.”

“Okay,” said Engler.

“And so if a position is left wing or right wing, then you can graph it. Which means you can describe it mathematically, which means it’s science.”

Engler smiled. I don’t think I really convinced him, but he didn’t goad me any further. At least in my mind, I’d won the day’s debate.

Within political science a small industry exists to do the “science” that I described to Engler: to calculate precise, numerical measurements that describe the liberalness or conservativeness of politicians. In fact, at the time Engler made his quip, I was working on such a project. Indeed, the political quotients that I describe in this book are based on that research.

A person’s PQ is a number, generally between 0 and 100, that describes how liberal he or she is. The higher the number, the more liberal a person is. I have computed PQs for members of Congress by observing their record on roll call votes. (To compute your own PQ, go to http://www.timgroseclose.com/calculate-your-pq/)

The PQ for politicians is constructed from roll call votes in Congress. This means that simply by noting how members of Congress voted on those roll calls, I can calculate their PQs, and you can compare your PQ to theirs.

The following are the PQs of some well‑known politicians:

Michele Bachmann   -4.1

Newt Gingrich            11.4

Ron Paul                     31.8

Susan Collins             44.2

Olympia Snowe         47.9

Joe Lieberman           74

Hillary Clinton           87.6

Nancy Pelosi              100.7

Barney Frank             103.8

Perhaps the main contribution of the book is that it uses PQs to judge media bias. To do this, I conduct the following thought experiment. Suppose you were given a set of stories that a media outlet reported. But suppose, instead of knowing that they were news stories, you were told that they were speeches by a politician.

After reading the would‑be speeches, what would you guess to be the PQ of the would‑be politician?

I define the slant quotient, or SQ, of an outlet as the solution to that thought experiment. In the article that my colleague and I wrote for the “Quarterly Journal of Economics,” we developed a statistical technique that calculates a precise, numerical SQ for the 20 news outlets that we examined. We found, for instance, that The New York Times has an SQ of 74, which is approximately the PQ of Sen. Joseph Lieberman. 

The primary data that we used were citations to think tanks. This means that The New York Times’s citation patterns to left‑wing, centrist and right‑wing think tanks were very similar to the patterns that Joe Lieberman adopted when he made speeches on the Senate floor.

What is important is that we can describe numerically the political views of politicians and the slants of various media outlets. Further, we can map these two sets of numbers to the same scale.

Despite what my hard‑science friends might say, it is possible to analyze politics, including media bias, objectively, numerically and, yes, scientifically.

For more information about the book, see www.timgroseclose.com.