February 28, 2010
What drives media slant?
University of Chicago professors Matthew Gentzkow and Jesse Shapiro propose an answer.
The study by Gentzkow and Shapiro has been around for a while in working paper form, though I only came across it on Friday when the latest issue of Econometrica in which it was finally published arrived in my mailbox.
Gentzkow and Shapiro propose to measure the slant of a particular newspaper by searching speeches entered into the Congressional Record and counting the number of times particular phrases were used by representatives of each party, mechanically identifying phrases favored by one party over the other. For example, a Democrat is more likely to use the phrase "workers rights" whereas a Republican is more likely to use the phrase "human embryos". They then counted the number of times phrases of each type appeared in a particular newspaper to construct an index of the political slant of that newspaper. The Gentzkow-Shapiro index of slant (shown on the vertical axis in the diagram below) has a reasonable correlation with subjective measures such as ratings assigned by users of Mondo Times (horizontal axis). For example, both measures agree that the Washington Times is one of the most conservative papers and the Atlanta Constitution is one of the most liberal newspapers.
Gentzkow and Shapiro then asked what other factors help explain a newspaper's slant. They found that the most important variable is the political orientation of people living within the paper's market. For example, the higher the vote share received by Bush in 2004 in the newspaper's market (horizontal axis below), the higher the Gentzkow-Shapiro measure of conservative slant (vertical axis).
On the other hand, the politics of the paper's owner seem to matter much less. For example, once one controls for geographic factors, there is no statistically significant correlation between the newspaper's own slant (horizontal axis below) and average slant of papers in other communities owned by the same owner (vertical axis).
Gentzkow and Shapiro conclude that papers to some degree are just giving their readers what the readers want so as to maximize the newspapers' profits.
One detail I'm curious about is how slant gets implemented at the ground level by individual reporters. My guess is that most reporters know that they are introducing some slant in the way they've chosen to frame and report a story, but are unaware of the full extent to which they do so because they are underestimating the degree to which the other sources from which they get their information and beliefs have all been doing a similar filtering. The result is social networks that don't recognize that they have developed a groupthink that is not centered on the truth.
Peter Gordon marvels that
There are intelligent adults in the world who go to bed each night believing that the other side tells lies, but their side is above all that.
Posted by James Hamilton at February 28, 2010 07:03 AMdigg this | reddit
What of national media such the mainstream media? I can totally understand the above results but how some networks continue to spew the same thing when their market is almost non-existent seems to beg the question concerning "giving the people what they want." Since FOX News is blowing all of the networks away you would think that the networks would be moving toword their format if they were appealing to a larger audience. It appears that their is another dynamic at work in this national media separate from the dynamic of local newspapers.
I would like to see results for national media that seems immune from customer preference.
Posted by: RicardoZ at February 28, 2010 07:48 AM
"Gentzkow and Shapiro conclude that papers to some degree are just giving their readers what the readers want so as to maximize the newspapers' profits."
From my perspective the above statement is correct. I've worked in the media all my life. The ongoing collapse of traditional media's business model will further accelerate the dissemination of infotainment.
Posted by: Rickk at February 28, 2010 10:06 AM
I find it interesting that the median score seems to be about 0.45, and even famously conservative newspapers like the Wall Street Journal still lie on the Democratic side of the fence (eyeballing the graph, it looks like the WSJ is around 0.48).
I guess there's something to this whole "Mainstream media bias."
I wonder how various blogs would score on this index. Or talk-radio programs.
Posted by: Jon at February 28, 2010 10:17 AM
There sure seems to be something of a node of a dozen significant newspapers (eyeball estimate) clustered about the 'conservativeness rating' = 2, 'slant rating' = .44 area. The graphic as shown cannot be said to disprove chronic media bias, that's for certain.
Posted by: persiflage at February 28, 2010 11:26 AM
"Gentzkow and Shapiro then asked what other factors help explain a newspaper's slant. They found that the most important variable is the political orientation of people living within the paper's market."
This doesn't explain the New York Times and the Wall Street Journal or New York Post in the same city or the Washington Times and the Washington Post in the same city.
"On the other hand, the politics of the paper's owner seem to matter much less. For example, once one controls for geographic factors, there is no statistically significant correlation between the newspaper's own slant (horizontal axis below) and average slant of papers in other communities owned by the same owner (vertical axis)."
Rupert Murdoch? London Times, Sydney Daily Telegraph, Melbourne Herald Sun, New York Post, Wall Street Journal -- same slant worldwide.
Posted by: Joseph at February 28, 2010 12:07 PM
One of these days I'll hear someone say: "The media is supposed to be biased, biased toward the evidence."
The day I see an analysis on a news media's bias toward the evidence is the day I'll have some respect for that analysis.
As a final point, I'd say that the fact that there is a 'slant' which corresponds to the paper's market tells us something interesting. Primarily that: Newspapers are untrustworthy sources of information that tend to tell the readers only what they want to hear.
Posted by: S.R.Barbour at February 28, 2010 12:18 PM
Are the newspapers giving their readers what they want, or does a long term exposure to the newspapers create the readers? And what does the existence of the Washington Post and the Washington Times in the same area mean for the correlations? And Obama's home paper is more conservative than McCain's?
Posted by: Bill Harshaw at February 28, 2010 12:52 PM
The Washington Times is Based on Washington D.C. isn't it? Thus disproving it feeds it's market. No, more likely It's CRACKPOT FOREIGN ownership is the issue, just like the Saudi-Foreign owned FOX news network, which I assume is such an outlier, it's off the graph.
Posted by: MikeRINO at February 28, 2010 12:56 PM
The question is raised - how do reporters end up matching their writing to the political leanings of the local audience? How's 'bout through their editors? Both when assignments are handed out and when the reporter gets a gander at what the editor has done to the initial text, the reporter will get an indication of what is wanted. In addition, we need to consider that the first test, comparing the frequency of certain expressions to the politics of the readership, has a passive component. Newspapers will cover local politicians, and if local politicians are of a particular stripe, a neutral editorial bias will turn up more expressions on one or the other side of the political spectrum, simply because the politicians they cover cough up those expressions more often.
I think you have a good point about newspapers creating a bias in readers, but I doubt it's a case of one-way causation. Rather, it is a self-reinforcing arrangement, in which the newspaper reflects the community, and the community's bias is hardened by being reflected. Note also that readers of the newspaper will also drink in the same bars and attend the same churches are the rest of the community, so that newspapers will be but one influence on local political bias.
Faux News isn't a newspaper, so it doesn't qualify to be in the study. You want to look at the WSJ, which is owned by the same guy as Faux. WSJ is both way to the right on the graph, and above the line. It is more "biased" by the expression-use measure than seen by its subjective raters.
Posted by: Anonymous at March 1, 2010 05:05 AM
One thing not clear from the study description above is whether editorials and op-eds are included in the key-phrases search. For example, the Wall Street Journal editorial page is famously conservative, but the news pages' bias is liberal, if anything. Susan Faludi, a feminist firebrand (author of Backlash) and Sharon Begley, now science editor at the NYT, both got their start at the WSJ. Other examples aren't hard to find. How many NYT reporters went on to write pro-market or conservative books?
Posted by: emerich at March 1, 2010 08:26 AM
This Democratic versus Republican dichotomy is so passe. How about something that gets at the colonial attitudes of various media and the willingess of media to present or tolerate a wide variety of news and analysis? Something that tries to sort out willingness to pay in blood for foreign booty?
In Canada, the controversial Asper-owned media empire is in bankruptcy proceedings. The Winnipeg-based conglomerate went too far, too fast in acquiring new assets. That's the primary reason. The Aspers also vigorously supported Israel's efforts to colonize more territory and resources in what-is-left-of-the-Palestinian mandate by filtering embarrassing and otherwise negative news and by only publishing pro-Israeli opinion-editorial pieces.
I'm sure that there are many people like me who would have contributed op-ed material to pro-business, pro-market newspapers but didn't because the anti-pluralistic editorial policy displeased us. I tend to think that some suppliers, advertisers and subscribers reacted in a similar fashion.
Posted by: GNP at March 1, 2010 08:58 AM
Rather than media having a liberal bias, this would mean the public has a liberal bias reflected in the media. I assume the reason this is called the conservativeness rating is conservatives focus on fewer issues but more devotedly and more often use phrases crafted for signaling making their phrases more indicative, but it also suggests liberals may be more diverse in their interests which may appeal to more diverse audiences so equating this to bias may be overly strong. Since they would more likely report on what their local representatives would be saying, this is another reason to be skeptical of calling this bias.
Posted by: Lord at March 1, 2010 09:41 AM
There are major cause and effect issues involved in measuring a paper's bias by its market, particularly, when the measure seems to be based on the particularly paper's readership, rather than the geographic area in which the paper is sold.
Surely, no one is surprised that the Washington Times has a more conservative readership than the Washington Post that serves the same geographic area. But, the conclusion that readers tend to choose the newspapers most aligned with their beliefs is vacuous in multi-newspaper markets.
Posted by: ohwilleke at March 1, 2010 10:30 AM
There is an older and very nice theoretical paper by Mullainathan and Shleifer, "The Market for News", (A.E.R. Sept 2005) which makes the same points, and a far, far shorter and less elegant "treatment" by me: http://serendip.brynmawr.edu/local/scisoc/emergence/resources/kuperberg18jan06.html
Posted by: M Kuperberg at March 1, 2010 10:32 AM
Some people have made the comment that the results don't make sense when you have multiple providers in the same area providing different viewpoints. I would disagree.
I think its a classic example of the Nash equilibrium. Given that the FNC has the largest viewership among those on the right nationally, it makes sense for MSNBC to go after the left rather than try and chip away at FNC's market share. The same would apply for the Washington Times and Washington Post.
I would bet that the paper's conclusion is more true when there is a single provider in a geographic area. In that situation the best strategy would be to give the majority what it wants. As soon as additional providers come into the market, it would be in there best interests to pursue the minority viewership.
Posted by: Dustin Cannon at March 1, 2010 11:03 AM
What's a newspaper?
Posted by: DWhite at March 1, 2010 12:01 PM
Posted by: DWhite at March 1, 2010 12:01 PM
"What's a newspaper?"
Posted by: Rickk at March 1, 2010 01:08 PM
Is this not more a story about how to "measure" slant rather than "what drives" it? The latter sounds more like a causal story, why does a particular newspaper have a particular slant rather than simply how to measure it. Surely it is not the use of Dem or GOP catch phrases of the moment that causes the slant. Rather these usages reflect what the slant is.
Posted by: Barkley Rosser at March 1, 2010 02:16 PM
Also worth noting that the fit isn't very good.
Posted by: ohwilleke at March 1, 2010 02:29 PM
So which came first? The newspaper or the reader. I suspect I know which causal link Galbraith would have made.
To read my local rag and the many AP articles recently about Toyota, one might suspect the company is now dead and forgotten, a victim of poor quality and a management that cares more about profits than its customers' lives.
The average journalist probably knows far less about any particular subject being written about than the average public school teacher has in the subjects being taught in class.
"My guess is that most reporters know that they are introducing some slant in the way they've chosen to frame and report a story, but are unaware of the full extent to which they do so because they are underestimating the degree to which the other sources from which they get their information and beliefs have all been doing a similar filtering."
Ya, and that covers a lot of ground. I love reading the daily explanations for stock market movements. Reading the articles gives the impression the journalist knows the absolute reason for any movement but, a cursory examination of such articles would give the impression that there are as many definitive reasons as sources.
Posted by: Hitchhiker at March 1, 2010 02:35 PM
Hitchhiker wrote at March 1, 2010 02:35 PM:
"So which came first? The newspaper or the reader."
The dollar came first. The media sells information/infotainment like McDonalds sells hamburgers, it is no different.
"To read my local rag and the many AP articles recently about Toyota, one might suspect the company is now dead and forgotten, a victim of poor quality and a management that cares more about profits than its customers' lives."
Live long enough and you'll learn that profits are valued above human life. It's just not something they are keen to talk about.
"The average journalist probably knows far less about any particular subject being written about than the average public school teacher has in the subjects being taught in class."
"My guess is that most reporters know that they are introducing some slant in the way they've chosen to frame and report a story,..."
"... but are unaware of the full extent to which they do so because they are underestimating the degree to which the other sources from which they get their information and beliefs have all been doing a similar filtering."
They don't think that deep.
"Ya, and that covers a lot of ground. I love reading the daily explanations for stock market movements. Reading the articles gives the impression the journalist knows the absolute reason for any movement but, a cursory examination of such articles would give the impression that there are as many definitive reasons as sources."
Astute observations. My compliments.
Posted by: Rickk at March 1, 2010 04:19 PM
The media's slant? MONEY YOU FOOL! They chase the money, and help those that give them the money.
Check out the number of drug and insurance ads on any Sunday morning talk show or other media that has high numbers of old people as customers (i.e. print newspaper). Now with those people and other Wall Street corps paying the bills of the media being consumed, do you think you'll hear a lot of anti-corporate, pro-insurance reform points of view from those places?
"It's all about bucks kiddo. The rest is conversation." - Gordon Gekko.
Posted by: Jake M. at March 1, 2010 05:44 PM
I'm not sure what media bias is if we are talking about reporting.
Is it leftist/liberal to say 1=2, right wing/conservative to say 2=3, and moderate to say 2=2?
When I was a loyal WSJ reader of several decades ending a decade plus ago, the reporting was not liberal, conservative, or moderate, but informative, factual, and instructive.
I began to see a unique character to many articles on financial matters which were almost tutorial, ensuring the reader had the grounding in such things as options, shorts, puts, and many other more and less complex concepts. For an article on the deadlock on legislation, I'd expect WSJ articles to provide historical context for statements quoted on such things as filibuster and reconciliation. I would have been shocked to see an article report without such explanatory context the quotes like "they are trying to ram this unpopular bill through Congress", pointing out both the normalcy of the legislative procedural process, and the actual broad acceptance of all the major provisions.
The WSJ editorial page was clearly slanted, but the reporting was not. That was in a day when the WSJ was not even trying to be your local news or even world news other than business, but very much devoted to informing its readers about the changes in the world.
I have long liked science and technology reporting, so I'm a fact centric guy.
I find the kind of reporting common today everywhere to be more along the lines of personality driven, with facts included to illustrate the portrayal of the personality. I know this has long been a part of the news, but it was less favored when I was younger, and thus a firmer line drawn between the reporting in each camp.
It seems to me that the great change has been to shift the reporting from facts to personalities, and personalities are almost always biased. The way you describe the way a person thinks reveals your hopes and fears about them.
In the WSJ of two decades ago, I'd expect articles to go into the definition of liberal, left, right, conservative, socialist, and so on, as a means of trying to provide an objective basis for interpreting what is now common speech. In my view, most of the political speech is empty of any meaning, in that you can not identify the beliefs or policies of someone based on the descriptions of the reporter or the people they quote.
For example, when Obama is said to be yielding to the extreme left liberal wing of the Democratic party, what the hell does that mean today? Compare the policies of that faction with those of Nixon, LBJ, Truman on an actual factual basis. I've been around long enough to be able to compare position statements of all the major pols with those of my youth from both parties, with iconic figures like Milton Friedman and WFBjr.
I don't think the media is biased, but instead merely muddled and piss poor.
Posted by: mulp at March 1, 2010 09:00 PM
I don't even know what bias means. Everything has a perspective. I have certain beliefs. You have certain beliefs.
Based on those beliefs, aren't we more or less likely to use certain phrases?
Reports have no ago abandoned the goal of telling just the facts. Many/most entered publishing to a social impact.
Rather than being "objective," I say let's just be clear on our positions and perspectives. If you're a free-market newspaper, say it. If you believe in a powerful central government, say it.
Posted by: sara at March 2, 2010 04:00 AM
M Kuperberg: You beat me to the punch.
JDH: Does it seem plausible to you that, in addition to the slant of the paper the views of the readership are also endogenous? (i.e. echoing Hitchhiker's thought "So which came first? The newspaper or the reader.")
I am thinking of a quote from Time Series Analysis p253 "Finding valid instruments is often extremely difficult"
Posted by: Robert Bell at March 2, 2010 07:27 AM
You mean a kazoo is not a "valid instrument"?
Posted by: Barkley Rosser at March 2, 2010 12:21 PM
Posted by: Jake M. at March 1, 2010 05:44 PM
"It's all about bucks kiddo. The rest is conversation." - Gordon Gekko."
Posted by: Rickk at March 2, 2010 03:31 PM
sara at March 2, 2010 04:00 AM wrote:
"I don't even know what bias means. Everything has a perspective. I have certain beliefs. You have certain beliefs..."
Walt Disney wrote:
"I dream, I test my dreams against my beliefs, I dare to take risks, and I execute my vision to make those dreams come true."
(1901 - 1966) The Disney Way
Posted by: Rickk at March 2, 2010 05:38 PM
I wonder how many of you, offering such insightful comments about how editors and reporters slant their stories, have ever been inside a real newsroom. How many of you have ever spoken to a real journalist?
A conservative community will have conservative politicians and conservative community groups. When those politicians and groups get news coverage, conservative buzzwords will be in the newspaper, reflecting conservatism back to a conservative community. A good newspaper will also cover the individuals and the activities of the political minority or minorities, since they too are part of the community.
Posted by: John Stark at March 3, 2010 11:39 AM
Look at all of the regression lines you like, the idea that our newspapers and other media outlets are honest (lets use that word instead of the nicey-nice 'biased') has to make you laugh. They hide facts. Is that bias or dishonesty?
And if they were as smart as they think they are (they never say 'We just don't know the answer' do they?) why are they falling off into the abyss where they should be going?
Posted by: Norman at March 3, 2010 01:12 PM
there is no news outlet that is off-message, at this point, if such a thing existed, nobody would understand it. That's truth.
Posted by: biker at March 3, 2010 03:22 PM
I think the audience is just as responsible for media bias as the medium itself. Conservative readers see liberal bias in everything. Liberal readers see conservative bias in everything. Intelligent readers realize their political predispositions and try to avoid them, just like good journalists do.
Unfortunately, there are many more unintelligent readers than there are unintelligent journalists.
Posted by: Ian at March 3, 2010 05:28 PM