Groseclose’s Left Turn

Murdoch’s actions have focused attention on media influence.  Tomorrow my UCLA colleague Tim Groseclose’s new book Left Turn will be published.   Good timing?   I’m looking forward to the book party on thursday night.   Tim’s work has nudged economists to take a close look at the media’s impact.  Jesse Shapiro has done some important work on media bias.   For some of Jesse’s technical papers on this subject, click here.

When I taught at the Fletcher School at Tufts,  one of my students argued that we all should be required to watch the PBS Lehrer Newshour each night.   He believed that it was the exception as it offered “fair and balanced” coverage.   On the first page of his article, Tim reaffirms my student’s claim.  

I hope that my diverse UCLA colleagues celebrate Tim’s coming success in the same way that we must have celebrated Jared Diamond’s success with the publication of Guns, Germs and Steel.

Author: Matthew E. Kahn

Professor of Economics at UCLA.

32 thoughts on “Groseclose’s Left Turn”

  1. Groseclose seems like just a colossal jerk, who gets an especially big charge out of pissing off liberals.

  2. Professor Kahn,

    What was your student’s proposed penalty if one of his fellow citizens failed o watch the PBS Lehrer Newshour every night? Surely a student of public policy knows that laws don’t enforce themselves. What punishment would fit this dastardly crime? And what would happen to repeat offenders?

  3. So a PoliSci professor with conservative leanings conducts a study on media bias and finds that the media is largely biased to the left. Hmm, persuasive.

    He seems to draw an equivalence between metrics that rank politicians on a left-right continuum and a metric to do the same for a news organization. I don’t know the specifics of his methodology. On it’s face it seems like a reasonable proposition but, it also seems to me, the problem is vastly more complex for a news organization. Most of the metrics I’m aware of for politicians are based largely on how they vote which are pretty straightforward, objective data points. Not so much for news stories. Ferreting out the data points is bound to be more subjective. Also, as politicians tend to be more singular in their political leanings, a metric on them is likely to be more meaningful than on mainstream news organizations which tend to have multiple personalities.

  4. Yes, the Murdoch scandals have focused attention on media influence. And made complete idiots of people concerned about “liberal media bias.” Notice that Groseclose’s careful social science omits any reference to the truth or falsehood of a proposition: a false claim that supports a conservative position is precisely equivalent to a true claim that supports the opposite. How postmodern!

  5. Fish, meet water.

    The issue of liberal bias versus conservative bias is theater for the rubes. The relevant divide in politics is between those inside the established power structure versus those outside. Where do you suppose academics at elite universities fit?

  6. Not to be overly 20th century or anything, but hasn’t Chomsky answered, in depth and, well, repeated detail, the “liberal media” canard? Can’t wait to dive into a ‘conswervative’ expose’ of an inherently and repeatedly proven conservative national media as, ummm, like, pro-abortion and stuff. And just off the cuff: any ‘study’ that identifies only the Washington Times and a single Fox PropagandaNews show (Hume) as “left of center” has a priori splayed its bias and typically weak and skewed methodology.

    Perhaps, as Mr. Kleinman points out, facts have a Liberal bias. And one last thing, Mr. Kahn. I have, I think along with many far more erudite and informed readers here at RBC, stuck with you as an RBC poster as a point of honor and signal of intellectual rigor and inclusion. But for you to attempt to link, without any evidence or reasonable justification, patently obvious remainder table dreck like “Left Turn” to a brilliant, paradigm changing text like “Guns, Germs, and Steel,” no less in a patently transparent and bathetic attempt to deflect attention from the astoundingly corrupt, criminal, and disgusting behavior of Murdoch’s right wing propaganda empire, does it for me. You’re a hack. A shameless, apologist hack with no place in a community of logical, reasoned debate.

    One other thing, son – it’s “Thursday,” capital “t.”

  7. As Stephen Colbert said a while ago, “reality has a well-known liberal bias”. If one accepts that hypothesis, it’s not terribly surprising that organizations that report about reality also reflect that bias.

    Of course, Stephen Colbert was making a joke, but in all seriousness I’d contend that the major problem that the Republican party has at the moment is extreme ideological purity; while I’m not exactly a huge fan of the Democratic party, either, an emphasis on ideological purity is about the last criticism I’d level at them (I remember hearing Obama described as “cat-herder in chief” somewhere).

    Thus, I do not consider it to be surprising to find not much factual support for a party motivated by an ideology (no matter how well-meaning) in the media; I’m just not sure I’d call it “bias”. I am pretty sure that a similar study would show a strong anti-Marxist “bias” of the American media, for similar reasons. You would find, using the same methodology, that pretty much every newspaper in America is well to the right of Karl Marx. Shockingly enough, I don’t see such a clear and unambiguous “bias” as a problem. Karl Marx was by all accounts a well-meaning individual who was seriously looking for the betterment of humanity, but he also happened to be very wrong.

    Of course, hypothetically speaking, if the United States had a serious conservative party, then such a study might be interesting; but it’s just not possible to perform such a study based on counterfactual scenarios.

  8. Given that liberalism is the most centrist of political outlooks, positioned precisely halfway between the endpoints of communism and fascism, and also the most invested in reality, being the least ideological, a liberal bias is exactly what the media should have.

  9. Tim wrote:
    > I don’t know the specifics of his methodology.

    I’m sure it’s in the book, but you can also read it in an article Groseclose and Milyo published in QJE a few years ago. Long story short, it’s based on which sources the media outlets quote.

    Specifically, it’s based on which think tanks and policy NGOs the media-outlet quotes and comparing this to the same procedure for Congressmen. Table 1 gives a list of the most commonly quoted organizations and how they are scored. In reading the list you should focus on the orgs with scores far from the center (“50”) that are close to the top of the list as these are the ones that are largely driving the index. So even though Brookings is the single most cited think tank, it doesn’t have much effect because it has bipartisan citations in the Congressional Record. Conversely, Hangun Control is coded as pretty far left but it has only got quoted a few dozen times. The frequently quoted organizations that fall outside the partisan 40 yard lines are (in order of how often they get quoted in the press): NAACP, Sierra Club, AEI, AARP, Heritage, Common Cause, Center for Responsive Politics, Consumer Federation of America, Christian Coalition, Cato, and NOW.

  10. From the link: “Also, we found that many supposedly far-left news outlets were not that far left. For instance, we found that National Public Radio was no more liberal than the Washington Post, Time, or Newsweek. And we found that it was less liberal than the average speech by Senator Joe Lieberman.”

    So, NPR is less liberal than the average speech by a Democratic senator whose politics are famously middle-of-the-road to conservative. And it is a left-leaning media outlet.

    Really, this dude’s serious?

  11. What an odd post. I wish Prof. Kahn would tell us what aspect of this book he found meritorious.

    Certainly Prof. Groseclose’s surprise at the controversy surrounding his work is either feigned, or a demonstration that he hasn’t spent much time reviewing the controversy over media bias.

    Prof. Groseclose picks an odd measure of bias. Apparently, the media should not merely parrot the median view at a given moment, but rather the median Senator’s view. An unbiased media would, therefore, over-represent the views of Wyoming, under-represent those of California, and not represent the views of nonvoters at all. Of course, as Prof. Kleiman and others point out, concern for factuality is right out the window in this scenario.

    Meanwhile, Prof. Kahn approvingly links to Prof. Shapiro, who tells us that newspapers are, in fact, representative of their readership.

    What do these studies have in common? They are each designed to substantiate right-wing talking points. The media takes into account public perceptions, showing that owners don’t influence the news, Shapiro tells us, and the media doesn’t properly account for public perceptions, showing that it has a liberal bias, Groseclose tells us.

    We can only guess what Prof. Kahn’s point is here. I’m stumped.

  12. Mark Kleiman: “Notice that Groseclose’s careful social science omits any reference to the truth or falsehood of a proposition: a false claim that supports a conservative position is precisely equivalent to a true claim that supports the opposite. How postmodern!”

    Mark, why not? The Chicago School is the quintessential postmodern school, and they’ve flourished, even as reality takes a nasty dump on their favored theories. If you think of this work as ‘even a liberal professor says….’ (because all professors are liberals), which will get sold heavily on the right, it makes sense.

  13. I’m intrigued to look at the conservative prof’s book. If he asked people what they thought about which newspapers, television shows, radio shows were “liberal” or “conservative,” then what is he think he is proving?

    I have always found it funny that people who are right wing say the media is “liberal.” I ask, “How do you know that?” Essentially, they tell me they learn that…from the media. And then I ask, “And when you learn the media is ‘liberal’ from the media, do you learn it from people who say that being a “liberal” is a good thing or a bad thing?” and at that point, there is just silence because they realize they’ve learned “liberal” is something to be derided and avoided from the same media they are describing.

    I then ask them to break down “liberal” and “conservative” into the Daniel Bell formulations of cultural, economic and political categories. We then go through the NY Times and Washington Post, and see if they can see what I see, which is that the NY Times tends to be culturally moderate to liberal (tends, not always; weakly moderate on abortion, but strongly culturally liberal on say gun control), moderate to conservative on economics (the Times rarely supports labor unions as a bulwark for democratic values, and loves every anti-worker trade treaty that comes down the pike), and is “liberal” on political matters, in the sense they want more open government more often than not. If the “conservative” is willing to talk with me, I then talk about the “values” of corporate executives, such as Jack Welch, and ask, “Isn’t it interesting that corporate executives tend to be liberal about abortion and gays, but really don’t like unions or socialized medicine, love trade treaties that protect intellectual property, but not workers’ wages, tend not to know about science and therefore are easily confused about climate change, and think there should be free speech except for their trade secrets and anyone criticizing them directly?”

    And then I try to shift the conversation to commercials, and how ad revenue plays a major role in allowing Glenn Beck to initially build up his audience (it eventually led to his downfall, showing there might finally be a limit to some right wing ravings on our television or radio), while Donahue, in 2002-2003, had the highest ratings on MSNBC at the time, but was let go because…not just his anti-war stance, which was cited in an infamous memo, but also because he was not drawing sufficient revenue. Joe Scarborough replaced him, had low people ratings, but excellent ad revenue. MSNBC later learned that there are some advertisers who may still advertise a show that might lean “liberal,” but even then, execs were very worried about skewing liberal, so that there were memos about having two “conservatives” for every “liberal” on the shows, which is less so at MSNBC, but more so at CNN, and other outlets. We don’t really have to deal with FoxNews too much here…:-) As I then say, “If television wants sensation and its executives ‘liberal,’ shouldn’t Bernie Sanders be on t.v. at least as much as John McCain? Shouldn’t you actually get to hear Noam Chomsky speak on a regular basis, instead of saying ‘Who?’ or thinking he is a devil? Shouldn’t Michael Moore just get a show and not be relegated to the rare appearance, or promoting a film every two years?”

    The upshot if I can get that far is that I help them realize the media in this nation is largely corporate owned, and it largely reflects corporate executives’ values.

    However, there is one thing left out of this often initial and last discussion: foreign policy. When we analyze foreign policy, we see how often the NY Times, Washington Post, television and radio reflects the most Establishment oriented thinking. We as a nation are always imbued with the best of motives, unlike other nations. We always trust “experts” except when those experts are against war, and the burden of proof in a militarized situation is nearly always on the side of those who seek diplomacy and peace, and rarely belligerence and war. There was little if any “liberal” media in the build up to the Iraq War, for example, and more cheerleading, acceptance of administration statements as emotionally as well as factual true, even when there were leaks from the CIA saying otherwise. The only papers of record who published those leaks at the end of articles were the Washington Post (though it’s editorials were largely boisterously supportive of the Bush-Cheney administration) and the LA Times. The NY Times was in the tank for Bush, i.e. Judith Miller, and well…such things are familiar to people on this blog.

    Again, I’m rather intrigued as to how this professor will prove the corporate owned media is “liberal” and what criteria he will use…

  14. Gabriel,

    Thanks for the information. I note that:

    The frequently quoted organizations that fall outside the partisan 40 yard lines are (in order of how often they get quoted in the press): NAACP, Sierra Club, AEI, AARP, Heritage, Common Cause, Center for Responsive Politics, Consumer Federation of America, Christian Coalition, Cato, and NOW.

    Does Groseclose make any effort at all to compare the integrity of these organizations, in terms of how likely what they say is to be true? I don’t think it helps objectivity to report on what the Christian Coalition says, for example. This study would be more interesting if Groseclose had fact-checked a reasonable sample of statements made by all these groups, before concluding that reason for citing, say, the Sierra Club rather than AEI was bias rather than accuracy.

  15. Kahn: “On the first page of his article, Tim reaffirms my student’s claim.”

    Eh, no, he doesn’t. He writes: “To compute this, we count the times that a particular media outlet cites various think tanks and policy groups, and then compare this with the times that members of Congress cite the same groups.”

    That is an insane way of defining bias. So if one side of the aisle in congress cites evidence from both liberal and conservative “think tanks”, and the other only from conservative ones, then the effect is that it makes the media look biased? Right. Because obviously the media’s main job is to reflect what 600 rich old people in DC think.

    Reminds me of the claim that republicans are more bipartisan since democrats are more willing to cross the aisle during votes, and thus any pieces of “bipartisan” legislation that get votes from both parties usually originate from the republican side.

  16. A twist on “reality has a liberal bias”: Liberalism is biased towards expertise, towards scientific inquiry, critical deconstruction of cultural norms and dominant paradigms. The extent to which any of these are the paths to truth, then truth can be said to have a liberal bias. Although that’s not really accurate. Better said, liberalism has a bias towards truth. So, for instance, when a journalist points out that a business is polluting a river, is it liberal bias? When most illegal aliens are found to be exploited when all they wanted was a chance at a better life, is it liberal bias? When global warming is found to likely have devastating effects, is it liberal bias? When evolution is found to be absolutely true, is it liberal bias? When gays are – newsflash! – found to be normal, healthy people, is that liberal bias?

    Conservatism is ultimately about common sense. And sometimes common sense is right; even a broken clock is right twice a day. But to the extent that conservatism is biased against expertise, or critical analysis, or relativism, or the deconstruction of tradition – in other words the machinery of free thought – then conservatism is biased against truth. As Buckley put it, to “stand athwart history and yell stop”, even alas, when that history is truth.

    In the end, there is no such thing as a bias towards truth, only away from it. To be biased is to be operating outside the parameters of truth-finding. To the extent that conservatism rejects the very process of truth-finding, preferring instead to rely on such subjective and non-rational epistemologies as tradition or common-sense, it is biased against truth. And to the extent that the media is concerned with truth, then conservatism is often biased against the media itself.

  17. I think Shapiro’s little snippet about the term “estate tax” being evidence of a liberal bias while the term “death tax” evidences a conservative bias says all we need know to be able to regard the study as filled with conservative bias and intellectual vacuity. That and other examples in the linked piece are yet another case of where conservatives tart up their propaganda by accusing use of terms traditionally employed for decades as “bias” even though those terms are also objectively more accurate than the conservative lie. Death is not taxed, and most who die pay no estate tax.

    Garbage in, garbage out. feh.

    I pity his students.

  18. Bernard,

    No, Groseclose and Milyo did not attempt to fact check whether think tanks that are more frequently quoted by liberal Congressmen provide information that is better or more reliable in some objective sense. This is obviously a major way to critique their method, though it’s also hard to establish convincingly and systematically.

  19. On an unrelated note, but since a lot of you are here, can anyone remember the name of that organization Zasloff cited as being very good at finding grassroots groups in Africa that are already doing good work? I think it was American Jewish Something Something, but it was an acronym?

    I have a friend trying to figure out how to send money to Somalia.

    I searched RBC under “relief organization” and “Africa” but I couldn’t find it. Maybe I’ll just look under Zasloff but that might be too many.


  20. Just my two cents: I find Kahn to be sort of knee-jerk sometimes too, but on the other hand, his best topic of green economics is pretty important. Maybe RBC should recruit a lefty green economist and let them go mano-a-mano. Sell popcorn too.

  21. And while we’re b****ing about the lineup, y’all need more women. I can suggest some excellent people if you don’t know any.

  22. The weaknesses of this study have been well documented in this comment thread. As such, I’d like to point out something more anecdotal.

    We have a national press that is going out of its way constantly to treat lunatic, fringe GOP pols like Palin, Ryan, and Bachmann as though they are serious and their points are valid. Wouldn’t a left-biased press call out the GOP a bit, given the extent to which the GOP has slid to the right of its arch-conservative elements of just 25 years ago? Instead, we get a media bending over backwards trying to create false equivalences between the problems in both parties.

    For crying out loud, we have a media that reports on climate change as though there really is a 50-50 scientific debate!

  23. Gabriel,

    This is obviously a major way to critique their method, though it’s also hard to establish convincingly and systematically.

    No doubt it’s hard. But this is one of those situations where you can’t avoid some sort of underlying assumption. Without the fact-checking, the implicit assumption is that there is no difference in the accuracy of the cited statements across groups. I don’t think that’s correct.

  24. Is the Balanced Budged Amendment a “conservative” idea? It represents a bold innovation, vacating some of the powers of Congress that were established by the Founding Fathers, and it would, for the first time in our history, place a specific economic policy into the supreme law of the land. Sounds radical to me.

    It is far from clear that the very able, learned, and distinguished professor has a robust set of criteria for the classification of this proposed amendment as conservative.

  25. I just got my copy of Groseclose’s book today and have been reading through it; so far I’m not impressed. The methodology he uses to measure “media bias” is problematic, to put it mildly, and I have some definite reservations about the quality of his statistical analysis. I’ll be posting a review at in a few days, but I wanted to mention one very basic problem I have:

    Groseclose bases his conclusions about “media bias” on his estimated values of what he calls the “slant quotients” of various media outlets. These “slant quotients” are generated through that questionable statistical analysis I mentioned. One basic problem that stuck out like a sore thumb was the absence of any sort of standard error estimates. Any statistical analysis of the sort done by Groseclose–from his paper with Milyo it appears that the quotients are generated as some kind of regression coefficients–should give standard error estimates. Yet, I can’t find them, either in the book or in the original Groseclose-Milyo paper. When Groseclose tells us that the “slant quotient” of the New York Times is 73.7. he has an obligation to also tell us the standard error, but he doesn’t. That failure, alone, really ought to disqualify his conclusions from serious consideration.

  26. Jestak,

    In my reading of G+M, the index is not calculated through regression but a mechanical algorithm that takes as its inputs the ADA scores of Congressmen, cites to think tanks by Congressmen, and cites to think tanks by media outlets. You are right to point out that they implicitly assume there is no error in these three types of data. Ideally they should have figured out a way to estimate confidence intervals, but (unlike a regression) it’s not obvious how they would have gone about doing so.

  27. I’m a solid liberal, but as I see it, the MSM on balance does lean to the left in its reporting. Dramatically–no. And nowhere near (!!) as much as Fox does in the other direction. But to the left, nonetheless. Since any attempt at an empirically-based debate would be absurd here, I’ll present a thought experiment:

    Imagine that 80% of reporters/editors/etc. were male. Don’t we think that the sex imbalance would impact news coverage? And it would do so in a way that would favor a ‘male understanding of the world’ and place a greater weight on ‘male concerns’. (With some issues a ‘male understanding of the world’ or ‘male concerns’ make no sense; but with other issues they do.) This almost always would not be the result of explicit or conscious sexism, but that wouldn’t make the bias less real.

    I’m pretty sure the polls consistently show that a solid majority of reporters/editors/etc. say that they are liberal. Why won’t these reporters/editors/etc. tend to produce work that reflects a more liberal understanding of the world, and places a greater emphasis on liberal concerns? (Again, I’d guess that this would be almost entirely the result of unconscious cognitive processes.)

    An unfortunate result of all of this seems to be that reporters/editors/etc., out of fear that they actually are guilty of unconscious bias (and sometimes just out of fear of being called biased), sporadically engage in overcompensation. Evidence: the 2000 election. The MSM was clearly harder on Gore than on Bush, and the world will forever be the worse because of that.

  28. Dan puts his finger on the problem; reporters overcompensate in accommodating to conservatives because they fear being criticized for liberal bias.
    We have a test case in the form of the out-of-control e-mail tirade that Rep. Allen West of Florida launched against his colleague, Debbie Wasserman-Schultz. She, following standard courtesy rules of the House, referred to him as “the gentleman from Florida,” and criticized his vote on Medicare, which would increase costs to many of his South Florida constituents. He responded by writing, “You are the most vile, unprofessional ,and despicable member of the US House of Representatives. If you have something to say to me, stop being a coward and say it to my face, otherwise, shut the heck up. Focus on your own congressional district!” He followed with “You have proven repeatedly that you are not a Lady, therefore, shall not be afforded due respect from me!” is a link to her statement on the House floor and to his response. (Posted yesterday in a comment under Mark’s ” ‘Play nice’ rules, again” column.

    Here is the test: NPR had a brief reference to this incident at the end of Talk of The Nation. Neal Conan and Ken Rudin mentioned it at the tail end of the broadcast, saying only:

    CONAN: Interesting, during the debate, the Representative Debbie Wasserman Schultz took to the floor and ripped the GOP’s cut, cap and balance bill and did not spare her fellow Floridian, Congressman Allen West.
    Representative DEBBIE WASSERMAN SCHULTZ: Incredulously, the gentleman from Florida who represents thousands of Medicare beneficiaries as do I is supportive of this plan that would increase costs for Medicare beneficiaries, unbelievable from a member from South Florida.
    CONAN: And today, well, Congressman West seems to have taken that personally.
    RUDIN: Yeah, well, he does. I mean, first of all, we should point out that Debbie Wasserman Schultz is the chair of the Democratic National Committee and a lot of what she says is political by nature. But, again, he really went on a tirade against her and called her vile and all these things…
    CONAN: Despicable?
    RUDIN: …despicable and vile. It was just…
    CONAN: Cowardly.
    RUDIN: It’s got ugly. First of all, you don’t mess around with Allen West and you don’t mess around with Debbie Wasserman Schultz. So this is, I think, a little skirmish that will continue.
    So they quoted West’s language calling Wasserman-Schultz vile and despicable, and Rudin did go so far as to call it a tirade, but Conan said that Wasserman-Schultz “ripped” the GOP plan and did not spare her fellow Floridian, implying that she had “ripped” him as well. Then Rudin sets up a symmetry between the two statements, saying only that you “don’t mess around” with either representative.

    I did not hear this story on the other mainstream media today, but perhaps other people did. Let us follow this for a while and see if Allen West receives any criticism from any of them.

    Conservative crybabies have been whimpering about liberal bias of the MSM since the time of Spiro Agnew, and have been getting unlimited mileage out of their bellyaching. Dan is correct to point out that this has had tragic consequences. Perhaps Prof. Kahn has a way to show that this test case actually proves liberal bias. Let us see.

  29. Another anecdote:

    We have one side in the debt ceiling debate essentially saying they will blow up the global economy if they don’t get their political goals. In essence, this is no different than strapping a bomb to yourself and threatening to light the fuse if your wishes are not accommodated. This is extortion, nothing more.

    Yet, our liberal media are not only not calling the GOP out on it, they are treating the GOP demands and tactics as legitimate actions of a responsible political party. It’s not helping, of course, that the Dems are too.

  30. With one exception, this comment thread is truly an impressive display of how to react against an uncomfortable fact without having to truly address it. At least Jestak has a critique based on his reading (some) of the book. His criticism of the sources and confidence intervals of the SQ is actually just a question which is most likely answered already— after all, the peer reviewers at the Quarterly Journal of Economics which published the paper probably dealt with the confidence intervals as well.

    As far as whether or not the right-wing think tanks are factually incorrect– come on, if they were, there’d be massive documentation from partisans such as yourselves.

Comments are closed.