September 13, 2004

Why Oh Why Can't We Have a Better Press Corps? (The New York Times Drops One of The Balls Edition)

Ah. AEI hack John Lott surfaces once again, claiming left-wing bias in journalism. Eduardo Porter of the New York Times writes about it, but he drops one of the balls that he is juggling.

Porter does quote Chris Carroll as saying, politely, what Lott's work is worth: next to nothing:

Eduardo Porter: While the researchers of the American Enterprise Institute claim to expose the political bias of the reporting, Mr. Carroll said, it was unlikely that they succeeded in stripping out other factors. He said the reporting of economic statistics depended on broad perceptions of the state of the economy, which are influenced by many variables. The fact that the economy did better under Mr. Clinton than either of the Bushes probably affected the coverage more than the researchers allowed for. Moreover, Mr. Carroll pointed out that the results had large statistical margins of error. "I'm not persuaded that the results have any statistical significance," he said.

But he also writes:

Eduardo Porter: ...[John] Lott... acknowledged that he assumed a pseudonym - Mary Rosh- to write his own praise and defend his positions in online debate on that subject from 2000 through January 2003. Mr. Lott said that the things he had said in the guise of Ms. Rosh were, indeed, truthful....

Eduardo Porter does not quote from any of the messages John Lott sent as "Mary Rosh." He doesn't tell his readers that John Lott wrote things like:

I [i.e., Mary Rosh] have loaned out my copy [of John Lott's book] a dozen times and while it may have taken some effort to get people started on the book, once they read it no one was disappointed....

I [i.e., Mary Rosh] had him [John Lott] for a... class... at the Wharton School.... [H]e was the best professor that I ever had.... Lott taught me more about analysis than any other professor that I had and I was not alone. There were a group of us students who would try to take any class that he taught. Lott finally had to tell us that it was best for us to try and take classes from other professors....

Academic after academic has replicated [Lott's] results... [which have] survived remarkably well.... This classic has stood the test of time....

[A]ll the attacks on [Lott's book] are completely bogus.... [T]hey are afraid that those who have lied about Lott’s research will lose their credibility.... Those attacking [Lott's] book will stop at nothing....

Needless to say, all the attacks on Lott's book are not completely bogus, Lott's results have not stood the test of time, Lott's work is not a "classic," Lott did not take his own class, Lott did not teach Lott more about data analysis than any other professor, Lott was not a member of a group of students who took all the courses Lott taught, Lott did not tell himself that he needed to take classes taught by other professors as well, Lott did not have one single copy of his book that he loaned out a dozen times, and Lott's statements made under the pseudonym of "Mary Rosh" were not truthful.

Porter should have quoted at least one of these messages from "Mary Rosh": he owes his readers at least a clue that Lott's claim that "the things he had said in the guise of Ms. Rosh were, indeed, truthful" was another lie.


UPDATE: Eduardo Porter says three things:

  1. "The story was not about Mary Rosh," so there was no need for him to check out whether Lott's statement that "the things he had said in the guise of Ms. Rosh were, indeed, truthful" was itself truthful.

  2. "There is enough in the story" for readers to form accurate judgments of the credibility of Lott's statements.

  3. While Porter declines to say whether or not he regards Lott's statement that "the things he had said in the guise of Ms. Rosh were, indeed, truthful" as truthful, he does say that his "story as a whole" does give a truthful picture.

One ironic thing is that a week and a half ago I was telling Porter that the standard operating procedures of America's elite press corps give an edge to those who are ethics-free in what they tell reporters. And now Porter shows us exactly how this works.

In my view, the New York Times is going to continue to spiral downward until its editors teach its reporters that it is not enough for the story as a whole to be more-or-less true: the details have to be true as well, even if they aren't the story's central focus.

Posted by DeLong at September 13, 2004 04:40 PM | TrackBack
Comments

Thank you. This is a hoot!

Posted by: DonPedro at September 13, 2004 05:06 PM

Oh lordy, I almost soiled myself reading that...

Seriously, what sort of twisted #$&@ pretends to be a college-age girl smitten with his own academic "credentials"? And more importantly, what sort of think-tank hires someone wh... oh wait, we're talking about AEI here, nevermind...

Posted by: Brad Reed at September 13, 2004 05:07 PM

Tim Lambert (University of New South Wales) loves commenting on the activities of John Lott. You might find his blog interesting:
http://cgi.cse.unsw.edu.au/~lambert/cgi-bin/blog/guns/Lott/

Posted by: Richard in Cambridge at September 13, 2004 05:10 PM

It's almost insulting that we have to take these people seriously.

Can someone please explain to me how the hell these lunatics took over the country?

Posted by: Dragonchild at September 13, 2004 05:18 PM

"Can someone please explain to me how the hell these lunatics took over the country?"

I'll answer that one:

"Senator John Kerry said that he would have voted to give the president the authority to invade Iraq even if he had known all he does now about the apparent dearth of unconventional weapons or a close connection to Al Qaeda.

'I believe it's the right authority for a president to have,' said Kerry, who has faced criticism throughout his presidential campaign for that October 2002 vote."

Posted by: Brad Reed at September 13, 2004 05:25 PM

Of course, we'll want to hold Lott to much higher standards than Dan Rather...I'm still waiting to see a "Why Oh Why Can't We Have A Better Press Corps?" on this one. True, the Rather stuff is fairly trivial in assessing the merits of Kerry...but so too is the Rosh stuff in assessing the merits of Lott's empirics....

In any case, Tim Groseclose's piece on media bias is the more interesting. He maps from Congressional ADA scores to an imputed ADA score for think tanks using frequency of citation by members of Congress. Then he looks at the frequency with which news outlets cite the various think tanks and maps them onto the same ideological scale as Congress. He finds Fox is right of the median member of Congress, but that it is closer to the median than any other news source, and all others are to the left. A pretty neat methodology for assessing media bias...and a pretty provocative result....

Posted by: Eric Crampton at September 13, 2004 05:26 PM

Lott's lying about Mary Rosh is actually quite relevant to whether we believe Lott's study. To believe his study requires our total trust that he has considered and included all confounding factors.

Posted by: Tim Lambert at September 13, 2004 06:03 PM

Eric Crampton:
I'm not sure that I have that straight. Let's do an example. Let's say each Senator gives the same number of speeches a year and cites exactly one think tank in each speech, and there are 60 members of Congress who cite only John Birch Society reserach in all of their speeches. So media is biased unless the it cites John Birch Society studies 60% of the time? Is that right?
Seems like the study you described just determines whether the media trusts the same think tanks that members of congress do.
Would you trust the think tanks that a randomly selected member of congress prefers? Would you be biased if you said "no, thankyou".
Citing think tank results from congresspeoples' citations is a bit iffy, since it is considered a wonderful thing a liberal to quote to a conservative the result the liberal likes from a conservative think tank -and vice versa.
This kind of research requires even more assumptions than economics.

Posted by: jml at September 13, 2004 06:04 PM

Mr. Lambert, why trust him? Why trust anyone? Why wouldn't you just ask for the data and run the test for yourself?

The reason we have resources like SSRN is to get other academics to look at research and critique it as part of the peer-review process. Had his gun book been on SSRN, maybe some of your critiques could have been brought forward and addressed in advance of publication, and perhaps any mistakes that were there would have been addressed.

Posted by: kb at September 13, 2004 06:45 PM

Lott's study is completely unnecessary. So many newspapers have lied to or mislead their readers about the so-called Assault Weapons Ban that the bias is obvious.

When the Chicago Sun-Times lied in a recent editorial and implied that machine guns were covered by the AWB, did any of you care? Nope, lies in the cause of the greater good are always forgiven.

Posted by: Classic Liberal at September 13, 2004 07:10 PM

How about some classic lies:

“During my service in the United States Congress, I took the initiative in creating the Internet.” (Al Gore, CNN’s “Late Edition with Wolf Blitzer,” 3/9/99)

"Three thousand young people in America will start smoking tomorrow. One thousand of them will die a death not unlike my sister’s. And that is why until I draw my last breath, I will poor pour my heart and soul into the cause of protecting our children from the dangers of smoking.” (Al Gore, The Democratic National Convention, 8/28/96)

Yet, 4 Years after the Death of Gore’s Sister the Press Reported: ‘“Throughout most of my life, I raised tobacco,’ the Tennessee senator hollered. ‘I want you to know that with my own hands, all of my life, I put it in the plant beds and transferred it. I’ve hoed it, I’ve dug in it, I’ve sprayed it, I’ve chopped it, I’ve shredded it, spiked it, put it in the barn and stripped it and sold it.’” ([New York] Newsday, 2/26/88)

HeeHee


Posted by: Howdy Partner at September 13, 2004 07:59 PM

What jml said.

The study is idiotic for another reason: the authors make no effort to control the type of citation or the context. Many liberal and centrist think tanks (e.g. CBPP and Brookings) put out lots of (descriptive) numbers, in addition to any (normative) recommendations. But if a journalist cites a (descriptive) number from CBPP, this stupid study will count that as a possible indicator of liberal bias.

Chalk up yet another embarassment for the social sciences...

Posted by: liberal at September 13, 2004 08:16 PM

To Tim Lambert:

I agree with you, but I'm less worried about Lott's results given that he makes his data publicly available for all comers. To the extent that results are based on private data, lack of integrety on minor margins is quite worrying as we have no better estimate of integrety on the things we can't observe that would affect the major results. When the results are verifiable and the data is openly available to be criticized, though, I've got far fewer worries.

To JML:

If there are 100 Senators and 60 of them only cite the John Birch society, and if the media doesn’t cite the John Birch society pretty frequently, the media will be biased relative to the median member of the Senate. And, unless political markets are seriously uncompetitive or something very strange is going on, it’ll also be biased relative to the median voter.

Trust would be a different matter, but you’d need to specify that a Congressman has more incentive to cite an untrustworthy source than does a news agency, and that seems unlikely.

As for folks citing folks of the opposing ideology, this will only cause a problem if one group is more prone to cite its opponents than is the other. Then we’d get a biased measure of the ADA score of particular think tanks. So, if Ted Kennedy were more likely to cite Heritage than Strom Thurmond was to cite the Children’s Defence Fund, we’d get a bit of bias that way. But you’d need to specify that one side is more prone to cite its opponents. And, Groseclose checks for that: Republicans cite Conservative groups about 83% of the time; Democrats cite Liberal groups about 82% of the time.

The paper is available here http://www.yale.edu/isps/seminars/american_pol/groseclose.pdf and is worth a read.

Posted by: Eric Crampton at September 13, 2004 08:34 PM

To Liberal:

Don’t like Groseclose’s study? Do a better one. But you might start by actually reading his paper. At the top of page 6 you’ll find the following:
--
“Also, we omitted the instances where the member of Congress or journalist only cited the think tank so he or she could criticize it or explain why it was wrong. About five percent of the congressional citations and about one percent of the media citations fell into this category.

In the same spirit, we omitted cases where a journalist or legislator gave an ideological label to a think tank (e.g. “Even the left-wing Urban Institute favors this bill.”). The idea is that we only wanted cases were the legislator or journalist cited the think tank as if it were a disinterested expert on the topic at hand. About two percent of the congressional citations and about five percent of the media citations fell into this category.”
--
Some of us think it’s generally a good idea to read things before criticizing them and calling them an "embarassment". Others disagree. I side with the former. And, I tend to discount the views of the latter.

Posted by: Eric Crampton at September 13, 2004 08:51 PM

Eric Crampton wrote, "Some of us think it’s generally a good idea to read things before criticizing them and calling them an 'embarassment'. Others disagree. I side with the former. And, I tend to discount the views of the latter."

I took enough of a look at the paper and commentary on it on the web to conclude that it was idiotic.

Note that the authors wrote, "The idea is that we only wanted cases were the legislator or journalist cited the think tank *as if* it were a disinterested expert on the topic at hand." [Emphasis added]

The fact is that on some topics, certain claims *are* unbiased statements of fact, whereas others are more contentious. Do they control for this? No.

For example, suppose you're interested in the distributional impact of taxes---that is, how changes in the tax code affect people of different incomes. To do that, you have to have a model. Until recently, when the Brookings-AEI Tax Policy Center opened, the only private organization with a prominent tax model was the (very liberal) Citizens for Tax Justice, which even right-wing groups concede has honest numbers. (Previously the Treasury Department put out such numbers, but under Bush 43, it has ceased doing so.)

Of course, the authors claim that citing honest numbers is still a possible indicator of bias, because (in my example) it shows that the party citing the distributional numbers is interested in that topic, and because liberal groups---like CTJ itself---are also interested in that topic, one can somewhat conclude evidence of bias on the part of the speaker. (Whereas right-wing groups aren't so interested in distributional impacts.) I don't buy their attempt to translate rhetoric into numbers.

Posted by: liberal at September 13, 2004 09:15 PM

Eric Crampton wrote, "I agree with you, but I'm less worried about Lott's results given that he makes his data publicly available for all comers."

If you read Tim Lambert's work on John Lott, you'll see why a reasonable heuristic is to disregard everything Lott says:
http://www.cse.unsw.edu.au/~lambert/guns/lott98update.html

Posted by: liberal at September 13, 2004 09:18 PM

Eric, Lott publishes the data and the model, but you don't know how many models he considered and does not tell you about. In most famous study on concealed carry he has kept changing the model. Why? Because new data analyzed with the old model does not show carry laws associated with crime reductions. So he comes upu with a new way of analyzing it to make his results come back.

Posted by: Tim Lambert at September 13, 2004 09:24 PM

To Liberal:

I'm not sure how controlling for the accuracy of the think tank studies cited by members of the media and members of Congress would improve Groseclose's results. If it were the case that the media simply didn't cite Conservative sources much because those sources were systematically worse than those provided by Liberal sources, then I suppose that could generate Groseclose's results. But it would require that the media care more about accurate reporting than do members of Congress (doubtful) and that we have evidence that the Liberal think tanks in the Groseclose study are systematically better than the Conservative ones (possible, but doubtful; in any case it's an empirical question and I urge you to refute my null of no systematic bias).

I'm far more sympathetic to the idea that citing honest numbers can indicate bias. Suppose that both of the following facts are true. 1. If the status quo continues, global temperatures will rise by 1 degree over the next hundred years and sea level will rise by 3 feet. 2. Preventing a 1 degree rise in global temperature will cost 50% of GDP. Wouldn't it seem likely that more liberal members of Congress will cite the think tank producing study #1, and conservative members will cite the think tank producing study #2?

Tim: I followed this pretty closely when it was all coming out. I wished then and wish now that you and he could find an impartial econometrician, who you both trust, to resolve this. As it stands, I'm confident that concealed carry legislation doesn't really increase crime (which itself is an important finding, given the hysteria that surrounded the adoption of concealed carry). Whether it causes decreases in some categories of crime is something I'm far more hesitant to say anything about without running the numbers myself at this point, and I don't want to do that.

Posted by: Eric Crampton at September 13, 2004 10:35 PM

Didn't Lott have a bit of a 'dog ate my howework' problem with his supposedly groundbreaking results from "More Guns, Less Crime"? I vaguely remember something about his claiming he didn't back up his data, and it all got wiped out. And there didn't seem to be the usual stuff that goes along with a scholar doing a survey - grants to fund it, grad assistants to do the grunt work, to verify that his survey had even taken place, IIRC.

Posted by: RT at September 14, 2004 03:01 AM

Why is the New York Times giving over important space to the likes of Kevin Hassett and John Lott, and giving an economics column to Ben Stein? What is happening that this precious paper should be so intent on reflecting fringe radical right voices?

Posted by: Ari at September 14, 2004 03:26 AM

Eric Crampton wrote, "I'm not sure how controlling for the accuracy of the think tank studies cited by members of the media and members of Congress would improve Groseclose's results."

My point isn't that the result can be improved, but rather that the entire methodology is suspect.

Another problem: suppose we assume the scores assembled by Groseclose and his coauthor
(1) constitute valid rankings for politicians, and
(2) constitute valid rankings for the media.

They appear to further conclude that one can compare the scores for politicians and the scores for media.

How can that be? I can understand that they could attempt to make some kind of internal validity check for, say, the rankings of politicians, but how can they ever validate comparing the score of a politician against the score of a media organization?

Posted by: liberal at September 14, 2004 07:11 AM

do you think he wore a dress while posting?

Posted by: cali_ at September 14, 2004 07:12 AM

RT wrote, "Didn't Lott have a bit of a 'dog ate my howework' problem with his supposedly groundbreaking results from 'More Guns, Less Crime'?"

See the link I cited above. Tim Lambert is the primary source for this type of issue.

Posted by: liberal at September 14, 2004 07:13 AM

Eric Crampton,
I'll look at the paper, thanks for posting the link. But I am suspicious of the conceptual framework for defining bias. From your comments, it looks like it must be bias relative to the wisdom of some "political market" -assumptions piled on assumptions. I don't think that is what most people think of immediately when you say "bias."

Posted by: jml at September 14, 2004 10:14 AM

Eric Crampton wrote, "I'm far more sympathetic to the idea that citing honest numbers can indicate bias. Suppose that both of the following facts are true. 1. If the status quo continues, global temperatures will rise by 1 degree over the next hundred years and sea level will rise by 3 feet. 2. Preventing a 1 degree rise in global temperature will cost 50% of GDP. Wouldn't it seem likely that more liberal members of Congress will cite the think tank producing study #1, and conservative members will cite the think tank producing study #2?"

Perhaps. But again, note that their claim is much stronger---that you can then make comparisons between politicians and media organizations.

I could see a plausible dynamic where members of Congress behave as you describe, and yet at the same time the media appear to give the same citation pattern, but for different reasons, invalidating the comparison between politicians and the media:
(1) Right-wing outlets like FOX quote number your 2. because they're vehemently opposed to global warming abatement. [Aside: I myself am much more agnostic on global warming than my political positions would seem to indicate, so this is just for sake of discussion---any reasonable planning for global warming abatement would have to balance costs and benefits, IMHO]
(2) CBS, NBC, etc focus on your 1. because 1. is sensationalistic, invokes images of global destruction, etc, whereas 2. is too cerebral and economics-oriented, and the media much prefer to deal with the former than the latter, just as they often cover stories of local destructive weather, crime, etc.

Posted by: liberal at September 14, 2004 10:51 AM

To conclude, this picture I concocted is compatible with Fox being right-wing and CBS etc being middle-of-the-road, whereas if you look at their methods and mapping, one might conclude the latter are liberal.

Posted by: liberal at September 14, 2004 10:54 AM

Posting about Lott without linking Lambert is like talking about the Lincoln-Douglas debates without quoting Lincoln.

Yes, that simile will do. I should find a way to make it stronger.

Posted by: Nash at September 14, 2004 12:45 PM

Back now…on New Zealand time.

Liberal: Groseclose and Milyo’s methodology seems pretty sound to me. I described the conditions under which your worries about the accuracy of cited sources might cause bias and they seemed fairly unlikely.

To your second point regarding validity now. The ADA scores used by Groseclose are about as close as we come to a gold standard in assessing the ideology of legislators; they’re used everywhere in the literature. I’m not sure how one might go about validating the results for media organizations, other than by having lots of folks attacking the problem using different methodologies and seeing if the results are broadly comparable. I can see no a priori reason why we wouldn’t think that media organizations shouldn’t be able to be placed on the same ideological spectrum as politicians. Poole and Rosenthal show fairly convincingly that US politics reduces to a single right-left ideological dimension; if political markets are competitive, then this should be reflecting the underlying ideological preferences of voters. Voters buy news, and prefer news that more closely reflects their ideological preferences to news that make them angry. Seems entirely plausible that this should all be on the same dimension.

Moreover, the rapid rise of Fox News really suggests that Groseclose and Milyo are correct. Think of a Hotelling model. If all media outlets other than Fox were broadly to the political centre, Fox would never have gotten the market share it now has. If voters are normally distributed along an ideological spectrum and people choose the news source closest to themselves on that ideological spectrum, there is no way that Fox could have risen to lead in the Neilson ratings unless there were a vast space left open on the right.

To your story explaining the citation pattern, you’d need to specify that members of the media have a greater preference for sensationalism than do members of Congress. And, you’d have to specify that sensationalism leads to a particular kind of bias where accurate statistics on the left are systematically more sensationalistic than accurate statistics on the right. Crime statistics seem an obvious counterexample. Or numbers on welfare fraud. Or numbers on illegal immigrants. But, this mechanism is at least plausible. Let’s specify now that you’re correct; the results are entirely an artefact of sensationalism. That doesn’t overturn the results of the Groseclose study, it just says that the bias isn’t intentional. The mapping from Congressional ADA scores to think tank ideology remains sound; we’ve identified think tank ideology. Media outlets don’t care about ideology, they just look for good sound bites, and the better sound bites come from the think tanks that we’ve identified as being on the left. This leads there to be a left wing bias in the presented news, though there may be no left wing bias at all among individual journalists. The bias is still there, it’s just unintentional.

On the other hand, look again to the rise of Fox News. They’ve not been immune to sensationalism. If everything were due to the public having a preference for sensationalism in their news stories, and there being a bias such that left wing issues are inherently more sensationalistic, Fox News wouldn’t have had room to come into the market.

Posted by: Eric Crampton at September 14, 2004 02:10 PM

Liberal:

It's a sad fact that in the social sciences we have few natural experiments and lab experiments can only do so much. We make the best with what we have. The Groseclose methodology, to me, seems an entirely appropriate, and ingenious, way of approaching a difficult problem. And, the results are consistent with other observed phenomenon like the rise of Fox News, which your alternative hypothesis of no bias cannot explain.

As to whether political markets are efficient, well, there's actually a literature on this. Don Wittman argues they are; a bunch of public choice people disagree (but many of the criticisms fall short of the mark). My take on it is that they're efficient at giving voters what they want on things voters actually care about. And, ideology seems to be one of the things they care about. Lots of papers looking at whether legislator ideology matches that of constituents, whether legislators shirking on ideological dimensions are hurt at re-election time, and so on.

I'll have to disagree with you strongly re sensationalism. I was giving you your best possible case. After listing reasons why I disagreed with your sensationalism story, I went on to see whether granting sensationalism as an explanation would invalidate the results, and it doesn't. Simpler example. Say that all think tanks that are conservative start with the letter C and all liberal ones start with an L. Suppose then that journalists don't care about ideology but really do like the letter L, and so they only cite things from places starting with the letter L. You'd call that a media bias favouring the letter L. And it would be. But the result of it would also be a liberal bias in the press.

I suspect that your methodological critiques here run deeper than Groseclose ("yet another embarrasment for the social sciences" I think you'd called it) and that you take issue with how the social sciences and economics are conducted in general. I would be willing to put money on the Groseclose paper being published in a decent economics or poli sci journal. It meets my view of the standards. And I haven't the time or the inclination to engage in debate about the general state of the social sciences.

Posted by: Eric Crampton at September 14, 2004 07:40 PM

There are some strong assumptions in the statistical analysis of the Groseclose paper. I didn't see a justification of using the Weibull distribution for the error term of an important equation. That distribution can look like a lot of things depending on the parameters. How robust are the results to that assumption? The authors seem to have changed their statistical methodology in the middle of the research because the computer would take a week to calculate the estimates of the full model. It would be interesting to let it run and see what happens. That kind of thing has been done. You let the PC run at night and hit pause in the morning. You run a spare for as long as it takes.

But that kind of thing is OK, when you are trying something new. But I could see no attempt to determine how sensitivte the results are to alternative assumptions or methods of estimation or approaches to hypothesis testing. So I think, statistically, it is more of an exploratory analysis than anything else.

But I can't get past the conceptual issue. If 10% of Congressional references were by Marxist Congressman to work by Marxist thinktanks, and 10% by the Druids to Druid thinktanks and 10% by Flatearthers to Flatearth thinktanks, then the idea is that the media should rely on those thinktanks 30% of the time for reliable factual information to report? Is that right? Because some ill defined political market determines what is true?

I think Eric Crampton's example of a person citing objectively accurate information but still showing bias is not good. If a person is reporting on the facts of global warming (whether pro or con) what do studies of the cost of GNP growth have to do with that?

Posted by: jm at September 14, 2004 10:44 PM

Oops... meant to say
"what do studies of *the cost of control in terms of* GNP growth have to do with that?"

Posted by: jml at September 14, 2004 10:51 PM

jml wrote, "But I can't get past the conceptual issue. If 10% of Congressional references were by Marxist Congressman to work by Marxist thinktanks, and 10% by the Druids to Druid thinktanks and 10% by Flatearthers to Flatearth thinktanks, then the idea is that the media should rely on those thinktanks 30% of the time for reliable factual information to report? Is that right? Because some ill defined political market determines what is true?"

Exactly.

Another example, this time from the right, is the age of the earth. I'll wager that if you look at polls, a substantial fraction of the populace thinks the earth's age is of the order of magnitude given by the Bible. Now suppose that a substantial fraction of members of Congress had the same view. Then the conceptual assumptions of G & M would say that all science reporting involved with geology was "biased".

"I think Eric Crampton's example of a person citing objectively accurate information but still showing bias is not good. If a person is reporting on the facts of global warming (whether pro or con) what do studies of the cost of GNP growth have to do with that?"

EC's point is actually well-taken. One issue of press bias, that the press won't admit to, is the "meta-issue" of deciding which stories to cover. A recent example is the Swift Boat campaign; once the documentary record showed zero support for the anti-Kerry side, the media could have stopped coverage. Their decision not to is a *choice*.

Posted by: liberal at September 15, 2004 04:54 AM

Eric Crampton wrote, "It's a sad fact that in the social sciences we have few natural experiments and lab experiments can only do so much. We make the best with what we have. The Groseclose methodology, to me, seems an entirely appropriate, and ingenious, way of approaching a difficult problem."

Perhaps, but as I've shown, it's wrong; there appears to be no consistency check that allows one to invalidate my counter model that the mapping should be
media ADA score --> congressional ADA score + x_0

"And, the results are consistent with other observed phenomenon like the rise of Fox News, which your alternative hypothesis of no bias cannot explain."

That's because you're not familiar with the history of media in the US. The recent rise of right-wing media in general is partly due to the repeal of the FCC's Fairness Rule.

Posted by: liberal at September 15, 2004 04:58 AM

Eric Crampton wrote, "Simpler example. Say that all think tanks that are conservative start with the letter C and all liberal ones start with an L. Suppose then that journalists don't care about ideology but really do like the letter L, and so they only cite things from places starting with the letter L. You'd call that a media bias favouring the letter L. And it would be. But the result of it would also be a liberal bias in the press."

Again, wrong. You're conflating two definitions of "bias". You're using something akin to the statistical definition. The definition people are using when they debate "media 'bias' " is this one, found in my dictionary: "an inclination of temperment or outlook; esp: prejudice".

Posted by: liberal at September 15, 2004 05:04 AM

Eric Crampton wrote, "I suspect that your methodological critiques here run deeper than Groseclose ('yet another embarrasment for the social sciences' I think you'd called it) and that you take issue with how the social sciences and economics are conducted in general."

Not all of it, but a lot of it. Econometrics in particular is a very useful technique, but one that can easily be abused. Tim Lambert, above, gave you an example: John Lott doing model hunting.

Here's a prominent example (not re econometrics, but economics in general): when Clinton raised taxes on upper incomes, Martin Feldstein---current president of NBER!---predicted that the cuts would be damaging to the economy. A prediction that singularly failed to materialize. Feldstein is an excellent example of an economist whose descriptive (positive) claims are influenced by his normative outlook.

"I would be willing to put money on the Groseclose paper being published in a decent economics or poli sci journal."

Again, the logical fallacy of appeal to popularity:
http://www.intrepidsoftware.com/fallacy/pop.php

Posted by: liberal at September 15, 2004 05:13 AM

Eric Crampton wrote, "I would be willing to put money on the Groseclose paper being published in a decent economics or poli sci journal. It meets my view of the standards. And I haven't the time or the inclination to engage in debate about the general state of the social sciences."

The other thing of it is is that you have a misperception of academic research.

Specifically, I wouldn't say that the G & M paper shouldn't be *published*, even in a prominent journal. But that doesn't logically or even substantially imply that one should therefore actually agree with their conclusions in the context of the debate as to whether the media is biased.

This isn't just true of the social sciences but also e.g. medical science. There's all sorts of stuff published all the time of varying degrees of quality, but one would be foolhardy to accept results just because they were published in a prominent journal.

My broadside against social science isn't really unfair. The fact is that (a) social science has (to its credit) tackled a much more complicated task than "traditional" science because its tackling a much more complicated system (humanity), and (b) it's just going to be the case that even the most fair, careful economist or other social scientist is going to be more influenced by his normative perceptions than, say, a physicist. That's not an indictment; it's just the nature of the subjects.

Posted by: liberal at September 15, 2004 05:25 AM

Bob H wrote, "'Didn't Lott have a bit of a "dog ate my howework" problem with his supposedly groundbreaking results from "More Guns, Less Crime"?' Hardly. ..."

I don't think that (and the ensuing claims) are a fair summary of this history. Interested readers are encouraged to look at all the evidence collected by Tim Lambert at
http://www.cse.unsw.edu.au/~lambert/guns/lott98update.html

"Academic after academic has replicated [Lott's] results... [which have] survived remarkably well.... This classic has stood the test of time...."

That's a complete distortion of the history of the issue. Many esteemed researchers haven't replicated Lott's results. Furthermore, much of his methodology has been called into question. Interested readers should, again, consult the link above.

Posted by: liberal at September 15, 2004 09:45 AM

Bob H wrote, "That site is hardly a useful or accurate source."

And I see you provide no evidence for this claim.

"Did not the Journal of Law and Economics publish a special issue with papers using the data from Lott's book?"

Gosh, then his results *must* be valid.

Posted by: liberal at September 15, 2004 02:36 PM

Credula vitam spes fovet et melius cras fore semper dicit - Credulous hope supports our life, and always says that tomorrow will be better. (Tibullus)

Posted by: cam girls at September 23, 2004 06:10 AM

5 card stud
7 card stud
7 card stud hi lo
card game
free texas holdem
free texas holdem poker
holdem poker
internet poker
live poker
multi player poker
omaha hi
omaha hi lo
one on one
online
online poker
Online Poker Games
poker games
texas holdem
texas holdem poker
texasholdem
world poker tour

Posted by: texas holdem at September 27, 2004 11:00 PM

best casino online
best online casino
blackjack online casino
buy online casino
casino betting online
casino careers online
casino gambling online
casino game online
casino games online
casino games online free
casino online
casino online free
casino online free online casino
casino online gambling
casino online gambling games com
casino online games
casino play for fun online
casinos online
free casino games online
free casino online
free casino online games
free money online casino
free online casino
free online casino bonus
free online casino game
free online casino games
free online casino money
free online casino slots
grand casino online
grand online casino
how to start an online casino
hoyle casino online
internet casino online
largest online casino
las vegas casino online
las vegas online casino
lasseters online casino
lasvegas online casino
lucky jacks online casino guide
lucky nugget online casino
mac online casino
mgm grand online casino
mgm mirage online casino
mgm online casino
new online casino
no download online casino
online casino
online casino affiliate program
online casino american express
online casino betting
online casino black jack
online casino blackjack
online casino bonus
online casino bonuses
online casino business
online casino com
online casino complaints
online casino connection
online casino craps
online casino deals
online casino directory
online casino for sale
online casino fraud
online casino free
online casino free bonus
online casino free cash
online casino free money
online casino gambling
online casino game
online casino games
online casino games free
online casino gaming
online casino guide
online casino law
online casino laws
online casino let it ride
online casino mac
online casino macintosh
online casino marketing
online casino news
online casino no deposit
online casino no download
online casino odds
online casino payouts
online casino paypal
online casino poker
online casino portals
online casino promotion
online casino promotions
online casino rankings
online casino rating
online casino ratings
online casino review
online casino reviews
online casino roulette
online casino scams
online casino slots
online casino software
online casino sport
online casino sport book
online casino sports betting
online casino sportsbook
online casino tips
online casinos
online free casino
online free casino games
online gambling casino
online internet casino
online sportsbook and online casino games
own an online casino
own online casino
own your own online casino
play casino games online
play casino online
river belle online casino
riverbelle online casino
sands online casino
sportsbook online casino
the online casino
top online casino
vegas casino online
vegas online casino
vegas towers online casino internet gambling virtual
www casino online gambling games com

Posted by: online casino at September 27, 2004 11:00 PM

In spiritu et veritate - In spirit and truth. (Versio Vulgata)

Posted by: beastiality bbs at September 28, 2004 01:22 PM

If you come looking to learn the rules or the basics of texas
holdem, check out our Poker
Basics, or our Texas
Holdem strategy section. You be able to find out both how the game is played,
and how to play the game better.

Posted by: texas holdem at September 28, 2004 03:09 PM

online poker

Posted by: online poker at September 29, 2004 12:09 AM

Looking for phone chat? try our juicy chat lines!

Posted by: Phone Chat at September 30, 2004 11:50 PM

Never let your sense of morals prevent you from doing what's right.

Posted by: Benton Janet at October 1, 2004 05:12 PM

7809 How can this all be as nice? Check out my site http://www.pai-gow-keno.com

Posted by: keno at October 8, 2004 04:24 AM

3049 http://www.texas-hold-em-i.com play texas hold em online here.

Posted by: texas hold em at October 12, 2004 06:34 AM

Your safe Online Pharmacy

Posted by: Your safe Online Pharmacy at October 12, 2004 09:52 PM

Buy cheap fioricet online

Posted by: Buy cheap fioricet online at October 13, 2004 09:35 PM

6851 http://www.e-texas-hold-em.com

texas hold em

Posted by: texas hold em at October 14, 2004 03:16 PM

home mortgage loan

Posted by: home mortgage loan at October 15, 2004 12:00 AM

Online Masters Degree Programs

Posted by: Online Masters Degree Programs at October 15, 2004 11:54 PM

1179 http://www.rapid-debt-consolidation.com

consolidate debt

Posted by: debt consolidation at October 17, 2004 12:17 AM

783 http://www.e-roulette.info

Posted by: roulette at October 17, 2004 10:21 PM

4947 You only get one set of teeth. Take care of them with a good
dental plan. Dental
insurance is
money well spent. I sleep better since I signed up for my new dental insurance
plan.
Get yours at: http://dental-insurance-plan.freeservers.com/
http://www.online-sports-betting-source.com

online sports betting
Sports Betting Odds
Sports Betting Line
Football Sports Betting
NCAA Sports Betting
NFL Football Betting Line
American Football Betting
Online American Football Betting
Online Football Betting Lines
Online Sports Book
NFL Football Odds
NCAA Football Odds
Las Vegas Football Odds
2004 Super Bowl Odds
Sports Book Betting
Las Vegas Sports Book
Sports Book Odds
Internet Sports Gambling
Sports Betting Information
College Football Betting Odds
College Betting Line
NFL Point Spread
NCAA Point Spread
Super Bowl Point Spread
Vegas Point Spread
Legal Sports Betting
Online Betting Odds
Online Sports Book Betting
Online Sports Betting Site
home based business
home based business
opportunity

Posted by: dental plans at October 20, 2004 09:22 AM

3465 online casino games

Posted by: casino games at October 21, 2004 11:42 PM

1654 You know anti wrinkle cream
can work Did you know online degree gets you jobs? online degrees

Posted by: anti wrinkle cream at October 22, 2004 02:59 AM

Online Casino Games

Posted by: Online Casino Games at October 24, 2004 01:06 AM

mortgage leads

Posted by: mortgage leads at October 26, 2004 12:30 AM

mortgage leads

Posted by: mortgage leads at October 26, 2004 01:19 AM

1618 http://www.blackjack-1000.com
PlLAY the best blackjack-1000.com
debt consolidation only.

Posted by: free online blackjack at October 26, 2004 08:10 PM

6787 http://www.i--cialis.net
Offering Cialis with overnight delivery. Also, If your looking for generic cialis this is a good site to visit.

Posted by: Cialis at October 26, 2004 10:47 PM

refinance mortgage

Posted by: refinance mortgage at October 29, 2004 12:29 AM

1765 Kona Coffee Starbucks Coffee Jamaica Blue Mountain
Coffee
coffee maker gourmet coffee green mountain coffee kenya coffee organic coffee specialty coffee folgers coffee coffee brewers costa rica coffee Tullys Coffee Millstone Coffee coffee grinder http://www.coffee-delivered.com

Posted by: gourmet coffee at November 3, 2004 03:37 AM

4683
directv
direct tv directv satellite direct tv satellite directv dvr direct tv dvr direct tv tivo directv tivo directway
direcway
directway internet directway satellite direcway internet direcway internet free hbo free cinemax free dvd player satellite radio http://www.satellitetvboutique.com

directv
direct tv
directv satellite direct tv satellite directv dvr direct tv dvr direct tv tivo directv tivo directway
direcway
directway internet directway satellite direcway internet direcway internet free hbo free cinemax free dvd player satellite radio http://satellite-tv.cjb.net

Posted by: direct tv at November 10, 2004 08:15 AM

A payday loans secured by a personal check - such as a payday loan - is very expensive credit. Let's say you write a personal check for $115 to borrow $100 in quick cash for up to 14 days. The check casher or payday lender agrees to hold the check until your next payday loan cash. At that time, depending on the particular plan, the lender deposits the check, you redeem the check by paying the $115 in cash, or you roll-over the check by paying a fee to extend the payday loans for another two weeks. In this example, the cost of the initial loan is a $15 finance charge and 391 percent APR. If you roll-over the loan three times, the finance charge would climb to $60 to borrow $100 payday loan cash advance.

Posted by: Quick Cash at November 10, 2004 08:43 PM

or without Merck over free Bayer get FDA The night http://fioricet-i.741.com supplies AstraZeneca of Order Fioricet delivery buy The largest now approved. variety vitamins. and Pfizer 30 days of largest Ortho-McNeil

Posted by: Fioricet at November 21, 2004 07:16 PM

The or delivery AstraZeneca of and Order Fioricet Ortho-McNeil days vitamins. get variety Bayer night free http://fioricet.redi.tk buy approved. of 30 Merck without FDA supplies largest Pfizer over largest now The

Posted by: Fioricet at November 21, 2004 08:23 PM

get 30 The and largest or night delivery http://fioricetweb.bravehost.com Merck FDA vitamins. The Order supplies largest AstraZeneca Bayer Ortho-McNeil now Fioricet approved. of free without buy over variety of Pfizer days

Posted by: Fioricet at November 21, 2004 09:04 PM

Going back to finish your online college degree or get more education can be a difficult, time consuming and sometimes impossible proposition. But with new advances in technology, earning additional online degrees has never been quite so convenient. Our online degree programs offer bachelors, masters and doctoral online college degrees programs that are designed for working adults and aspiring professionals like yourself. So don't delay, earn your college degree online today: online or from a campus location near you. Going back to finish your Associate Degree Online or get more education can be a difficult, time consuming and sometimes impossible. But with new advances in technology, earning Bachelors Degree Online has never been quite so convenient. Our Masters Degree Online program are designed for working adults and aspiring professionals like yourself. So don't delay, earn your Doctoral Degree Online today. Going back to finish your Business Degrees Online or get more education can be a difficult, time consuming and sometimes impossible. But with new advances in technology, earning Design Degrees Online has never been quite so convenient. Our Technology Degrees Online program are designed for Healthcare Degrees Online working adults and aspiring professionals like yourself. So don't delay, earn your Culinary Arts Degrees Online today. Simply review our Education Degrees Online selections and choose the option that best fit what you are looking for. Each Criminal Justice Degrees Online has detailed online information on their Travel & Tourism Degrees Online and certificate programs as well as student services, and more.

Posted by: Online College Degree at November 21, 2004 09:41 PM

8444 Kona Coffee Starbucks Coffee Jamaica Blue Mountain
Coffee
coffee maker gourmet coffee green mountain coffee kenya coffee organic coffee specialty coffee folgers coffee coffee brewers costa rica coffee Tullys Coffee Millstone Coffee coffee grinder
http://www.coffee-delivered.com

You only get one set of teeth. Take care of them with a good
dental plan.
Dental
insurance is
money well spent. I sleep better since I signed up for my new dental insurance
plan.
Get yours at: http://dental-insurance-plan.freeservers.com/

individual
dental
plans
You only get one set of teeth. Take care of them with a good
dental plan.
Dental
insurance is
money well spent. I sleep better since I signed up for my new
dental insurance
plan.
Get yours at:
http://www.dental-plan-source.com
individual dental
plans
You only get one set of teeth. Take care of them with a good
dental plan.
Dental
insurance is
money well spent. I sleep better since I signed up for my new
dental insurance
plan. Get yours at:
http://www.e-dental-insurance-plans.com/
individual dental
plans

Posted by: dental insurance at November 24, 2004 05:15 AM

You just go out and preach the gospel!

Posted by: Tranny Hunt at November 24, 2004 12:49 PM

Excelente Blog

Posted by: Upskirts Mania at November 26, 2004 09:32 PM

Some believe in God but some not it is not right. What a shame!

Posted by: Upskirt Sniper at November 27, 2004 12:43 PM
Post a comment









Remember personal info?