Sloppiness and its enemies
« previous post | next post »
Paul Krugman ("The Sloppiness Syndrome", NYT 5/22/2013):
So what is it with New Republic alumni? First Michael Kinsley, then Charles Lane, weigh in with defenses of austerity that aren’t just wrong, but painfully ill-informed. Kinsley not only makes a really bad analogy between current events and the 1970s, he seems not to know anything about what happened in the 1970s either. Lane attacks stimulus advocates for failing to address an argument that I actually discussed, at length, in my last column but one.
Whence cometh this epidemic of sheer sloppiness?
I’m not really sure, but in these cases I suspect it has a lot to do with the famed TNR/Slate premium on being “counterintuitive”, which in practice meant skewering supposed liberal pieties. (Kinsley himself joked that TNR should be renamed “Even the liberal New Republic”).
Of course, economics is not the only field where pundits are guilty of astonishing carelessness in the service of a story line; nor is clever counterintuiveness the only rhetorical frame that motivates such bullshitting (to use the correct philosophical terminology); nor are TNR and Slate the only publications where such material can be found.
For example, David Brooks writes for the New York Times, as Prof. Krugman does, and Brooks' motivation for shockingly sloppy presentations of "facts" and theories is usually to reinforce an all-too-familiar point of view, like "boys and girls need to be educated differently", or "social roles are determined by 'patterns that nature and evolution laid down long, long ago'", or "Western societies have an individualist mentality and Eastern societies have a collectivist mentality", or "individualism and governmentalization are rising, and morality is declining", or some combination of stuff like that.
(Though I guess that these long-familiar points of view can also be seen as "counterintuitive … skewering [of] supposed liberal pieties", although the skewering consists of presenting earlier supposed conservative pieties.)
Anyhow, Krugman concludes that 'twas ever thus:
[H]ere’s my guess: if you went back through all the clever counterintuitiveness of past years, you’d find that a lot of it was every bit as sloppy and ill-informed as what we’re seeing now. The difference is the existence now of a policy blogosphere (in economics, of course, but in a number of disciplines too), which makes bluffing harder. In the past grotesquely ill-informed articles on, say, the Clinton health plan could sit out there for years, with only a handful of specialists aware of just how bad they were; now the pundit emperor’s nakedness is all over the web within days if not hours.
I suspect that a lot of punditry was sloppy and ill-informed long before the founding of Slate in 1995, with Kinsley as its first editor — and for that matter even before the founding of The New Republic in 1914.
But I agree that the world is better off with more of what Ben Goldacre has called "The noble and ancient tradition of moron-baiting", and that the blogosphere, for all its many faults, is the best method so far invented for "[making] bluffing harder".
(Obviously, all of the individuals under discussion are in fact highly intelligent — and for that matter so are the people that Goldacre critiques in his column, or the people that Martin Gardner critiqued. This issue is not intelligence, but willingness to bluff in the service of making an argument.)
BlueLoom said,
May 24, 2013 @ 12:06 pm
" (Kinsley himself joked that TNR should be renamed “Even the liberal New Republic”). "
I (a formerly loyal TNR reader) gave up in about the 1970s. It really hasn't been liberal for decades.
Philip said,
May 24, 2013 @ 12:21 pm
Is there a Language Log angle here? I thought you were going to discuss Krugman's "whence cometh."
[(myl) The LL angle is the link to eight earlier posts…]
Paul Mulshine said,
May 24, 2013 @ 12:22 pm
Krugman might have an argument there if he himself weren't so sloppy in his writing: http://blog.nj.com/njv_paul_mulshine/2010/10/new_yorkers_can_stick_those_tu.html
[(myl) Actually, your 2010 post (and your comment here today) seems to me to be striking confirmation of Krugman's argument. He made a confident assertion that there's just one rail tunnel now connecting New Jersey with New York City; your Star-Ledger blog post corrects this, asserting that there are in fact six such tunnels (and I'm assuming without checking that you're right); you've thereby called his bluff, in a form that interested parties in present and future times can read.
But if I were posting about this, I'd do some further fact-checking, and present links to sources. What and where are the six tunnels? Aside from your assertion, how do we know that they exist?
In Krugman's 2010 column (which you don't link to — does the Star-Ledger have a policy against hyperlinks?), he does write that
right now there’s just one century-old rail tunnel linking New Jersey and New York — and it’s running close to capacity.
And Krugman wrote again about this topic in 2012, but in that column he asserts that
Demand for public transit is rising across America, reflecting both population growth and shifting preferences in an era of high gas prices. Yet New Jersey is linked to New York by just two single-track tunnels built a century ago — tunnels that run at 100 percent of capacity during peak hours.
So he's given two different numbers at different times, without citing any sources in either case.
The Wikipedia article on "Bridges and Tunnels in New York City" lists five tunnels across the Hudson river, of which only three (Downtown Hudson Tubes, Uptown Hudson Tubes, and North River Tunnels) appear to carry rail traffic.
Can you cite some sources now for your claim that there are six such tunnels?]
Bob Ladd said,
May 24, 2013 @ 1:26 pm
@MYL: From the names of the three tunnels you list, I suspect that they're all paired, with traffic in each direction going in one of the pair. So strictly speaking there are six tunnels. Which only goes to show that distinguishing between lying and bullshitting, or even just fact-checking, is not a straightforward. If you want to show that public transport is inadequate, you can say there are three, and survive a fact-check. If you want to show that public transport is doing just fine, you can say there are six, and also survive a fact-check.
[(myl) Or, in this modern age, you can provide a link to a detailed discussion of the facts — and perhaps summarize the facts briefly in a useful fashion, e.g. here by saying that "there are three sets of paired tubes, carrying six tracks in total, three in each direction" (if that's what the facts are). Anyhow, it certainly looks like Prof. Krugman loses this round.]
J.W. Brewer said,
May 24, 2013 @ 1:42 pm
I expect Krugman's internal inconsistency is the same thing – he's counting a particular pair of tunnels (the "North River" ones that go under the Hudson f/k/a North between NJ to Penn Station) as one in one setting and two in another. Krugman is simply ignoring the other two sets of tubes that carry what are these days called the PATH trains, and I think that's at least as likely to be sheer ignorance/oversight rather than an idiosyncratic meaning of either "rail tunnel" or "public transit" that excludes the PATH trains and the tubes through which they cross the Hudson. In the pre-PATH days, my impression (which is not 100% guaranteed accurate) is that those were often simply called something like the "Trans-Hudson Tubes"; which might mean that each pair was being referred to in the plural by the general public, not just engineering/infrastructure geeks (if memory serves, pre-9/11 there was still some ancient signage in the subway stations adjacent to the WTC consistent with this former usage). This contrasts to the popular usage whereby e.g. the Lincoln Tunnel is referred to in the singular even though it's a set of three parallel tubes. Not sure if that's a rail v. road distinction, or just local-toponym arbitrariness.
Note that from an engineering perspective (including resiliency of the system to disasters such as flood or explosion), the difference between two smaller tunnels built in parallel as part of the same system and one larger tunnel with the same peak capacity as that of the smaller two put together is not necessarily trivial.
iaom said,
May 24, 2013 @ 1:50 pm
I'd call the tunnel thing nitpicking — the canceled ARC tunnel was meant to be for NJ Rail trains, which currently share only the North River Tunnels connection (one way in each direction) with Amtrak. The other rail tunnels are specific to PATH trains; they're basically third-rail subway tunnels and aren't usable for the longer-distance trains. So there's only one connection for the kind of trains under discussion in the Krugman column.
Counting linked tubes as separate tunnels is pretty unusual anyway; the Lincoln Tunnel has three tubes and the Holland Tunnel has two tubes, but people would look at you like you had five heads if you therefore insisted that there are five automobile tunnels connecting NYC and NJ instead of just two.
[(myl) The nitpicking would be unnecessary if either Krugman or Mulshine had spent an extra 10 words or so to describe the transhudson rail situation clearly. The details do matter to Krugman's argument, so he ought to have been more specific; and since Mulshine is accusing Krugman of careless peddling of nonsense, it seems even more important for him to be make a specific case for what Krugman got wrong and why it matters (if it does).]
Keith M Ellis said,
May 24, 2013 @ 1:54 pm
This is relevant to linguistics, though, because there are large swaths of policy discussion that rely upon economics — except it's economics as "commonsense" and "what everyone knows" in the way that many language discussions are.
For example, in Kinsley's rebuttal to Krugman, he cites Krugman's recent blog post discussing how the deficit has been falling and contrasts that with Krugman's gloomy economic message, and then claims that this is an inconsistency. But the deficit is not the economy! (Kinsley's doubly wrong, too, demonstrating that he's cherry-picking without even bothering to pay attention to what he's claiming to have read: Krugman is quite clear, in that falling deficit blog post, that he thinks a falling deficit now is bad for the economy. So no contradiction in that sense, either.)
In Kinsley's TNR piece, he gets almost all of the economics wrong, both theory and matters of fact. In his response to his critics, he attributes the criticism to ideologically-aligned defenders of Krugman but, in fact, the criticism has been fairly ideologically broad because his errors are that egregious.
The other similarity to discussions about language is the confusion between being descriptive and being normative. A lot of language peevish prescriptivism is built upon an implicit set of normative values, usually a declinist narrative of cultural, even moral, decay. The public debates about the financial collapse, the recession, stimulus, monetary policy, unemployment, deficits and debt have been drenched in moral intuition. Left or right, both sides discuss economic policy in light of the recession as primarily about identifying villains and punishing them. And various past and present and proposed policies are primarily evaluated within the context of whether they reward or punish villains, and only secondarily about whether they would raise employment and growth.
With both language and economics there's vast amounts of public discourse by people with little knowledge of the subject matter, with little or no attempt to fact-check, built upon conventional wisdom and moral intuition and this is especially so because this reflects the audience's relationship to the subject matter. My observation is that most people simply don't believe that there's much to know about either linguistics or economics, other than fiddly details, that they don't already know intuitively.
(Note: everything I wrote is much more true with regard to macroeconomics than microeconomics, although it's true with regard to the latter, as well. There's probably a parallel in linguistics — historical linguistics versus phonology?)
iaom said,
May 24, 2013 @ 3:54 pm
myl: Sorry, I meant I felt Mulshine (not you) was nitpicking — as you put it, "accusing Krugman of careless peddling of nonsense" where I would say Krugman's error would be better described as imprecision — although you're probably right that Krugman ought to have been more clear about the situation.
What makes it more amusing is that Mulshine's entire column is based on a glaring error: Mulshine brings up Krugman as a chief example of the "New Yorkers" trying to push their will on New Jersey in the form of tunnels and gasoline taxes, even though Krugman is a fellow resident of New Jersey who resides and teaches in Princeton and travels to Manhattan by train when necessary for business.
Paul Mulshine said,
May 24, 2013 @ 9:23 pm
There are six tunnels, as I said.
[(myl) It's good to see that your mastery of supporting references has remained consistent.]
Alex Blaze said,
May 24, 2013 @ 11:53 pm
Obviously the problem is caused by lots of things, so that big caveat here.
First of all, some pundits *are* stupid. Agreed, most aren't. But David Broder comes to mind.
Second, many are lazy. It's hard to discern from stupid. But there are some real-world differences here.
Third, many are dishonest, as MYL points out. Sometimes that dishonesty is in service of the laziness (it'd be too hard to look this up, so let me make up some other fact to cover my tracks), but mostly it's just propaganda.
Last, especially when it comes to "centrist," "even the liberal/conservative," "unbiased" punditry, it's a lack of expertise. Krugman basically calls this laziness (that they didn't do their homework), but it's more than that. Most aspiring journalists get hired based on writing ability (if they don't have connections). Logic, math skills, and technical knowledge are unimportant. Since editors started in the same positions, they assume that you can learn anything you need to know so long as you can write.
And then they get promoted to writing about more important topics to wider audiences, and who has time to actually learn anything at that point?
For example, what is Kinsley supposed to do? Go back to college and get a MA in economics before he writes about the economy? That'd take him out of work for several years, and his job would not be waiting for him when he finished.
Another example is the Simon Heffer-esque peever. While plenty of people will acknowledge that trained economists exist but continue to follow their prejudices because they just assume they must be confirmed by the science, I've noticed that self-proclaimed grammarians rarely even know that grammar is an actual field of study and know-nothing pundits can make a quick buck off the rubes' ignorance.
But what can Simon Heffer do – go back to school and study grammar for 6 years?
This isn't meant as sympathy to the writers of the world who don't know much but find themselves with a platform intended for someone a whole lot more knowledgeable than they are; it's pointing the finger at the editors who, for too long, impose the standards under which they were hired/promoted on the writing industry much to the detriment of their readers.
And the internet is part of the cure – readers can bypass the newspaper editor and read from actual experts. Washington Post readers interested in religion had to put up with Sally Quinn's column on the topic for several years when she's not an expert on religion (and probably not religious) just because the Post didn't want to fire her. Now those who are interested in religion can find many experts who are writing on the topic.
Paul Mulshine said,
May 25, 2013 @ 5:19 pm
There's no need to support an assertion that I documented ion the piece.
[(myl) I presume that by "the piece" you mean this blog post for which you provided a URL, in which you wrote:
In fact, there are six rail tunnels connecting New Jersey and Manhattan. Two run to Penn Station and are shared by NJ Transit and Amtrak. Four others are used by PATH trains. The ARC plan would add an additional two for a total of eight. But who’s counting?
That's all that you say about them — and it certainly counts as "an assertion", but I guess I don't see where the "documented" part comes in.]
The existence of those six tunnels support it. As for my further point about the sloppiness, Krugman made no effort to learn why the Access to the Region's Core project was canceled. It was originally supposed to provide acces to the "core" i.e. grand Central, which is not now linked to Penn. Political considerations killed that by 2002. Then further political considerations kept those two tunnels from going to Penn Station. They were going to go to a dead-end terminal of a design obsolete for a century. That in turn meant they no longer could be shared by Amtrak, which was the other goal of ARC. Krugman was blissfully unaware of any of this because he did no research. If he had, he would have realized that the very people he construed to be his allies, the rail passenger groups, opposed ARC for all the above reasons. He write, "And, on Thursday, Chris Christie, the governor of New Jersey, canceled America’s most important current public works project, the long-planned and much-needed second rail tunnel under the Hudson River." He then shows he has no idea whatsoever of where those plans went off the rails. If he had made on call to the rail passenger groups, he could have found that out. As for me, I took two extended tours of the area with an expert and know every detail of the project. that's the difference between a real journalist and a dilettante.
[(myl) This is interesting and relevant, and I'm glad to learn about it. It wasn't discussed in the piece that you linked to earlier — have you written about it more extensively elsewhere?]
Richard Hershberger said,
May 25, 2013 @ 7:08 pm
"There's no need to support an assertion that I documented ion the piece. The existence of those six tunnels support it."
I can only sit back and admire.
"As for me, I took two extended tours of the area with an expert and know every detail of the project. "
Wow! Two tours–even "extended" ones–is all it took to know every detail? Who is this expert? If he can teach every detail of a project this big with just two tours, he clearly is wasted in civil engineering. He would be the greatest educator in history.
Kaleberg said,
May 25, 2013 @ 8:50 pm
Krugman is probably just the victim of a common New York City blindness. If you ask a New Yorker if he has ever ridden on a train, odds are that he will say that he hasn't, even if you are having this discussion while shouting in a subway car. Most New Yorkers don't think of riding the subway as traveling by train, it's taking the subway, much as one would take an aspirin, but much noisier and smellier.
The handful of New Yorkers who know of the existence of the PATH tubes, possibly a solid 28%, just ahead of the 25% who realize that the Empire State Building is an office building, not an amusement ride for tourists, would just consider them to be New Jersey subway tunnels, but not railroad tunnels. In an amazing example of "category error", they would regard none of these transportation modes as involving trains which are bulkier, provide seating, and travel between major cities.
(I can almost imagine the New Yorker logic. If the PATH tunnels were railroad tunnels, a New Yorker would reason, they would go to a railroad station like Grand Central or what is left of Penn Station. Since they don't, they are not railroad tunnels.)
Paul Krugman may be wrong on this, but he is far from alone.
P.S. The factor of two problem is different. Most people consider something like the Holland Tunnel to be one tunnel, but tunnel engineers count the structural elements, so it is, by their lights, two.
J.W. Brewer said,
May 26, 2013 @ 1:39 pm
Would myl demand documentation in a piece by a journalist who asserted (even against another journalist who had falsely claimed the contrary) that "New York City has five boroughs" or "Pennsylvania is known as the Keystone State"? Why is this any different? New Yorkers who are ignorant of the various different ways in which people who live in New Jersey travel to Manhattan may perhaps be indulged in their ignorance, but not if they are going to opine in public as to whether or not a particular proposal to spend bajillions of tax dollars changing the existing infrastructure is a good idea. Similarly, not knowing how many states Australia has (or how many provinces Canada has) is no big deal for the average American, but you probably shouldn't get it wrong if you're on the NYTimes op-ed page talking about Australian or Canadian regional politics.
Perhaps a more interesting question is how, if Krugman's research assistant had been competently using wikipedia during the relevant time period in order to reduce the risk of this sort of blunder, Krugman could have with the minimum number of words possible acknowledged that there are three pairs of rail tunnels running between NJ and Manhattan while simultaneously explaining why only one pair is in his estimation directly relevant to the issue at hand. Heck, there are even ways in which one could describe that single pair of tunnels with enough qualifying language that you could truthfully assert they were solitary/unique because you would have defined the PATH tubes out of the subject matter under discussion. (E.g. from most of NJ the PATH tubes do not at present offer a "single-seat" ride into Manhattan, which is a significant desideratum in some circles.)
KevinM said,
May 28, 2013 @ 3:59 pm
No, he didn't say there was only one tunnel; all he said was that there was only one "century old rail tunnel" – true, if the other ones are 100 years old, or if they aren't rail tunnels.
Opiniões de especialista | Blog Pra falar de coisas said,
July 16, 2013 @ 9:00 am
[…] Slopiness and its enemies. […]