31 December 2011

What do quallys know, anyway?

Don't worry, Robert Caro,
I'm still pre-ordering your
books in hardcover.
Well, let's write a post guaranteed to alienate some friends and granting agencies! You should know upfront that I'm partly venting here and partly trying to work through some ideas about what "qualitative," "quantitative," and "multi-method" research are.

 I'm reading Daniel Kahneman's Thinking Fast and Slow, which brought to mind some latent fears about qualitative methods. In particular,
  1. What do narrative historians know, anyway?
  2. How are they sure of what they know?
  3. If narrative evidence conflicts with statistical evidence, how do we adjudicate between these accounts?
These ideas were prompted by a gushing piece about Apple's archives at Stanford University. There's lots of happy stories about how "researchers" will be able to learn more about Apple's history, and the history of Silicon Valley, by delving into these records.

And I'm sure they will! Really! But how much more will they learn from this sort of thing than they would from:

  1. The (fairly extensive) existing secondary record, including newspaper and magazine articles
  2. The (fairly extensive) existing primary record (see e.g. and e.g. #2)
  3. Something that I've never actually seen an academic or popular historian do: taking apart quarterly and annual reports and investigating them in the same way a forensic accountant would.
  4. Something that I'm really sure I've never seen an academic or popular historian do: learn how to program or read circuit diagrams in order to understand exactly what it was that, say, Woz was doing when he designed the Apple ][ that made it so efficient, or how innovative the Macintosh's handling of certain features (like rendering the mouse arrow using software instead of hardware) were.

I bring these points up, and particularly the last two, because the more country studies and biographies I read the more I'm torn between the fact that narrative historians clearly know a heck of a lot of facts and Sartori's dictum that "that he who knows only one country knows none."

Yes, I'm using this not to illustrate the depths of human
suffering but the challenges of methodology. 30 Rock
was right: Grad students are the worst people.
The depth of knowledge that Adam Tooze or Richard Evans brings to their discussion about Nazi Germany, for instance, is pretty much matchless. But neither Evans nor Tooze could invest the same amount of effort into knowing, say, Imperial Japan or Fascist Italy; both of them would do the same thing I would do, namely buying some books by Herbert Bix or R.J.B. Bosworth, mine their bibliographies, and spend six months reading through them. Add the UK, France, or God help them the USSR and they're 2.5 years away from feeling comfortable about beginning primary source research.

And, inter alia, they'll have to learn Russian, Japanese, and Frankish, which may take some time.

And what happens after this investment of time? Once they've done this, they're still apt to be subject to both the availability heuristic, based on the ideas they've been exposed to during their research, and the narrative fallacy, because they're in the business of crafting narratives. That presupposes that there is a narrative to be found, that the narrative is meaningful, and that the narrative is not simply the product of randomness.

More to the point, as I alluded earlier, historians do not have a specialized functional area of expertise aside from research methodology. Now, that's a lot! But it's not a field that seems to have vastly become more sophisticated over the past century compared to, say, economics, sociology, political science, or psychology.

I'm not insisting that historians would be better
off if they read this, just rudely insinuating it.
Some may object that I am now conflating sophistication with quantification. And to a large degree I am. Why? Because of work like this study of Nazi voting patterns by Gary King. In reading this article, I learned more about the bases of support for the Nazis than I have in several years of avocational reading about the NSDAP. Then there's work like this Larry Bartels article on the irrationality of retrospective voting, which makes a point that I will go so far as to say that qualitative researchers in any field would be unable to prove: that voters are subject to myopic retrospective voting. And even if you think that the Bartels article, the Achen and Bartels working papers, or the Healy et al classic on football games and incumbent vote share don't prove it, you can return to the classic "shark attack" paper on the Wilson 1916 election. Yes, qualitative researchers might be able to suggest that, as they have since approximately Plato's time, but they are simply incapable of making the argument as persuasively or as conclusively as quantitative researchers.

This is not an argument about whether it is "harder" or "easier" to do qual or quant work. The bad news for quallys is that it is frequently easier for quantitative researchers to write papers, certainly once they've invested in learning methodology, writing code, and putting together a dataset. The worse news for quallys is that quantitative workers benefit from every advance in computer power vastly more than they do; consider recent advances in computer-assisted textual analysis, which is beginning to encroach directly on what historians' core competency used to be: analyzing giant corpuses of text.

Next week: What do quantoids know, anyway? The same half-baked level of rambling "analysis" deployed against number-crunchers who think that knowing how to program R code is a substitute for knowing what Watergate was.

29 December 2011

Why do people use giant puppets in street protests?

His fur isn't the only thing that's red.
So, what's up with the puppets? When Occupiers do their thing, even in Lincoln, Neb., they always make sure to bring their puppets. They don't emerge spontaneously: the puppeteers and the costumes are coordinated in advance, at least to the same degree that a potluck is coordinated. Puppets aren't free, so somebody's funding them; at least one puppeteer draws money from a Kickstarter project. (The puppeteer in question is the notorious guy who wanted to get an MFA in puppetry and blamed the economy for his inability to get a job.)

Not being in the target audience for protesters or puppetry (except) (also except), it's unsurprising that the point is lost on me. But unlike other things that I don't get--Sex and the City, Miller Lite, Seventh-Day Adventism--I can't even see the appeal. Putting the puppets together is hard work, I'll grant; this video certainly suggests that there's planning that goes into each puppet.

But what is it supposed to mean? Why did this become the signature form of leftist protests? Even when Tea Partiers hoist their quasi-literate signs, I at least know immediately what they're for (white people) and against (non-white people). With the puppets, though, I'm not sure what the semiotics are. (Is the Statue of Liberty sincere or ironic?)

The best guess I can hazard is that the puppets are a manifestation of the self-reinforcing insularity of the protests. (This interview with a leading street protest puppeteer certainly suggests that the artists are severely out of touch with the apparent targets of their agit-prop.) Puppeteers communicate only with each other; the spectacle they produce is not the puppets, but the incomprehension of the audience.

27 December 2011

Richard Nixon Was Not Having a Gay Affair With Bebe Rebozo

Nevertheless, this article from the Daily Mail is worth a read, if for no other reason than for this paragraph:
A new biography by Don Fulsom, a veteran Washington reporter who covered the Nixon years, suggests the 37th U.S. President had a serious drink problem, beat his wife and — by the time he was inaugurated in 1969 — had links going back two decades to the Mafia, including with New Orleans godfather Carlos Marcello, then America's most powerful mobster. Yet the most extraordinary claim is that the homophobic Nixon may have been gay himself. If true, it would provide a fascinating insight into the motivation and behaviour of a notoriously secretive politician.
Indeed! I'm writing a new book right now that alleges that George W. Bush was a secret Soviet sleeper agent activated by Vladimir Putin. If true, it would provide a fascinating insight into the motivation and behavior of a notoriously blundering politician.

Similarly, a friend of mine is finishing an article that alleges that Calvin Coolidge was actually carved from an oak tree. If true, it would provide a fascinating insight into the motivation and behavior of a notoriously silent politician.

And let's not forget the conspiracy theory that William Henry Harrison never died of pneumonia but is, instead, still the living president of the United States. If true, it would provide a fascinating insight into the motivation and behavior of a notoriously deceased politician.

21 December 2011

Models, Modeling, and Social Science

This is on the first page of
Google Images results for
Kaiser Fung writes about the difficulties of modeling social behavior, particularly in economics; Andrew Gelman picks up the thread here.

Well, it's always encouraging to hear people much smarter than me say they find modeling behavior to be hard, since I do too. It's much more discouraging to hear people much smarter than me say that modeling is going to be well-nigh impossible. But I think that this is an overly pessimistic conversation. Yes, detecting causal inference is hard; yes, it is probably epistemologically impossible for us to uncover the "real" drivers of human behavior; and, yes, the measurements of human behavior and of the motivating forces of those actions that we routinely deploy are pretty bad.

Nevertheless, the question is not whether we should ask political science to perform as well as physics. (It won't.) The question, rather, is whether we can reasonably expect social science to outperform our intuitions and our folk wisdom, and to become more sophisticated and more certain about at least some propositions over time.

That, I think, is likely. Think of the implications just of (say) the Arrow theorem or the Hibbs bread-and-peace model for understanding American politics. Arrow takes a lot of high-school civics course pabulum off the table, while Hibbs should remind us that much of the churn of polling in the general election is irrelevant even as campaigns are determined, at least in a sense, by the fundamentals that we wish that they would be determined by.

It's odd to see Rodrik cast as the defender of economic orthodoxy, and in particular of quant-led orthodoxy. (Really? That Dani Rodrik?) Yet it's equally strange to see Fung and Gelman so glum about the chances that knowledge can lead to better estimates of both the current social scene and also of the broader regularities of human behavior. Without taking too much of a swipe at social science before, say, 1975 (and I mean from ca. 10,000 BC to AD 1975), I think it's pretty clear that we've learned a lot, and that there are good grounds to think we'll learn more.

20 December 2011

When the movie is better than the book

No, this post wasn't an excuse to link to this picture.
But having written this post, I wanted to link to a pic
of Olivia Wilde, and so why not?
This is always a fun cocktail party game, at least until the pedantic jerk who insists "You can't compare them, because they're meant to be different" shows up and you're all like, "Yeah, we know, jackass, but they can't be completely different, because the book and the movie have the same title, and the same characters, and the same plot, mostly," and then someone quotes McLuhan and then someone else quotes Woody Allen and you realize that you're having the same conversation you had five years ago with a completely different set of people.

All of this, as a preamble for an unequivocal declaration that Cowboys And Aliens, the movie, is better than Cowboys and Aliens, the graphic novel.

I know C&A wasn't a great movie. I saw it, and I didn't regret that I'd spent the money, but I wasn't thrilled. But next to the source material it is an act of unparalleled genius, a tribute to the ability of editorial insight to find the hidden masterpiece, or at least journeyman's work, inside even the roughest chunk of marble.

What was alluded to in the movie (see, the aliens are displacing the indigenous Earthlings just like the cowboys displaced the Indians!) is blatant in the novel. What was interesting in the movie--why are the aliens here?--is rendered completely dry by some exposition given by the alien captain, who not only speaks (a lot) but (a) picks a feud with some local cowboys and (b) talks like a third-rate imitation Lensman villain. Think George Lucas dialogue, but less realistic.

Oddly, at least two Philip K. Dick works ("We Can Remember It For You Wholesale" and "Second Variety") were similarly improved by their translation to the big screen. So maybe it's a genre effect.

19 December 2011

Kim Jong-Il and other people are dead

KJI's death, of course, is the most important thing to happen this month. For a great many people (at least 40 million people on the Korean peninsula), it is the most important thing to happen in 15 years. I assume it will get vastly less press than Christopher Hitchens' death. The Hitch could drink, and write, and sleep around. He was the kind of public intellectual that journalists think public intellectuals should be: glib and impressive to people who had a B average (these days, an A-) in their English major at a "good" school. Update: iOz does it better.

11 December 2011

Fun With Stata and Jogging

This morning, I ran a longer distance than I ever had before. Just an 8k, for those who jog regularly. But the prospect of running five miles--five miles--filled me with the sort of dread that I assume SEC defensive linemen have when they approach final exams in Diff EQ.

What was my time? It was very bad, by the standards of people who run: 10:29/mile. By my standards, this is very, very good---not far off from what I ran in the single mile when I was in middle school but for much, much longer.

Anyway, I wanted to have some fun, so I grabbed the race results from a Web site and decided to do some analysis.

Table 1 displays the results of ordinary least-squares estimates of running times per mile for the participants in the Jingle All the Way 8K Race this morning. Scanty information is available, so the models are pretty spare. Nevertheless they do have some pretty strong results.

Table 1. OLS models of time per mile in seconds
for the Jingle All The Way Race, December 2011.

(1) (2) (3)
Overall Men Only Women Only

age 1.642***1.264***1.955***
(11.52) (5.84) (10.29)
female 78.24***
_cons 484.3***497.8***552.2***
(85.19) (60.89) (84.93)

N 4726 1709 3017

t statistics in parentheses
* p < 0.05, ** p < 0.01, *** p < 0.001

As you can see, women are significantly--significantly--slower than men. For a runner of the same age, women are 78 seconds slower than men. Moreover, women slow down with age somewhat faster than men, losing nearly 2 seconds from their mile time for each year they get older. Compare that with men, who slow by only about 1.3 seconds per year.

The point is reinforced by the figure, which displays kernel density estimates for women's and men's time. On the other hand, there are way, way more women than men running, so it is possible that men just don't run if they're not competitive. (Please, God, I'm not trying to do causal analysis here ... )