Wednesday, April 7, 2010

Does this sound right?

From David Brooks (via a disbelieving Brad Delong):
Over the last 10 years, 60 percent of Americans made more than $100,000 in at least one of those years, and 40 percent had incomes that high for at least three.

Positive Reviews

Have you ever reviewed a paper and thought "wow, this paper is excellent". It's well written, relevant and obviously crafted with care. Well, know I have.

My question is what do you do with it? Do you find items to nitpick so you can show that you were rigorous or do you simply recommend that the paper proceed to publication?

Thoughts?

“Gives new meaning to company town.”

From the Asilomar International Conference on Climate Intervention Technologies reported by Jeff Goodell and brought to our attention by Brad Plumer:

It was generally agreed that for CO2-sucking technologies, private investment was not a problem [n.b., assuming we have some sort of cap-and-trade system in place]. Sunlight-reduction technologies, however, are another issue. if some company (or entrepreneur) is able to develop a new way of injecting particles into the stratosphere that becomes indispensible to the survival of the human race, well, that gives that company or person a lot of leverage.

“I’m not interested in selling my soul to some company who is going to control how much sunlight hits the planet,” said Phil Rasch, a climate modeler at the Pacific Northwest National Laboratory in Washington state. (As one audience member quipped, “Gives new meaning to company town.”) Granger Morgan, the head of the department of engineering and public policy at Carnegie Mellon University, argued that the creation of a profit motive would inevitably lead to a geoengineering lobby: “Lobbying is the last thing we need on this.”

Tuesday, April 6, 2010

I know there's a Schrödinger's cat joke here somewhere

Andrew Gelman has an excellent post up on the contradictory findings of two teams of researchers on the effect child-gender has on parents' political leanings.

Tyler Cowen reports the following claim from sociologists Dalton Conley and Emily Rauscher:

Using nationally-representative data from the [1994] General Social Survey, we [Conley and Rauscher] find that female offspring induce more conservative political identification. We hypothesize that this results from the change in reproductive fitness strategy that daughters may evince.

This surprised me, because less than a year ago, we reported here on a study by economists Andrew Oswald and Nattavudh Powdthavee with the exact opposite finding:

We [Oswald and Powdthavee] document evidence that having daughters leads people to be more sympathetic to left-wing parties. Giving birth to sons, by contrast, seems to make people more likely to vote for a right-wing party. Our data, which are primarily from Great Britain, are longitudinal. We also report corroborative results for a German panel.

Understanding the results (possibly) in terms of "family values"

This is a fun problem: we have two different studies, both by reputable researchers, with opposite results! I took a look at both papers and can't immediately see a resolution, but I will offer some speculations, followed by some scattered comments.

Andrew speculates that the differences in the findings can be explained by the fact that one study looked at the U.S. while the other looked at Britain with some additional data taken from Germany. Commenter DN elaborated further:

Seems to me that this is more likely a "law and order" effect in the US. Families with daughters are more likely to be concerned about 'protecting' their daughters against violent crime. These impacts are either mitigated in Germany and the UK (because of perceptions of the parties) or swamped by other perceived advantages of left-wing parties.

This brings up one of my least favourite practices in bad statistics reporting: generalizing conclusions about attitudes drawn from a specific culture or social group. One example that stayed with me (reported by the ever credulous NYT under the headline Bicycle Helmets Put You at Risk) was that of Ian Walker, a psychiatrist at the University of Bath. Walker, an opponent of helmet laws, put a sensor on his bike and rode with and without a helmet until he had been passed 2,500 times (see the curse of large numbers). To control for potential gender effects he sometimes donned a long wig (to get the full comic effect, check out Walker's picture below).

Walker found vehicles came on average 3.35 inches closer when he was wearing a helmet (for context, the average passing clearance was over four feet).

Putting aside for the moment questions about the methodology of the study and the sweeping conclusions Walker from it, the New York Times article works under the implicit assumption that despite major differences in traffic laws, road conditions, driver etiquette and education, vehicle type and biking culture, findings from that small stretch of English road are equally applicable to American highways.

Subtle issues

This recent study in Pharmacoepidemiolgy and Drug Safety suggests that long term warfarin use leads to more aggressive cancer (as opposed to never users). However, the same authors have shown that warfarin reduces the incidence of prostate cancer in this population.

Is it not just as likely that the warfarin use is more effective versus less agressive cancers as a mildly anti-cancer agent?

I'm curious as to why this interpretation of the data was not pursued

Monday, April 5, 2010

The Long Goodbye -- the networks are taking a long damned time to die

Yesterday I posted a link to Clay Shirky's essay comparing the big media empires to societies like the Roman Empire that collapsed under the weight of their own complexity. I commented that I doubted his conclusions. Here's one of the reasons why.

Back in the early Nineties, when I was still picking up some extra money as a freelance writer, I did a long feature on the surprising vitality of broadcast television in Northwest Arkansas. The vitality was surprising because broadcast television had been under a death watch for more than a decade.

These ominous prognoses had started started sometime in the mid to late Seventies, prompted by the arrival of home video and the rise of cable and satellite-based cable stations. By the beginning of the Eighties, most observers agreed that the outlook was dire. At least one columnist predicted that one of the networks (his guess was NBC) would be gone by the end of the decade. (The columnist obviously didn't have a lot of faith in that Tartikoff kid the network had just hired.)

By 1990 there were exactly twice as many networks as the columnist had predicted and one more than anyone had expected. What happened? Entertainment executives had been trying to launch a fourth network since DuMont went dark in 1956. Why was an industry facing an explosion of new competition suddenly able to support a major new player?

What industry watchers were overlooking was that broadcast stations fell into two distinct groups: VHF and UHF. For the VHF stations, cable really was a threat to a virtual monopoly. it marginally improved their picture quality but it brought in numerous new channels. UHF stations, however, had poor picture quality and low range. As businesses, they were only viable in large, heavily populated areas (and then only marginally). With cable, though, a channel 39 could compete on an equal footing with a channel 2.

This was what made Fox and later the CW viable businesses.

But to me the most remarkable thing about this story is the way the standard narrative has changed so little over the decades. For more than thirty years, there has been a steady stream of stories about reduced schedules, mergers, and deaths of networks and with the exception of the creation of the CW, every single one of them has been incorrect. Over three decades of wrong.

Just to put this in perspective, ABC, the last of the big three, started broadcasting in 1948. That was also about that the other networks started moving past the experimental stage. If we pick that as our starting point, network television has spent more than half of its life under a prognosis of imminent death. During that time the number of networks has gone from three to four to five (and might have been six if not not for some disastrous strategic errors by the heads of WB and UPN) and CBS, the network that reacted the least to the changes in the market (sticking mainly with sitcoms and hour dramas and keeping a very small web presence) is arguably the healthiest of the bunch.

Of course, this could all change in the next year or so. There is no reason to assume the current networks are any more immortal than DuMont was. All the same, we've been told since the late Seventies that this section of the sky is just about to fall. Perhaps it's time to question Mr. Little's credibility.

All others things being equal

Regression analysis is often interpreted as being the effect of changing one variable while holding all other factors constant. Sadly, when working with complex human behaviors, like in nutritional epidemiology, factors rarely stay constant. A change in one parameter can shift other parameters making inference difficult. A very good example of this was reviewed by Travis Saunders of Obesity Panacea.

Now obviously randomized experiments allow an estimate of the average causal effect of an intervention but they are expensive to run and these types of research questions raise grave issues of equiposie and adherence in the design of these experiments!

Decline and Fall of the Media Empires?

Clay Shirky has a post up comparing the current state of the big media companies to the declining days of the Romans, the Lowlands Maya, the inhabitants of Chaco canyon. I think he gets a lot wrong (hell, I think he gets most of it wrong), but he gets it wrong in a really interesting way.

Give it a look:

Bureaucracies temporarily reverse the Second Law of Thermodynamics. In a bureaucracy, it’s easier to make a process more complex than to make it simpler, and easier to create a new burden than kill an old one.

In spring of 2007, the web video comedy In the Motherhood made the move to TV. In the Motherhood started online as a series of short videos, with viewers contributing funny stories from their own lives and voting on their favorites. This tactic generated good ideas at low cost as well as endearing the show to its viewers; the show’s tag line was “By Moms, For Moms, About Moms.”

The move to TV was an affirmation of this technique; when ABC launched the public forum for the new TV version, they told users their input “might just become inspiration for a story by the writers.”

Or it might not. Once the show moved to television, the Writers Guild of America got involved. They were OK with For and About Moms, but By Moms violated Guild rules. The producers tried to negotiate, to no avail, so the idea of audience engagement was canned (as was In the Motherhood itself some months later, after failing to engage viewers as the web version had).

The critical fact about this negotiation wasn’t about the mothers, or their stories, or how those stories might be used. The critical fact was that the negotiation took place in the grid of the television industry, between entities incorporated around a 20th century business logic, and entirely within invented constraints. At no point did the negotiation about audience involvement hinge on the question “Would this be an interesting thing to try?”

Saturday, April 3, 2010

Business models and secret sauces

As mentioned in the last post, today's New York Times has an article of the good people at Talx. One thing that caught my eye was this account of the company's rise:

Talx entered the field brashly, buying the industry’s two largest companies on a single day in 2002. In the next few years, it bought five more. Until then, Talx had never handled an unemployment claim, and skeptics wondered how well it could blend seven companies in an unfamiliar industry.

The Federal Trade Commission argued in a 2008 antitrust complaint that the acquisitions, which cost $230 million, had allowed Talx to “raise prices unilaterally” and “decrease the quality of services.” Talx modified some contracts to settle the case, but admitted no legal violations.

Financially, the gamble paid off: Talx was acquired three years ago by Equifax, the credit-rating giant, for $1.4 billion. But work once done locally became centralized — at a loss, critics say, of responsiveness and expertise.

If this article had been written in 2007*, I suspect the proportions would have been reversed, with two or three paragraphs on complaints and ethical lapses and the rest focused on the company's rise to dominance. Shortly before the collapse, even sober institutions like the New York Times were more inclined to focus on success and less inclined to pay attention to the man behind the curtain.

These success stories were filled with wide-eyed descriptions of innovative business plans and nimble management. There were often vague, portentous references to sophisticated analytic tools and powerful proprietary models and algorithms that could only be appreciated by a handful of Nobel Prize winners.

But when you looked closely at these companies or, better yet, sat down and talked with the people who actually came up with this stuff, you usually found yourself in the position of the trainees at a fast food restaurant in Fast Times at Ridgemont High who learned that the mysterious secret sauces of two rival chains were, respectively, thousand island dressing and ketchup & mayonnaise.

On close inspection, most innovations and business models, including the heavily hyped ones, turn out to be mundane and unimpressive: switched to cheaper suppliers, dropped expensive features or services, introduced some basic statistical tool like logistic regression, marketed a high-fee product to an underserved segment. More than a few are based on creative accounting or a cozy relationship with various regulatory agencies.

Talx' strategy was based on two simple components:

1. Most big companies will be reluctant to work with a small companies so if you buy up most of your competition you can set prices like a monopoly;

2. The penalties for breaking the law are small and the penalties for bending it are negligible relative to the potential benefits.

Neither idea was profound. Neither required a team of Ph.D.s for implementation. They were, however, wildly successful for Talx, just as they have been for any number of companies with much better P.R. departments.



* Very few were. The company doesn't even have a Wikipedia page.

Sleaziness as a business model -- first of two posts on Talx

Contesting Jobless Claims Becomes a Boom Industry

WASHINGTON — With a client list that reads like a roster of Fortune 500 firms, a little-known company with an odd name, the Talx Corporation, has come to dominate a thriving industry: helping employers process — and fight — unemployment claims.

Talx, which emerged from obscurity over the last eight years, says it handles more than 30 percent of the nation’s requests for jobless benefits. Pledging to save employers money in part by contesting claims, Talx helps them decide which applications to resist and how to mount effective appeals.

The work has made Talx a boom business in a bust economy, but critics say the company has undermined a crucial safety net. Officials in a number of states have called Talx a chronic source of error and delay. Advocates for the unemployed say the company seeks to keep jobless workers from collecting benefits.

“Talx often files appeals regardless of merits,” said Jonathan P. Baird, a lawyer at New Hampshire Legal Assistance. “It’s sort of a war of attrition. If you appeal a certain percentage of cases, there are going to be those workers who give up.”

When fewer former workers get aid, a company pays lower unemployment taxes.

Wisconsin and Iowa passed laws to curtail procedural abuses that officials said were common in cases handled by Talx. Connecticut fined Talx (pronounced talks) and demanded an end to baseless appeals. New York, without naming Talx, instructed the Labor Department staff to side with workers in cases that simply pit their word against those of agents for employers.

...

Advocates for the unemployed cite cases like that of Gerald Grenier, 47, who spent four years as a night janitor at a New Hampshire Wal-Mart and was fired for pocketing several dollars in coins from a vending machine. Mr. Grenier, who is mentally disabled, told Wal-Mart he forgot to turn in the change. Talx, representing Wal-Mart, accused him of misconduct and fought his unemployment claim.

After Mr. Grenier waited three months for a hearing, Wal-Mart did not appear. A Talx agent joined by phone, then seemingly hung up as Mr. Grenier testified. The hearing officer redialed and left an unanswered message on the agent’s voice mail. The officer called Mr. Grenier “completely credible” and granted him benefits.

Talx appealed, claiming that the officer had denied the agent’s request to let Wal-Mart testify by phone. (A recording of the hearing contains no such request.) Mr. Grenier won the appeal, but by then he had lost his apartment and moved in with his sister.

“That was a nightmare,” he said.

In the case of Dina Griess, Talx and its client, the subprime lender Countrywide Financial, were involved in what a judge deemed an outright fraud. Ms. Griess worked for Countrywide outside Boston and quit as it collapsed in 2008, saying she was distressed by internal investigations of lending practices. People can receive unemployment benefits if they quit for “good cause,” like unsafe working conditions, but Talx argued that Ms. Griess’s reason did not meet the legal standard.

She won benefits at a hearing that Talx and Countrywide skipped, but Talx successfully appealed, saying the Countrywide witness had missed the hearing because of a family death. Later asked under oath if that was true, the witness said, “No, it’s not.”

A Massachusetts judge reviewing the case, Robert A. Cornetta of Salem District Court, denounced the deceit and gave Ms. Griess benefits. “The court will not be party to a fraud,” he said.

[click here for a history of complaints -- Mark]

Friday, April 2, 2010

Statistics is common sense raised to a science

And Nate Silver is a very good statistician.

From Five Thirty-eight:
That de Rugy has testified before Congress on the basis of her evidence, and never paused to consider why the top five congressional districts on her list overlap with Sacramento, Albany, Austin, Tallahassee and Harrisburg, is mind-boggling. The presence of a state capital is the overwhelmingly dominant factor it predicting the dispensation of stimulus funds. This could have been discerned in literally five minutes if she had bothered to look at the apparent outliers in her dataset and considered whether they had anything in common — a practice that should be among the first things that any researcher does when evaluating any dataset.
Chait has some nice comments on the exchange here.

More on social norming

Rajiv Sethi picks up the ball from Freakonomics and runs with it:
This question lies at the heart of the lifelong work of Elinor Ostrom, co-recipient of the 2009 Nobel Memorial Prize in Economics, whose contributions I discussed in in a couple of earlier posts. Using an eclectic mix of methodological approaches, including case studies, laboratory experiments, and game theoretic models, Ostrom managed to overturn conventional wisdom regarding the "tragedy of the commons." She demonstrated the possibility of self-governance when a well-defined group of users with collective rights to an economically valuable resource were at liberty to develop their own rules, and to ostracize, expel, or otherwise sanction each other for violations.

While Ostrom's focus was on natural resources such as forests, fisheries, and pastures, her basic insights have more general relevance. For instance, institutions of self-governance are critically important in the case of urban communities that lie largely outside the reach of the formal legal system in the United States. There are parts of the country where residents do not have recourse to the courts to adjudicate contractual and other disputes. Given the very high costs of violence as an enforcement mechanism, norms backed by limited community sanctions can therefore play a crucial role.

Thursday, April 1, 2010

In keeping with the season

Here's an April Fool's Day piece on joke-stealing, social norms and intellectual property (via Mark Thoma):
...Comedians have rules of their own about joke-stealing. And they impose their own punishments on thieves... Why do comedians do this? In part, because they live in a world where intellectual property law – in particular, copyright – does not help them much when a rival comedian steals a joke... lawsuits are simply too expensive and uncertain to work as an effective response... Today’s comics are intent on enforcing ownership rights. Yet they do so via social norms – informal but nonetheless powerful rules enforced by comedians on their peers... Comedians maintain a small list of commandments that every comic must follow – or risk being ostracized, boycotted, and sometimes worse. These norms track copyright law at times... More often than not, however, the norms deviate from copyright: for example, copyright protects expression but not ideas, but comedians’ norms protect expression as well as ideas...
I might quibble with the definition of social norms as rules enforced by fear of reprisal, but that's a minor point and, given my previous comments about Freakonomics, I have to give credit where credit is due.

Two thoughts on April Fool's Day blogging

1. I'd better be careful linking to any too-good-to-be-true stories.

2. Wait a minute! This is the blogosphere; how much less reliable can it get?

Skipping the math

From Felix Salmon:
Justin Fox sums up the overwhelming majority of economics papers in one sentence:

The basic form of an academic economics paper is a couple of comprehensible paragraphs at the beginning and a couple of comprehensible paragraphs at the end, with a bunch of really-hard-to-follow math or statistical analysis in the middle.

What he doesn’t (need to) mention is the way that journalists, myself included, read economics papers: we generally have no ability or inclination to try to understand the details of the formulae and regression analyses, so we confine ourselves to reading the stuff in English, and work on the general assumption that the mathematics is reasonably solid.

The problem of course is that we really have no basis for making that general assumption: we make it not because we think it’s particularly justified or justifiable, but because we don’t have any choice. What’s more, because we’re always interested in what’s new, and because we have easy access to the internet and little access to expensive journals, we gravitate to preprints at sites like SSRN, rather than papers which have gone through peer review.