Paul Krugman blogged on that, with initial impetus from Noah Smith.  Here is Noah:

If you and your buddies have a political argument, a vast literature can help you defend your argument even if it’s filled with vague theory, sloppy bad empirics, arguments from authority, and other crap. If someone smart comes along and tries to tell you you’re wrong about something, just demand huffily that she go read the vast literature before she presumes to get involved in the debate. Chances are she’ll just quit the argument and go home, unwilling to pay the effort cost of wading through dozens of crappy papers. And if she persists in the argument without reading the vast literature, you can just denounce her as uninformed and willfully ignorant. Even if she does decide to pay the cost and read the crappy vast literature, you have extra time to make your arguments while she’s so occupied. And you can also bog her down in arguments over the minute details of this or that crappy paper while you continue to advance your overall thesis to the masses.

…My solution to this problem is what I call the Two Paper Rule. If you want me to read the vast literature, cite me two papers that are exemplars and paragons of that literature. Foundational papers, key recent innovations – whatever you like (but no review papers or summaries). Just two. I will read them.

If these two papers are full of mistakes and bad reasoning, I will feel free to skip the rest of the vast literature. Because if that’s the best you can do, I’ve seen enough.

Those are both interesting posts, but my perspective is different, probably more as a matter of temperament than thinking they are objectively wrong.  Here are a few comments:

1. The best two papers on ethics are not very convincing.  Nonetheless people who have worked their way through a good amount of that literature are much better at ethical reasoning than those who have not.

2. The best two papers on global warming are not very convincing.  What is convincing is how many different perspectives and how many different branches of science point toward broadly similar conclusions.  In fact the aggregate effect here is quite overwhelming (don’t debate gw in the comments, not today; I’ll delete).  It is a question of many moats, not all of them being entirely muddy.

3. I see the Smith-Krugman standard as fairly economistic, and fairly MIT-late 20th century at that.  It is one vision of what a good literature looks like, and a fairly narrow one.  It will elevate simple answers in status, whether or not that is deserved.  It discriminates against dialogic knowledge, book-based knowledge, historical knowledge, and knowledge when the answers and methods are not very exact.  There is the risk of ending up too certain about one’s knowledge.

That all said, I do understand that specialized top researchers, including Nobel Laureates, often may do better holding relatively narrow methodological visions.  Look at all the Nobel Prizes that have been awarded to Chicago.  It might be entirely correct to insist that Becker’s treatise on the family pay more attention to anthropology, but that doesn’t mean he should have followed that advicee.

4. The standard seems to discourage reading, and I would not want to teach it to my students.  I teach something more like “always read more, unless you are writing or doing relevant quantitative work.  And one reason you write is to improve the quality of your reading.  Read more and write more, all the time.”  I still think that is better advice for most (not all) people.

5. Isn’t there a lot to be said for deferring to the opinions of those who have read through the “muddy moat”?  By no means are they all partisans, and the non-partisan ones care most of all about the truth.  After all, they did all that reading!  Defer, rather than trust so much in your ability to pick you the right two papers, or have someone pick them out for you.  I have a much more positive view of survey articles than does Noah, while understanding they do often leave you fairly agnostic on major issues.

6. If the truth of the matter is in fact muddy, you may need to dip into the muddy moat to learn that.

7. The difference between total value and marginal value may be relevant.  You might conclude a field literature has low total value, but the marginal value of learning more about that area still could be quite high.  That is in part because muddy fields and results don’t spread so readily, and so dipping into the muck can yield some revelations.  That is another reason why I would not offer the “two paper standard” as practical advice.

8. If anything, I would put the reading pressure on the other side, namely more rather than less.  Rather than encouraging readers to dismiss or downgrade fields, I would urge them to consult different disciplines altogether, including political science, sociology, and anthropology, others too.  This is much easier to do if you take a more positive attitude toward survey articles.

9. This is quite a subjective impression, but I worry that the dogmatic will use the two paper standard to dismiss or downgrade particular lines of investigation.

10. I don’t know if Noah and Paul were referring to my colleague Garett Jones, who frequently tweets “…if only there were a vast empirical literature” when he sees claims that he regards as empirically false.  Now, I am not the Garett Jones oracle, but I always took his use of the word “vast” to be slightly sarcastic.  Usually these are cases where even a fairly cursory knowledge of the literature in question would indicate something is wrong with the claim at hand.  In my view, Garett is not demanding “vastness” of effort, rather he is criticizing those who don’t grasp what the effort space looks like in the first place.

I calibrate the [spatial] model to the U.S. economy and find that the rise in regulation accounts for 23% of the increase in wage dispersion and 85% of the increase in house price dispersion across metro areas from 1980 to 2007. I find that if regulation had not increased, more workers would live in productive areas and output would be 2% higher. I also show that policy interventions that weaken incentives of local governments to restrict supply could reduce wage and house price dispersion, and boost productivity.

That is from Andrii Parkhomenko (pdf), a recent job candidate, there are 62 pp. at the link, and for the pointer I thank Tyler Ransom.

Here is a related column by Noah Smith.

Thursday assorted links

by on May 18, 2017 at 11:52 am in Uncategorized | Permalink

BMJ: In a national sample of elderly Medicare beneficiaries admitted to hospital with medical conditions, we found that patients treated by older physicians had higher 30 day mortality than those cared for by younger physicians, despite similar patient characteristics. These associations were found among physicians with low and medium volumes of patients but not among those with high volumes.

…Our findings suggest that within the same hospital, patients treated by physicians aged <40 have 0.85 times the odds of dying (1.00/1.17) or an 11% lower probability of dying (10.8/12.1), compared with patients cared for by physicians aged ≥60 (table 2⇑). This difference in mortality is comparable with the impact of statins for the primary prevention of cardiovascular mortality on all cause mortality (odds ratio of 0.86)39 or the impact of β blockers on mortality among patients with myocardial infarction (incidence rate ratio of 0.86),40 indicating that our observed difference in mortality is not only statistically significant but arguably clinically significant. In addition, if our results are causal, an adjusted risk difference of 1.3 percentage points suggests that for every 77 patients treated by doctors aged ≥60, one fewer patient would die within 30 days of admission if those patients were cared for by physicians aged <40.

The paper has data on over 700,000 Medicare admissions and over 18 thousand hospitalist physicians. Physicians are assigned to patients more or less randomly depending on admission time so there are no significant differences between patients assigned to younger and older physicians. Older physicians are more likely to be male and, of course, to be trained during a different time period so the paper can’t fully distinguish age effects from cohort effects. The authors do find that older physicians who work a lot perform well–perhaps these physicians update their training or perhaps they are a self-selected vigorous sample. Continuing medical education and assessment requirements are probably very valuable.

Hat tip: Eric Topol.

Sentences to ponder

by on May 18, 2017 at 3:17 am in Current Affairs | Permalink

National Security Council officials have strategically included Trump’s name in “as many paragraphs as we can because he keeps reading if he’s mentioned,” according to one source, who relayed conversations he had with NSC officials.

This second bit I am fully in accord with:

Trump likes to look at a map of the country involved when he learns about a topic.

Here is the Reuters article, via Brad Jaffy and Brendan Nyhan.

That title made me think of the woodchuck…anyway, here is the abstract:

Fact-checking has gained prominence as a reformist movement to revitalize truth-seeking ideals in journalism. While fact-checkers are often assumed to code facts accurately, no studies have formally assessed fact-checkers’ performance. I evaluate the performance of two major online fact-checkers, Politfact at Tampa Bay Times and Fact Checker at Washington Post, comparing their interrater reliability using a method that is regularly utilized across the social sciences. I show that fact-checkers rarely fact-check the same statement, and when they do, there is little agreement in their ratings. Approximately, 1 in 10 statements is fact-checked by both fact-checking outlets, and among claims that both outlets check, their factual ratings have a Cohen’s κ of 0.52, an agreement rate much lower than what is acceptable for social scientific coding. The results suggest that difficulties in fact-checking elites’ statements may limit the ability of journalistic fact-checking to hold politicians accountable.

That paper (pdf) is by Chloe Lim, political science at Stanford.  For the pointer I thank Andrew Hall, some interesting political science papers on his home page.  Here is his very interesting book manuscript on how the devaluing of political offices drives polarization, worthy of a top publisher…

Wednesday assorted links

by on May 17, 2017 at 1:48 pm in Uncategorized | Permalink

Vix is up 16% today, a sign that a Trump presidency is now seen as having a much more uncertain future.  I agree with Charles Cooke that the 25th amendment is not really an option, nonetheless investigations will be proceeding, with the FBI and many Republicans not really on Trump’s side.  It is not obvious that Trump will handle himself well during that process.  The chances for tax and health care reform are dwindling.  Many Republican leaders are pondering the logic of Timur Kuran, namely when they should flip out of their preference falsification and state their real views.

I think also that Trump’s instructions to Comey to halt the Flynn prosecution are significant.  I view much of the press coverage as overstated or sometimes even hysterical, including for the Russia leaks, but the Comey business fits into the category of “impeachable offense.”  A normal president would not be impeached for it, but Trump is not a normal president.  The instructions to Comey would not be the actual reason he would be impeached, but they create a path along which an impeachment inquiry could proceed, nudged along by other “non-impeachable but unpopular and objectionable actions” Trump might take in the meantime, and what information might be revealed in the meantime.  There are many shoes yet to drop.  So my estimate of the chances of a Trump impeachment or resignation have gone up from about 5% to about 25%, in less than a two-day span.

Addendum: Do consider the remarks of Philip Wallach.

Here’s an excellent story about Chris Carr who played in the NFL for 10 years and is now about to graduate from law school. That’s unusual but not so unusual, the late U.S. Supreme Court Justice Byron R. White also played in the NFL. What makes this story special is that Carr will specialize in immigration law. Why?

Carr grew interested in immigration law a few years ago, after reading Thomas Sowell’s “Ethnic America.” (“A really cool book,” he said.) That made him reflect on the country and “just how unique the American experiment was.” He read blog posts by Bryan Caplan, an economist at George Mason University, and the writing of Michael Huemer, a philosophy professor at the University of Colorado.

Hat tip: Fabio Rojas.

Hemingway: 80

Twain: 81

Melville: 126

Austen: 128

J.K. Rowling: 140

E L James: 155

That is from the new and interesting Nabokov’s Favorite Word is Mauve, by Ben Blatt.  The Hemingway book with the highest usage rate for -ly adverbs, True at First Light, was released only after his death and is considered one of his worst works.  The same pattern is true for Faulkner and Steinbeck, namely that the most highly praised works have relatively low rates of -ly adverb usage.  Among other notable authors surveyed, D.H. Lawrence seems to be the most obvious exception to this regularity.

In the novel The Hobbit, J.R.R. Tolkien used the word “she” only once.  In The Prime of Miss Jean Brodie, however, she relative to he is used 79% of the time, the highest ratio of the classics surveyed.  Female authors are very strongly represented on that side of the curve, let me tell you.  And male authors do the “he” far more, in relative terms, than female authors do the “she.”

You also will learn from this book that David Brooks starts more sentences with “The” than any other word, whereas for Paul Krugman that place of honor goes to “But.”  And, for better or worse, Krugman uses far less anaphora.

D.H. Lawrence leads for the number of animal similes.

That is the query motivating my latest Bloomberg column, here is one bit:

Second, the higher health-care spending for women is partly because of services related to childbearing. Society may have an obligation to help out babies (and mothers), plus they will someday finance our retirement, so let’s make childbearing easy. That said, governments have numerous means of subsidizing childbearing — direct payments, tax credits, free clinical services and public education — and it’s not obvious that regulating insurance pricing is this best way to achieve this end.

And:

Uniform pricing also gives insurance companies less incentive to attract female policyholders. To be sure, as a matter of law the companies cannot turn women away. But if writing policies for women is less profitable, or perhaps unprofitable altogether, the insurance companies will allow or encourage their provider networks to evolve in a way that is more attractive to men than to women. Services for women, including for childbearing, might end up underprovided or stagnate in quality. That also would be a kind of differential treatment, with potentially dire consequences.

There is much more at the link, controversial throughout.  You’ll find plenty of overwrought reactions on Twitter, simply because I am saying there is a trade-off, and we do not yet know what is the right margin to seek.

Tuesday assorted links

by on May 16, 2017 at 2:26 pm in Uncategorized | Permalink

A British prison has become the world’s first to use a new system designed to stop drones flying over perimeter walls to drop contraband into jails.

The device creates a 2,000ft (600m) shield around and above a prison that will detect and deflect the remote-controlled devices.

It uses a series of “disruptors”, which are sensors to jam the drone’s computer, and block its frequency and control protocols. The operator’s screen will go black and the drone will be bounced back to where it came from.

Drones have become a major security problem in Britain’s prisons and are increasingly used to smuggle in drugs, weapons, phones and other valuables.

The new system, called Sky Fence, is being introduced at Les Nicolles prison on Guernsey, where around 20 “disruptors” will be installed on the perimeter and inside.

The Channel Island jail was initially going to install a drone detection system, but went a step further to put in the technology that stops drones in-flight.

Here is the article, via Tyro.  By the way, the newly available BBC TV show, Planet Earth II, is an amazing illustration of the use of drones to track and film nature (that includes us!).

In The Failure of Solanezumab –How the FDA Saved Taxpayers Billions, an article in the NEJM, Sacks, Avorn and Kesselheim (SAK) defend the FDA by arguing that its high standards prevented the Alzheimer’s drug, Solanezumab, from being approved and thus saved the taxpayers billions in Medicare payments.

It is, of course, not the job of the FDA to approve or fail to approve a drug based on its effect on taxpayers. The FDA has historically stood independent of this kind of politics and that has been all to the good. But the SAK article is a reminder that under socialized medicine every FDA decision moves money from one patient group to another and between patients and taxpayers thus FDA decisions become a tempting leverage point to control allocation through collective choice.

I do not favor collective, which is to say politicized, choice and find much else objectionable in the SAK article–it attempts, for example, to evaluate a rule by examining a single decision when a system-wide, long-run analysis is called for. Rather than go into detail, however, let’s instead point to a much better article by Vradenburg, Fillit, Morgan, Sabbagh, Aisen, and Mohs, a group of leading physicians and scientists who treat Alzheimer patients and research the disease.

Rather than support or criticize an isolated FDA decision, Vradenburg et al. call for a change to the rule/norm currently used to evaluate Alzheimer’s drugs:

The analysis…recommends that the FDA approve new medicines that demonstrate a proven benefit on at least one therapeutic endpoint – either cognition or function. The current FDA standards require a new drug to show benefits on both proven endpoints, an unnecessarily challenging hurdle the authors say may be inhibiting investment in new Alzheimer’s treatments.

The authors make three excellent points about such a change. First:

…the success rate of drugs tested for Alzheimer’s disease has been extraordinarily low when compared with drugs in other therapeutic areas. Of the 244 compounds that were tested in 413 clinical trials between 2002 and 2012, only one resulted in approval of a new chemical entity, in 2003. No others have been approved since that time; the failure rate in clinical trials conducted over the last decade exceeds 99.6%. This staggeringly high failure rate has adversely impacted investment in Alzheimer’s disease research at precisely the time when new advances are most needed.

The failure rate reflects how difficult the problem is but also policy. Either way, when firms look at the billions of dollars in research and development that haven’t led to a single approved drug they are naturally wary about spending more. Breakthroughs don’t happen randomly, however, they happen after lots of trial and error. To stimulate such trial and error firms need revenues and thus to stimulate more swings at the bat it may be justified to approve drugs with relatively small benefits.

Second, as I have noted previously, the FDA needs to be careful not to commit an error of composition. Three ineffective drugs need not add up to an ineffective treatment.

Many drugs in development for Alzheimer’s disease have complementary mechanisms of action. Even if each of these might, individually, deliver a modest clinical benefit, when used in combination or adjunctively, the benefit could become more substantial. If the FDA were to reject, individually, several safe and well-tolerated therapies with complementary mechanisms of action that each demonstrate a modest clinical benefit, it would unwittingly deprive patients of potentially substantial advances in the quality of treatment over the long run with a combination of therapies.

Third, it is ultimately the patient that matters, especially with regard to Alzheimer’s where so much depends on the patient’s internal experiences, and thus we ought to be careful before rejecting their perspective:

The ultimate perspective on clinical meaningfulness, of course, comes from the patient….Efforts to identify what matters, what matters most and how much change matters to patients should become a priority for the field, focused on all stages of the disease. The requirements of the recently-passed 21st Century Cures Act are instructive in this regard.

Hat tip: Abhay Moghekar

Crunching data from disparate states, Mr Chinoy says state borrowing rose by a whopping 32% in the year to March 2017, after a 25% rise in the previous year…

Bihar, the country’s poorest, with a budget deficit of nearly 6% of its state GDP last year and a hole in its finances after it banned alcohol sales…

That is from The Economist.