Science vs. expertise

In the a recent episode of EconTalk, where Russ Roberts interviews the great Bill James, the father of sabermetrics, they discuss the distinction between science and expertise as James delineates it in a great essay called “The Note”. It begins:

Handwriting expertise, more clearly than any other field that I know of, illustrates the difference between science and expertise. Science uses general rules and principles, understood by millions of people, to work toward increasing our shared understanding of the world in which we live, a key word being “shared”. What is learned by the scientist is in no sense the property of the scientist, if it is truly science and not commerce. If in our field we were to discover, for example, that tall hitters do best against tall pitchers and short hitters do best against short pitchers, the value in this would not be to me, but to a baseball team by way of its manager. The manager would be as much the owner of this knowledge as the analyst who discovered it. Crucial to that fact—to the shared nature of its ownership—is how it is known. If I were to say that I know that tall hitters hit best against tall pitchers and you should believe me because I am an expert in this field, no one would believe me and no one should believe me. Others would believe me only if (1) I were to present evidence demonstrating that it is true, and then only if (2) others were able to study the same subject and reach the same conclusion. Studying the same subject and reaching the same conclusion or a different conclusion does not require expertise limited to a small cadre of persons; it merely requires that we apply the rules of scientific enquiry which are universally owned and widely understood. Shared principles yield shared knowledge.

Expertise, on the other hand, is the property of the expert. The only way that WE know that Document 1 and Document 2 were or were not written by the same person is that an expert says so. The only way to get a second opinion is to ask another expert. If the other expert cannot get access to the documents, you can’t get another opinion.

James is not the first to identify this division, but he puts it very well.

Advertisements

How should we deal with newcomers?

The Atlantic had an interesting piece recently about how gentrifying neighborhoods attract more police resources, due to calls from the newer residents, and how this disproportionately affects racial/ethnic minorities. I think it reads more productively as a criticism of policing rather than of gentrification per se, though the author doesn’t seem to, but more interestingly it evokes a widespread social issue. To begin my explanation, let me highlight these passages:

When low-income neighborhoods see an influx of higher-income residents, social dynamics and expectations change. One of those expectations has to do with the perception of safety and public order, and the role of the state in providing it. The theory goes that as demographics shift, activity that was previously considered normal becomes suspicious, and newcomers—many of whom are white—are more inclined to get law enforcement involved. Loitering, people hanging out in the street, and noise violations often get reported, especially in racially diverse neighborhoods.

In 2013, the city of San Francisco launched Open311, a mobile app that allows residents to easily report public disorder like loitering, dirty sidewalks, or vandalism by snapping a photo and sending their location. The app can feel altruistic; residents, for example, are able to report the whereabouts of homeless people who seem to be in need of assistance. But some worry that the dispatches can result in unnecessary citations or harassment. And while broken-windows policing remains controversial, a 2015 poll suggested that it’s still largely accepted by the general public, so when people see something, they’re likely to say something. After the app launched, 311 calls increased throughout the city, and one study showed that gentrifying neighborhoods saw a disproportionate spike.

Butler, who recently wrote the book Chokehold: Policing Black Men, believes that this is a result of newcomers refusing to assimilate to longstanding neighborhood norms. [emphasis added] “Culturally, I think the way that a lot of African American and Latino people experience gentrification is as a form of colonization,” he said. “The gentrifiers are not wanting to share—they’re wanting to take over.” One of the tools they can use to take over public spaces, he argues, is law enforcement.

This is interesting, and it reminds me of the neighborhood where I lived in Alexandria, Virginia, only in reverse. There was a lot of very recent immigration, largely Salvadoran, as there has been all over the DC metro area. Granted, I was a newcomer to the neighborhood, but of similar SES and background to the longer-term residents. Being from the US-Mexico border, this immigration doesn’t bother me, but I noticed it bothered a lot of other people because the Salvadorans, to use the language of the article, ‘refused to assimilate to longstanding neighborhood norms’ regarding noise, cleanliness of public spaces, junk cars parked on the street, etc. Indeed, this is a common occurrence all across the country, and for the last decade or more has been a simmering issue. This cultural friction is one of the reasons for the surprising result of the 2016 presidential election. The Atlantic’s political wing tends to regard people who are upset by immigration as, well, deplorable. No doubt some are motivated by racism, which is indeed deplorable, but cultural frictions exist that aren’t motivated by racism. (For the record, in my ideal world immigration rules will be greatly liberalized relative to current rules.)

So which is it? Are newcomers wrong for “refusing” to assimilate to longstanding area norms, or are older residents wrong for not accepting the changes? There’s no simple answer, and I don’t mean to suggest the writer of the article or anybody else thinks there is. I don’t think we’ve had much public dialogue on the issue, and I’d like to, especially as so much governance happens outside of formal systems, and better public understanding of good rules and norms would help us resolve more social issues with nuance and wisdom before they became political issues resolved with neither.

JFK Assassination – Oak Cliff locations map

I’ve seen maps of the area between Lee Oswald’s boarding house and the site of Officer J.D. Tippit’s murder trying to figure out how the timeline works, but there is more in Oak Cliff that’s interesting to the puzzle. This map is not by any means exhaustive, but here are six locations:

The first odd thing to notice is the two Oswald residences, one where Marina Oswald lived with the children, and which Lee Oswald visited on weekends, and the boarding house where Lee lived during the week. It’s said this was due to the Oswalds’ marital problems, but it would have been very expensive for a minimum-wage worker whose wife did not work; Oswald’s $1.25/hr minimum wage job at the Texas School Book Depository in November 1963 would be worth $10.01/hr in November 2017. Where Oswald came up with money is a question that comes up frequently when looking at his life.

The second odd thing to notice is where Jack Ruby’s apartment is in relation the Tippit shooting site and Oswald’s residence. I’m not the first to notice this, of course. It might be pure coincidence, as it’s entirely plausible Oswald was Tippit’s killer, was in panic mode, and was fleeing with little regard to direction, or to a bus stop not so close to his residences. The housekeeper said she saw him waiting for the bus at Zang and Beckley, at the end of the block, which he obviously did not take. There has also been a lot of confusion about which direction Tippit’s killer was walking when Tippit stopped him. I don’t know enough to have a firm opinion on whether the killer was walking east or west on 10th Street, or if the killer and Oswald were the same person. But it’s strange either way. Whoever the killer was, he was oddly close to Jack Ruby’s apartment when the murder occurred.

I’ve also put the site of Officer Tippit’s mysterious phone call on the map. The location so near to the Texas Theatre, where Oswald was arrested, is not really noteworthy; that’s the main drag in the area, so plenty of places people would stop in on were in those few blocks. But the phone call is strange. At this point the president has already been assassinated, and every law enforcement officer in the Metroplex is on high alert. Rather than communicating by his police radio, Tippit tries (and apparently fails) to get in touch with somebody by phone only a few minutes before he is murdered. We’ll probably never know what was so urgent and unrelated to his official task at the moment that he had to make this phone call about.

Random coincidences happen. Nothing I’ve written here may really mean anything. But I like maps, and these locations are more interesting on a map than as textual abstractions.

How to blind yourself and others to good research

Via Twitter, this gem about ideologically-driven research:

I want to highlight two parts. First, the opening act:

The criteria of valid knowledge associated with quantitative research are ones that turn women, whey they are the focus of research, into objects. This means that women are again subjected to exploitation, in that knowledge and experience are extracted from them with nothing in return, even when the research is conducted by women (Mies 1993).

Because of the many things people have in common with other people, we can find out some kinds of information about people generally by doing research involving specific people. And because people are part of the natural world, many things about them can be described with numbers. Those numbers cannot capture everything interesting, important, or even relevant about human beings, but like carbon dioxide parts per million, sometimes putting numbers to the various features of the natural world can tell us interesting and useful things. If you want to call this treating women (and people generally) as objects, fine, but you thereby imply a complete separation between humans and the rest of the natural world. Humans are special, but it does not make sense as a blanket rule in social and physical science.

If you think of quantitative data gathering as particular women being treated as objects and thereby out of bounds, what is your alternative proposal for medical research? There are certain kinds of research that are unethical, of course, but beyond those cases the overwhelming majority of medical research works this way, and women (as well as men) have certainly benefited from medical research. Does this passage imply compensated participation in research is acceptable, i.e. not exploitative? Or at least informed consent? These are standard practices. What if research participants recognize the potential benefit to people other than themselves and participate for altruistic motives, even if they don’t know the specific purpose of a particular study and aren’t compensated? Are we saying only compensated participation is acceptable?

But this is all a warm-up for the really astounding main event:

The emphasis on controlling variables further exacerbates this last problem, and indeed the very idea of control is viewed as a masculine approach.

This is really egregious. Is the objection to trying to determine causal and probabilistic relationships among variables by the technique of holding some other factors constant? That seems absurd to write out, something nobody could posit even in jest. The only explanation that makes sense is that the mysterious, unnamed academics who with the passive voice view “the very idea of control” “as a masculine approach” simply have no understanding of the processes under discussion, and think because some forms of control are objectionable the word itself must entail them in all other contexts. (Perhaps some self-controldiscipline in studying statistical methods is in order.) If this is the case, we are dealing with an entire academic subfield premised on a complete misunderstanding of what words and numbers are and do. This is a group within the academy that literally and completely has no idea what the others are talking about, but commands pontificating about it anyway. Not only is using any resources for this purpose wasteful, it actually takes us farther away from reality.

Making sure women get a fair shake is a noble (and uncontroversial) goal. Tying the improvement of the female condition to dubious, even mystical research practices does not help the goal. Fortunately, almost all scholars, including the scholarly women I know, don’t follow this approach.

As directed by the JFK Records Act of 1992, many more files relevant to the investigation of Kennedy’s assassination were released today, following a batch in July, although the current president reversed course from his earlier statement and held some back. I don’t expect too many shocking revelations to come out of this latest batch, but you never know.

For those who are interested, .pdf versions are available here.

Henry George on trade:

Free trade consists simply in letting people buy and sell as they want to buy and sell. It is protection that requires force, for it consists in preventing people from doing what they want to do. Protective tariffs are as much applications of force as are blockading squadrons, and their object is the same—to prevent trade. The difference between the two is that blockading squadrons are a means whereby nations seek to prevent their enemies from trading; protective tariffs are a means whereby nations attempt to prevent their own people from trading. What protection teaches us, is to do to ourselves in time of peace what enemies seek to do to us in time of war.

Sadly, this insight keeps being relevant.

A short deductive inference

P1. Bumper stickers for political campaigns that have ended are evidence for the public choice idea that political participation is expressive rather than instrumental. What could be more about you and less about concrete objectives than a call to action for an event that’s already happened?

P2. Bernie Sanders bumper stickers survive all out of proportion to how much support he had.


C1. Supporting Sanders is expressive at an above-average level.
OR
C2. Sanders supporters are lazy and haven’t gotten around to taking off the stickers. (heh)

Underestimating Obama, overestimating the Democratic candidate

In a recent interview, the involuntarily-retired Hillary Clinton widened the circle of culprits in her defeat to include the Democratic National Committee, and DNC data people are not happy. But there’s another issue that’s also interesting: reports of the inevitability of the Democratic future have been greatly overestimated. It appears Barack Obama, not the Democratic Party’s presidential candidate per se, was the irresistible force.

One problem for Clinton as she was beginning her 2016 candidacy was a time warp about the way things were or should have been. Her perspective about the role of the DNC and the Democratic Party apparatus was shaped by experiences from Bill Clinton’s campaigns in 1992 and 1996 and his (and her) approach to the party.

In those days, the DNC was a more robust and battle-ready institution, and Bill Clinton as candidate and president paid attention to it. President Barack Obama did not. Under Obama, the DNC was neglected and left to atrophy.

“The DNC has not played any dynamic role except just on a rare occasion since Obama was president,” said a former party official.

Obama’s two campaigns were built largely separate from the DNC. Data produced by Obama for America and its various other names was proprietary and not readily shared with the party.

I don’t know how common this knowledge was, but I didn’t know it before. I suppose back in 2007 Obama and his campaign team, still outsiders to the Democratic establishment, saw the Democratic National Committee as a Clinton fiefdom and circumvented it as much as possible. The DNC data guy who returned fire said the real problem was the HRC campaign’s mishandling of the data, and I’m sure there’s some truth to that. All accounts I’ve seen show they assumed the victory and refused to update their strategy when local units began to notice cracks. But whatever the reason, the combination of the Democratic Party and the Clinton campaign was simply not optimized for victory. The Obama campaign was optimized for victory. Confounding these entities made the party seem much stronger than it was.

Someday we might find out why Obama didn’t drop the barrier between his 2012 campaign and the DNC. Perhaps he still didn’t trust it? It would be ironic if the Clintons’ legendary skills at politicking and maneuvering got them this close but prevented them from going that last step further.

The JFK Assassination, pt. 15: Oswald’s personal effects

It should be clear how I use this blog as a mental scratch pad rather than for polished writing. And it should also be clear how I approach the JFK assassination as a great mystery and puzzle rather than a case that urgently needs solving. Reviewing the most recent post on this topic, however, I really thought it was poor, so this is the revised version.


Items on Lee Oswald when he was arrested, copied exactly from the Dallas PD Property Clerk’s invoice dated 12/30/1963 (and numbered by me for convenience):

1. Eighty-seven cents in money (1 half dollar, 3 dimes, 1 nickel, and 2 pennies)
2. Thirteen dollars in money (1 five-dollar bill, 8 one dollar bills)
3. Dallas County bus transportation coupon or ticket for transfer dated Nov. 22, 1963
4. Marine Corps, silver color
5. Chrome color ID bracelet with expansion band with the inscription ‘Lee’
6. Brass key marked “Postoffice Department Do not Dup.” #1126
7. paycheck voucher from American Bakeries Company dated 8/22/60
8. top of a small cardboard box with “Cox’s Fort Worth” printed on top

Not mentioned are a .38 revolver or a wallet containing two IDs with different names, but these would have been taken separately, as evidence, rather than being stored with personal effects.

We know that after leaving the Texas Schoolbook Depository, Oswald’s next location was his boarding house, where he spent just a few minutes before leaving for the Texas Theatre, where he would be arrested. He’s known to have picked up a light jacket at the house, despite the warm weather. What else did he do there? He must have gathered up some of these items as well, unless he had taken them all to work, which is unlikely.

Items 1, 2, and 6 are pretty ordinary things to carry, so I’ll skip them. Item 3 is not necessarily very exciting. Roger Craig claimed that he saw Oswald leave Dealey Plaza in a station wagon, but Craig could have been wrong and/or Oswald could have been dropped off to catch a bus afterward, and in either case the bus ticket is perfectly natural.

Item 4 is interesting. The invoice entry is not complete, but elsewhere this is described more fully as “Silver color Marine Corps emblem ring”. It’s hard to know if this was something he wore often or something he specifically picked up at the boarding house. I can’t find a photograph of him wearing it, but most of the existing photos are from times he wouldn’t have had it on.

Item 5: uninteresting, except that while “Lee” was his real first name, one of the IDs he was carrying had a different name. Item 6 is perfectly ordinary.

Items 7 and 8 are where it gets really interesting, and where the previous version was lacking.

Item 7 is a paycheck voucher issued three years before, and not to Oswald. At that time, Oswald was working in an electronics factory in the Soviet Union, not for the American Bakeries Company. A strange thing for anybody to carry, and extra strange for Oswald.

I initially thought the list would be fairly reliable, i.e. not falsified, because the officials never made a big fuss about it. If it were manipulated, it would have been for some purpose and it seems like they would have then used it as evidence. However, the history of the pay stub is more than a little curious. It was issued to one James A. Jackson of 214 W. Neely St. The date in the invoice is given as 8/22, although elsewhere it’s listed as “dated 8/22/1960 or 8/27/1960”—why the confusion? (The 22nd was a Monday, and the 27th was a Saturday, by the way.)

The W. Neely St. address is significant because it’s the location of the backyard photos, which supposedly show Oswald holding the Mannlicher-Carcano rifle used to shoot Kennedy. However, Oswald said the backyard photos were faked, denied owning the rifle, and denied ever living at that address, and no clear picture comes out of interviews with other residents at that apartment complex. James A. Jackson’s pay stub is significant because it’s supposed to connect Oswald to the Neely apartment, and thus to the photos and the rifle.

You can see the problem: why would he own this item at all, and why would he pick this item up during his brief stop? It doesn’t make sense. And there’s another problem. It’s not clear when the pay stub was discovered. It’s listed on the invoice, but other sources report it was found later, during a search of his home. This has an explanation: there were two pay stubs, both in Jackson’s name, discovered at different times.

This looks fishier and fishier.

Item 8, the box top, is also very strange. This seems to be some kind of spy novel stuff where two people can identify each other by putting together their torn box tops; this can’t be faked, so you know the other person is who he is supposed to be. I’m agnostic on this. I already thought Oswald was involved in intelligence work, so it would make sense, but it seems a little too much like fiction.

Taken all together, this collection of items is just too strange to be what a lone gunman would carry. I admit this is a probabilistic rather than definitive conclusion.

The social problem of low conscientiousness

Let me preface this with an example I’ve given in classes. When we talk about social problems, we have to be careful what we mean. We don’t simply mean something a lot of people don’t like. In terms of public health, for example, one person’s case of heart disease is not a public health issue in the way one person’s case of ebola is a public health issue. Heart disease becomes a public health issue when it’s a widespread issue with systemic causes that can be fruitfully addressed in a systemic way.

This weekend, as I sat waiting to be called to renew my vehicle registration, I overheard the following exchange (this is inexact, but roughly corresponds):

Receptionist: Can I have your name, sir?

Guy: [J]

Receptionist: And what can we do for you today, J?

Guy: I need to update my registration.

Receptionist: Do you have all the forms? ID, title, emissions inspection, two proofs of address?

Guy: Does a W-2 count as proof of address?

Receptionist: Unfortunately it doesn’t. It can count for proof of your Social Security number. We need two other proofs of address.

Guy: (Loudly) Man, I ain’t got two proofs of shit! Fuck! (storms out)

Now, I agree with J, the rules of motor vehicle registration can be a hassle to comply with, even before you come in and wait. And with only a little bit of imagination we could probably come up with a better system. Most of us have probably thought of tweaks while we waited. But come on, J, them’s the breaks. Yelling at the receptionist won’t change anything. A little self-control goes a long way in life. By itself the interaction wasn’t really a big deal, but people who yell at receptionists tend to express their low conscientiousness in other ways that can be. (I’m sure the receptionist wasn’t happy, but she didn’t seem overly fazed.)

In what follows I may be unfair to J, but from my limited sample of his behavior I am going to generalize him into an archetype we’re all familiar with, and hey, I doubt he will read this anyway. (Plus, is it more likely that the moment you encounter somebody is when they’re acting consistently with their long-term personality or when they’re acting inconsistently with it?)

I think about people like J a lot, people who get low scores in the five factor model’s conscientiousness category. The general population is full of them, and they’re greatly overrepresented in sub-populations of interest to policy makers and people who study rules: the unemployed, welfare recipients, jail and prison inmates, etc. All evidence suggests that major personality traits are partly heritable, i.e. there’s little we can do about them, but there is plenty of room for environment to influence them as well, and of course we know people respond to incentives.

We have a lot of money in formal social institutions that deal with low-conscientiousness people. I recently saw a news story about two men arrested for auto theft while on a roundabout drive to the local courthouse, where one of the men was to pick up an ankle monitor for a previous auto theft arrest. It’s hard to imagine what could be done to deter this kind of bad apple but more of the criminal justice system. People like J, however, aren’t beyond the pale.

I give the Left credit for thinking about systemic influences on behavior, but I often think its prescriptions are lacking. For example, people like J may benefit in the short term from having a social safety net to fall back on when their lack of self-control makes maintaining gainful employment difficult, but in the long term this weakens the incentive to develop self-control. I know the counterargument: the social safety net helps poor children and other deserving poor, and it’s better to cushion too many rather than too few. Creating institutions that penalize low-conscientiousness behavior without unduly impacting others, and that aren’t retroactive criminal justice institutions, is a very difficult task. I readily acknowledge this, but the goal should still be one of the guiding lights of policy. At the very least, failure to acknowledge the tradeoff should be viewed as highly suspicious.