Wednesday, April 14, 2010

Market Efficiency and Dating

Katja Grace has a great post at Meteuphoric called "How Does Information Affect Hook-ups", where she looks at the impact of online dating and the free up-front exchange of information about partners that it engenders, as well as the impacts thereof.

In a comment I argued that this was one example of how increased market efficiency has changed the romantic behavior of humans. Sometimes increased market efficiency is the result of information technology as in Katja's example, but it can also be the result of urbanization, as I argued previously.

In response to Katja's post, I argued that if we want to see the effect of efficient markets on dating, we didn't need to wait for online dating - single people in dense urban centers were and are exposed to many, many other potential partners over the course of a day, and even though they don't get the same up-front information that online dating allows, the effect is that people have a much better idea of the overall market (even if not necessarily an immediate read on the quality of the individual across the dinner table from them). In a little town in a rural area, you can't take individual dates too lightly, because how many single people in your league in your age group are there going to be out there? Contrast with Manhattan or Tokyo, where on every date you can tell yourself well yes, s/he is attractive and pleasant and interesting, but that guy/gal you just met at a professional mixer last night seems even better, and if neither of them work out there are still (literally!) a million other potentials out there. Given that dating is a usually required step before marriage, one would expect this is contributing to later marriage ages in dense urban centers, although there are confounding factors: a) urban centers in the developed world have large numbers of highly educated high-earning women; women often seek men with greater education or earning than themselves; therefore, they may hold out longer for this reason, and b) that education takes time, and people haven't even started the earning phase of their careers when they're 24, much less are they ready to get married. Comparing urban to rural marriage age while controlling for education and income would tease this out.

A further confounding factor is the partner-finding-process itself, which varies considerably across countries - hence the mention of Tokyo (arranged marriage is still more common in Japan than Westerners realize). Consequently it would also be interesting to compare arranged marriage rates in Tokyo to outlying areas of the country. A personal observation of Asian friends who've accepted an arranged partner is that they seem to view one of the biggest benefits of accepting an arranged partner as having been relieved of the task of the search. This theory's prediction is that in a large and efficient dating market, rationally optimizing singles should be less inclined to listen to bothersome parents who want to arrange a marriage with the down-the-street neighbors' nice but homely son or daughter from back home. (Also known as "Mom, are you actually serious?")

Notably, Katja's post is about hooking up, not dating/marriage. She states an apparent conundrum, that people want partners equal to or better than themselves, which leads us to wonder why people hook up at all. One answer is that people have different valuation criteria and they look for different qualities (otherwise very little exchange of goods and services in general would occur!) Hence, romantic partner's frequent "joking" that s/he got the better end of the deal in the relationship (one hopes the other partner is not so self-deprecating that they agree!) This argument extends to whether people are looking to hook up or to date seriously. In particular, people looking only to hook up might not be so concerned with having a higher quality partner. This is where your own anecdotal experience of dating in a big city may be helpful to understand the phenomenon: you might have had a relationship with a good-looking successful guy that you thought was going somewhere and was quickly ended by him, for no reason that you could figure out (reason: he was never serious and it was doomed from the start). Or, you might be a young guy pleased with yourself that you were able to get a really hot, smart professional woman (often a few years older) only to realize after a few nights together (and a few oblique comments about a recently departed ex) that she's not returning your calls anymore (I dub thee "transition man", and I share your pain or at least your ambivalent feelings about the situation). It is likely that these encounters would not occur if both partners were looking for the same level of commitment. Traditionally women are more likely than men to express interest in romantic relationships that go somewhere, rather than in flings, although the latter situation has become more common.

It's worth mentioning that in tolerant large cities there are also questions about same-gender hook-ups and relationships, but my comments are restricted to heterosexual pairings since as a straight male my experience and reading has been about that realm. The underlying assumptions of the dynamic (male vs. female reproductive strategies, dad vs. cad, short- versus long-term mating females, seeking genetic quality for future reproduction) may not apply or have very different effects in same-sex relationships, though these relationships may also provide a unique venue for testing these theories.

If we expect that access to information and market size affects dating and marriage, it should affect reproduction and gene frequency and distribution, which is what I was posting about before. I think that exactly this effect can already be observed to have worked more in some areas of the world than in others.

As a final speculation - because our romantic lives are far different than those of our paleolithic ancestors, and there are probably paleolithic-diet/mismatch hypothesis type arguments to be made about widespread relationship and family problems that we see in modernity; i.e. is divorce rate in industrialized countries the obesity and diabetes of modern relationships?

Monday, April 5, 2010

U.S. Life Span, Obesity, Elevation and Temperature

I had a conversation with a friend a few months back about the effect of cold temperatures on life expectancy; bottom line, she thought the colder the place you live, the longer you live. She theorized that this might have something to do with brown fat storage and/or thyroid function. I was skeptical, but curious.

Consequently I did a (very) little digging into questions of thyroid levels and brown fat and their impact on body mass index (BMI) and life expectancy. That adult humans even have brown fat at all was only accepted in the last few years, so there won't be brown fat maps floating around as yet. As for THR, TSH, or T3/T4 levels, I would guess that people living in cold climates would have higher T3/T4, at least seasonally (this is probably what drives the "blood thickening" that you lose when you move to Florida, for example). However after a brief search I wasn't able to find any solid studies relating thyroid levels to life expectancy and/or climate.

But we still have amateur epidemiology! Recall, the original discussion was about whether there is an effect of climate on obesity and/or life expectancy; let's not worry for now about how it might be mediated. (The following hyperlinks go to data sources.) I threw together some scatter plots comparing average U.S. state temperature and average U.S. state elevation to average state BMI and average state lifespan. I also used per capita income and % black population as a comparison for the effects of poverty. Instead of tarting up some boring scatter plots I'll just show the R-values:


RelationshipR-valueCorrelation
Obesity-Ave Temp0.47833Positive
Obesity-Elevation0.642028Negative
Obesity-PCI0.225389Negative
Obesity-% Black0.648228Positive


RelationshipR-valueCorrelation
Life Expect-Ave Temp0.631585Negative
Life Expect-Elevation0.292404Positive
Life Expect-PCI0.583952Positive
Life Expect-% Black0.724155Negative



So, what do we see here?

1) For both obesity and life expectancy, % black state population by state is a better predictor than PCI, average temp or altitude.

2) However, for obesity, % black population is just barely a better predictor than altitude. Low pO2 makes you skinny. No surprise there.

3) Interestingly, while altitude almost tied for best predictor of obesity, it is the weakest predictor here of life expectancy. People at altitude are a lot skinnier on average but don't live that much longer on average.

4) The situation is reversed for income. The richer a state, the less obese it is, although the relationship is weak. However, the richer a state, the longer-lived it is, with a substantially stronger relationship than with obesity.

5) There is a positive correlation between obesity and temperature, and a negative one with life expectancy and temperature. My friend would argue that this supports her position. However, people are not distributed evenly among states: Southern and South-central states (that both have a higher black population and are not as wealthy) make up 17 of the 20 warmest states, so it's difficult to tell what's going on, especially since those were better predictors anyway. A real study would look at economically and ethnically-matched individuals at least. While I'm ordering up studies from real epidemiologists to satisfy my curiosity, I'd still like to see average T3/T4 levels by average temperature and average January temperature.

Added later: from Marginal Revolution - I'd always been curious why obesity would suddenly become a problem in the second half of the twentieth century, when a lot of the production problems had made large volumes of calories cheap before. Turns out there's no mystery and it was already a problem a century ago.

Sunday, March 28, 2010

A Database of Socioeconomic Innovation

No, I don't have one or know where one is, but such an index would be useful to voters and policymakers around the world. The tragedy is not when these innovations are attempted, and fail; the tragedy is where they are tried, and work, but no one knows about it, and so we think all these questions are mere political theory when in fact we have cold hard practical data.

- For example, who knew (at least in the U.S. where I'm writing) that in 2001 Portugal decriminalized all drugs? In 2001. Suddenly California's November vote to legalize marijuana no longer seems so cutting edge.

- For another example, Uruguay is experimenting with competing simultaneous currencies. At the very least it will be interesting to see if a gold-standard currency re-emerges out of such a system. (I predict not; more on the the coincidental nature of the gold standard here.)

One innovation that has yet to be institutionalized is the adoption of automatically-assigned publicly accountable metrics on public programs; it continues to amaze that democratic publics are so willing to let their leaders avoid closing the loop. What democracies typically rely on are journalists, but in an increasingly complex age I don't think this assumption will hold much longer. One exception is civil engineering projects, which in some areas of the U.S. do publish goals and progress, but local governments don't often trumpet their successes and failures. For example: if a freeway ramp goes up ahead of schedule and under budget? Great! The civil engineering firm should become a local household hero! That's a real example, except unfortunately for the household hero part. When this happened with construction on a feeder ramp to the San Francisco Bay Bridge in 2007, no one seemed to notice or care, despite that hundreds of thousands used the freeway every day and it had a direct impact on the local economy and quality of life.

On the other hand, in the democratic feedback loop world I'm envisioning, what if some measure - educational, environmental, economic growth promotion, etc. - did not hit the targets it claimed it would? An informed public would demand that their representatives fix it or throw it out. A much stronger version of this system is futarchy, proposed by Robin Hanson; in such a system, not only do you mandatorily attach metrics to new law and policy, you adjust how much each voter's voice counts based on how well the candidates and policies s/he has voted for achieve their goals. If you keep voting for measures and candidates that say they'll achieve X, but don't, you keep losing clout. If this seems undemocratic, recognize that we already partly do it, although clumsily and with proxy indicators; felons and some mentally disable people aren't allowed to vote in the United States.

Not only would clearer public accountability (clearer by virtue of objective metrics) orient voters and officials more heavily toward results, it would give good-faith minority opposition leaders some concrete grounds on which to base objections (see Megan McArdle's request to build metrics into the healthcare reform package that just passed the U.S. Congress.) Your party thinks that law X won't achieve outcome Y? Fine. Let's all bet on it. Legitimate objections to majority policy are realized when Y doesn't happen, and bad faith negotiators who don't have a real objections (or don't have one that the voting public supports) now have a more difficult job.

In any event, the outcomes of others' experiments are worth watching. Decisions involving millions of people are obviously important, and it would be great for a political science department somewhere to put together a one-stop shop for such innovations - along with, of course, objective results.

Sunday, February 28, 2010

Name That Heuristic!

Five years after a major and mediagenic tsunami, there's a tsunami warning along the west coast of North America as a result of the earthquake in Chile. If there's a heuristic besides "be really stupid" that describes this, I don't know what it is:

After I got back to the car [a reporter] came over and pointed out the apparent low tide. Woh! I had never seen anything like it. It looked lower than the lowest tide I had ever seen at La Jolla shores. After watching it for about 10 minutes, I did exactly what you are warned not to do during a suck out during a tsunami. Yeh, that’s right, I walked out there.

At least the people in Thailand could say they didn't know what the low tide meant, and didn't have an official warning. This person (and others like him) didn't have that excuse, and still went for a walk on the seabed. What would his friends have said if he'd been swept out? That he didn't realize the danger? I suppose that's true.

Tuesday, February 23, 2010

Sarcasm and Theory of Mind

The British Psychological Society Digest has posted a paper inexplicably investigating the onset of sarcasm in child development. Kids get sarcasm at least by 10, but not yet by 5. [Added later: no word on whether British children develop sarcasm earlier than normal children.]

Why? For sarcasm you need a theory of mind. For me to appreciate (and enjoy) the incongruity of your saying something that you don't believe, I have to understand that a) you're not saying it to deceive me (that is, your intent) and b) you don't believe it (that is, I have to use what I know about you to form a model of your true beliefs, independent of the statement that you're making right now).

Blattman's post ends with the inevitable crack on the supposed lack of sarcasm of earnest Midwesterners, though in my experience I find this to be more true of Southern Californians. In my own anecdotal experience, Midwesterners understand
sarcasm just fine, they're just clear communicators averse to saying the opposite of the truth, and furthermore they're not so self-centered as to expect you to internalize a model of their belief states in order to get their brilliant jokes.

But while we defend geographic sarcasm stereotypes - since sarcasm relies on theory of mind, i.e. on the ability to quickly form and evaluate models of other human beings (who you may have just met within the hour), is there a link between sarcasm ability and number and breadth of humans you're exposed to? In other words, in a dense cosmopolitan area, are people more likely to be sarcastic? Certainly New York is the sarcasm capital of the United States. Meanwhile, Southern California is not dense, and diverse though it may be, people here have an unfortunate tendency to hunker down in ethnic colonies, both of which conspire to keep them from "practicing" modeling a wide range of other humans.

So are London and Paris notoriously more sarcastic than the rest of their respective countries? What about Hong Kong relative to rural China? Another possible explanation is that people in cities tend to be more educated. You could control for this by finding a dense city with low education levels and compare against more educated people in the suburbs; this model predicts the people in the city would still be more sarcastic.

Saturday, February 13, 2010

That Geographic Facebook Map

...that's been making the rounds (original article here) is interesting because of the social boundaries it shows:


Two perennial questions I have about American demographic boundaries are reflected on this map:

1) Why the boundary between western Stayathomia and northern Greater Texas? Really this just revisits the Upper Midwest vs. Lower Midwest. Iowa and Minnesota are better educated, less religious, and have higher per capita incomes than Kansas and Oklahoma. Echoes of slavery are often raised, not unreasonably, as the source of many of the cultural differences between Northerners and Southerners - but that's really an argument for the Eastern Seaboard, since it was never well-established in Kansas or Oklahoma. So is the reason for this that settlers of the Upper Midwest came from the Northeast and settlers of OK and KS from the Old South or Texas? Something to do with crops there or the early discovery of oil and the subsequent impacts of those resources? The much larger Native American populations in Oklahoma and longer history of conflict in Kansas? I'm at a loss.

2) The boundary between Dixie and Stayathomia tracks the Bad Stripe fairly well. The Bad Stripe is an area of poverty and general unhappiness trending basically along the western side of the Appalachians from West Virginia down through eastern Kentucky and Tennessee, then trending west through Arkansas and to the edge of Oklahoma. (It's not just economics; for crying out loud even the weather is worse - it's cloudier there. Most recent article here). The Stripe first came to my attention when I noticed that people there voted more strongly GOP in the 2008 presidential election than they had in 2004. This could be a coincidence, or it could be that someone in North Carolina is unlikely to be friends with a lower-status, less-educated, less-happy person in Kentucky and Tennessee, and likewise somebody in Ohio, so the chance for a connection from OH to NC through the Stripe is low. (Sorry KY and TN, those are the unpleasant realities.) It's also interesting to ask whether there's a general problem with border regions, i.e. two regions of roughly equal parity have a region between them of lower parity. Is there a similar bad stripe between, say, northern and southern Japan, or northern and southern Germany?

Monday, January 4, 2010

Rational vs. Traditional Warfare: the Mongols and the Shawnee

The historical origin of the phrase "on the warpath" in American English is underappreciated; in fact, you can go for a stroll on literal, physical warpaths around the eastern United States. Prior to the arrival of Europeans in North America, there already existed in a network of footpaths used for trade, communication, or warfare between nations. Certain of the paths were traditionally used more often than others for war, especially by the well-traveled Shawnee - hence the name, in Shawnee, Athawominee - path where they go armed.


The Great Warpath ran along the Great Valley of the Appalachians (the light pink band on this map).

It's critical to recognize that the warpath was not necessarily the easiest path between two warring nations, but a traditional path used during military conflict. It was a system of war that even enemies agreed on. Traveling the warpath signaled to friend and foe alike that you were going to war. By obeying the same set of mores, this means that enemies had made an inter-cultural agreement of sorts, a Geneva convention for pre-contact America.

It should seem strange that such a tradition would persist. War results when the interests of two nations have become so estranged that at least one party abandons the sub-optimizing but non-zero-sum agreements of civilization to achieve its ends through zero-sum violence. At this stage, there is no negotiation or mutually agreed restraint to be had; adherence to rules when rules are being discarded in favor of brute self-interest would seem to make survival more difficult; and those so naive as to follow some code of conduct would swiftly perish from the Earth. Sometimes in violent conflict, both parties behave themselves (sub-optimize) in ways that obviously and immediately benefit them both; for example, trading prisoners, not shooting at medical personnel, etc. - but even those are quickly tossed out in times of desperation. So how could the idea of a traditional warpath survive?

To a modern military mind, it's bad enough that your army might have to travel through a single pass or over a single bridge, because that means your route will be predictable to your enemy, and they can attack you at your most vulnerable. But to voluntarily use a traditional route only because it's the traditional thing to do seems like sheer folly. If you're a Shawnee chief, it seems the best tactic would be to spread false rumors that your men have been sighted on the warpath, but meanwhile sneak over a ridge and attack your Cherokee foes in their predictable location on the path. The Cherokees might be scandalized by your dishonorable behavior, but who cares? They're dead, and you won.


Traditional Warfare vs. Rational Warfare

Traditional warfare may best be summarized by anthropologist Warren Morrill's comment regarding "wars" in the New Guinea highlands: "It's more like football. They won't fight in the rain because then their make-up would run." You would think that any neolithic people with such a preoccupation with cosmetics would quickly perish from the Earth; not only is this untrue, but until recently in history, traditional warfare was the norm. [Added later: another place where traditional warfare is clearly exemplified, precisely because it can be contrasted against the innovations that changed local norms to rational warfare, was early nineteenth century South Africa, i.e. Shaka Zulu - who "changed the nature of warfare in Southern Africa" which was previously "a ritualised exchange of taunts with minimal loss of life into a true method of subjugation by wholesale slaughter."]

And that is exactly the mystery. For the majority of human history and prehistory, the conduct of warfare probably resembled the puzzlingly cooperative military endeavors on Appalachian warpaths more than it did the rational war that modern armies conduct. By "rational" I mean warfare that is characterized by clear goal-setting, critical thinking to achieve those goals, and disregard of traditional mores in the pursuit of achieving them. Rational war can be, but is not necessarily, total war (involving the deliberate destruction of civilian infrastructure as military objectives). A hallmark of rational war is the adoption (even from the enemy) of innovations based on their effectiveness, and the elimination of military practices involving systems of custom, tradition and honor. It should be emphasized that as "rational" is used here, rational and moral will often be opposed - World War I was a major step forward toward rational war. More often than not rational warfare results in more death and suffering.


Rational War Comes to the Americas All On Its Own

It's important to emphasize that rational war is not an invention of evil Post-Enlightenment Europeans who dragged the world out of Edenic traditional innocence. In eastern North America, immediately to the north of the aforementioned warpath was the Iroquois Confederacy, who at least twice in the seventeenth century practiced total war and genocide in the pursuit of territorial expansion and elimination of adversaries (the Huron and the Susquehannock). It's clear that the Iroquois did not feel any need to avoid dishonor by keeping to the warpath, which is why they earned harsher-than-usual epithets from their neighbors, helpful nicknames like "black snakes" and "penis heads". And they didn't care, because they won.

Sometimes indigenous North Americans did have tricks that the European colonists didn't. When a 22-year-old Colonel Washington fought the French at Fort Necessity, he faced a combined force of French and Indians who used guerilla tactics - hiding behind trees, jumping out to take a shot, then hiding again. While this seems like common sense today, the French had just learned it from the Indians, and as a good British officer, Washington was of course shocked at their cowardice. After all, the honorable European way to fight in the eighteenth century was to stand in rows, fire, then kneel while you reloaded. The French and Indians didn't care, because they won (at least that battle). A little over two decades on, now-General Washington had learned his lesson, and fought his former commanders using the same tactics that had cost him Fort Necessity. The British considered his behavior cowardly, but you know where this is going. While this is far from an exhaustive survey, it certainly seems - and makes sense - that if all other things in battle are equal, if you cheat and throw out honor and tradition while rationally pursuing a clear goal, you win.


Rational War in Eurasia

The Aztecs had wheels on toys, but never thought to use them for actual work. Reading the military history of ancient Eurasia presents innovations that to us seem blindingly obvious but that nonetheless had to start somewhere. In the West at least, there was no such thing as the use of, and training in, standardized weapons and tactics (at least consistently) until they were introduced by Gaius Marius, Julius Caesar's uncle. Accounts of warfare in the ancient world often read more like confused gang-fights with 20,000 on each side than like battles.

Critical thinking about warfare was still surprisingly slow to get off the ground, and like the Eastern American woodland nations, Eurasians were ripe for exploitation. It's been said by military historians that the thirteenth-century Mongols behaved strikingly like a modern army set down in medieval Eurasia. Among other things their officer-equivalents were promoted based on - who knew - meritocracy, rather than family connections or whether they were popular with the local witch doctors and medicine men. There were no warpaths for the Mongols, though they did allow themselves to become sentimental when they executed royal enemies. When the armies of this same people encountered medieval Europeans, encumbered as they were with notions of Christendom and battlefield etiquette, the results were predictable. The Mongols didn't care, and not only did they win, they established the largest land empire ever to exist on Earth.


Game Theory: Traditional Warfare and the Long Term



Hulagu Khan, grandson of Jenghiz, whose army sacked Baghdad in 1258.

One criticism of game theory is that it tends to focus on "winning" isolated contests in the here and now at the expense of the long-term. In extreme cases these are called Pyrrhic victories. This is one limitation of individual reason in general - we employ simplifying abstractions that allow us to solve problems considered semantically in short-term memory. This has proved powerful for humans but results in unintended consequences over longer time periods than the individual's lifespan (and which are therefore more difficult to anticipate.). It's necessary to make simplifying assumptions to develop theories; but a simplifying assumption that a game is one-round or defined-length is anything but trivial to applications of game theory. In real life, games with a clearly defined number of "rounds" are very rare.

For many organisms including humans, most competitions for resources (money, food, territory mates) not only go on continuously for life, they involve frequent contact and even cooperation with conspecifics. Furthermore, game outcomes are not either-or. An aggressive animal always willing to fight may in fact win every fight, but will still accumulate injuries that a more docile animal might not. Always optimizing every round will not necessarily win the game.

The difference between traditional and rational warfare is the application of critical thinking. Humans are the only animal with a brain powerful enough that we can confuse ourselves by being able to solve short-term problems rationally instead of performing instinctive ("dumb") sub-optimization. War is a high-stakes instantiation of game theory, and players who follow traditional warfare rules would seem to handicap themselves by limiting their moves and making themselves more predictable; then again, they're behaving more like our pre-rational ancestors.

So if the Mongol model is so much more effective than the Shawnee model, why do we ever see warpaths, in eastern North America or anywhere else? Why retain any traditions in warfare when it's plausible that they're actively pernicious to the retainers?

To this point there are four possibilities that leap to mind explaining the persistence of traditional warfare. First is that traditional vs. rational models of warfare have no long-term outcome on the success of the culture that holds the associated values; they're just noise. Second, it's possible that this small survey is non-representative and there there is no problem, that just as game theory predicts there really is a global species-wide bias toward rational warfare, and that this small sample is masking it. Third is that it was just a matter of time, and we happen to be in a transition period toward rational warfare; that is, all those remaining traditional-warfare cultures in New Guinea will either have to adapt, or will be assimilated/destroyed by neighbors in the near future.

The most intriguing possibility is that there is actually pressure against rational warfare for reasons which aren't immediately obvious. An immediate parallel to game theory comes to mind, that it's better in Prisoner's Dilemma to forgive single defectors (but not double or worse defectors); otherwise the game turns into a string of mutual recriminations, to the benefit of neither player.

A mechanism for this is more difficult to construct but if these four hypotheses exhaust the possibilities and the first three are incorrect, than this must be it. Could it be (for example) that any culture willing to ignore inter-cultural military conduct codes inevitably implodes due to its disrespect for tradition (a kind of "rational warfare Fermi paradox")? If there actually is some pressure against rational war, it's difficult to imagine how such regulation is being imposed on its practitioners (that is, the Hurons and the Caliphs weren't in any position to make the Iroquois and Mongols start following the rules of honorable warfare). It's also tempting to draw an analogy between the distribution of the rational warfare meme and the self-limiting spread of a too-virulent pathogen, but the virulence mechanism is entirely different. In any event, this homeostasis-like pressure is likely visible over the extreme long-term, in complex ways, after countless game-theory rounds have elapsed. But even given the surprising persistence in modern times of traditional modes of warfare, in terms of casualties from warfare, far and away the predominant character of warfare on Earth today is rational, and it's unlikely that this will change. This is new in history as of the past century and if there will be long-term effects - ones which tend to eliminate the practitioners of rational war from circulation - we should, obviously, try to find out what they are.