Ranking of University Attended Does Not Correlate Well With Adult Well-Being

More on this in a Q&A format in the post right above.

It's worth understanding the relationships between the multiple statistics about a university,[1] and with its output - that is to say, life satisfaction. There are numbers for average early career salary, but precious little on satisfaction: Forbes, for example, uses a very dubious metric of percentage of alums donating to a university, and how much they donate on average. (Dubious, because this likely has to do with the institution's proactivity in soliciting donations, as much as with alumni gratitude.) There is almost nothing on overall life satisfaction - which is curious, because this is the main output we're concerned with when we apply, or send our kids, to universities - isn't it? What I could find, strongly suggests that the differences in life outcomes caused by universities are minimal, if any; the strongest is a ~10% premium in income at the most selective institutions. But the only reason to care about money is because we think it would make us happier - and if it's not, which is what the few numbers we have are showing, why are we bothering with this?

It's scary to buck trends, especially for parents, in a way that you worry might damage your kid's future. Here's what I'm telling you, parents: you're causing your kids definite harm with the current get-into-college rat-race, for benefits which are at best slight, and maybe - probably - are entirely illusory. That is to say, you may well be taking away a happy childhood for nothing. When you tell your kids not to cave to peer pressure, set an example for them to follow. If you're not brave for them, no one else will be. If you go all-out in college admissions madness from preschool on, you owe it to your kids to know why you're really putting them through this. For their sake, I hope it isn't just because you're trying to win status points or avoid judgment from your family or neighbors or coworkers.

The most important conclusion is for stressed out kids and their stressed out parents: you will be fine. You will get in somewhere, and you will get a good education, and have a good life, and it will be fine. The ranking numbers are often based on very arbitrary decisions, cardinal rankings are not good bases for statistics because they often imply gigantic differences, and the makeup of the individual students is far and away the most informative driver of choice of university. Kids: try to get into the best school you can for what you're interested in, but don't kill yourself to do it, and don't despair if it doesn't happen - because it really doesn't make much of a difference. Also consider where you want to live, and what kind of people you want to date and be friends with for life.


FOONOTES

[1] If you're interested in relationships between those statistics, they're below. I have to emphasize, again, for something that causes so much stress and consumes so much time, when we choose colleges, we really don't know what we're buying. There is amazingly little literature on outcomes, which suggests that whatever is driving the college admissions Olympics, it's not how much getting into good college benefits our lives. Even though that seems absurd, it's also obviously true.

For the relationships between SAT, acceptance, and endowment size, here's how it came out. SATs are more closely correlated with the other two, about equally with both in fact. Endowment per student correlates markedly less well with selectivity.

• SATs vs percent of applicants accepted: R^2 = 0.5896. The ones that are much more selective than their SATs would predict - the four military academies (which are obviously selecting on something besides SAT), followed by CUNY-Baruch and Babson College. The ones that benefit students - they are NOT as selective as SATs would predict - are Villanova, University of Maryland-College Park, and University of Minnesota-Twin Cities.
  
  
• Endowment per student vs SATs: R^2 = 0.5711. I even chose a logarithmic curve, since the SAT approaches but cannot pass 1600, but it didn't improve the goodness of fit.
  
  
• Endowment per student vs percent of applicants accepted: R^2 = 0.3388. The curve looked like endowment might have an increasingly marginal effect beyond $500,000/student, but even taking out the institutions above that or trying to fit exponentially or logarithmically didn't do much to the fit. The ones that are off-trend in a way that benefits students (lets higher amount of applicants in than their endowment would suggest) are Grinnell, Wellesley, University of Richmond, Texas Christian, and believe it or not Princeton. "Benefit" assumes that the endowment actually affects student experience.

[2] I had also been quite curious about the effect of nationally prominent athletic programs, especially football, on academic rankings. This is from personal experience, since I recall how the yield (% of accepted students actually matriculating) went up after Penn State's almost-number 1 1994 season. The Flutie Factor (cited in this paper by RT Baker, which was submitted to an academic institution but doesn't look like it's peer-reviewed) shows that the effect was known prior to that (the year after Doug Flutie won the Heisman, Boston College had a 25% increase in applications.) While there is very little literature about the effect of sports performance on academic ranking, this paper argues that in fact increasing football ranking does increase academic ranking.

To Remind People of Status Hierarchies, Especially Opt-Outable Ones, Is Immoral

It's a shame that humans are so sensitive to status hierarchies (a zero sum game of relative value) as opposed to the "absolute value" of one's life experience. If I'm living in a shack in the woods getting by doing odd jobs and my family is fed and healthy, why should I ever be resentful? My life is fine!

Even if I'm going to compare myself to others, compared to the vast majority of people who ever lived, I'm doing great. But it's very hard not to let that status awareness among your here-and-now peers sneak in and start making us less content with our actually just-fine lives. This is the great irony and tragedy of living in the West in the twenty-first century. Avoid zero-sum games if you can, and remember, status is positional and therefore always zero sum.

There are many non-opt-outable status hierarchies (house, job, money, family, looks), and these are the most frustrating ones. One way that people in wealthy societies cope with status hierarchies is by voluntarily inhabiting multiple overlapping status hierarchies, but even then, you still live somewhere, look a certain way, and have a certain amount of money.

It is for this reason that people who go out of their way to make others MORE aware of status hierarchies and their position in them, whether non-opt-outable ones or consumption-based ones, are destroying contentment and are profoundly immoral. I'm sure you've already heard it a million times, but it's worth thinking about how your brain considers those rich good-looking leisurely people you see on TV to be your here-and-now peers, not to mention those enhanced pictures and narratives of your friends' and colleagues selfies and vacations. Robin Hanson goes into more detail on this here. Meanwhile - yes, I have quit Facebook, and maybe you should think more about it.

Tasting Freedom: Happiness, Religion and Economic Transition

Paper is here, by Orsolya Lelkes (2006.) Emphases mine.

Abstract:

Economic transition lowered happiness on average, but did not affect everyone equally. This paper uses Hungarian survey data to study the impact of religion and economic transition on happiness. Religious involvement contributes positively to individuals’ self-reported well-being. Controlling for personal characteristics of the respondents, money is a less important source of happiness for the religious. The impact of economic transition varies greatly across different groups. The main winners from increasing economic freedom were the entrepreneurs. The religious were little affected by the changes. This implies that greater ideological freedom, measured by a greater social role of churches, may not influence happiness per se.

The interpretation I take from this is that in the positive psychology triad of the things the produce happiness (pleasure, meaning, and flow), each of these things is a variable in each person's overall happiness equation, and each component's importance varies between individuals. Money is mostly something we exchange for pleasure, unless you're an entrepreneur, then it produces meaning and flow. If you're religious, the meaning component is bigger so again money is not such a big term. As well, I would imagine that the starting point and absolute difference in the transition makes a difference. That is, if you're starting out second-world (like Hungary, the source of the data in this paper) you're probably not going without food, shelter, or public safety. But if you're in Botswana or for that matter China over the last few decades, you might well have gone from famines and no housing or police to a more developed social environment, there may be a greater impact on happiness.

Happiness by State in the US, 2018

A study done by Wallethub (their image below) using their own 31-factor happiness index shows the Bad Stripe, along with a few other interesting patterns.


1) The Bad Stripe (West Virgina, Kentucky and Tennessee, Missouri, Arkansas, and Oklahoma) is evident as a negative outlier as usual. These make up 6 of the bottom 11. For fans of Albion's Seed, this is where the Reavers are, i.e. Greater Appalachia. This also shows the limitation of a state-level analysis. There is significant structure within the states. Pennsylvania's southwest if taken separately would very likely look like West Virginia. The southeast if taken separately would be much more like New York and New Jersey. Same thing for Missouri - the northern part of the state is likely more like the Upper Midwest, and the southern part is the Ozarks, part of the Bad Stripe and more like Arkansas.

2) The Upper Midwest and Utah stand out as positive outliers, as usual. Moynihan's Law - is it the result of Yankee settlers (again Albion's Seed), non-British Isles North European immigrants, or some combination? (Map below from Wiki on German Ancestry in the USA and Canada.)



3) Very interesting that two demographically similar states like the Carolinas could be so different in this rating. North Carolina has done better economically than South Carolina, and culturally does tend to move in sync with Virgina (perhaps most famously in the 2008 presidential election predicted by Nate Silver) - which makes sense because South Carolina was largely settled from Georgia initially, and North Carolina from Virginia. Still, they're not THAT different, and they have a very different happiness outcome.

4) I can't argue for similar historical links to the Reavers for Mississippi, Alabama and Louisiana, who as always fare very poorly. You'll have to develop your own theory for that!

What Are the Things That Could Derail the Improvement of Life on Earth?

We assume that wealth will grow, violence will decrease, and in general life will get better for life on Earth, if only things keep going roughly as they are now. What might happen to interrupt this process?

On a Timescale of Minutes to Years:

1. Weapons of mass destruction/acute ecocide. This breaks down into three categories:

a) nukes: there are a lot of them left. This one is frustrating because it's still an existential threat but it's been around for a long time, so people have become used to it.

Once one is is used deliberately or even explodes by accident, the taboo is broken, and more will follow in much shorter order than the interval between that event and Nagasaki.

b) any other such weapons, especially biological weapons.

c) The Singularity. Superhuman general artificial intelligence would still not necessarily be intelligent enough to be evolutionarily stable, and instead of a blossoming of ultra-intelligence, could just result in ultimate ecological castastrophe.

2. Natural events that could destroy parts of the infrastructure the modern economy relies on. Earthquakes weren't a big deal when you could rebuild your bark longhouse in a few annoying hours. Not so San Francisco. 9.0 earthquakes aren't that frequent, but they occur. Similarly, we don't even know how often Carrington-level events happen because until there were electrical lines to be affected, there was no way of knowing and no reason to care. Also related to Carrington events, refer back to item 1a; the result of a nuclear high-altitude EMP would be catastrophic.



Decades to Centuries:

3. Selection against intelligence by economic development, both within and between countries. That is, idiocracy. The most fertile countries are often the most disastrous. Related: the world and technology will not change any slower than they are now (unless one of the disasters in this list occurs) yet there are people (the majority?) who appear constitutionally unable to adapt to this level of change and think in abstract terms. The modern world ironically appears to make these people regress into more and more of a fundamentalist, tribal state where they assume, correctly, that they will not be able to understand the world at all, so they cling to tribal authority.

4. Consumerism and collapse/transparency of status hierarchies, making people unhappy with otherwise stable productive lives.

a) Consumerism: it has been argued that above a certain amount (usually given as US$70,000), gains to income translate increasingly marginally to happiness. However it is increasingly impossible to escape images of houses, mates, experiences that you're not getting even with your $70,000 and your nice home, spouse and vacation. Hence this principle runs up against the human irrationality that we would rather live in a neighborhood of $100k earners and make $120k, then a neighborhood of $200k earners and make $180k.

b) This same media technology also means that increasingly, we are de facto in a world culture where there are few isolated laboratories for meme innovation. In the past, even in neighboring dictatorships, at least the more flawed dictatorship might lose on the battlefield, with the slightly better practices of the other dictatorship winning out. But what will ever fix your shitty institution now? The Mongols aren't about to overrun the DMV for being too slow. Also, increasingly we cannot preserve the independence of our multiple overlapping status hierarchies and "healthily" isolate our social spheres from one another - so your boss, or that girl that was prettier than you in high school, can make fun of your for being the president of a local organization which otherwise would've give you a nice status boost to increase your QOL.


Therefore, to avoid acute events, we should focus on continuing nuclear disarmament, start taking biology more seriously before CRISPR gives us the equivalent to Rosenbergs working with ISIS, and increase focus on AI safety. We should also try to understand how to predict and protect against Carrington solar events and similar century-frequency geological and astrophysical threats.

To avoid the longer-onset ones: no solution here would seem palatable, but otherwise we face death by a thousand cuts. To fix #3, it would seem only eugenics by licensing reproduction would work within a country, but this is abhorrent in Western politics to liberals and conservatives alike. (A certain China does come to mind and they seem to have done "okay", and by "okay" I mean "the greatest developmental triumph in human history". Despite or because of one-child?) There is also the Brave New World style solution of a big reservation or favela for all the people who can't hack it in the future economy, but in Brave New World it was just a few misfits, as opposed to all their Epsilons and Deltas.

For #4, we could adopt cultural norms about media use - while this is already happening to some degree, it takes both personal discipline and is easily eroded by non-cooperators, i.e. your co-worker who you suspect will check their email on the weekend even if they said they wouldn't. Also, solving the "tyranny of territory" would speed the diffusion of good memes, even in a connected mono-culture world. ("Tyranny of territory" is that humans have to live on the surface of the Earth so organizations from families up to government have static boundaries. Charter cities are a nice idea but fall flat as long as they are within territories held by the cartel system of mutually-recognizing violence monopolists, i.e. states. That is, I can't tell my DMV, California, that I choose to use Minnesota's DMV, because at bottom guys with guns will come and make me cooperate.)

Economic Freedom and Happiness

I was recently looking at the economic freedom numbers for countries around the world and I wanted to know what the connection to actual outcomes was; in particular, the happiness of the people in those countries. All the rest are surrogates. Political and economic debates sometimes lose focus on this fact.

I looked at the following data, for all countries that had them (evident in the datasets); for each, it was always a clear majority of countries on the planet.

- The Heritage Foundation's Economic Freedom Index for 2011

- White's Life Satisfaction Index

- GDP per capita (IMF data, supplemented by CIA for small countries, non-reporting countries, or inaccurately reporting countries)

- Gini index (from the U.N., or for some countries in the midst of conflict the Global Peace Index statistic)

- Economic growth for countries 1990-2007 (data from United Nations Statistics Division)

As I've done previously, rather than show a bunch of scatter-plots, I'll give you the statistical highlights. I'm happy to share the spreadsheet if anyone is interested, although this is not quite graduate-study-level QC'd data.


Map of White's life satisfaction index in 2006. This is what counts.

Highlights:

1. Interestingly, economic freedom correlates more closely with life satisfaction than with economic growth (R^2=0.239 vs 0.1075). This suggests that economic freedom adds to utility other than through direct material gain. Freedom does correlate better with per capita GDP than with growth (more below).

2. Economic freedom is associated (albeit weakly) with a decrease in Gini, that is with a more equitable income distribution.

3. There were a number of outliers in the plot of economic growth vs. life satisfaction. Most of these were very high Gini countries.

4. In the "unsurprising" cateogry: there were two R^2 that rose above 0.3 were the correlation between per capita GDP and life satisfaction (0.3106, stronger without outliers). The correlation between economic freedom and per capita GDP was even stronger at 0.3873. The curve appears to flatten at the high end of per capita GDP (removing three outliers raised the R^2 to 0.4665) reinforcing the conclusion that once basic needs are met there is a diminishing return. It bears emphasizing that this is a correlation, not a cause; economic growth may CAUSE economic freedom, or they may both be caused in parallel by the same thing.

Correlation	R^2	Relationship Means?
Econ free & PCI	0.3873	Every point increase in econ freedom (range 0-100), raise PCI US$70
PCI & life sat.	0.3106	Raise PCI US$1,000, increase life sat. 1.14 pts (range 100-250)
Econ free & life sat.	0.239	Add a point in econ freedom (range 1-100), increase happiness 1.67 points (range 100-250)
Econ free & growth	0.1075	Add a point in econ freedom (range 1-100), increase growth 0.001%
Econ free & Gini	0.0837	Add a point in econ freedom (range 1-100), decrease Gini 0.2685
Growth & life sat.	0.0544	Every 10% increase in growth, get 25 points happier (range 100-250)
Gini & life sat.	0.0258	Increase Gini by 1, life sat. drops by 0.5 (range 100-250)

CONCLUSIONS:
Unsurprising, but interesting to see in this form. Adopt policies that expand economic freedom in order to make people happier, partly by increasing growth. The best way to produce happiness is to reach a target high per capita income, but there is a diminishing return. Economic freedom has a weak beneficial effect on Gini, but Gini can offset the happiness effects of high PCI and good growth.


Future questions:
- There are a number of countries that, looking merely at per capita GDP, aren't as life-satisfied as they should be. These countries usually have large Gini; a surface in a 3D scatterplot would show this distortion. It might be informative to see which countries are "off the surface" in terms of how much we expect their Gini to distort their happiness:PCI ratio, and then ask how this effect is transmitted - a first guess to investigate would be degree of media saturation. Prediction: countries with more media and a high Gini will tend to be less happy that those with less media but the same Gini. Seeing how the other half lives forces everyone into the same status game.

- Are certain regions of the world off these curves in predictable ways because of cultural commitments? (See the World Values Survey.) E.g., are Confucianist countries less happy per dollar of PCI? Or are cultures with more family-oriented, traditional values differentially susceptible to the effects of Gini distortion?

- (Added later: personal economic freedom in the 50 U.S. States can be found by category here.)

Quote of the Day

"...it is not a sign of intolerance for us to notice that some cultures and sub-cultures do a terrible job of producing human lives worth living."

- Sam Harris, talking about his forthcoming book The Moral Landscape

The Bad Stripe Continues: Overall Well-Being

In an overall well-being survey, the Bad Stripe that jumps off of maps of the U.S. of well-being, health indicators, and economics is again unfortunately represented.

Read more about the Bad Stripe here, here and here. Looking at the reverse of the map above, the photonegative of the Bad Stripe stands out in this map of frequent mental distress (Kentucky comes up worst.) In addition to the relations there, the Bad Stripe is also the boundary between three geographical social networks built by Facebook users, and it also tracks the boundary between Baptists and Methodists judging by geography-associated tags on the internet. It's tempting to speculate about the link between being a cultural boundary zone and an emotionally depressed area and what the causality might be.

Maximization: The Heroin Problem of Happiness-Based Morality

A problem of any happiness-maximizing theory of morality is "the heroin problem". If the point of morality is to increase pleasure (for self or for the greatest number, i.e. utility theories), shouldn't our goal be to create as much pleasure as possible, even if that pleasure is created by gaming the system (e.g., with heroin or a Matrix-like simulation)?

There are a few ways to think about this problem.

Resolution #1: Destroy values that obstruct pleasure maximization. In a parallel development, I have endeavored to destroy my taste in wine because by developing a taste in wine or anything else, you're working against yourself - you're actually making your marginal unit of pleasure more expensive. You have to choose which is the better option: having a refined taste, where you drink an expensive wine, experiencing X pleasure points, and signal your refinement to peers; OR tasting a cheap wine and not knowing any better, so you also experience X pleasure points, and also you have $50 dollars in your pocket to buy 5 more bottles of wine (so you actually get 6X pleasure points; unless the admiration you get from peers has some combination on the spectrum between being worth $50 or is 5 times more pleasure points, you're better off with no taste in wine.

Expanding this approach to happiness-maximizing morality in general, certainly it's the rare human whose moral intuition drives them toward hedonic excess at the expense of all other values. However, perhaps we're still laboring under bad and unquestioned moral assumptions which are after all not innate to human beings, and in fact it should be our goal to identify and do away with all values, beliefs and behaviors that get in the way of optimum utility. For example, most of us recoil at the suggestion that empathy should be eliminated because it conflicts with the pursuit of utility, but perhaps our outrage at such a suggestion is an example of a bad moral assumption. The anti-ascetic of the future will gladly pluck out (for example) the brain circuits that create a desire to care for his/her offspring, as clear offenders to the unflinching goal of increasing happiness.

Resolution #2: Heroin and orgasms aren't the only things that bring about happiness. There are certainly multiple types of experiences that lead to happiness beyond that of physical pleasure; again, if morality is really and only about happiness, the goal should be to identify those types of experiences, the realization of which conflict with each other, and destroy the desire/capacity for conflicting experiences which are in the minority. And presumably a simulation could provide not just heroin rushes and orgies with supermodels, but all the higher hedonic forms, up to and including a sense of meaning: professional achievements, family ceremonies, etc.

Resolution #3: It's the capacity for current and future happiness that matters. Neurologically gaming the system leads to an organism vulnerable to predation and disease; is ten years in heroin-simulation land until your body dies of dehydration better than sixty years in the real world? If asking to be hooked up to a pleasure simulation and left there until you die is wrong, why? This resolution is suspect because it usually conveys some degree of knee-jerk disgust for the incapacitated agent who has diminished their contact with reality in exchange for the equivalent of neurochemical masturbation. Again, see #1 above: this disgust for voluntarily putting oneself in such a passive position is certainly an obstacle to realizing a life of greater pleasure. Also test claimants of this resolution by asking: what if the simulation were built by aliens who guarded it and made absolutely sure the deteriorating, drooling simulation zombies inside the simulation (you!) were 100% safe - i.e., capacity for future happiness is now not a problem. Do you still have a problem with giving yourself over to such a simulation? If so, then future happiness capacity is not your real demand.

Resolution #4: Our moral sense is not entirely predicated on happiness. Our behavior is certainly not 100% rational or conscious-principle-guided by any means, so why do we think that our moral sense would different?