The Cumulative Power of Small Effects

Caplan posted a link to this paper on the "Power of Personality" by Roberts et al.  It includes some valuable passages on the history of psychometry: 

... Walter Mischel(1968) argued that personality traits had limited utility in predicting behavior because their correlational upper limit appeared to be about .30. Subsequently, this .30 value became derided as the ‘‘personality coefficient.’’ Two conclusions were inferred from this argument. First, personality traits have little predictive validity. Second, if personality traits do not predict much, then other factors, such as the situation, must be responsible for the vast amounts of variance that are left unaccounted for. The idea that personality traits are the validity weaklings of the predictive panoply has been reiterated in unmitigated form to this day (e.g., Bandura, 1999; Lewis, 2001;Paul, 2004; Ross & Nisbett, 1991). In fact, this position is so widely accepted that personality psychologists often apologize for correlations in the range of .20 to .30 (e.g., Bornstein, 1999).

    Should personality psychologists be apologetic for their modest validity coefficients? Apparently not, according to Meyer and his colleagues (Meyer et al., 2001), who did psychological science a service by tabling the effect sizes for a wide variety of psychological investigations and placing them side-by-side with comparable effect sizes from medicine and everyday life. These investigators made several important points. First, the modal effect size on a correlational scale for psychology as a whole is between .10 and .40, including that seen in experimental investigations (see also Hemphill, 2003). It appears that the .30 barrier applies to most phenomena in psychology and not just to those in the realm of personality psychology. Second, the very largest effects for any variables in psychology are in the .50 to .60 range, and these are quite rare (e.g., the effect of increasing age on declining speed of information processing in adults). Third, effect sizes for assessment measures and therapeutic interventions in psychology are similar to those found in medicine. It is sobering to see that the effect sizes for many medical interventions—like consuming aspirin to treat heart disease or using chemotherapy to treat breast cancer—translate into correlations of .02 or .03. Taken together, the data presented by Meyer and colleagues make clear that our standards for effect sizes need to be established in light of what is typical for psychology and for other fields concerned with human functioning.

The paper goes on to make an important point about the cumulative power of small effects:

... Moreover, when attempting to predict these critical life outcomes, even relatively small effects can be important because of their pragmatic effects and because of their cumulative effects across a person’s life (Abelson, 1985; Funder, 2004; Rosenthal, 1990).In terms of practicality, the À.03 association between taking aspirin and reducing heart attacks provides an excellent example. In one study, this surprisingly small association resulted in 85 fewer heart attacks among the patients of 10,845 physicians (Rosenthal, 2000).

I have heard effect sizes around the 0.2 level from reseach in psychology and sociology dismissed by engineers as "useless". Medical practitioners are not so cavalier. An effect size at the 0.05 level in the field of finance would be a powerful money making machine. 

China, America, and Abortion

China is the country of the One Child Policy, a policy that has been in place for a generation. From that, you might expect that Chinese people are more positively disposed toward abortion than Americans are. Au contraire. The World Values Survey asked people in America and China to report their opinion of abortion on a 10 point scale, with 1 as "Always Justified" and 10 as "Never Justified". The American distribution of replies has three humps at 5, 1, and 10 (in that order): the middle and the two extremes. Replies in China, on the other hand, lump overwhelmingly at one end: "Never Justified".   

The Emergence of China in the Global Economy

In mid 2004 Lester Thurow gave a talk on the emergence of China in the global economy. He raises a number of important points. Half way through the talk, he dissects the growth numbers. China has a bottom up reporting system and a top down promotion system. Administrators advance in the ranks by posting good numbers. This is a recipe for over reporting. Unsurprisingly there is a mismatch between the amount of economic activity reported to Beijing and the amount measured through electricity use. The rule of thumb is that an increase in GDP of 1% requires an increase in electricity use of 2%. Japan is the most efficient. It can squeeze a 1% increase in GDP from a 1.7% increase in electricity, but that is only because, since the energy shortage of the 70's, Japan has systematically shut down industries, such as the production of aluminium,  that require a lot of electricity. The amount of electrical use in China suggests a rate of growth of about 4% to 6%, rather than the 8% to 10% that gets reported. 4% is a very healthy growth rate. Most finance ministers would kill for a number like that, but the difference between 5% and 10% is significant if you project those numbers over 50 years. 

This brings us to why press freedoms are so restricted in China when compared to the Hong Kong or Taiwan. The Party's legitimacy is based on performance. If the numbers don't look good, the Party is in trouble. With a free press comes greater third party review. That's a threat. It follows that we will see little movement in the direction of greater press freedoms until the Party can find a different way to sustain its legitimacy. China is a big country. The Party could use geography to their advantage. Chinese now can not move freely. A system of internal passports restricts the movement of people from one place to the next. Relaxation of those restrictions would allow people to move away from administrators they see as less legitimate toward ones they see as more.  

China vs. United States: Economy

Source: Mint.com

China vs. United States: People

Source: Mint.com

China vs. United States: Geography

Source: Mint.com

China vs. United States: Military

Source: Mint.com

Eurozone GDP

The Wall Street Journal has an interactive map of Eurozone GDP performance by quarter. Countries such as Cyprus, Greece, and Malta have fared the best. Previous high flyers such as Ireland and Luxembourg have fared the worst. Slovokia experienced a stomach-churning 11% decline in the first quarter. Caplan has written about how bad times beget bad policy. Let's hope he's wrong. 

Source: WSJ 

Happy Women by Ideal Family Size

The GSS asks the question:

All in all, what do you think is the ideal number of  children for a family to have? Please just write a number in the box below.

The answers range from 0 to 7. I truncated the graph to 4 because the answers for 5 through 7 attract extremely low N-sizes, which produces flakey results.  The most popular answer is 2, but we can see a clear pattern of greater happiness among women who view larger famlies as ideal. The marginal return in happiness for each additional child above 2 is small. The women who think 0 children is the ideal are the least likely to report being very happy and the most likely to report being none too happy. 

Source: GSS

Declining Female Happiness

The Paradox of Declining Female Happiness

Betsey Stevenson, Justin Wolfers NBER Working Paper No. 14969 Issued in May 2009

By many objective measures the lives of women in the United States have improved over the past 35 years, yet we show that measures of subjective well-being indicate that women's happiness has declined both absolutely and relative to men. The paradox of women's declining relative well-being is found across various datasets, measures of subjective well-being, and is pervasive across demographic groups and industrialized countries. Relative declines in female happiness have eroded a gender gap in happiness in which women in the 1970s typically reported higher subjective well-being than did men. These declines have continued and a new gender gap is emerging -- one with higher subjective well-being for men.

Source: NBER

An ungated copy can be found [here]

 

Declining Female Happiness: The Graphs

The first graph comes from the paper by Stevenson and Wolfers. The second one comes from the GSS. They both display the same data. 

Lluís Domènech i Montaner

Tyler was tweaking Gaudi and offered Lluís Domènech i Montaner as an alternate.

Yummy. 

Red Vote, Blue Vote

[click to enlarge]

If we look at regions that relate to Red States and Blue States, the relationship matches up with the stereotypes we often find in pop culture (which is disproportionately produced in Blue States.)  If you look at the state level data, the relationship weakens, as the averages of Red States and Blue States converge. At the level of individual voters, the relationship both disappears and re-emerges. The average Kerry voter scored no different than the average Bush voter (6.48 vs. 6.49 -- Bush voters were microscopically ahead.) But the variation among Kerry voters was greater than the variation among Bush voters (a standard deviation of 1.97 for Kerry vs. a standard deviation of 1.79 for Bush.) The Kerry distribution had fat tails. At the top end, there were more high scoring Democrats than Republicans. Blacks filled out the ranks at the bottom of the Democratic distribution. 

IQ by Region

WORDSUM is a proxy for IQ in the General Social Survey. Like IQ tests, few do very well at it. Few do very poor at it. Most are somewhere in between. We can sort the results of the WORDSUM quiz by region. The results are much like you would expext from what you see on prime time TV and  at the movies. Brains are over represented in the Northeast and West Coast. Folkwise cowboys roam the Moutain West. The Southern regions show an over representation at the other end  of the scale. This pattern does not disappear if you look at whites alone, although the magnitude of the brain scarcity in the South declines.  

Nonsense Multiplied is Still Nonsense.

Students of rhetorical fallacy learn about the argumentum ad populum.  Democracies are often practitoners of it. Sometimes it is important to take a step back and remember that opinion polls are often an aggregation of error, not the stenography of truth.  In 1993, the NORC asked Americans five questions about science. The results ranged from bad to worse. About a quarter did not believe that antibiotics kill bacteria or do not kill viruses.  About a third believed that humans make all radioactivity.  More than half believed that astrology has some scientific basis, that humans did not develop from some earlier sort of animal, and that all man made chemicals cause cancer in sufficient quantity. 

20 countries took the test, which included these questions in additon to seven others. 13 other countires scored worse. 

Happy People Are Less Likely to Feel Sad

Sometimes it's just good to know that the GSS is not just random nonsense. 

Democrats More Likely Than Republicans to Blame "The Jews"

The Boston Review reports the data:

Interestingly, Democrats were especially prone to blaming Jews: while 32 percent of Democrats accorded at least moderate blame, only 18.4 percent of Republicans did so (a statistically significant difference). This difference is somewhat surprising given the presumed higher degree of racial tolerance among liberals and the fact that Jews are a central part of the Democratic Party’s electoral coalition. Are Democrats simply more likely to “blame everything” thus casting doubt on whether the anti-Jewish attitudes are real? Not at all. We also asked how much “individuals who took out loans and mortgages they could not afford” were to blame on the same five-point scale. In this case, Democrats were less likely than Republicans to assign moderate or greater blame.

The story does not give us a link to the original study with the raw data (a search through Google Scholar using the authors' names yeilds nothing), does not tell us how jews responded, nor does it tell us how people assesed the blame of other ethnic or religious groups.  Nevertheless, it is striking that Democrats were more likely than Republicans to blame "the Jews" for the financial crisis -- by a wide margin. 

The General Social Survey asked respondents to asses their feelings toward jews using a thermometer from 0 to 100. We can sort those assesments by party identification.

The middle column reports the responses of Independents. The left wing of three columns reflect the opinions of people who lean Democratic. The right wing of three columns reflect the opinions people who lean Republican. We can see that "50" was the most popular response by a plurality among all of the catagories of political identification. Republicans, however, were more likely to report warmer feelings toward jews (greter than 50.) Democratic respondents were more likely to report colder feelings (less than 50.)

The two lines of evidence point in the same direction.

One might say this is just a reflection of black anti-semitism. Blacks identify with the Democratic Party at a rate of around 80%. But if you exclude blacks from the mix, the results are still largely the same. Republicans are still more likely to report warm feelings. Democrats are still more likely to report cold ones. 

Source: Boston Review 

Source: GSS

Headless or Heartless

We've heard the quote before--often attributed to Churchill--, "If a man is not a socialist at the age of 20, he proves he has no heart. If a man remains a socialist at the age of 30, he proves he has no head." Substitute "liberal" for "socialist" and the data confirm a trend: people grow more conservative as they grow older. According to a very large data set compiled by the NORC during more than three decades, people are more likely to call themselves "slightly conservative", "conservative", or "extremely conservative" than they are to call themselves "slightly liberal", "liberal", or "extremely liberal". It is only among the youngest age cohort, from 18-30, that the reverse is true. 

In 1990 the average American was 32.9 years old. As of 2007, the Census Bureau reported that the average American was 36.6 old -- well past the inflection point at which the balance of opinion flips from liberal to conservative. 

Source: GSS

Source: Census Bureau

The China Model

Look at Singapore.  Look at Hong Kong.  Look at Taiwan.  You will get a sense of where China will be 30+ years from now (assuming a constant -- and explosive --  8% growth rate, the mainland will be where Hong Kong is today in 36 years) .  The rise of China will be one of the biggest themes of the 21st Century.  To get a good sense of where China is going, it is important to get a clear picture of where China is today.  Glasshouse Forum recently hosted a debate on the nature of the China model of development.  The conclusion seems to be the sociological equivalent of Heisenberg's Uncertainty Principle.  China is change.  To the extent we know the direction and speed of China, its precise model is hard to pin down.  To the extent we have a clear picture of the model, we loose sight of how things were done yesterday and how they will be done tomorrow.  If you look at China over the last 50 years, the Chinese Way will differ from the Way you see if you look at only the last 30.  The Way you see over the last 30 years (since Deng) will differ from the Way of the last 20 years -- or 10. 

The discussion raised some important points on how the government in China gains and sustains legitimacy. In the West, this takes the form of elections. In China, this takes the form of "performance legitimacy". Goals are set in Beijing or the provincial capitals and the rise or fall of administrators depends on achieving those goals. Opinion polls were mentioned in the discussion as a substitute for elections. Legitimacy is easy when the economy is booming. The true test comes during hard times. Times have been very good for very long.  But China, with the Party at its helm, was also able to weather the Asian Crisis of  '97.

Wang Shaoguang urged, borrowing from Seymour Martin Lipset, that a government is legitimate if there is no workably better alternative, pointing to Russia to illustrate his point. Democracy may be feasible, in this way of thinking, but it is not necessarily better. You can imagine governments better than the one in Beijing, but they are not necessarily feasible. But what about Taiwan? Taipei is now home to a democracy that has at least two passionately opposed political parties who have peacefully ceded power to each other at least once.  Democracy works there in a way that would be recognized in the West -- even under conditions of extreme stress (the subsequently peaceful transition of 2004 occurred even after mutual recriminations of assassination and conspiracy.) Prosperity in Taiwan is higher than on the mainland by a full order of magnitude times two and is therefore arguably "better" on a performance basis. 

There seems to be a reluctance to embrace Fukuyama's End of History -- that liberal democracy is inevitable. There seems to be a resistance to radical transformation.  Given the history of China in the 20th Century, this in understandable. China's road to capitalism was a marginal one, not shock therapy in the Eastern European mould.  Production in excess of the Plan was sold on the open market at a profit until private production dwarfed and consumed the public sector. The Party could use geography to their advantage. China is a big country. Allowing greater geographic mobility would be in keeping with a trend of marginal improvements that could shore up any slack in legitimacy. People unhappy with the performance of administrators in their city or province could pack up their bags and move to the greener pastures, voting with their feet rather than by ballot box. 

Source: Glasshouse Forum 

Wolfram Alpha - Google Trends

Wolfram Alpha is due to be unveiled on Friday, May 15 at 7:00pm Central Time. (Today!) Google Trends reports that "Worlfam Alpha" as a search phrase is picking up. There was an initial spike in interest in March. Interest fell, and is back again here on the eve of the unveiling. Strangely, Google trends also reports that the search phrase is most popular in Romania, Germany, and Ireland (in that order.) 

Busting the Moon Hoax Conspiracy Theory

According to a poll done by Gallup in early 2001, 6% of Americans (about 1 in every 16) believe that the Moon landing was a hoax.  I recalling being asked by a co-worker if I thought men had actually been to the Moon after Fox aired it’s TV special. It's the sort of question that can catch you off guard. My immediate response was, "I had an uncle who was involved with the space program. Either we went to the Moon, or he was a liar. I tend to believe the former, rather than the latter." I also suggested that doubters could use a telescope to see the remains of the Apollo mission. Mythbusters did an episode on the Moon Hoax conspiracy theory, in which they dissects and refuted some of the more common arguments that conspiracy theorists use to make their case: the photographs that would supposedly require more than one source of light; the shadows that don’t align; the flag that supposedly flaps in a breeze; the footage of space walks that could have been made by over cranking the film; that the crispness of the boot prints requires moisture. 

The result? Shadows that don’t align can be accounted for by the contour of the landscape alone. Photographs that supposedly could only be made with multiple sources of light can be made with only one. After being jostled, a flag in a vacuum will actually flap more than one in air. Video at normal speed of a man walking in a space suit on a Vomit Commet matches the footage from the Apollo mission better than anything that over cranking can produce. And the rough surface of Moon dirt creates crisp boot prints under a condition of a vacuum. The episode ends with Adam and Jamie going to a telescope and shining a laser at reflectors that were left on the surface of the Moon by the Apollo mission.

Source: Gallup Organization

Source: Mythbusters   

The UK Over Time

While looking for a time series graph of unmployment in Britian, I stumbled over this paper from House of Commons Library published in 1999. It's a collection of 100 year time series graphs for a wide variety of economic and social tends. Here are a few of the dozens of graphs from the paper. 

Source: House of Commons Library