Validity of national IQ
In IQ and the Wealth of Nations (2002; IQatWoN) and IQ and Global Inequality (2006; IQGI), Richard Lynn and Tatu Vanhanen (L&V) present measurements and estimates of average national IQ (national IQ). In IQatWoN, L&V argue that national IQ predicts per-capita GDP (sup Fig 1). In IQGI, L&V argue that national IQ predicts quality of life measures (sup Fig 2). A common criticism of both works is to question the validity of national IQ. This criticism is motivated in part by the very low scores reported for countries in sub-Saharan African. A look at the distribution of national IQ is instructive (Fig 1).
Figure 1. The distribution of national IQ values (192 countries from IQGI).
L&V address the issue of validity by comparison of national IQ values with international test scores in math and science such as TIMSS and PISA. IQGI presents data from 10 different tests, with different scoring scales, in the form of 3 tables. To get a better grasp on the question of the validity of national IQ, I reanalyzed the test score data from IQGI. For better comparison, I renormalized each set of test scores relative to the maximum test score for each assessment. This is an imperfect but sufficient technique. An unweighted average of the available test score data was used to calculate a composite national test score for the set of 62 countries for which at least 1 test score was available (Fig 2).
Figure 2. The association between national test scores and national IQ for 62 nations.
National test scores are available for a limited range of national IQ scores, with few test scores available for countries with national IQs below the mid 80s. I interpret this to mean that for countries with national IQs below ~85, the test score data is insufficient to inform the question of validity. However, for the available scores (i.e., mostly above ~85), the relationship between national IQ and national test scores is very strong (see Sup Table 1).
The validity of sub-80 national IQs is addressed in part by the finding that IQ correlates with GDP and QHC (Sup Figs 1,2) throughout the observed range of IQ.
Update: Although there are only four values, the sub-80 national IQs are outliers, all with positive residuals. While this is hardly informative, it trends in the direction of casting doubt on the validity of sub-80 national IQ values.
Supplemental Figure 1. National IQ correlates with GDP per-capita (192 countries from IQGI).
Supplemental Figure 2. National IQ correlates with a L&V’s quality-of-life index (QHC; 192 countries from IQGI).
Supplemental Table 1. Correlation matrix for national IQ (IQ), national test score (Test), L&V’s quality of life index (QHC) and log per-capita GDP (logGPD) for 62 countries.
| r | QHC | logGDP | IQ | Test |
| QHC | 1 | 0.898936 | 0.7933265 | 0.7803476 |
| logGDP | 0.898936 | 1 | 0.760138 | 0.7565582 |
| IQ | 0.7933265 | 0.760138 | 1 | 0.9008035 |
| Test | 0.7803476 | 0.7565582 | 0.9008035 | 1 |
Related papers:
* Earl Hunt and Werner Wittmann, National intelligence and national prosperity, Intelligence, In Press –examines PISA scores
* Richard Lynn and Jaan Mikk, National differences in intelligence and educational attainment, Intelligence, Volume 35, Issue 2, March-April 2007, Pages 115-121. –examines TIMSS scores
Labels: general intelligence, IQ
is that a fucking line, or wut?
Or it could be that sub-80 countries don’t have everybody in school…
Perhaps stating the obvious: This study tells us that ‘IQ’ and other similar test scores correlate with wealth, but implies nothing about causality. The study will be considered by some to imply that differences in innate intelligence cause the differences in wealth, and by others that differences in education cause the differences in wealth, depending on his or her politics. But differences in wealth are also causal, since to a certain extent high IQ reflects how practiced a person is in certain mental activities which are possible in a wealthy society. Seems to me that in a sense the study doesn’t really say anything that’s not already trivially obvious. Maybe there is an underlying confusion about what ‘IQ’ means.
What MJB said. I love GNXP, but I think that the “g theory of history” is an enormous, implausible overreach.
You have several cases of nations increasing their relative income level enormously (beyond the general increase in wealth in most nations) as the result of changed social forms. Two centuries ago Norway, Sweden, Ireland, and Italy were poor nations, but now they are in the front rank globally. (It is not generally known that Sweden and Norway was still bitterly poor in 1850, but they were). You didn’t have a massive increase in IQ during that period — or i you did, what it means is that “g” is not mostly genetic.
“You have several cases of nations increasing their relative income level enormously (beyond the general increase in wealth in most nations) as the result of changed social forms. Two centuries ago Norway, Sweden, Ireland, and Italy were poor nations, but now they are in the front rank globally. (It is not generally known that Sweden and Norway was still bitterly poor in 1850, but they were). You didn’t have a massive increase in IQ during that period.”
Virtually all countries were bitterly poor by contemporary standards before industrialization. (People generally underestimate just how important industrialization was for per capita incomes)
If anyone halfway reputable (on this site, or elsewhere) has proposed that industrialization was caused by a massive boost to G, please give references. I doubt you will find many.
What is proposed by some people on this site is rather that the pace of industrialization and economic development can be influenced by intelligence, with intelligence as one of many variables.
Lynn argues that two things are required for wealth: a national average IQ above 90 and a free market economy. There will be a time lag, of course, because wealth takes time to accumulate, but the changes will show in annual GDP. Therefore, what Emerson describes as “changed social forms” is entirely consistent with Lynn’s hypothesis. The best example is China: the genetics haven’t changed, but once the government was able to dump Mao in all but name, then the country became (is becoming) wealthy. And, by the way, most Americans were pretty poor in 1850. The health and longevity data in the American interior are similar to that suffered by Africans in the 1960′s.
And one more thing: wealth creates leisure, but I don’t think we have any evidence that it creates IQ. Certainly, giving money to people does not boost their IQ, and countries with petro-dollars or commodity-dollars have not shown a significant IQ boost, though they undoubtably have better social facilities, schools etc.
As I said, Dobeln, these countries changed their relative positions compared to the wealthier nations of the world.
The “g theory of history” has never been carefully stated. It just seems to be assumed. What I said was a reductio ad absurdum; no one actually said it, but it might be deduced from some statements.
From what I’ve seen, hereditary g has not been established as an independent variable yet. And while I realize that Wiki is never authoritative, their claim that the majority of Lynn and Vanhanen’s national IQ figures were extrapolations and estimates, if true, is pretty damning.
In Saudi Arabia you’ve had a financial windfall not accompanied by much positive social change (in terms of improved education, increased political participation, or strengthened rule of law. Usually economic advances are both the cause and effect of a more skilled workforce, but that isn’t what happened in Saudi Arabia. Windfall piles of money don’t teach much.
Wealth presumably boosts mental function at some point by improving nutrition and sanitation, and getting rid of some of the worst overcrowding and disease and such. I think it probably also increases expected IQ score given g, because of more education, familiarity with tests, familiarity with this kind of mental/paper-and-pencil task, etc.
their claim that the majority of Lynn and Vanhanen’s national IQ figures were extrapolations and estimates, if true, is pretty damning.
From J.P. Rushton, in his review of Lynn’s IQ and global inequality:
“they [Lynn & Vanhanen] address the argument made by several critics of the invalidity of the estimates of national IQs from the measured IQs of neighbouring countries. They show that there is a correlation of 0.91 between the estimated IQs for 32 countries given in their first book and the measured IQs for the same countries given in their new book. “
“As I said, Dobeln, these countries changed their relative positions compared to the wealthier nations of the world.”
But the theory is about, as Dobeln said, “that the pace of industrialization and economic development can be influenced by intelligence,” not about relative (historical or not) comparisons. They are not the same. That China is now far wealthier than a few decades earlier, and closer to being the wealthiest country by GDP per person, does not challenge the hypothesis.
” As I said, Dobeln, these countries changed their relative positions compared to the wealthier nations of the world.”
Yes – because they (we) industrialized. There was nothing mysterious about it. As industrialization spread, incomes rose. Not necessarily equally in every country (due to a myriad of factors) which sucessfully industrialized. It has, as far as I know, never been suggested that the primary driver of rising incomes in this time period was an IQ shift.
That of course does not negate IQ as a potential factor in explaining (for instance) remaining income differentials post-industrialization, differential pace of economic development post-industrialization, or divergent pace of industrialization (or lack of industrialization).
“The “g theory of history” has never been carefully stated. It just seems to be assumed. What I said was a reductio ad absurdum; no one actually said it, but it might be deduced from some statements. “
You need to name names for this to have relevance.
As for the actual argument, you can of course apply “reductio ad absurdum” to render any reasoning regarding a complex subject seem silly – I.e. “Oil was important during World War II” becomes, ‘reductio ad absurdum’, “Oil determines the outcome of all wars, always”, etc. That’s not really a fruitful way of discussing these kinds of things.
“From what I’ve seen, hereditary g has not been established as an independent variable yet.”
A good summary on IQ, G, etc:
http://www.gnxp.com/MT2/archives/002366.html
IQ certainly appears to have a very significant hereditary component. G seems to be even more hereditable according to several studies referenced in the article above. I’ll defer to Jason & Co in these matters however.
“Lynn and Vanhanen’s national IQ figures were extrapolations and estimates, if true, is pretty damning.”
While Lynn’s reputation is less than sterling, their metodology has never been a secret. The whole dataset, with sources, from the first book (IQ and the wealth of nations) can be found around the net, I.e.:
http://www.isteve.com/IQ_Table.htm
Finally, the point of this entire post was that L&V scores, despite the flaws in their methodology (extrapolation for some countries, etc.), corresponds closely to results from all major educational attainment studies. (I made a similar post way back on the correlation between L&V scores and the TIMSS) Which is interesting.
Steve Sailer: Or it could be that sub-80 countries don’t have everybody in school…
Yes, the validity question definitely cuts both ways. Also, most of the national test scores from the low-IQ countries come from the oldest test score assessment.
the point of this entire post was that L&V scores, despite the flaws in their methodology (extrapolation for some countries, etc.), corresponds closely to results from all major educational attainment studies.
Test__QHC_logGDP
Indeed, that was the point of the post, along with the fact that the test score data doesn’t really shed light on the question of sub-85 IQ scores.
Here’s the correlation matrix when you drop the bottom 5 IQ scores.
________IQ__
IQ…. 1.00 0.91 0.64 0.67
Test.. 0.91 1.00 0.67 0.68
QHC… 0.64 0.67 1.00 0.89
logGDP 0.67 0.68 0.89 1.00
While the range is now much more restricted, the correlations with QHC and logGDP remain approximately equal (compared to one another) for both IQ and Test.
The fact that guesswork data wasn’t too bad doesn’t justify reporting it. In the medical field they call that “drain testing” — you pour out the specimen and assume that the result was about the same as yesterday’s. (And while that does occasionally happen in hospital labs, it’s rare and it’s a firing offense.)
Dobeln, the link you gave talks about heritability of g within a culture (separated twins, etc.) It doesn’t say anything about cross-cultural comparison of IQ (a separate problem) or about average national IQ as an important independent factor in history.
To clarify my earlier comment: I didn’t mean to suggest that wealth causes intelligence. I’m saying that the mental behavior that accompanies wealth creation in most societies has a causal effect on ‘IQ’, as does genetics (such as might correlate with race), as does formal education, and these different factors have not been disentangled. So any signficant conclusions of causality that might be drawn based on the observed correlations would likely be fallacious.
John, there is some concern that L&V’s numbers are attenuated by Lynn’s unreliability with transcription and calculation, but they’re certainly not guesswork of the “drain test” variety, nor do they seem to lack validity (at least for the scores in the top half of the distribution).
In IQGI, L&V report on scores from 25 countries which were merely estimated in IQatWoN, but have measured values in IQGI. The correlation between the estimate and the measure is r=.91. Notably, this is the same correlation between IQ and Test that I found.
Here is the data table:
Estimated.IQ Measured.IQ
85. 87.
68. 64.
93. 90.
75. 67.
84. 82.
97. 99.
84. 81.
98. 101.
87. 84.
83. 86.
89. 89.
97. 91.
79. 82.
95. 97.
81. 89.
72. 64.
81. 84.
84. 83.
81. 79.
75. 62.
75. 71.
87. 83.
89. 84.
96. 94.
83. 85.
“The fact that guesswork data wasn’t too bad doesn’t justify reporting it. In the medical field they call that “drain testing” — you pour out the specimen and assume that the result was about the same as yesterday’s.”
The medical field has very stringent standards for a reason – otherwise people die.
When you are trying to put together a data set for the first time ever in a subject where lives aren’t on the line, on the other hand, cutting slack is usually in order. Especially when the authors are open about their methodology.
“It doesn’t say anything about cross-cultural comparison of IQ (a separate problem) or about average national IQ as an important independent factor in history.”
It is certainly highly biologically heridatable as well as the most potent sociological variable around. If that doesn’t qualify it as an independent and heridatable variable, I don’t know what would. Jason can do a data dump here when he shows up.
Now, it is true that cross-cultural IQ measurement is a different subject, but that’s not really the point I was trying to adress with the link. Again, though, it is rather interesting that a measurement that many claim breaks down in cross-cultural comparison is highly correlated to cross-cultural economic development, as well as educational progress.
If indeed IQ had no validity in cross-cultural comparisons (Despite, empirically, many asian counties doing cross-culturally well before taking off economically, etc.), wouldn’t we rather expect those correlations to be a big fat zero, rather than on blockbuster levels in social science terms?
(I.e. compare Sala-i-Martins regressions, variables such as School spending, etc.) There are plenty of caveats, as usual, – but I hardly find much basis for such a dismissive stance towards the approach.
LWKA: “drain testing” can be accurate — a lot of patients have daily tests which don’t vary much. But reporting non-data as data is extremely troubling. The guy had a data set to report. Why didn’t he report it? Reporting the data you have is the default.
Perhaps stating the obvious: This study tells us that ‘IQ’ and other similar test scores correlate with wealth, but implies nothing about causality.
This is correct. Causality is its own argument and set of data points. Data points are not analyzed in isolation.
The study will be considered by some to imply that differences in innate intelligence cause the differences in wealth, and by others that differences in education cause the differences in wealth, depending on his or her politics.
We can test education theory, since inputs (education) are not the same as outputs (intelligence). For instance most Americans graduate from 12 years of schooling, and yet have dramatically different intelligence levels. Education and educational spending is not associated with intelligence within nations. Jones and Schneider (2006) found that barely one educational variable had even a weak effect on crossnational growth 1960-1992, while IQ had one of the strongest effects of any variable. If IQ is a “proxy”, then a) why aren’t other variables making the same predictions, b) what is this proxy?
since to a certain extent high IQ reflects how practiced a person is in certain mental activities which are possible in a wealthy society
Citation please?
Seems to me that in a sense the study doesn’t really say anything that’s not already trivially obvious.
Actually it says a lot of things you obviously wouldn’t predict. Like the fact that there are people in very poor countries right now, who are much smarter than people in rich countries. (e.g. China) When predicting future wealth, people like you bet on Ghana and the Philippines over Korea in the 1950s, because the variable of importance was not “trivially obvious”.
Two centuries ago Norway, Sweden, Ireland, and Italy were poor nations, but now they are in the front rank globally.
John do you have any serious quantitative additions or counterarugments to this post? A specific data point/topic (validity of comparison) was posted and you derailed about its ‘g theory of history’ subtext. Not helpful.
And while I realize that Wiki is never authoritative, their claim that the majority of Lynn and Vanhanen’s national IQ figures were extrapolations and estimates, if true, is pretty damning.
Since the conclusions were not predicated on the estimated nations I fail to see what’s ‘damning’.
113 countries is a good and informative sample size, and covers most nations of significant population size. Lynn and Vanhanen correlated IQ-GDP with both their limited sample/actual data, and complete sample including estimated data. (the correlations were close but smaller for the complete sample – so its not like it was used sneakily or in their favor)
The estimation method was also sound as shown above.
I’m not talking specifically about this post, but about the general argument of which it is often part here. No one else is being prevented from discussing the statistical specifics. The serious doubts I have about comparing statistics between nations are relevant to the post and not limited to me.
Is there no one else here who is skeptical of someone who reports conjectural and extrapolated statistics? I’m far from a statistical whiz, but that’s on the short list of statistical errors, and it is not at all confidence inspiring.
The “Wealth of Nations” question is hundreds of years old and it has generated tons of literature. All the evidence is that it is a complex multi-factor process, and the best explanations so far have to do with social organization (free market, rule of law, no parasitical ruling class, work ethic.) For IQ or g to be recognized as one of the major factors, I think that there will have to be a lot better argument than I’ve seen so far, and it will have to be integrated into the overall argument.
Jason, drain testing is not sound even when the results reported are OK.
John, the “drain testing” analogy is lost on me. What in the case of L&Vs data or methods is analogous to “drain testing”? The estimated IQ scores?
“I’m not talking specifically about this post, but about the general argument of which it is often part here.”
Then producing quotes should be pretty easy.
“Is there no one else here who is skeptical of someone who reports conjectural and extrapolated statistics?”
The original sources are most certainly not “conjectural and extrapolated” and are easily checked. (relevant link above)
If the method of “conjecture and extrapolation” (I.e. for filling in the blanks) is out in the open, and results both with and without the filled-in blanks are shown, there is certainly no “statistical error” being committed.
Trying to work around less-than-perfect data is not a mortal sin, as long as you are open about your methods.
“The “Wealth of Nations” question is hundreds of years old and it has generated tons of literature. All the evidence is that it is a complex multi-factor process”
Agree so far. The L&V dataset is connected with a subset of the issue, however – contemporary wealth.
“and the best explanations so far have to do with social organization (free market, rule of law, no parasitical ruling class, work ethic.)”
All fine and dandy, but why shouldn’t IQ be on that multi-factor list? I’ve read a fair chunk of the growth litterature, and national IQ is right up there with “material investment” as an explanatory variable in the regressions. (It is surprisingly difficult to find reliable explanatory variables for growth. Which is part of the reason why the L&V dataset is so interesting.)
The “g theory of history” has never been carefully stated. It just seems to be assumed.
No. There is no “g theory of history”. All that is being argued is that there is a strong correlation between averagfe IQ and per capita income in the modern world. No one is claiming that such a correlation existed throughout history. In fact, it clearly did not, and it’s not that hard to think of reasons why it shouldn’t have.
>>since to a certain extent high IQ reflects how practiced a person is in certain mental activities which are possible in a wealthy society
>Citation please?
This should be clear to any educated person who has taken an IQ test and examined their own thought process.
>>Seems to me that in a sense the study doesn’t really say anything that’s not already trivially obvious.
>Actually it says a lot of things you obviously wouldn’t predict. Like the fact that there are people in very poor countries right now, who are much smarter than people in rich countries.
Well, you are not at all understanding my point then, even though I posted an attempted clarification. I would definately predict what you say I ‘obviously wouldn’t predict’, in fact I would be shocked if it weren’t true.
If a study or citation is needed to show that smart people can be poor, are that measured IQ tends to improve with practice, then I admit I was unfairly dismissive of the study. And I was unreasonably rude in any case, so I apologize for that.
But in some of the above postings, including yours, I still see A implies B confounded with B implies A. Nowhere did I ever say that race or other factors are not a cause of IQ differences. And furthermore implication seems to be confused with causality in a couple of postings, though not necessarily in yours. So it seems that the doubt I expressed about people jumping to conclusions about causality was well founded. And as I indicated, if you need a statistical analysis to establish what is obvious from considering how a system works, then have fun with it.
“When predicting future wealth, people like you bet on Ghana and the Philippines over Korea”
First, what does “people like you” mean here.
Second, I don’t know of anyone who bet on Ghana over Korea. People were more hopeful about Ghana than they should have been, but Ghana had a very poor institutional model (African socialism).
are that measured IQ tends
^^^
‘or’
I am willing to grant the strong correlation. I don’t think that that’s all that’s being claimed.
In the end, IQ might turn out to be an important independent variable. There will have to be a stronger and more complex argument than I’ve seen so far, though.
>First, what does “people like you” mean here.
He’s an expert in studying human intelligence, and he obviously has concluded that we must be idiots :(
“I am willing to grant the strong correlation. I don’t think that that’s all that’s being claimed.”
Correct – there is strong correlation and several plausible potential causal mechanisms for why IQ should impact GDP.
“He’s an expert in studying human intelligence, and he obviously has concluded that we must be idiots :(“
I’m not Jason, but I’ll take a shot: “Like you” refers to those who, like you, are very sceptical of IQ as an important factor behind contemporary growth and economic achievement.
(It is a reference to a Jared Diamond quote: “. . . around 1950, when South Korea, Ghana and the Philippines were equally poor, most economists predicted that resource-rich Ghana and the Philippines were on the verge of wealth, whereas South Korea was doomed to remain mired in poverty. The result, of course, has been the opposite . . .”
It was in the link right after “like you”.)
“This should be clear to any educated person who has taken an IQ test and examined their own thought process.”
Trying to rotate little squares in your head, etc. is only possible in a wealthy society?
Jason, drain testing is not sound even when the results reported are OK.
I’m not clear on what the error is. If another researcher had made the IQ estimates using a method of similar validity (0.91), and L+V used his independent data in the same way, would you still consider it invalid?
If Diamond was right about that, it means that most economists overestimated resources and underestimated trade, infrastructure, and labor productivity. economists haven’t talked that way for at least 3 or 4 decades.
OK, let me start over. Coastal China is booming and interior China is not, and it’s a fairly sharp gradient. There are simple reasons for that having to do with transport (and China’s wealth now is dependent more on trade than on the internal economy). On the one hand, no one would have anything manufactured in Sichuan if they could have it manufactured nearer the ports. On the other hand, Szchuan products have to go through the ports anyway to get out to the world, and will contribute to the coastal economy and the Szchuan economy both.
Now, Sichuanese and Shanghainese are ethnically very similar, as far as I know it’s not claimed that IQ is a factor here, except insofar as IQ results from prosperity (which is not the main claim here).
One large batch of very poor nations is in the Sichuanese (landlocked) category, and they’re worse off than Sichuan because they are not part of a nation with a coastal outlet at all (Sichuan gets a benefit from being part of the same nation as Shanghai.) Examples would be Paraguay, Bolivia, Laos, Mali, Chad, and perhaps some of the Central Asian countries (Kyrgyzstan and Tajikstan). Most of these also have unfavorable institutional and educational histories, and none of them are adjacent to thriving countries where they could get a spinoff (the way Mexico does from the US, for example).
Some of the countries I named might be a bit more or less prosperous than they might be because of IQ, but the other factors I named (plus a lack of resources in most cases; resources aren’t sufficient but they do help) make it highly unlikely that any of them could be a lot more prosperous than they are. I can’t see IQ being more than a nudge.
Singapore is a big success and its people are very sharp, but a landlocked Singapore in the Andes with equally sharp people would not do nearly as well.
GNXP tends to be a genetics-and-IQ forum, and I’m not terribly sharp about those topics. I’m closer to the GNXP norm now than I was when I first came, but I’ll always be a devil’s advocate here on those questions.
The reason I’m here, though, is that GNXP is also a history and politics forum, and while a lot of the stuff here is extremely interesting and valuable, I also often feel that the historical argumentation needs a lot of improvement.
I’m going to drop the “drain testing” point because it’s somewhat irrelevant. The guy had a data set to analyze and report out, and even if he was explicit about what he was doing, I don’t see the scientific value or validity of also reporting out the extrapolated data, which added no information to his actual data. This does not strike me as normal scientific practice at all. (There are 5 pairs where the discrepancy is 5 points or more, though given that everything is just averaged into the total that may not be significant.)
There’s a clear relationship between intelligence and economic growth. Growth is, in a sense, entrepreneurs finding better ways to satisfy consumer desires. In general, smarter entrepreneurs will have more success in anticipating desires or transforming productive processes than dumber entrepreneurs — speaking on average of course.
The only interesting issue is whether intelligence actually does vary among countries. To the extent that it does, it is highly likely that such variation affects economic performance. Which is not to say that other factors (such as natural resource endowment, legal institutions, cultural mores) may not play an equal or larger role at particular points in time.
“Like you” refers to those who, like you, are very sceptical of IQ as an important factor behind contemporary growth and economic achievement.
I am not skeptical about IQ being an important factor behind growth and economic acheivement.
I am skeptical about what the presented correlations imply, because of some of the factors that IQ depends on. I’m not even skeptical that inherited, genetically based IQ differences can be the cause of differences in economic acheivement. But I also think we should be fair about what data does and does not suggest.
Reminds me of something written relatively recently by Murray, of Bell Curve fame. He was arguing that cultural differences can not account for IQ differences. As an example, he spoke of a test where the subject was required to repeat a sequence of letters or numbers backwards, and claimed that there was absolutely no way culture could impact that outcome. Does he have no imagination? The test depends highly on concentration. How can not giving an F- about jumping through white man’s IQ test hoops not impact the outcome? Heck, by that standard, my IQ probably dropped several points last summer after I was arrested and prosecuted for a property crime I didn’t commit. Yes, there are statistically significant differences between races. But I don’t trust what people like Murray are really up to.
A similar fallacy is at play when conservatives tut tut about black and hispanic out-of-wedlock birth rates. Blacks and hispanics, being less invested in America’s system, are naturally less inclined to file for a slip of paper representing the government’s blessing of their copulation. This is not the same as how well they are willing to take care of their children. Absent fathers is a big problem, but that’s a different statistic. And cultural alienation is a problem, but that’s not entirely the same either. Certainly, a lot of the problems experienced by black Americans are caused by black American culture, and by genetics. And its important to point this out. But it still seems to me that people are using this dishonestly and unfairly.
(I’m not saying that anyone who has posted at this site is guilty of any of this.)
And now for some comedic relief. Doreen Kimura, in an article on Recollections of an accidental contrarian, has this to say:
Table 1 is a summary, not exhaustive, of sex differences reported in the literature. I am not going to elaborate on these since they are well known to almost everyone, except perhaps the faculty at Harvard. (Emphasis in the original.)
>”This should be clear to any educated person who has taken an IQ test and examined their own thought process.”
>Trying to rotate little squares in your head, etc. is only possible in a wealthy society?
Well, this exactly illustrates the kind of BS that inspired my initial post. Nobody said it is possible only in a wealthy society. But it can be a lot easier if you’ve grown up playing with legos, and if you have received much positive reinforement from parents and teachers about skills that are useful in a vibrant engineering job market.
John, by which measure would Sweden have been “bitterly poor” in 1850? I don’t have the patience to Google right now, but IQ and the Wealth of Nations happens to have some GDP per capita data for 1820 and 1900 (they refer to some book which probably would have data for 1850). In 1820, Swedish per capita GDP was 93 % of US GDP; in 1900, it was 63 % of US GDP. If Sweden was “bitterly poor”, then so was the United States. It just doesn’t make sense to consider Sweden a poor country for the era. For comparison, if you look at this Wikipedia list…
tries_by_GDP_%28PPP%29_per_capita
http://en.wikipedia.org/wiki/List_of_coun
…and put the limit of “bitterly poor” at 63 % of US GDP, you’ll end up classifying New Zealand as bitterly poor and New Zealanders don’t seem that bitter to me. Sweden was one of the wealthiest countries in the world at the start of the 19th century; the English-speaking countries had a head start in industrialization and were a bit ahead for some time (and the US still is) and that’s it. Sweden has been one of the really rich countries for a very long time. (Naturally, I have to point out this, because I’m from a formerly poor and still bitter neighbour and it mortally offends me that it’s still not known around the world that Swedes are born with silver spoons up their asses.)
Of course, between 1820 and 1900 the United States wasn’t the wealthiest major country in the world; a better measure might be a proportion of UK GDP. For 1820, here are some hand-picked examples for the list
US 73 %
Sweden 68 %
Italy 62 %
Norway 57 %
Ireland 54 %
Mexico 43 %
Finland 43 %
Russia 43 %
Japan 40 %
Indonesia 35 %
India 30 %
Of your list, even when measured this way, Italy and Sweden are today at a POORER relative position than in the early-mid 19th century: in the full lists, there would now be more countries ahead of them than there used to be. This includes a couple of risers who were a bit poorer than them in the early-mid 19th century (eg. Norway, Ireland) but hardly “bitterly poor” in relative terms (the Irish reputation for special poverty must have something to do with Ireland getting so easily compared to the UK and the US) and a couple of risers that really were only barely ahead of the poorest of the poor in the early-mid 19th century (eg. Finland, Japan) (oh, yeah, and Qatar…).
Jaakeli, I don’t have my source. A history of Sweden assured me that Sweden has only become prosperous since 1900, and before then was on a level with Poland’s or Ireland’s. It was a standard text AFAIK but I don’t have it any more.
If those stats are good, in 1820 Sweden was halfway between England and Indonesia, which strikes me as pretty poor.
I said that because people tend to assume that Sweden has always been prosperous. (Except for freemarketers who think that Sweden is a Third World country, of course.)
As an example, he spoke of a test where the subject was required to repeat a sequence of letters or numbers backwards, and claimed that there was absolutely no way culture could impact that outcome. Does he have no imagination? The test depends highly on concentration. How can not giving an F- about jumping through white man’s IQ test hoops not impact the outcome?
No time to comment at length now, but this is one of the subtests on which blacks do best relative to whites.
“Some of the countries I named might be a bit more or less prosperous than they might be because of IQ, but the other factors I named (plus a lack of resources in most cases; resources aren’t sufficient but they do help) make it highly unlikely that any of them could be a lot more prosperous than they are. I can’t see IQ being more than a nudge.”
If you are going to be really sensitive to others jumping to conclusions, oversimplifying, etc. perhaps it would be best to refrain from then immediately drawing up a grand, largely unsupported model, even specifying effect sizes (a “nudge”!), on really complicated subjects where the academic consensus can roughly be summarized as “boy, this is hard!”? Or am I missing something here?
“Jaakeli, I don’t have my source. A history of Sweden assured me that Sweden has only become prosperous since 1900, and before then was on a level with Poland’s or Ireland’s. “
Pretty much everyone was dirt poor on a per-capita basis before they got hit by industrialization. Sweden only started really taking off by the end of the 19:th century, as industrialization spread and industrialization elsewhere created demand for raw materials (steel and timber, mostly).
“Well, this exactly illustrates the kind of BS that inspired my initial post. Nobody said it is possible only in a wealthy society. But it can be a lot easier if you’ve grown up playing with legos, and if you have received much positive reinforement from parents and teachers about skills that are useful in a vibrant engineering job market.”
I assume this is stuff that “you just know”? I can also make stuff up. First effort: “mental rotation is much easier if you’re a third-world kid out playing with real objects all day, instead of staring at a brain-rotting 2D TV”.
“A similar fallacy is at play when conservatives tut tut about black and hispanic out-of-wedlock birth rates. Blacks and hispanics, being less invested in America’s system, are naturally less inclined to file for a slip of paper representing the government’s blessing of their copulation. This is not the same as how well they are willing to take care of their children. Absent fathers is a big problem, but that’s a different statistic.”
So, you are betting that the variable “out-of-wedlock births” and the variable “missing fathers” are uncorrelated for African americans and hispanics? If so, that’s one bet I’ll take you up on.
“GNXP tends to be a genetics-and-IQ forum, and I’m not terribly sharp about those topics. I’m closer to the GNXP norm now than I was when I first came, but I’ll always be a devil’s advocate here on those questions. “
Hey, makes for more interesting discussion.
hey guys, wow, a lot of comments quickly. let’s keep this focused and value added (seems OK right now). no point in wasting all of our time in pointless arguing for the sake of arguing.
>I assume this is stuff that “you just know”? I can also make stuff up. First effort: “mental rotation is much easier if you’re a third-world kid out playing with real objects all day, instead of staring at a brain-rotting 2D TV”.
Yes, that would be another possible example illustrating my point. My original claim was, and remains, that the causal relationships are not quantified. I am not trying to argue that higher scores in wealthy countries can in some proportion be attributed to wealth, or to anything else. Play is good for IQ, and may very easily be better than TV watching, but its just one factor among many.
” hey guys, wow, a lot of comments quickly. let’s keep this focused and value added (seems OK right now). no point in wasting all of our time in pointless arguing for the sake of arguing.”
Well, the thread seems to be reasonably on topic so far, so no need to sound the alarm. :)
if we can rise about the vituperative back and forth for a minute, i’ll try to draw attention back to the point of the post.
at least through the range of 85-105, national IQ scores appear to be a strongly valid measure of national average cognitive ability. scores below 85 are an open question in relation to this data set, but the only real question appears to be: how much below 85 would a suitably valid measure of IQ place these countries?
thus, when a scholar writes we are aware of Lynn’s work and … we “consider it so scientifically flawed as to be not worth taking seriously”, I have to argue that this data calls into question the grounds on which such a conclusion can be reached.
Dobeln, there are a lot of well-known factors impacting economic development and wealth. This is not a new topic, and there’s been a lot of progress in the last 50 years.
As far as I know, if you asked anyone in that area of study the conditions under which Bolivia, Paraguay, Laos, Mali, Chad, Kyrgyzstan, or Tajikstan could be wealthy nations, they would have trouble thinking of them. The only exception I can think of is that if Russia were prosperous, Kyrgyzstan and Tajikstan would be more prosperous than they are. Those countries have a lot of odds against them. (An oil strike would help, but recent history has shown that without a polticial infrastructure, etc., resource windfalls don’t help much.)
In other words, while there is no adequate model of economic development saying for sure why Portugal isn’t doing as well as Italy, for example, the areas is not a big unfathonable mystery.
All of the countries I named except Paraguay (which is at the top of the bottom half) are in the bottom quarter of the poorest countries. There are plenty of explanations why. Throwing IQ into the pot wouldn’t explain things better.
>So, you are betting that the variable “out-of-wedlock births” and the variable “missing fathers” are uncorrelated for African americans and hispanics? If so, that’s one bet I’ll take you up on.
Woa. Of course they are strongly correlated. But the statistics are used in argument as if the correlation is close to 100%. I’ll take Chet’s advice. I’m out of here.
“Yes, that would be another possible example illustrating my point. My original claim was, and remains, that the causal relationships are not quantified.”
Okies, I interpreted this exchange:
—since to a certain extent high IQ reflects how practiced a person is in certain mental activities which are possible in a wealthy society.
–Citation please?
-This should be clear to any educated person who has taken an IQ test and examined their own thought process.
in a stricter fashion.
jesus, at the rate u guyz are commenting you should take this to a chat room ;-)
“Dobeln, there are a lot of well-known factors impacting economic development and wealth. This is not a new topic, and there’s been a lot of progress in the last 50 years.”
bstract_id=31213#PaperDownload
Most of the well-known factors are lightweight. For a rundown a good place to start is:
http://papers.ssrn.com/sol3/papers.cfm?a
And that’s the “optimistic” view on empirical growth theory!
Sure, factors like efficient and stable institutions are good for growth. But how do you get them? It’s not terribly useful for understanding why some countries grow and other don’t.
Also note that most variables that are strongly related to growth are regional and religious dummy variables. (Often virtually the same thing)
The only strong practical conclusion is that some degree of market economics is good for growth. That’s not really a whooper these days. It’s really a field that could use some strong correlations, credible and very basic possible mechanisms, etc. being brought into the discussion.
I’m going to drop the “drain testing” point because it’s somewhat irrelevant.
bstract_id=552481
Ah, so it was not, in fact, ‘damning’, after all. It just didn’t add much now. That is a big difference.
All the evidence is that it is a complex multi-factor process, and the best explanations so far have to do with social organization (free market, rule of law, no parasitical ruling class, work ethic.) For IQ or g to be recognized as one of the major factors, I think that there will have to be a lot better argument than I’ve seen so far
In the end, IQ might turn out to be an important independent variable. There will have to be a stronger and more complex argument than I’ve seen so far, though.
John, it takes a lot of arrogance to dismiss a topic you admit you know nothing about. You act like this is a matter of taste, when in fact it is a quantitative question, and the numbers have already been crunched. Namely Jones & Schneider have already checked the robustness of this variable against the small handful of variables demonstrated to predict growth:
Finally, for an overall assessment of how IQ compares to other common growth variables, consider Sala-i-Martin?s original results, which used combinations of 62 growth variables in over two million regressions. Among his top 21 regressors?the ones which he considered robust?the median regressor was statistically significant in 76.4% of cases, with a range from 100% (for fraction Confucian) to 2.81% (for revolutions and coups). Fraction Confucian was the only regressor that passed an extreme bounds test at the 5% level. Only eight of his top 21 had coefficients over three standard errors from zero, while in our full-sample results using his top 21 growth variables, IQ?s coefficient is over four standard errors away from zero. IQ would thus appear to fit comfortably in the top half of Sala-i-Martin?s top 21 growth variables.
http://papers.ssrn.com/sol3/papers.cfm?a
Where are your numbers?
FWIW, here are S-i-M’s top 8:
_________Fraction Significance
_____________100.00% ____________92.04% ____________88.67% _____________33.37% ______88.26%
bstract_id=31213
bstract_id=552481
Tested Variable____
1. Equipment Investment____________99.97%
2. Number of Years Open Economy_____99.97%
3. Fraction Confucian___
4. Rule of Law_________
5. Fraction Muslim______
6. Political Rights______
7. Latin America Dummy_______
8. Sub-Saharan Africa Dummy________76.38%
http://papers.ssrn.com/sol3/papers.cfm?a
cf: “Out of these 1330 regressions, IQ is statistically significant at the 95% level in 99.8% of the regressions, and positive in all regressions.”
http://papers.ssrn.com/sol3/papers.cfm?a
7. Latin America Dummy
8. Sub-Saharan Africa Dummy
lest anyone get the wrong idea, maybe you should clarify that those are dummy variables, not a comment on anyone’s intellectual capabilites :)
If those stats are good, in 1820 Sweden was halfway between England and Indonesia, which strikes me as pretty poor.
Almost EVERY OTHER COUNTRY ON THE PLANET was at least that far behind the UK back then. The ONLY exceptions are the Netherlands and Australia who were at slightly less than 90 % of UK GDP. There was the UK, then those two trailing countries and a wealthy club at about 65 +- 10 % of UK GDP; then a huge gap with just a few countries followed by a huge mass of countries at 30-45 % of UK GDP (and nothing after that, no one was that poor). That is, the distribution of GDP per capita is rather similar to what it is today: there’s a leader with a few small countries at a similar level or even a bit ahead, followed by a mass of countries a decade’s or two decade’s worth of economic growth behind the leader, a few random countries in random phases and then the poorest of the poor who are stuck at similar GDPs right there on the bottom (any country that isn’t in mass starvation has to have some minimal level of economy up).
Note, BTW, that today in GDP “half-way between the UK and Indonesia” means slightly behind South Korea and Portugal. Not rich, but not bitterly poor, either.
The point is, the countries in those groups are remarkably similar. The leader has changed from the UK to the US, but the UK has only dropped to the club of the moderately wealthy. Considering Ireland as the poorest of the rich makes the list of the rich back then Australia, Austria, Belgium, Denmark, France, Germany, Ireland, Italy, the Netherlands, Norway, Spain, Sweden, the UK and the US. (L & V only have data for some countries, there are many missing that probably would be there and not all of these regions were actual countries back then, of course.) Sounds familiar? Taking Mexico as the richest of the poor makes the list of poor countries to be Brazil, China, Finland, India, Indonesia, Japan, Mexico and Russia. (There are fewer of these, presumably because there aren’t good estimates for most undeveloped countries.)
The list is, more or less, exactly the same today. Ireland has moved from being the poorest of the rich to being one of the really rich. With those limits, exactly two countries have made a transition from being relatively poor to being among the really rich in the past two centuries: Finland and Japan. They don’t have much data for Eastern Europe, but countries there have certainly dropped in relative ranking during communism (Finland was pretty typical for a place in the Russian Empire at the beginning of the 19th century and has nearly exactly the same GDP as the rest of Russia in this data, too; the current GDP difference to the rest of the ex-empire comes from the earlier industrialization in Finland (that’s due to more liberal politics) and the century/half-century of communism that froze development there).
As for Sweden being “like Poland or Ireland”, there are two points here: you’re still not getting that by global standards the 19th century Irish were not poor!!! so yeah, the Swedes were like them, and even considering that, Sweden was significantly ahead of them (again, Sweden has been one of the really wealthy countries for, like, forever, especially if you cut out what’s now Finland from pre-19th century data). The reason the 19th century Irish seem so very poor to you is that you’re comparing them only to what were the wealthiest countries on the planet (and because 19th century life in any country sucked ass compared to life today, even if it was in a relatively well-off country like Ireland).
Jaakelli, conditional on the accuracy of the data you’ve cited, I grant your point about Sweden and Ireland.
Jason, I wasn’t conceding anything — if you won’t drop the point, I won’t. Reporting conjectural and extrapolated data which adds no information to the actual data is bad science. Just report and analyze the data you have — as far as I know, this isn’t a revolutionary, avant-garde, pomo idea. And because the two authors reported fictional data, I don’t trust them.
Jason, is IQ a dependent variable or an independent variable? Cause or effect?
And please explain to me how the nations I named would have become wealthy if they had had smarter populations. Those nations were not going to become wealthy. Basically I was saying why trade and resource geography plus history was sufficient to cause poverty in these particular cases, regardless of IQ.
And please don’t tell me I don’t know what I’m talking about.
John, if you don’t concede then please answer this question: If another researcher had made the IQ estimates using a method of similar validity (0.91), and L+V used his independent data in the same way, would you still consider it invalid?
And please explain to me how the nations I named would have become wealthy if they had had smarter populations.
Ok. You asserted that countries like Paraguay, Bolivia, Laos, Mali, Chad are poor because they are landlocked and not “adjacent to thriving countries”, and that Singapore would be just another poor country in the middle of the Andes. Meanwhile I linked to a paper using quantitative models that did not consider these variables relevant, and another finding IQ extremely relevant. Here is another econ paper using more evidence than verbal assertion:
“In summary, our results suggest that most of the new variables that have been introduced into growth regressions in the 1990s survive a rigorous test against alternative models. The ones that do not (landlockedness, growth of neighbouring countries) are arguably those with the weakest theoretical basis.”
You stated that “For IQ or g to be recognized as one of the major factors [in economic growth], I think that there will have to be a lot better argument than I’ve seen so far”. I provided research that tests this and finds it more important than all the variables you listed. You have linked no research, no data, no numbers. This feels very one sided.
Jason, I would not accept any drain testing at all as valid, regardless of how accurate it is. Most drain testing is valid, because it’s normally done to fill in blanks when the same test is repeated over and over on the same patient with similiar results. Is there something radical and difficult about the idea that you just report the data you have?
I was not asking you to link to a statistics paper. I was asking you to tell me in which conditions the actual countries I named would have become rich. I was asking a path-dependent historical question and asking you to describe an alternative path leading to a distinctly different outcome. I was asking you to function as a historian.
So tell me — how would it happen if you transferred Singapore into the Andes, it would remain wealthy? What specifically would the new Andean Singapore become wealthy on?
Note that the least poor of these countries is adjacent to Brazil and Argentina, which by Latin American standards are economic powerhouses.
Note also that Switzerland is a landlocked country, but in the center of Europe — Switzerland has been a trade center since the Roman empire. “Landlockedness” is not an adequate description of what I was talking about.
I clicked your link and did not find during what time period “growth” was measured.
John, I’ll give it a shot. If these countries had such smart people, they would negotiate free trade treaties with neighboring countries. Illinois is landlocked, but not particularly poor.
I’m not exactly clear if your arguing that landlocked regions will be poorer than coastal regions (disproved by comparing the Midwest to Lagos)or, that for a given IQ, a landlocked region will be poorer than a coastal region.
I just looked at an atlas. The only prosperous landlocked nations were Switzerland and Austria. Switzerland is surrounded by prosperous nations. Austria borders several semi-prosperous nations and some prosperous nations. Both nations have been part of the relatively developed world (Roman empire onward) for about 2000 years, and part of the most developed part of the world for 500 years.
It’s a fact that most of Africa is miserably poor, Latin America runs from moderately poor to poor, and Asia runs from wealthy to fairly wealthy to moderately poor but developing to poor but developing to poor and not developing. Europe and N. America are wealthy, the European Soviet bloc is moderately poor to poor. There are various ways this can be explained, but I think that a historical explanation is better than one based in a sophisticated attempt to do a statistical analysis of selected variables, because any selection of variables will be tendentious. The stats are presumably good, but the selection of variables, as far as I can tell, was common-sensical, and you the variables you choose determines the results to a large degree.
Illinois is not a nation. It is also not landlocked (St. Lawrence Seaway). It is also part of a nation with an interstate highway system suitable for trucking freight.
Rob’s post brings up another point: all nations live under conditions imposed by other nations. Paraguay and Bolivia are dependent on the decisions of their neighbors. Even maritime nations have to live under the conditions of the international system in place.
tries_by_GDP_(PPP)_per_capita
Nations are not a good unit of measure if IQ is what you’re looking at. There are tiny quasi-nations like Luxemborg (and many much smaller), and enormous continent-sized nations like China and India. It would be more persuasive to deal in geographical population chunks of 10 million or so scattered randomly over the globe.
http://en.wikipedia.org/wiki/List_of_coun
“Nations are not a good unit of measure if IQ is what you’re looking at. There are tiny quasi-nations like Luxemborg (and many much smaller), and enormous continent-sized nations like China and India. It would be more persuasive to deal in geographical population chunks of 10 million or so scattered randomly over the globe.”
Erm, excuse me? Belonging to a certain nation is *one* very important variable that can impact growth heavily. Case in point: The Korean peninsula – a nearly perfect, albeit somewhat extreme natural experiment in just how important which political entity you belong to can be.
That shouldn’t prevent us from looking at other subdivisions, of course, such as regional data (often of worse quality, however), and ethnicities across countries, etc. But there really isn’t any fundamental obstacle(aside from resources, etc.) to taking multiple approaches to this issue.
“Jason, I wasn’t conceding anything — if you won’t drop the point, I won’t. Reporting conjectural and extrapolated data which adds no information to the actual data is bad science. Just report and analyze the data you have — as far as I know, this isn’t a revolutionary, avant-garde, pomo idea. And because the two authors reported fictional data, I don’t trust them.”
They did report and analyze the data they had, in addition to trying to fill in blanks. As stated above, both approaches yielded very similar results.
Second, do you classify all (non-clandestine) attempts at filling in gaps by using related data to extrapolate new data points as “fictional” and “bad science”? Or should we ask you on a case-by-case basis? And why should trust suffer when they were open about what they did?
“I was not asking you to link to a statistics paper. I was asking you to tell me in which conditions the actual countries I named would have become rich. “
Ok, so in short he should make up a data-less just-so-story, actual data be damned? Where did your indignated demands for super-stringency and the utmost respect for data go all of a sudden? Hell, I could make up excellent stories as for why Sweden should be dirt poor. (It’s a vast, barren country, with a hostile climate, short summer, removed from traditional major European trade centers, etc.)
“Note also that Switzerland is a landlocked country, but in the center of Europe — Switzerland has been a trade center since the Roman empire. “Landlockedness” is not an adequate description of what I was talking about. “
In short, “landlockedness” means “landlockedness” unless you want it to mean something else, and in that case the definition shifts to “landlockedness minus traditional trade center role”, or “landlockedness plus decent infrastructure”.
Nothing wrong with taking multiple variables into account, but discussing this on a case-by-case basis will rapidly become pretty useless.
“So tell me — how would it happen if you transferred Singapore into the Andes, it would remain wealthy? What specifically would the new Andean Singapore become wealthy on?”
I dunno – the Finns are wealthy in part because they are the home of the largest mobile phone maker in the world. The Japanese turned an overpopulated, resource-starved island into an economic great power, etc. – the span of possible development paths is so great that you need actual data to discuss these things fruitfully, not just stories made up on the fly.
“It would be more persuasive to deal in geographical population chunks of 10 million or so scattered randomly over the globe.”
Yea, let’s throw out every single existing dataset and create a new and incredibly expensive and difficult-to-gather one. Excellent idea.
“As an example, he spoke of a test where the subject was required to repeat a sequence of letters or numbers backwards, and claimed that there was absolutely no way culture could impact that outcome. Does he have no imagination? The test depends highly on concentration. How can not giving an F- about jumping through white man’s IQ test hoops not impact the outcome?
No time to comment at length now, but this is one of the subtests on which blacks do best relative to whites.”
.
Could you please elucidate…what I recall reading was that blacks and whites did about the same in recalling numbers going forwards. It was in recalling them BACKWARDS that the discrepancy showed up, blacks not performing as well.
My recall is that the obviously culturally-dependent aspects of IQ tests, such as vocablulary, identifying well known people or historical incidents, have generally been shown to be much less race dependent, blacks and whites of similar education performing more similarly. It is on the more abstract, “culturally” independent areas that more differences show.
But that brings up a question for me: the FORMAT of IQ testing is certainly western and “white” although Asians seem to have no problem with them. But what is being “measured” is not singular to western, modern countries. People in undeveloped countries certainly need the kinds of judgment skills being assessed in IQ tests, and I understand that non-written, non-read, tests are used. I have read that people in undeveloped countries who do especially well in such IQ tests, are indeed people who are considered exceptionally intelligent by their own standards, within their own culture. The concept of intelligence transcends cultures, and is not even all that different. It’s just that the chance or necessity to use “g” differs according to the environment.
). The reason the 19th century Irish seem so very poor to you is that you’re comparing them only to what were the wealthiest countries on the planet (and because 19th century life in any country sucked ass compared to life today, even if it was in a relatively well-off country like Ireland).
Agreed, but I do remember reading a quote by a traveler, (possibly Dostoevsky but I am not certain), who had been to places all over Europe. His impression of the Emerald Isle: “I thought I had seen true poverty, but I had not yet visited Ireland.” He hadn’t been to Sweden either. I’m paraphrasing, but that’s the idea. Certainly the Irish in America had a unique reputation for accepting the poorest accommodations imaginable, the first couple generations.
OK, I’ll backtrack and restate.
I still think that the extrapolation methodology was amateurish and weak. If there are sciences where that kind of method is acceptable please tell me about them. Working with a dataset seems to be the norm.
If the purpose of all this is to put g and IQ on the table as a causal factor to be considered, I can’t object to that. However, it seems to me that much more than that is being claimed. My phrase “g theory of history” is a caricature, but I think that there is a strong tendency here to overestimate the causal significance of g.
I do not have a complete alternative explanation of history, and no one does, but I have a lot of familiarity with the enormous body of literature on the industrial revolution and contemporary development. People here seem to have a relatively weak familiarity with that literature. (Brad DeLong is a place to go on the internet; he frequently posts bibliographies).
I think that path-dependent historical, institutional, and geographical explanations are stronger than the kind of statistical approach taken here, even though the use of statistics makes ideas seem more scientific. This is a general problem in social science method, and economics seems to be becoming more historically oriented and less lawlike, and that’s a good thing.
This all reminds me of several attempts of physicists to revolutionize historical linguistics without any knowledge of linguistics. One such method, going in blind, reported Portuguese and Spanish as much more distantly related than they actually are (they’re pretty much two dialects of one language). It was hard to have any confidence in the original results reached by a method which misrepresented known facts, but the physicists presented their conclusions with great pride, as though they were finally solving problems that linguists had been bumbling around with for centuries.
This is a big question, but I think that the kind of statistical approach (“single most important variable”) given by Jason is less powerful than it seems. It’s highly dependent on choosing the right variables to begin with; if you didn’t, the statistics won’t work. On the question of landlockedness you can say something like “On poor continents, the landlocked nations are the poorest”. As far as I know that’s true, and there are only six continents and four of them have no landlocked countries. If you throw Austria and Switzerland into the statistical pot, you can make landlockedness seem unimportant, but in the context of a historical-geographical understanding of third world landlocked countries (or landlacked areas of large developing nations), you’ve really confused the issue. That’s a factor which should be stressed and not minimized.
The independence of g / IQ as a variable is a problem for me. I understand that GNXP regards this question as settled but I don’t think that American twin studies are adequate to prove the point. A comparison between Taiwan and Mali will show that Taiwan has higher IQ. Is this independent of history and social organization (or a cause of differential history?) Taiwan (as Chinese) inherits a 3000 year tradition of literacy and education and 2000 years of alienable land and a cash economy. It was also modernized by the Japanese starting in 1895. It also had a favored status with the US starting, and a modernizing government, in 1945. It also is favorably located for trade. Mali had none of these advantages except a degree of literacy. Without g, there’s no mystery about the difference, and there’s a suspicion that Mali’s multiple deficiencies impact not only economic development but also IQ.
In the long run, g and IQ may turn out to be the major independent factor they are claimed to be. I don’t think that the methods used so far have made the case. For the case to be persuasive, it must also take into consideration the other developmental factors, including the path-dependent ones of geography and history, and become part of the larger debate on development issues.
That’s as much as I can say. It’s not an alternative theory, or a rigorous criticism. It expresses my dissatisfaction with this particular line of explanation as well as I can. I don’t claim to have won the argument or to have proven anything. Some here may find some value in my points. Others probably won’t.
The reason I’m here, though, is that GNXP is also a history and politics forum, and while a lot of the stuff here is extremely interesting and valuable, I also often feel that the historical argumentation needs a lot of improvement.
I feel that the political argumentation needs a lot of improvement. :-)
Trying to forcefully push an agenda introduce a LOT of noise, “counter-noise” and nitpicking on details to the detriment of more considered investigation of the ideas proper.
“I do not have a complete alternative explanation of history, and no one does, but I have a lot of familiarity with the enormous body of literature on the industrial revolution and contemporary development. People here seem to have a relatively weak familiarity with that literature.”
Frankly, your cite-to-conjecture ratio has not been that great in this discussion. You now follow that up with an assertion that “people here” have a “weak familiarity with [growth] litterature.”. Please.
Now, I’m just as much in favor of free-wheeling speculation as the next guy, but when you enter a discussion playing the role of the avenging angel of strict academic stringency, flipping into storytelling mode after approx. five minutes is not a smooth move.
Back to the actual subject – if anything my own (“weak”) familiarity with growth litterature has tought me that there is no really good empirically founded general understanding of economic growth, (apart from controversial statements like “industrialization is good” or “a bit of market economics help”) and that there is thus huge room for improvement. That’s why I disagree with your assertions that observed economic growth patterns are easily explained by identified factors.
“I think that path-dependent historical, institutional, and geographical explanations are stronger than the kind of statistical approach taken here, even though the use of statistics makes ideas seem more scientific.”
I would go so far as to say that statistics is generally a scientifically far more rewarding tool compared to unsupported conjecture. Of course, all conclusions that are produced using the “historical” approach are not unsupported – many are also largely useless platitudes. We may all agree that “good institutions” help development along – but why do some countries have lousy institutions? You mention a “2000 year tradition of literacy”. Is that tradition something that springs from the soil?
“The independence of g / IQ as a variable is a problem for me. I understand that GNXP regards this question as settled but I don’t think that American twin studies are adequate to prove the point.”
Luckily, some recent papers appear to indicate that you won’t have to rely on “american twin studies” for much longer. Also, your definition of “independence” used above would be interesting to hear. Virtually none of the variables proposed by you in explaining growth could be said to be “independent” in any meaningful sense. (We can discuss “landlocked” as a variable, possibly. But “literacy”? “institutions”? Please.)
Gregory Clark’s forthcoming book, A Farewell to Alms: A Brief Economic History of the World, tries to shed some light on these questions. Clark’s thesis is that for most of our species’ history we’ve been stuck in a Malthusian trap, where populations respond to improved living standards with population growth until the point where average living standards are back where to the equilibrium (low) level. Clark attempts to explain when, where and why populations have escaped from the Malthusian trap.
Dobeln, is it OK to criticize a gross statistical blunder in the course of making a non-statistical presentation? What I said to begin with was something like “I’m not much of a statistician, but isn’t reporting extrapolated data a big no-no?”
To satisfy you, I would have to write a well-documented full-length book. I didn’t do that! What I wrote was a blog post sketching a number of specific criticisms. If you want to call what I did unsupported conjecture,OK. What I was doing was pointing you to the disciplines of history and economic history.
As far as independent variables go, in path-dependent historical and geographical description you rarely have them. I wasn’t asking for an independent variable to be produced, but was questioning the claim that g was one.
I don’t claim that there’s a perfect global theory of economic development available, but I also don’t think that the underdevelopment of Mali is hard to explain without a global theory, and without g. Not to deny that g might be one of the many factors, but g is not the last piece in the puzzle triumphantly explaining the mysterious underdevelopment of Mali.
If you really believe that the historical literature is mostly a mix of useless platitudes and unsupported conjecture (which you plan to scientize), we have succeeded in locating our key disagreement. I say that for what’s being said about g in history to be taken seriously, it will have to be integrated into a historical argument which also takes into account the other, already-known factors. I also say that the stuff I read at GNXP often seems historically sketchy and needs to be beefed up a bit. Perhaps you disagree.
Clark’s basic thesis, as I understand it, is that the Industrial Revolution and the escape from the Malthusian trap occurred when the rate of technological progress crossed some threshold value. On the question of why the rate of technological progress crossed the threshold where and when it did, I’m think Clark is open to a role for both cultural and biological evolution.
“Dobeln, is it OK to criticize a gross statistical blunder in the course of making a non-statistical presentation? “
You appear to be pretty much alone in considering that part of their methodology “a gross statistical blunder”. A fib, for certain, but a blunder? Hardly – they knew what they were doing and were open about it.
Of course, you did not stop at criticising their methodology. You then took the “blunder” and invoked the good ol’ “poisoned well” to dismiss the parts of the dataset that did not involve any extrapolation. Now that’s stringency!
“To satisfy you, I would have to write a well-documented full-length book.”
I’m far from that demanding, but providing at least a single source beyond “Google Brad de Long’s site” would have been cool before:
a) Dismissing the cites providing evidence significantly challenging some of your assertions. (Without counter-cites, naturally)
b) Accusing those you disagree with of having a “weak grasp of the litterature”, etc.
“As far as independent variables go, in path-dependent historical and geographical description you rarely have them.”
I’m all with you on that – indeed, that’s partly why I’m not willing to dismiss a potential large role of g in some developmental contexts.
A variable that shows strong correlation to contemporary levels of economic development that has also been shown to have a strong genetic component (this does not necessarily mean geographical variation in g is genetic, but it certainly raises the possibility) is very attractive (independence-wise) when compared to distinctly non-indepentent variables like “institutional quality” or “a tradition of literacy”.
“If you really believe that the historical literature is mostly a mix of useless platitudes and unsupported conjecture (which you plan to scientize)”
I’m more limited than that – I believe the litterature explaining contemporary differences in levels of economic development contains far too much platitudes (“good institutions”) and conjecture (“Mali is poor because it’s landlocked and had bad government!”)
“I say that for what’s being said about g in history to be taken seriously, it will have to be integrated into a historical argument which also takes into account the other, already-known factors.”
I agree completely – for instance, I tried to explain above, based on my understanding of economic history, why I personally don’t give much of a potential role to g in the majority of cases when it comes to historical patterns of industrialization and growth. (You brought up Swedish industrialization as a topic, on which I have a fairly good grip on the to a significant degree Swedish-language litterature).
“I also say that the stuff I read at GNXP often seems historically sketchy and needs to be beefed up a bit. Perhaps you disagree.”
People get gene-blind sometimes around here, true, but I don’t agree this post and thread is a very good example of that.
Charles Murray
“How to Accuse the Other Guy of Lying with Statistics”
Statistical Science
2005, Vol. 20, No. 3, 239?241
DOI 10.1214/088342305000000250
Samples offer a rich source of smoke. Something is wrong with every sample. Start with that assumption, which has the advantage of being true, seek out that something, and then announce that the data are uninterpretable. If the sample is representative, argue that the data are outdated. If the sample is recent, argue that it is unrepresentative. If it is both recent and representative, you may be able to get some mileage out of missing data. If the author drops cases with missing data, argue that doing so biases the sample. If instead the author uses imputed values, accuse him of making up data.