What is the Most Valuable Trait?

[Nachtwolf]

The only 4 hits that are returned by OVID for the PsycINFO database when queried for the keyphrase Equatorial Guinea are chimpanzee and lowland-gorilla studies.

Hahahaha!

--Mark

 

[Nereid]

hitssquad, quoting Lynn and Vanhanen:A factor that needs to be taken into account in making these calculations is that the mean IQs in economically developed nations have been increasing since the 1930s. An adjustment needs to be made for this increase when calculating the mean IQs obtained in countries from tests that were adminstered some years before or some years later than the British test with which it is being compared... Mean IQs on the Wechsler tests increased by approximately 3 IQ points per decade from the mid-1930s to the 1990s (Flynn, 1984, 1998) *SNIP

Translation: we don't have any data on secular trends in Equatorial Guinea, or any nation which is undergoing a demographic or economic transition. However, if we apply the Flynn effect (which has been documented in some economically advanced countries), we'll be able to adjust the "National IQ" of Equatorial Guinea downwards, and that strengthens our case, so we'll do it.

ditto A problem in estimating some national IQs is that the samples have scored below the first percentile in relation to British norms. The first percentile is equivalent to an IQ of 65. Where national samples have scored below the first percentile, They have been assigned an IQ of 64.

Translation: we have clear evidence that the data from the 1984 study is incompatible with measures we use in the UK. Rather than investigate the cause of this incompatibility, we've chosen to scale the data in a fashion which supports our case (and hope that no one notices our unscientific sleight of hand).

Question for hitssquad: Here are the questions I asked originally about any studies re Equatorial Guinea's "National IQ":
Would you please list the studies of IQ in Equatorial Guinea? Please ensure that the following elements of each study are mentioned:
1) the mean (or median, please specify which) from each source
2) the standard deviations about the mean/median
3) tests used to show Gaussianity of the distributions
4) the sample sizes
5) age/gender/population group distribution of each sample
6) sample selection methods
7) tests and test protocols
8) year when the research was done
9) the method(s) used to account for secular effects
10) the size of each such correction

So far, you have given us (partial) answers to:
1) (partial; you didn't state what measure was reported)
4) {and just how meaningful is a sample of 48, in a population of >400,000?}
5) (partial; "10- to 14- year olds")
7)
8) (partial; "around 1984")
9)
10)

... that's ~55%; sorry, in an Australian university, you fail (maybe US universities have lower standards?)

 

[hitssquad]

Flynn effect adjustments for Equatorial Guinea

Originally posted by Nereid

hitssquad, quoting Lynn and Vanhanen:
the mean IQs in economically developed nations have been increasing since the 1930s. An adjustment needs to be made for this increase when calculating the mean IQs obtained in countries from tests that were adminstered some years before or some years later than the British test with which it is being compared

Translation: ... if we apply the Flynn effect (which has been documented in some economically advanced countries), we'll be able to adjust the "National IQ"

Flynn effect adjustments were not peformed on national IQs. Flynn effect adjustments were performed on the means of the results of individual IQ studies from which national IQ were subsequently calculated. Some were adjusted downward, and some were adjusted upward. Since some nations' national IQs were calculated from the results of more than one study, and since each study was performed chronologically neither necessarily equidistant, nor necessarily equi-directional, from the year of the relevant test's closest British norming, performing a Flynn effect adjustment merely on national IQs wouldn't be possible in at least some cases.

...and that strengthens our case

Merely knowing that Flynn effect adjustments were performed on the results of the studies used to calculate the national IQs, it wouldn't be possible to positively state whether such adjustments weakened or strengthened Lynn's and Vanhanen's case. More information would be needed. If Lynn's and Vanhanen's methods were reproduced without the Flynn effect adjustments, then it could be determined which of either the non-Flynn-adjusted or the Flynn-adjusted version produced the strongest case.

Originally posted by Nereid

hitssquad, quoting Lynn and Vanhanen:
A problem in estimating some national IQs is that the samples have scored below the first percentile in relation to British norms. The first percentile is equivalent to an IQ of 65. Where national samples have scored below the first percentile, They have been assigned an IQ of 64.

Translation: ... we've chosen to scale the data in a fashion which supports our case

Similarly to to the above case, it can't be said merely from knowing the facts presented here whether this adjustment weakened or strengthened Lynn's and Vanhanen's case. A reasonable guess would be that it weakened it, since the result of the adjustment was in each case likely a compression of the score range, and hence likely an overestimation of the national IQs of the nations in question.



--
A test in which the easiest items are failed by some persons is said to produce a "floor" effect; a test in which the hardest items are passed by some persons is said to produce a "ceiling" effect. Floor and ceiling effects truncate the range in test scores, which then underestimate the full range of individual differences in the population.
--
The g Factor. p310.
http://www.questia.com/PM.qst?a=o&d=24373874



--
probably because of the "floor" effect on the distribution of test scores (and hence restriction of range) created by most of the SPM items being too difficult for most of the subjects.
--
The g Factor. p416.
http://www.questia.com/PM.qst?a=o&d=24373874



Again, a restriction of range resulting from a floor effect implies an overestimation of the mean IQ of the sample.

Would you please list the studies of IQ in Equatorial Guinea?
1) the mean (or median, please specify which) from each source
2) the standard deviations about the mean/median
3) tests used to show Gaussianity of the distributions
4) the sample sizes
5) age/gender/population group distribution of each sample
6) sample selection methods
7) tests and test protocols
8) year when the research was done
9) the method(s) used to account for secular effects
10) the size of each such correction[/color]

So far, you have given us (partial) answers to:
1) (partial; you didn't state what measure was reported)
4) {and just how meaningful is a sample of 48, in a population of >400,000?}
5) (partial; "10- to 14- year olds")
7)
8) (partial; "around 1984")
9)
10)

For question 7, tests and test protocols, the test used was the WISC-R, which stands for the Wechsler Intelligence Scale for Children - Revised. The WISC-R was published in 1974. Presumably, the British norms were collected in years preceding that date.


For questions 9 and 10, secular corrections, the present author quoted Lynn and Vanhanen as writing:


--
Their IQ was 63. Because of the 12-year interval between the two data collections, this needs to be reduced to 59.
--
IQ and the Wealth of Nations. p203.



Again, the test was a Wechsler test, and Lynn and Vanhanen stated that:


--
Mean IQs on the Wechsler tests increased by approximately 3 IQ points per decade from the mid-1930s to the 1990s (Flynn, 1984, 1998)
--
Ibid. p197.



Taking their word for it that the distance between the sample collection and the British norming of the test was 12 years, we can calculate the degree of the Flynn adjustment by multiplying 3 points by 1.2 decades and getting the answer of a 3.6 point adjustment. We will assume that the test was normed previous to the Equatorial Guinean sample collection, since Lynn and Vanhanen imply that the correction was in a downward direction. Subtracting 3.6 points from the given unadjusted mean score of 63 gives a Flynn-effect-adjusted score of 59.4. Rounding that off gives a score of 59, which is what Lynn and Vanhanen reported.

As for why the uncorrected mean score was reported as 63 and not 64, after Lynn and Vanhanen explained that scores below 65 would be assigned a handicap score of 64, presumably -- based upon many other examples of both cases in the Calculation of National Intelligence Levels appendix of their book IQ and the Wealth of Nations -- they meant that IQ data in the form of raw scores would be assigned handicap scores of 64 when the data for individual scores was available, whereas IQ data in the form of a mean score below 65 and without the accompanying individual sample data wouldn't be assigned the handicap score of 64.

Another example of Lynn and Vanhanen using a mean score below 64 is in the case of the nation of Guinea, where the first sample they cite is a 1933 collection of IQ data from village schildren that resulted in a mean score of 61 relative to American norms (Lynn and Vanhanen, p206).





-Chris

 

[Nereid]

And the answers to questions 1) (measure), 2), 3), 5) (distributions, not just an age range), 6), 8) ("around" is exactly very precise)?

hitssquad: Flynn effect adjustments were not peformed on national IQs.

I'm not sure that I said they were; but that makes the shortcomings worse.

hitssquad: Flynn effect adjustments were performed on the means of the results of individual IQ studies from which national IQ were subsequently calculated.

Which in the case of Equatorial Guinea was a test on a single sample of 48, done 'around 1984'.

Lynn and Vanhanen used a average secular adjustment based on data obtained from several advanced economies. AFAIK, within these 'national' results, there is considerable variation (and little reporting of systematic errors?) - there are some links to reported results on Nachtwolf's website. This suggests that there may be several factors at work. Since there are no studies on the Flynn effect in sub-Saharan Africa (or any developing economy?), it's a stretch to apply a 3/decade factor. Surely a less biassed approach would be to estimate the errors, and add them (in quadrature), to give a result expressed like this (I am making the numbers up, to illustrate):
63 +15/-10 (formal 1 $$\sigma$$)?

 

[sandinmyears]

I think the Apostle Paul said it well:

And now abide these: faith, hope and love. But the greatest of these is love.
~1st Corinthians 13:13

 

[hitssquad]

Lynn's data and sources for Japan's IQ

(in message 153227)

ten sources used for the book's assumption of IQ 105 for Japan

Would you please list them

• Japan
  
  
  1. The first calculation of the IQ in Japan appeared in Lynn (1977a). This paper presented a calculation of the Japanese IQ obtained from the Japanese standardization of the Wechsler Intelligence Scale for Children (WISC). The WISC was standardized in the United States in 1947 and in Japan in 1951. The verbal tests were altered in the Japanese standardization so they could not be used for a comparison, but the performance tests remained the same. On these tests, the Japanese standardization sample obtained an IQ of 103. Because the Japanese standardization was made four years after the American standardization, one IQ point has to be subtracted to give a Japanese IQ of 102. The American standardization was made on whites only, so in relation to the whole American population the Japanese IQ becomes 104. The American IQ in relation to the British IQ of 100 is 98 (see under United States). Thus, in relation to a British IQ of 100, the Japanese IQ becomes 102.
     
     
  2. In 1980, a new method for calculating the Japanese IQ was devised (Lynn and Dziobon, 1980). This entailed the administration of the Japanese Kyoto NX test and the American Primary Mental Abilities to a sample of 213 9- to 10-year-olds in Northern Ireland and calibrating the Kyoto test against the American test. The result was that the average Japanese child, with an IQ of 100, obtained an IQ of 110 on the American test, and therefore that the Japanese had an average IQ of 110. Because the Japanese test was standardized in 1972 and the American test was standardized in 1962, adjustment for the 10-year interval between the two standardizations requires the reduction of the Japanese IQ to 108, and to calibrate the Japanese IQ against the British IQ of 100 requires a further reduction to 106.
     
     
  3. The next study of the Japanese IQ was made by Misawa, Motegi, Fujita, and Hattori (1984), who analyzed the performance of Japanese children on the American test, the Columbia Mental Maturity Scale. This is a non-verbal reasoning test for 4- to 9-year-olds that was first published in 1954. A revised version and standardization appeared in 1972 and this version was standardized in Japan on a sample of 780 around 1980. The calculation of the Japanese IQ on the test in relation to the American norms shows that the Japanese attained an IQ of 113. Because of the 8-year interval between the two standardizations, this needs to be reduced to 111. To calibrate this IQ against a British IQ of 100, it needs to be reduced further to 109.
     
     
  4. In 1985, Stevenson, Stigler, Lee, Lucker, Kitanawa, and Hsu (1985) published a study comparing the IQs of 240 6-year-olds and 240 10-year-olds in the American city of Minneapolis and the Japanese city of Sendai. They constructed their own test consisting of a number of subtests of various abilities. These tests did not include a test of non-verbal reasoning ability such as the Progressive Matrices. It did include a vocabulary test and a spatial test. The results were that in relation to the American children, the Japanese 6-year-olds obtained IQs of 89 on vocabulary and 105 on the spatial test, which can be averaged to 97 for a measure of general intelligence, and the Japanese 11-year-olds obtained IQs of 98 on vocabulary and 107 on the spatial test, which can be averaged to 102. Combining the two results gives a Japanese IQ of 100. This result led Stevenson and his colleagues to conclude that there is no difference between the IQ in Japan and the United States. This study is defective because Minneapolis is not representative for intelligence of American cities. A series of studies have shown that intelligence in the in state of Minnesota, in which Minneapolis is situated, is higher than in the United States as a whole. In the military draft in World War I, whites from Minnesota obtained the highest score on the Army Beta Test out of all American states (Ashley Montagu, 1945) (the scores of blacks are not given). In the draft for the Vietnam war, the percentage of draftees (blacks and whites) who failed the pre-induction mental assessments was the second lowest in Minnesota among the American states (Office of the Surgeon General, 1968, p. 45). On the basis of these data, Flynn (1980) has calculated that the average IQ in Minnesota is 105. Thus, in order to equate the Japanese IQ to that of the United States, 5 IQ points need to be added to the mean of 100 found by Stevenson et al., bringin the IQ to 105. To equate this to a British IQ of 100, one IQ point needs to be subtracted, giving an IQ of approximately 104.
     
     
  5. An analysis of the Japanese IQ in terms of the American WISC-R has been made by Lynn and Hampson (1986a). On the Japanese standardization of this test on 1,100 6- to 16-year-olds, the Japanese obtained an IQ of 105. The American test was standardized in 1972 and the Japanese test in 1975, requiring the subtraction of one IQ point from the Japanese mean. To calibrate this figure against a British mean of 100 requires the subtraction of a further 1 point, bringing the Japanese mean to 102.
     
     
  6. An analysis of the Japanese IQ in terms of the McCarthy Test has been made by Lynn and Hampson (1986b). The McCarthy Test is for children between the ages of 2.5 and 8.5 years. It was standardized in the United States in 1971 and in Japan in 1975 on a sample of 550 children. In relation to the American standardization sample, the Japanese obtained a mean IQ of 101.7. To adjust for the 4-year interval between the two standardizations, the Japanese mean needs to be reduced to approximately 101. To equate this figure to a British mean of 100, two IQ points need to be subtracted to give a Japanese IQ of 99. This figure is probably lower than the other calculations because half of the test is for young children aged 2.5 to 5.5 years, during which the Japanese children perform poorly when compared with the older age group, suggesting that Japanese children are slow developers during early childhood.
     
     
  7. An analysis of the Japanese standardization sample of 600 on the American Wechsler Preschool and Primary Scale, an intelligence test designed for 4- to 6-year-olds, has been made by Lynn and Hampson (1987). Their mean IQ was 107.8. The American test was standardized in 1964 and the Japanese test in 1967. To adjust for the 3-year interval between the two standardizations, the Japanese mean needs to be reduced to 107. To equate this figure to a British mean of 100, two points need to be subtracted, giving a Japanese IQ of 105.
     
     
  8. Data for the Standard Progressive Matrices of 444 Japanese 9-year-olds were collected in 1989 by Shigehisa and Lynn (1991). In relation to the 1979 British standardization, their IQ was 112. To adjust for the ten-year interval between the two data collections, this needs to be reduced to 110.
     
     
  9. A study by Takeuchi and Scott (1992) reported the performance of a sample of 454 5- to 7-year-old Japanese children in the city of Nagoya on the Canadian Cognitive Abilities Test. The Japanese IQ on the abstract reasoning subtest was 106.9. The Canadian test was standardized around 1989. Because we have estimated the Canadian IQ at 97, three IQ points need to be deducted from the Japanese IQ to calibrate it against a British IQ of 100, bringing the Japanese IQ to 104.
     
     
  10. A further study of intelligence in Japan compared with that in the United States has been published by Li, Sano, and Merwin (1996). The samples consisted of 239 14- to 15-year-olds in the Japanese city of Toyama and 318 adolescents of the same age in Minneapolis and St. Paul. Six tests were given including verbal and non-verbal reasoning tests. In relation to an American IQ of 100, the Japanese obtained an IQ of 103. Because the IQ in Minnesota, in which Minneapolis and St. Paul are situated, is 105 (Flynn, 1980), five IQ points need to be added to the Japanese mean to bring it to 108. To equate this figure to a British mean of 100, this needs to be reduced to 106.
  
  The average of the ten calculations of Japanese intelligence gives an IQ of 105.

(Richard Lynn and Tatu Vanhanen. IQ and the Wealth of Nations. Appendix 1: The Calculation of National Intelligence Levels. pp210-212.)

 

[selfAdjoint]

I notice that all these standardizations and comparisons were made with tests for children. Since g for adults behaves differently from g in children, it seems regrettable that no adult tests were included. Also it seems a stretch to assume a causal relationship from childhood IQ to adult-oriented economic performance, without some data to split out the latent variables.

 

[Aquamarine]

I notice that all these standardizations and comparisons were made with tests for children. Since g for adults behaves differently from g in children, it seems regrettable that no adult tests were included. Also it seems a stretch to assume a causal relationship from childhood IQ to adult-oriented economic performance, without some data to split out the latent variables.

This is a very interesting point that seems to cast serious doubt on most of their data. As can be seen here, most of it are from tests on children:
http://www.isteve.com/IQ_Table.htm

Also, some have suggested that the different writing system used by East Asians may train spatial ability in children. Could this explain why East Asian children score higher in the non-verbal parts of IQ tests?

 

[Nereid]

Further:
- what data is there to show that the Flynn effect (secular variation in mean IQ, using the same test) is the same in all countries (specifically, the US and Japan)?
- if Minneapolis is an atypical US city, why do we think Sendai isn't similarly atypical?
- on his website, Lynn makes much of regional differences in average IQ (UK, Spain, IIRC); all but one of the studies cited by hitssquad lack info on the regional variation (or not) of their samples
- IIRC, Lynn is on record as saying that men and women have different inherent IQs; to what extent did the above studies control for gender differences? What studies does Lynn cite to show that the gender differences he accepts apply to equally to the US and Japan?
- there are numerous US studies showing an SES-IQ correlation; what do similar studies in Japan show? How were the Japanese studies of IQ (cited by hitssquad) constructed to ensure any SES contributions were neutral?
- pseudoscientists such as Rushton claim that different 'races' mature at different rates; studies of IQ changes among children (from ~age 3 to early adulthood) show that the IQ of any individual is not stable. Do the changes in IQ from early childhood to adulthood vary from one population group to another? How do the studies cited by hitssquad take account of any such changes?
- Lynn's own data show a strong correlation between the degree to which a country's population lives in urban areas and his 'national IQ' values (stronger than his derived correlations between 'national IQ' and economic success measures). On the surface, the Japanese and US data appears to confirm this correlation (Japan is more urbanised than the US, and has been for several decades).

I think I'll stop here.

 

[hitssquad]

Lynn is on record as saying that men and women have different inherent IQs

What is an inherent IQ?

 

[hitssquad]

what data is there to show that the Flynn effect (secular variation in mean IQ, using the same test) is the same in all countries (specifically, the US and Japan)?

http://www.gnxp.com/MT2/archives/001502.html

 

[Nereid]

What is an inherent IQ?

I was (slightly mis-) quoting Mandrake:

Lynn has convincingly demonstrated that the mean IQ for women is 4 points below the mean for men.

 

[Tom McCurdy]

Empathy perhaps i will have to think about this