3) Right, but examined in a different way.
7) Removed the URL.
8) I will leave this watermark so that people can be certain that it is not the final version. You can copy text from the .tex file.
New version attached with the above fixes.
This would imply 3 approvals. 1 short of publication.
This is an excellent paper, conceptually and statistically. It does have some stylistic problems in the intro e.g.,
"This is because the different datasets do not contain just the same countries, and not in the same order, and they are also often not spelled exactly the same"
Also, you might rewrite this one sentence in the conclusion:
"The issue now seems to have been settled by two studies that compared latent g factors using confirmatory factor analysis from different batteries to each other and found them to be very close to unity."
It doesn't read smoothly, at least to me. (I, of course, know what you are saying.)
There is one nitpick. Did you get permission to copy your figure 1 -- or was there no copyright? If there was, you might just say:
"The SPI is very comprehensive; it is based on 54 components. The structure is complicated and is best shown visually; for a diagram, readers are referred to Figure 1.A on page 56 of the SPI methodological report"
A less picayune point. You said:
"Results are high but not near unity. They show again that, as long as one picks a reasonable number of them, e.g. 10., it is not so important which subset of national indexes one chooses since they measure to a large degree a common factor: S."
Your variables overlap with the ones Rushton et al. used:http://www.sciencedirect.com/science/article/pii/S0191886908001773http://www.sciencedirect.com/science/article/pii/S0160289609000592
From the last paper:
"A national K factor was computed from six life history indicators (for sources, see web references):
(i) Teenage childbearing is the proportion of children born to mothers aged 19 and below. Data are from the Demographic
Yearbook of the United Nations (2008). Missing data points were extrapolated from World Bank data.
(ii) Contraceptive prevalence among married couples is averaged from several sources including the United Nations’ Human Development Report (2004) and the UN statistics division.
(iii) Sexually transmitted diseases (STDs), from the World Health Report of the WHO (2004 edition), include syphilis, gonorrhea and chlamydia. HIV/AIDS is not included because of its recent African origin, which affects its present geographical distribution.
(iv) Homicide rate (last available date) is from the UN office of drugs and crime.
(v) Crime is a measure of crime victimization derived from the Gallup World Poll. It is the unrotated first principal component of the proportion reporting theft during the last year, proportion reporting assault/mugging, and proportion feeling unsafe on the streets at night.
(vi) Savings rate is gross domestic savings, 1975–2005 average, from the World Bank.
A national K factor was extracted as the unrotated first factor of a maximum-likelihood factor analysis of the six indicators described under Section 2. The correlations of this national K factor with its indicators as well as with intelligence (g) and log-transformed GDP are shown in Table 1. Most striking is the high correlation of .877 between K and g. This is about as high as the correlation between school achievement and IQ, the two variables from which g was averaged. The close relationship between K and g is also shown when the correlations of the six K indicators with K are correlated with their correlations with g (r = .831, N = 6, p = .041)...."
This might suggest that the S factor overlaps with the K factor. Maybe look into this in a later paper. At some point you should probably mention the (probable) overlap. You might later try to determine if and how the two factors differ.
Check the copyright issue (figure 1) and look over the intro for language issues.
["For example: "The entire dataset as well as [the] source code is available at"
Missing an article in front of "source code".]
After, I approve.