Title: The international general socioeconomic factor: Factor analyzing international rankings
Authors: Emil O. W. Kirkegaard

Abstract:
Many studies have examined the correlations between national IQs and various country-level indexes of country well-doing and wealth. The analyses have been unsystematic and not gathered in one single analysis or dataset. In this paper we gather a large sample of country-level indexes and show that there is a general socioeconomic factor which is highly correlated with national IQ.

Keywords: National IQs, social progress index, democracy ranking, intelligence, g-factor, group differences, general socioeconomic factor, method of correlated vectors.

Attached:
- PDF
- Zip file of all source data

This draft still needs language editing. I am posting this one here now so that reviewers can begin analyzing the content.

I have already pointed out some typos to you over email and I am sure other reviewers will catch more. The statistical analysis seems sound. However, I think there is an exclusive focus on IQ at the expense of other psychometric variables. Perhaps you could carry out SEM or MR to disentangle the underlying potential causal relationships. Since the correlation of IQ with your factor S is almost perfect, it is very unlikely that this is a genuine relationship, not mediated or moderated by other factors. Can you build a model that includes other variables, for example the Big 5? I also noticed that East Asian countries have large residuals, in particuar China's S score is much lower than predicted by its IQ alone. It would be interesting to see what drives this strong correlation with IQ. My impression is that the paper simply assumes that IQ alone is enough to account for social progress and democracy but I suspect there are other factors that come into play. I can name Openness to Experience as a good candidate or climate (latitude or mean temperatures)...etc.

I can't build a model for big five or similar because: 1) I'm not familiar with any good international datasets for those, 2) I don't know SEM, 3) I don't have any particular hypotheses about them. I'm not really arguing a case for causal roles in this paper, just pointing out the S factor and it's correlation with national IQ (proxy for G).

If one wanted to try cultural factors, one could perhaps try the Hofstede data, if they are public. But that would be for another paper IMO. http://en.wikipedia.org/wiki/Hofstede%27s_cultural_dimensions_theory

Then spell out clearly in the paper why you choose to correlate S with IQ, and that your study makes no assumptions about causal roles from IQ to S.

New version edited for language and adding a bit about the interpretation of the S x G correlation.

I think that when you have a theory guiding your factor analytic model, it is better to use Confirmatory Factor Analysis, because you are testing a model based on a priori ideas/hypothesis and you are not venturing into unknown territory. Probably the results of CFA in this case would not be much different but it would be good to compare EFA and CFA.

This is more of an explorative study than a model-fitting competition study.

New draft has mostly just language improvements.

Can you spell out clearly why you have decided to correlate these indexes to IQ? E.G. Why do you think it is interesting to know how they correlate with IQ, if you expected them to correlate positively and if so why.It's not enough to say that previous studies have correlated these with IQ, because these variables have been correlated to a bunch of other variables. Also make it clear that this study is only correlational and does not tell anything about causal links (from IQ to human development and viceversa).

I don't know what it is that you want. Recall that I already wrote this in the section:

As it is usually done, one finds some interesting country-level variable and then regresses it on national IQ, perhaps with some controls. Often authors will also argue for a causal connection from national IQ/G to country-level variables. The typical example of this is wealth (e.g. [17, 18, 19, 20, 21]). Since we know that g causes greater wealth at the individual level, and that nations can generally be considered a large group of individuals, it would be very surprising, though not impossible, if there was no causation at the group level as well.

We don’t want to argue at length for any causal role in this paper, so we merely present the correlations and leave the interpretation to the reader. We chose to look at two cognitive ability measures, the total scores from SPI and DR, as well as the first (PCA) factors from the datasets. For cognitive measures we used Lynn and Vanhanen’s 2012 national IQ estimates[5] and Altinok’s educational achievement estimates[22] as an alternative cognitive measurement. Results are shown in Table 2.

The reason we try national IQs is that this is the common thing to do in these kind of studies. I am merely integrating these kinds of results into a statistical theory.

See also the following section:

As expected, the correlations with the aggregated country-level measures have a very strong correlation with proxies for G. All correlations are significant beyond the p<0.001 level (N’s 100-132). Furthermore, the correlations between cognitive measures and the S factors from both datasets were stronger than the indexes as made by the authors. This indicates that it is the first factor that drives the relationship, not the remaining variance. Global hereditarians[23] may interpret this result as being in line with predictions, while non-hereditarians may interpret as showing that national differences in S cause national differences in G, or something else entirely.

When did you add these sections? It'd help if you marked your additions with a colour or wrote them in bold. I do not understand why you write "we" when it's clear you are sole author. Perhaps you include other collaborators? Look I will still approve this paper. The reason why I am asking you to make more clear why you chose to include IQ as a variable is because I have been harassed by editors before for including something just because it's a common thing to do. In academia they want you to justify everything. I do not know if it is the best thing to do or not.I think you should discuss in the introduction why you expect IQ may be correlated to the S factor (not just wealth). There are some authors (some very old, beginning of 20th century, others contemporary) that think IQ is correlated to all socially desirable outcomes, and you should find references for that.This would justify your inclusion of IQ in the study better than just claiming that it's been used before. It's perhaps just a formality but it will make the paper look better in the eyes of your academic readers.

"Furthermore, the correlations between cognitive measures and the S factors from both datasets were
stronger than the indexes as made by the authors. This indicates that it is the first factor that
drives the relationship, not the remaining variance"
In my opinion, this is the most interesting empirical finding of this paper. Your use of the method of correlated vectors supports this conclusion and I think it should be mentioned in the abstract and in the discussion.

I was surprised to find the MCV result to be so high, and in both datasets. .99 and .98 is pretty fantastic by MCV standards.

Yes, that is why you should emphasize it more in abstract and conclusion.Perhaps title too?

...and this suggests that latent S factor is responsible for the correlation between socioeconomic variables and IQ.

Then add this to the abstract

I approve publication