Tuesday, August 26, 2008

Higher a country's IQ, the more likely women are to want someone their own age

Agnostic has an interesting post on female age preferences for men by country. The data are from the late eighties, so dust is a potential problem, especially in countries like China that have undergone major changes in the last couple of decades. Varying sample sizes also present a problem, as he discusses.

Nonetheless, suggestive correlations are still to be found, suggesting that real relationships are stronger than the data suggests--if a statistically significant correlation shines through rough approximations, better data will almost certainly reveal that relationship to be even stronger than it initially appears to be.

Those surveyed were asked how much younger or older than themselves they preferred their mates to be. The average age of the respondents varies by country, so Agnostic corrects for this by looking at the preferred age differences as a percentage of the age of the average respondent by country. I'm not sure that's helpful. Ceteris paribus, is a woman who at 20 prefered a man three years older than her likely to shift his preferred age advantage when she's 23? If not, then Agnostic's method decreases the preferred age difference percentage 2% (from 3/20=15% to 3/23=13%) for no reason, translating to a four or five spot change on the comparative rankings of the countries. Anyway, the inverse correlation between PPP and desired age difference among females is stronger if the absolute difference in years is used rather than the percentage method.

The inverse relationship between estimated average IQ* and desired age difference among females (.58, p-value of .0002) is also stronger if the absolute age difference is used (using the percentage method, it is .50, p-value=.002). The more intelligent the place, the more likely women are to want a mate nearer themselves in age. The extent that I see this pattern 'on the ground' mostly comes from girls in their teens and early twenties who are not in school dating men several years older than they are. It seems to be more of a working- and underclass behavior, but I don't have any domestic data to back that up.

This is, not surprisingly, similar to the relationship between wealth as measured in terms of purchasing power parity and desired age difference among females, as Agnostic shows. He attributes this to the greater presence of middle class values in developed countries, although the lack of emphasis put on chastity in the first world relative to less developed countries in Africa and Asia might complicate that presumption. It also strikes me as plausible that the Western cultural emphasis on egalitarianism plays some part in minimizing preferred age differences in Western Europe and the Anglosphere.

Other interesting tidbits from the Buss paper Agnostic founds his post upon:

- Of the 37 countries**, in every single one women put a greater emphasis on a potential partner being a "good financial prospect" than men did. European-descended women are moderate in their gold digging tendencies relative to Asian, South American, and African women.

- Conversely, in all 37 places men put more emphasis on "good looks" than women did. The playing field seems more even for men, who widely have the opportunity to overcome an uninspiring appearance by applying their minds and busting their tails (although that requires some genetic 'luck', too, like high intelligence, low time preference, high executive function, etc). Women's prospects are less malleable. It's little wonder homely women in their twenties and beyond seem so bitter. This replaces some of the natural disdain I have for them with empathy. I should make it a point to acknowledge that they're not invisible for moral reasons, even though they'll remain so for practical reasons.

- "Ambition and industriousness" moves Americans, both men and women, more than it does Europeans. Of the 23 majority-European groups, only Estonians had a higher net score (men and women combined, the higher the score the more important the traits in a potential partner) for the category than the US did.

- In most cases, men prize virginity more than women do. There are a few exceptions among the Euro-descended, but they are marginal, as the Euro-groups care much less about chastity in general (with the exception of Catholic Ireland) than do Asians and Africans.

* IQ scores for South African whites and blacks come from Philippe Rushton. Yugoslavia's average IQ comes from Richard Lynn's finding of 89 for Serbia. Jordan's average of 84 is used for the Palestinians. Israeli Jewish IQ is estimated to be 99, arrived at by assuming the 23.6% of the population that is non-Jewish (mostly Arab) averages 84 and plugging 99 for the Jewish population (74.6% of Israel's total) to fit with Lynn and Vanhanen's national estimate of 95. Hawaii's IQ of 94.4 comes from my NAEP-based state estimates. All others come from Lynn and Vanhanen's IQ and Global Inequality.

** Actually, the data come from only 33 countries, but 37 groups. Two groups each come from the US (the mainland and Hawaii), Canada (French and English), South Africa (whites and Zulus), and Israel (Jews and Palestinians).


agnostic said...

If it's not too much trouble, I'd try to find other estimates of national IQ than Lynn -- he's notorious for having fudgy data, especially for southern Europe. You might email Jason Malloy of GNXP.

I don't think it's our egalitarianism that makes us want more similar partners -- egalitarianism correlates with stronger sex differences in personality (large in Europe, small in E. Asia and Africa). So it seems like it should cause the underlying preferences to really polarize -- fewer collectivist pressures.

I think it's "marrying for prudence" rather than "marrying for looks."

Do you have a scatter-plot, btw? It'd be neat to see who the trend-buckers are -- what highish IQ country has women who are cool with older guys. Greece for sure.

Also, the UN did a study on actual age diffs in marriage -- big samples, pretty representative, covering just about every country. "World Marriage Patterns 2000." So next we can see how these things relate to actual behavior, aside from tastes. Rough picture looks the same, though -- Greece still sticks out from the European trend, thank god.

Sleep said...

Lynn assigns Greece an IQ value of only 92. Hardly a surprise that they would score near the Third World in terms of marriage age patterns. I believe he justifies the 92 score by saying that the Turks more or less took over the gene pool, but on the other hand, most of the peoples to the north of the Greeks score even lower on IQ tests, such as 89 for the whole of Yugoslavia if I'm not mistaken. And I'm sure there's some truth in those numbers, although I really don't think it's likely that they reflect a true genetic low IQ, because that would mean the IQ gap between the highest and lowest European nations is as big as the gap between whites and blacks or whites and gypsies, and I just don't see evidence for that kind of a gap reflected in behavior. (When was the last time you heard about high crime in Greek neighborhoods in the USA? Or even in Greece itself?)

I'm kind of a heretic on race/IQ questions. I believe in the black-white gap, because it's been proven by enough studies to fill a warehouse, and I believe that Japanese and Koreans are a few points above the white average, but I'll call "bad data" or "insufficient data" on everything else. I think it's likely that g is a bit lower in Southeastern Europe than in Northern Europe, but as to how much that difference is, I don't believe we can make any reasonable guesses at.

Audacious Epigone said...


I'm taking that comment as a cover letter when I bug Jason.

Great food for thought, as usual. I'll gather my own, crunch some more numbers, and post in response.


They're not perfect, but that makes the relationships more intriguing. Crisper data will reveal most of the correlations to be stronger than Lynn and Vanhanen's numbers make them appear.

agnostic said...

The blogger Dienekes had a post awhile ago pointing out Lynn's error in calculating Greek IQ -- either he miscoded something, or did another Lynnism, but it's really 99 or 100 or something. Something similar is probably going on with Yugoslavia.

"The Turks took over the genepool" = proof that Lynn doesn't know what he's talking about, as far as biology goes, although he knows psychology.

JBS said...

"The Turks took over the genepool" = proof that Lynn doesn't know what he's talking about, as far as biology goes, although he knows psychology."

Er, the invading Central Asian Turkish peoples did not kill off the entire Byzantine/Anatolian peoples, they just converted the bulk of the population to Islam. Further, the true invading "Turks" did not leave a large genetic footprint.

Most modern people of Turkey are Indo-Europeans.

It would be far more true to say the Greeks took over the gene pool of Turkey/Anatolia thousands of years ago through migration flows/wars.

agnostic said...

Here's a link to the Dienekes post, to show I'm not making it up. Somebody else probably has a much better database of national IQ estimates.


Sleep said...

Lynn's figure of 92 is an average of two studies, a 91 and a 97, both adjusted down to account for Flynn and the US/UK gap. If the 97 was meant to be a 99, Greece would still only be 93 overall, not 99 or 100.

I may be making a Capt Obvious statement here, but I think Dienekes is just trying to better the image of Greeks because he is one.

We're still dealing with really low sample sizes here, though. Only 600 people representing the entire nation of Greece. Moreover, the two studies that were done produced scores 6 IQ points apart. Who's to say a third study might not produce a new score even higher? By contrast, if I'm not mistaken, some of the studies used to produce IQ scores in the USA have sample sizes in the millions because they're averages of tests given to the entire student population. So, I'll believe Lynn's data when I see it replicated a few thousand more times.

Jason Malloy said...

To the extent that I own many of these international IQ studies, and have compared the numbers, L+V are largely reliable. Most mistakes are small, do not affect their analyses, and get corrected in their papers over time.

Heiner Rindermann's 2007 paper, which I discussed on gnxp, has transformed TIMSS and PISA scores added to L+V's estimates, if that's considered an improvement. (See here) You could also just use the TIMSS and PISA scores, but since they correlate with the L+V scores pretty much perfectly, your results would be the same.

Jason Malloy said...

The Volkmar Weiss link, btw, is interesting, as well, because it contains PISA data transformed into IQ for different parts of Canada, England, Germany, Switzerland, Italy, Belgium, and Spain. I've also used this data to track different immigrant groups in different nations.

Note the gradation of scores from Southeast Europe into West Asia shows up in TIMSS/PISA as well. (Rindermann's numbers are every year from these tests averaged with L+V; the single year PISA transformations are from Weiss)

98 (IQ&GI)
99 (Rindermann 07)
99 (PISA 03)

94 (IQ&GI)
94 (Rindermann 07)
85 (PISA 06)

94 (Rindermann 07)

96 (IQ&GI)
94 (Rindermann 07)
98 (PISA 06)

90 (IQ&GI)
93 (PISA 06)

89 (IQ&GI)
91 (Rindermann 07)
90 (PISA 03)

93 (IQ&GI)
95 (Rindermann 07)
87 (PISA 03)

90 (IQ&GI)
87 (Rindermann 07)
78 (PISA 03)

89 (Rindermann 07)
80 (PISA 03)

92 (IQ&GI)
95 (Rindermann 07)
92 (PISA 03)

85 (IQ&GI)
93 (Rindermann 07)

90 (IQ&GI)
90 (Rindermann 07)
88 (PISA 03)

Audacious Epigone said...


Thanks, that is greatly appreciated. I didn't suspect any conscious fudging by L&V and I'm glad to see there isn't evidence of it given very similar results from PISA scores.

Out of curiousity, I wonder if anyone has an idea as to why Cyprus shows up lower than Greece or Turkey in L&V, while more predictably between the two in Rindermann. Just randomness in testing or something else?