Saturday, October 16, 2010

IQ and age differential at first marriage

A couple of years ago I posted on the relationship between estimated average IQ and desired age differential in marriage partners by women at the national level. The revealed correlation was a statistically significant .58. That is, the lower a country's average IQ, the older that country's women want their men to be.

How does this align with what actually happens? Extremely well. Using 69 countries for which data on both attributes are available, the correlation between estimated average IQ and the average age differential for first marriages at the national level is a statistically significant .61 (p=0), nearly identical to stated female desires. The data used for desired age differential are from the eighties and before, whereas the actually age differentials at first marriage are almost all from the 2000s (there a few countries for which the most recent estimates available are from the nineties), suggesting that the relationship is more than "just so".

The two biggest outliers are South Africa (low IQ, low age differential) and Taiwan (high IQ, high age differential). In the case of the former, I wonder how representative the marriage data are. If it is skewed white, that offers an explanation.

Excepting Taiwan, the best place in the developed, industrialized world for a guy wanting a younger woman is Poland, followed closely by Greece, Italy, and Germany. The Anglophone countries, in contrast, are among the most age-egalitarian.

I also looked at the relationship between age differential at first marriage and purchasing power parity. Predictably, it travels alongside IQ, but the correlation is less rigorous, at .51 (p=0).


pwyll said...

Not "deferential", "differential"!

Audacious Epigone said...


Anonymous said...

The real interesting question is: "what is the age difference at first marriage for high IQ men and women?"

Do high IQ men go for high IQ younger women or less smart women?

Anonymous said...

The data used ,,, from the eighties and before, whereas... age differentials from the 2000s. You cannot use parts of two data sets and then say the coorelation is 61. That makes the assumption that the coorelation is stable over time. Maybe from 1980 to 1982 but a twenty year difference in cohorts is really stretching your assumption. Now this does not mean your assumption is false. You did note the structure of the data sets so you are not falsifying data. But when you say there is a correlation my statistical beeper goes off.

Anonymous said...

Hi IQ women do not have the option of marrying until they are out of college. Their fathers and mothers demand that they get educations even if it amounts to nothing more than four years of liberal indoctrination. No where near 100% of these lovely young ladies want careers, but it is absolutely unacceptable in their social circle to just get married straight out of high school even to very eligible young men who are quite gainfully employed. Their parents were brainwashed by hippies and feminists and they enforce the college requirement with greater zeal than the most fanatical fundie. Many upper middle class women quit working and stay home with their kids as soon as they can because they never wanted to butt heads with men in the work world anyway. The crappy deal for the eligible men is that they miss those good years from 18-22 where their young wives would fall for their husbands instead of liberal bull.

silly girl said...

Hey, AE, I ran across this study. They were trying to make a point about income and voting, but it looks to me like they are trying to hide and downplay their actual data. Their thesis is uninteresting, however the data they collected is likely quite interesting but it is presented in such a way that you can't really get to the good stuff. I guess I am just used to the higher quality presentation here at AE.

Income inequality and partisan voting in the United States*
Andrew Gelman, Columbia University

Anonymous said...

I think the correlation with IQ is spurious. The real correlation is between the prevalence of poverty--real life-threatening poverty--and the marriage preferences of women. In the Philippines, women prefer to marry older men who have accumulated some resources and have demonstrated that they are stable, responsible members of the community. Perhaps they also feel that an older man will be safer. He will be less likely to leave her for a younger woman if she is also a younger woman. Anyway, the age disparities are driven by the need women feel for financial security in poorer countries. Since poor countries tend to be low IQ countries, there is that correlation as well. But if you could find a variable such as "percentage of the population below the poverty level," it would correlate better. Great disparities in wealth would also correlate with greater disparities in age at first marriage, since men who have the power to choose prefer younger women.

Audacious Epigone said...


Yes, I'm very curious as to what the answer to that question is. To the extent that this data suggests anything, it's that high IQ men and women will tend to be quite similar in age to one another, but I'm not aware of anything that would let us know for certain.


The correlations are b/w stated preference (80s) and IQ (various), and also between actual age differentials in marriage (mostly 2000s, a few 90s) and IQ (again, various).

Silly girl,

That's nice of you to say. I'll definitely look at the data Gelman (Mr. Red State, Blue State) presents, because it's in just those types of papers that I like to go gold digging :)


Could be, although the correlation with PPP is weaker than with estimated average IQ.

Audacious Epigone said...


The link.

Anonymous said...

I'm assuming that those with the high IQs met their S.O. in college, hence the little age difference.