Tuesday, April 13, 2010

What women want

A perennial topic in evolutionary biology is the determinants of mating success. A commonly held view is that men want good looking women [as a signal of a good quality genotype] and women want intelligent and/or economically successful men [who will support them bringing up baby]. Since this topic is also of some practical, as well as scientific, importance it seemed worth another look. Just trying to help you guys.
Using the trusty NCDS [the 1958 British birth cohort] I model the probability of being married at age 33 for males only. After a certain amount [though not exhaustive, its only a blog] of experimentation some interesting results emerge. While this is not a definitive statement, it shows how its not too difficult to come up with results that differ from conventional wisdom.
  1. Being aggressive helps
  2. Being withdrawn does not
  3. Being smart, hardworking or more educated makes no difference
  4. The higher your BMI is up to a value of about 20 the better: thereafter it makes no difference i.e. you can be too skinny but not too fat
  5. Height is not that big a deal
  6. If you were judged a wild child at age 11 [by your teacher], this is not a good omen
  7. Dont' be sad: if you show depressive tendencies at 23, this is also not a good omen
Some of these results are certainly contrary to intuition and received wisdom generally, especially 3. There may be other subtleties i.e. there may be limits to the attractions of male aggression.


Married @ 33













School after 16



Bmi <20



Bmi >20



Height @23



Bad boy @ 11



Malaise @ 23





Marginal effects; Absolute t statistics in parentheses

(d) for discrete change of dummy variable from 0 to 1

* p <.05>** p <.01>*** p <.001>


Enda Hargaden said...

Firstly, the tags are brilliant.

Secondly, height is not significant but the coefficient dwarves many of the others.

Peter Carney said...

without re-opening old debates, why are you using a bmi of 20?

Kevin Denny said...

Enda: I would need to have been clearer about units [it was late]. Its actually in feet. I was surpised there wasn't an effect but maybe if I allowed for some non-linearity something clearer might emerge.

Peter: I experimented with different knots and that seemed to be where the turning point was. There is nothing magical about it, so 19 or 22 would probably do the trick too.

Peter Carney said...

i see. it might be interesting to check the effect of religion?

Kevin Denny said...

I could certainly do that but in this data, religion tends not to do too much in general, not as much variation at the time, but worth checking.
At some point I will do the equivalent exercise for the females.

Leigh Caldwell said...

If more educated (smart, etc) men are in higher demand by women, it does not necessarily follow that they would marry earlier.

Compare quoted companies. Are companies with higher profits more likely or less likely to merge with another company? There's no a priori reason to believe in an effect either way. However, if they do merge, their shareholders will achieve a higher price.

Looking at it another way, education and intelligence have a "wealth" effect as well as a "price" effect.

Martin Ryan said...

I think you're onto something Leigh: in that women may tend to marry earlier than men, due to biological reasons. If men do not have the same "clock" to consider, then it may be the case than men tend to marry younger women, or vice versa. Could such a possibility be a plausible prediction even just from the point of view of economic intuition? That is, if women marry older men, then they will (on average) marry more successful men.

Kevin, is it possible to look at any ages higher than 33? If the Irish case is any guide, we know the following from CSO statistics.
In 1996, individuals reporting themselves to be single (on the Census) made up 65% of the cohort aged 25-29 years (32% of the cohort aged 30-34 years were single).
In 2006, 80% of individuals aged 25-29 years were single (50% of the cohort aged 30-34 years
were single.)

Finally, from CSO statistics it can be seen that the prevalence of singles substantially increased
between 1996 and 2006. What about those that do not marry at all? Is it possible to observe long-term co-habitation in the NCDS?

Kevin Denny said...

Leigh:I don't think the merger analogy is right. Merging for companies is not a necessity, you go on being a successful singleton. Biologically, if you don't "merge" & reproduce you are a failure in the sense that you don't pass your genes on. That's why, the argument goes, we have evolved to reproduce. In another words, being smart is adaptive.
On the other hand, my results are consistent with your hypothesis so who knows?

Martin: in general the age difference at marriage is about 2-3 years (except in Greece where its higher). I think its because girls mature earlier.I could look at these folks when they are 42 & you also observe people who were married or cohabiting & are no longer. So one could see whether intelligence predicts marriage failure. Scary thought.

For a paper with the standard view see: Intelligence and mate choice: intelligent men are always appealing
Mark D. Prokosch, Richard G. Coss, Joanna E. Scheib, Shelley A. Blozis
Evolution & human behavior , 29(1), 11-20, January 2009.