Using a well known,large representative dataset [NCDS] I showed that neither education nor cognitive ability nor a measure of whether the person was a hard-worker predicted whether the person was married (at age 33). This looks pretty counter-intuitive, to me at least.
I revisit this using a different dataset,SHARE. The data is not as rich in many respects but it catches people at a later age: 50+ in 12 European countries.
The dependent variable is whether they were ever married. The table shows marginal effects from a Probit. The covariates are years of education, measures of verbal & numerical ability [ standardized (0,1) ], height in cm and a quadratic in age [& country dummies,not shown].
As before, education does not kick in but, by contrast, cognitive ability does. So this provides mixed support for the standard theory. As before, height doesn't matter which is interesting as it is often seen as a proxy for biological "quality". Interesting, isn't it?
| | (1) |
| | Ever married |
| Yrs education | 0.00000709 |
| | (0.01) |
| | |
| Verbal_z | 0.0128*** |
| | (5.54) |
| | |
| Numeracy_z | 0.00726** |
| | (3.09) |
| | |
| Height_cm | 0.000426 |
| | (1.55) |
| | |
| age | 0.00858*** |
| | (3.89) |
| | |
| age squared | -0.0000486** |
| | (2.90) |
| | |
| | |
| Country d.v.s | |
| N | 13229 |
Marginal effects; Absolute t statistics in parentheses
(d) for discrete change of dummy variable from 0 to 1
* p <.05 ** p <.01 *** p <.001
7 comments:
Definitely interesting correlations, and for me surprising because I would have thought that the effects of cognitive ability are only important for the pairing aspect, where people tend to marry approx. within the 'same league' and GE effects also kick in, everybody cannot marry above them. I think this was explored in one of your other posts(?). Perhaps that's why you're not getting a stat sig result on the height, or the effect is collinear with the other measures?
What happens when you regress height on cognitive ability, yrs education, income et al? I think that's when you would see the biological superiority coming in!
Dr. Kev,.. you should look into getting a column to reach a wider audience with these insights.. interesting stuff.
just a thought - perhaps no more than a matter of semantics but given the major educational reforms that took place across Western Europe over the past fifty years should we be surprised that "cognitive ability" is out-pacing "years of education" in your regression? Would you expect to find the same strength in those variables today? ... could Mr O'Malley lend anything to the results here?
Alan: the GE possibility is interesting. Non-economists tend not to think about this. But not everyone marries so it could be that the poor schmucks who can't get a date are the not-so-smart?
In general height & ability are correlated (see Case & Paxson JPE), also height & income and health too - though the benefits tend to tail off: you don't want to be like 6'10.
Peter: Thats a smart idea, I might have to steal it. So in a world where education is supply constrained there may be a higher return on ability? The data here does not include Ireland, as it was only added recently & I haven't got round to adding it. Whether the sample is big enough to address that I don't know. Worth trying though.
I am hoping that the next instalment will use the GSS (or similar) to look at the effect of social values and risk preferences on the likelihood of marriage.
Working title: "I will do anything for love, but I won't do that."
Enda: another good idea that I may have to steal (what is it with the Meatloaf obsession around Geary?). The ESS would allow that I think & maybe the NCDS.
A possible snag is that the values would have been elicited ex-post so may be effect rather than cause. So lets say you find that people who are not married are risk-loving. Maybe its just regret that they didn't take the plunge? If personality, as measured, is stable perhaps its not a problem.
There is a website incidentally, www.soyouvebeendumped.com,
which lists reasons people have been given for being dumped. As I recall, some poor sod was told by his ex- that she "couldn't love a man who wasn't into the GAA". Better off without her, I say.
Dr. Kev,
You could extend your insights to the online dating market. Don't quote me on any of the figures below, btw!
There is a huge cohort in the online dating market. The number of persons aged 0-14 years reached a peak of 1,044,000 in 1981 (the so-called Pope's Children). This means that the largest cohort in the country is currently in their late twenties; an age when they are more likely than any
other to think about online dating.
According to ResearchandMarkets.com, the worldwide online dating market is worth 1 billion euro and online dating in Europe is expected to reach a worth of 650 million euro by 2012. The market for dating services in Ireland is comprised of 850,000 single adults aged 25 years and over.
I don't know much about on-line dating (honest). I have seen papers about speed-dating, perhaps I should discuss them with the one colleague I know who has engaged in this. People have also used data on personal ads from the newspapers.
There was an interesting paper, somewhere, that looked at the optimal balance of information in a personal ad. If its all about yourself, you look egotistical. If its all about what you are looking for ("WLTM Italian lingerie model and her sister") then it looks like you are hiding something. The optimal proportion is not 50% but I am afraid I cannot remember what exactly.
Incidentally, the example I gave above is one I actually saw in the Guardian once. No harm in wantin' I suppose.
Post a Comment