Do children need fathers to bring them up? David Cameron thinks they do. But MPs decided last week that children conceived by IVF don't. And this paper, which has just comes to my attention, suggests that the MPs are right and Cameron wrong.
Ian Walker and Yu Zhu confirm that there is "a large negative correlation" between living in a non-intact family and having low educational attainment.
But this is because there's a large negative correlation between being poor and having low education, and single parents tend to be poor. And Walker and Zhu found that, controlling for this removes almost all the link between single parenthood and bad educational attainment. They estimate that a child from a single parent home is 27 percentage points more likely to leave school at 16 than one from an intact family. But controlling for income reduces this probability to just 5.9 percentage points. And with the standard error of this estimate at 10.1 points, we can't rule out the possibility that the pure impact of single parenthood upon educational achievement is zero or even positive.
So, perhaps the problem isn't that dads leave home - but that their money does.
This corroborates my Bayesian prior - based on that most unreliable of evidence, personal experience - that if a violent or deadbeat dad walks out, kids gain from greater emotional security and the loss of a negative role model.
Those with different priors, however, might note that the coefficient on single parenthood and educational under-achievement is still positive, implying that it is probably bad for kids' schooling, but that the problem is rather that there's so much variation in the data that it's hard to be certain of anything. And you might also claim - ideally with other evidence - that educational attainment is not the only important metric of family success.
Or you might ask, as I did, why the evidence should matter. Maybe two-parent families are good for non-consequentialist reasons.
Ian Walker and Yu Zhu confirm that there is "a large negative correlation" between living in a non-intact family and having low educational attainment.
But this is because there's a large negative correlation between being poor and having low education, and single parents tend to be poor. And Walker and Zhu found that, controlling for this removes almost all the link between single parenthood and bad educational attainment. They estimate that a child from a single parent home is 27 percentage points more likely to leave school at 16 than one from an intact family. But controlling for income reduces this probability to just 5.9 percentage points. And with the standard error of this estimate at 10.1 points, we can't rule out the possibility that the pure impact of single parenthood upon educational achievement is zero or even positive.
So, perhaps the problem isn't that dads leave home - but that their money does.
This corroborates my Bayesian prior - based on that most unreliable of evidence, personal experience - that if a violent or deadbeat dad walks out, kids gain from greater emotional security and the loss of a negative role model.
Those with different priors, however, might note that the coefficient on single parenthood and educational under-achievement is still positive, implying that it is probably bad for kids' schooling, but that the problem is rather that there's so much variation in the data that it's hard to be certain of anything. And you might also claim - ideally with other evidence - that educational attainment is not the only important metric of family success.
Or you might ask, as I did, why the evidence should matter. Maybe two-parent families are good for non-consequentialist reasons.
"we can't rule out", but - using intuition, the numerous other surveys and research projects that reach the opposite conclusion, and our own experience - we probably should.
This isn't your own cognitive bias speaking, is it?
Posted by: Recusant | May 27, 2008 at 02:34 PM
This corroborates my Bayesian prior - based on that most unreliable of evidence, personal experience - that if a violent or deadbeat dad walks out, kids gain from greater emotional security and the loss of a negative role model.
And in what proportion of cases is that the case - compared with being stuck with an equally disfunctional mother with no alternative source of support?
Posted by: reason | May 27, 2008 at 03:33 PM
You might find this paper interesting as it finds that marriage is important for children (http://ssrn.com/abstract=1084562 ).
It's a paper by Marsha Garrison. My own view, based on this and other evidence, is that trying out social experiments for the benefit of adults where the guinea pigs are children is grotesque. When it turns out bad will those in favour of gay/lesbian adoption, or no father families, be prepared to say 'sorry'? And even if they do, who will repair the lives of the children who were the guinea pigs in this social experiment?
Posted by: tolkein | May 27, 2008 at 09:39 PM
"using intuition, the numerous other surveys and research projects that reach the opposite conclusion, and our own experience"
a and c are the last things we should be using (since intuition and generalisation from personal experience are both more-often-than-not wrong when scaled up to macro level); and the studies in b only reach the opposite conclusion when income effects are not taken into account.
Iceland provides some evidence in support of Chris's hypothesis - it has the highest rate of single parenthood in Europe, a strong welfare state combined with strict enforcement of child support so that single-parent-family kids aren't economically disadvantaged, and the happiest people in Europe.
http://www.guardian.co.uk/world/2008/may/18/iceland
Posted by: john b | May 28, 2008 at 11:34 AM
Yes single parents are disproportianately likely to be poor, but that is an argument for reducing the incidence of single parenthood, not incentivising it by giving them even more tax payers money !!
Your argument is like saying that because smokers are more likely to die of cancer, we should give them all free cigarettes, rather than helping them to stop smoking.
Economics aside, there are a wealth of srudies, going back to Bowlby in the 1950s to show that on almost every conceiveble psychological measure, outcomes for children, especially boys, are damaged by loss of contact with a father. It even holds true if the father is physically present but mentally absent i.e alcoholic/drug addict/mental illness. Read Banduras research from the 1970s on agression, as a starter.
If kids weren't supposed to have fathers, nature wouldn't have invented them......
Posted by: Matt Munro | June 02, 2008 at 06:05 PM