Well, yes, obviously

Data from the first two waves of the Fragile Family and Child Wellbeing study indicate that infants who look like their father at birth are healthier one year later. The reason is such father–child resemblance induces a father to spend more time engaged in positive parenting. An extra day (per month) of time-investment by a typical visiting father enhances child health by just over 10% of a standard deviation. This estimate is not biased by the effect of child health on father-involvement or omitted maternal ability, thereby eliminating endogeneity biases that plague existing studies. The result has implications regarding the role of a father’s time in enhancing child health, especially in fragile families.

Homo sapiens sapiens is one of those species where paternal investment in offspring is important.

Thus why the kafeeklatch of women surrounding the new mother continually bray “How much his father” he looks. Whatever the reality or the evidence.

It might also be a hint as to why serial motherhood isn’t all that good a solution. Despite the potential advantages of having children with a greater genetic mix.

44 comments on “Well, yes, obviously

  1. “An extra day (per month) of time-investment by a typical visiting father enhances child health by just over 10% of a standard deviation.”

    Anyone whose pre-existing assumptions are not supported by this datoid would point out that that’s basically no measurable difference.

  2. What’s their opinion of the high divorce rate and right of women to kick their husbands out of the house whenever they fancy a new bit of stuff?

  3. 0.1 SD is noise, unless it indicates something, and unless TEN extra days per month of fatherly involvement translates into 1.0 SD or something near it. Stimulation by an adult male (or female, but single mothers are often too busy) is better for a developing mind than stimulation by street noise, siblings, television, or lack of stimulation at all. One of the gifts of social welfarism is to empower mothers to rear their babies in a relative vacuum of stimulus.

  4. abacab,

    If you precision-engineer 1000 identical widgets you might reasonably expect those widgets to perform identically, within ± 0.1 σ, if not within an even narrower band. Humans are, however, not precision engineered, thus measurements of their biology and psychology are subject to wide variations at baseline, the δ due to treatment can vary drastically, as can the δ due to random effects over time.

    This is why, other than when desperate for a publication, biologists, psychologists, and so on, tend to ignore 0.1 σ effects (today’s random example: the regular declarations on here that blacks are inferior because IQ tests says so), in the absence of convincing and corroborating evidence. On top of which, some hint that the data has some value for or says something useful about individuals in the group, rather merely being something you can just about measure given a big enough group.

  5. “Random” should read “uncontrolled”.

    Talking about lack of control, how many absentee dads go “my child is healthy I shall put more effort in”, and how many go “my child is sickly, I shall put less effort in”. Hard to know which way the causality goes when it could be either direction. Or both simultaneously.

  6. BiG, do I infer that you state that the black v non-black IQ difference is too small an effect to be significant? I happen to be interested in this topic at the moment so if you had some further thoughts or a link to something I’d be very grateful. Cheers!

  7. Spike, to be fair to sciencedirect they are just the platform for the journal, and they probably don’t have editorial control. Elsevier has an absolutely vast stable of titles, thousands and thousands, some of them society or privately published, some are Elsevier-managed. Most of them are second tier at best, proved by the fact they have even published my drivel in the past!

    Almost everything they publish just vanishes unread into the borg of academe.

  8. @Marc,

    It’s not so small that it “isn’t significant”, since it is, measurable (if small) and repeatable. The problems are: what the hell are we measuring? How well controlled is it really? And most importantly: what use is a group mean to any individual in the group? The SD for any group is much wider than the difference between groups, so the actual real-world value of this information is basically zero.

    That of course isn’t what the racists want to hear, all they want to hear is “Science has proven Blacks are thick”.

  9. That’s a misunderstanding, BiG. Let’s say that blacks are better sprinters than whites, but that “the SD for any group is much wider than the difference between groups”. It still means that the finalists in the Olympic 100m are all black, because they’re all at 5 SDs (or something similar) above the mean. It’s not the average person that drives societal progress, it’s those that are well above average.

  10. An extra day (per month) of time-investment by a typical visiting father enhances child health by just over 10% of a standard deviation.

    Could well be correlation rather than causation. A father “investing time” in his kid is probably doing a whole lot of other stuff as well, which is what really makes the difference.

  11. “Let’s say that blacks are better sprinters than whites, but that “the SD for any group is much wider than the difference between groups”. It still means that the finalists in the Olympic 100m are all black, because they’re all at 5 SDs (or something similar) above the mean.”

    The problem with that is that it uses one characteristic as a proxy for another, only marginally related characteristic.

    Being good at sprinting is due to a cluster of genes related to oxygen transport, glycogen storage, and fast muscle growth. These genes have nothing whatsoever to do with skin colour. But the number of people with the sprint genes is small and geographically constrained, and happens to be in Africa, so the people with those genes all happen to be black. But most black people don’t have those genes, and there’s no reason why white people couldn’t acquire them by interbreeding.

    The scientific problem with racism is that it tests the correlation of every other characteristic with race, irrespective of whether there is any causal link. Science is about finding what the cause of an effect is. Irrelevant correlation studies just muddy the waters.

    IQ is really about poverty, education, and culture – and is frequently associated with recent immigrants to a new country. It takes a while to assimilate, be accepted, and to learn how to succeed. Everyone knows the jokes about the Irish being thick – but that’s because at the time the Irish were generally poor and immigrant. The genetic differences are not relevant (red hair has nothing to do with brain growth), and they had white skin. It just so happens that in America and Europe black people are more likely to be immigrants and poorer, and that impacts their educational attainment.

    It’s always a problem when in order to target characteristic A, we instead pick on a more easily identifiable characteristic B and treat them as if they were identical – every person in group B is assumed to be in group A; every person not in B is assumed to be not in A. And people not in group B can congratulate themselves on not being in group A without having to prove it. It’s just the sort of thinking that would, ironically, cause you to fail an IQ test.

    I’d be interested to see some surveys of the IQs of racists, compared to the rest of the population. I don’t know, but I suspect the results would not be congenial to certain parties. 😉

  12. Chris, have you noticed the preponderance of 100m winners and marathon winners are from different parts of Africa? Are they two different “races” (double entendre intended)?

    See Nov for an explanation of why race isn’t a useful marker of anything. It isn’t meaningfully definable.

  13. ” infants who look like their father at birth are healthier one year later…”

    Ahem: ” infants where the father is in residence and supporting the mother are healthier one year later…”

    ftfy.

  14. I’m not arguing about whether there are ‘racial’ (or whatever PC terminology you wish to use) variations between groups of humans. I’m merely pointing out the egregious nonsense that causes you to say: “The SD for any group is much wider than the difference between groups, so the actual real-world value of this information is basically zero.”

    Let’s suppose (for the sake of argument) that IQ is real and that some group A consistently tests slightly higher on average than some other group B (the groups being of equal size for this hypothetical example). If I’m told that “this individual has an above average IQ”, I’d be betting that they came from group A, even though the odds of that being the case may only be 51% (or even 50.01%). But if I’m told the individual has an IQ of 150, I’d be much more confident that they come from group A (although there is still a chance that they’re from group B).

    For me, such differences are a significant factor in explaining why there aren’t many white faces in the final of the 100m, however much that fact upsets you, and why there aren’t many black faces at the Nobel awards.

  15. @ BiG
    Yes, they are different races, just as Hungarians are different from Bretons and Basques are different from both.
    I will discuss the subject in a separate post.

  16. biologists, psychologists, and so on, tend to ignore 0.1 σ effects (today’s random example: the regular declarations on here that blacks are inferior because IQ tests says so)

    Except black/white IQ differential is on the order of 1 σ, not the 0.1 σ you posit. Agreed that 0.1 σ differences are generally noise, of course.
    To the subject at hand, I remember reading (Dawkins, maybe?) that relatives of the mother were much more likely to comment on how much a new-born looked like the (putative) father, while his relatives were much more suspicious of the paternity, or at least, less likely to comment on a supposed resemblance.

  17. @ dcardno
    Your last comment may reflect propaganda on behalf of the mother’s relatives but may just as well reflect their noticing the differences between the baby and the mother, with whom they are more familiar, than the father,with whom they are less familiar.

  18. @BiG. Intelligence isn’t a tournament? I normally find your comments insightful but that one is woeful. Sure, luck plays a part but in general intelligent people earn more than less intelligent people let alone people below the mean. Intelligent people (in general) have more stuff, lead healthier longer lives and do more with them. You can’t do many things without sufficient intelligence, like competing to be a surgeon for example. Life is a tournament that lasts decades, and you say intelligence doesn’t matter!

  19. @ Gamecock
    0.1 of a Standard Deviation is not necessarily “statistical noise” – it is just not “statistically significant” at the 5% level which idiots believe to be the only determinant to separate “noise” from “real differences”.

  20. The whole discussion is irrelevant to the normal family as it is only about “visiting fathers” who are a tiny minority.

  21. @ John77

    …may just as well reflect their noticing the differences between the baby and the mother, with whom they are more familiar…

    Yes, that hadn’t occurred to me, but it’s a possible explanation. My observation is that a newborn rarely looks like anybody – but friends and family feel compelled to search for and comment upon supposed similarities.

  22. Take two groups of people with some measurable characteristic (IQ, running ability, height – whatever you want). Question 1, why should the averages be the same, Question 2, why should the standard deviation be the same, Question 3, why should the variation within either group conform to the normal distribution?

    What happens if SD(a) = 10SD(b), in which case 0.1SD(b) is significant to population (a), isn’t it?

  23. “An extra day (per month) of time-investment by a typical visiting father enhances child health by just over 10% of a standard deviation.”
    I see a lot of complaints about this. Of course it’s not statistically significant with a sample size of one. But with a sufficiently large sample it is both significant and measurable, i.e. real, albeit probabilistic.

  24. @ ND Reader
    “Statistically significant” at what level?
    “Statisically significant” is JARGON that I use without thinking but prefer to explain when talking to non-statisticians. At the 10%/5%/1% level it means that this would only happen 10%/5%/1% of the time if the “null hypothesis” was true – so you should expect it to happen 10%/5%/1%of the time – not *never* which is what some journalists imply.
    A differential of 10% of a Standard Deviation is “probably” showing a real difference.

  25. “Mine is that they all look like Winston Churchill.”

    Absolutely! I have a picture of my son at 8 months in my wallet, and he looks EXACTLY like WC!

  26. Ignoring the whole 0.1 SD malarkey, can I just add that that whole kaffeeklatch thing is, frankly, fucking terrifying.

  27. @Tim Newman, Gamecock

    “Mine is that they all look like Winston Churchill.”

    That makes three of us who have all independently arrived at the same conclusion. My only addition is that when we leave this life we also all look like Churchill. It’s a certain circularity (at the 10% confidence level).

  28. Bloke in Germany –

    In the annals of really bad posts from BiG this is really very bad. Spectacularly bad in fact.

    “It’s not so small that it “isn’t significant”, since it is, measurable (if small) and repeatable.”

    So in other words is not an example of what BiG was claiming – it is not an example of an effect that is so small that it is insignificant. In fact it is a solid finding that has been found in pretty much every study ever done. And it is a large effect as well.

    “The problems are: what the hell are we measuring?”

    So he opts for the second fall back position – IQ means nothing. Who knows what we are measuring. Perhaps IQ tests measure something, perhaps they do not. What we do know is that they have an incredibly strong link with the real world. That is, a truck driver who has a higher IQ tests than the other truck drivers is invariably a better truck driver. IQ tests remain the best predictor of college results. So whatever we are measuring, it is important.

    “How well controlled is it really?”

    Third fall back position. Which BiG does not really develop because he is just flailing about wildly to avoid drawing attention to his mistake. How well controlled is it? Who knows? It is possible, I suppose, that the literally thousands of tests and studies done on this topic are all poorly designed. Not likely, of course, but it is possible.

    “And most importantly: what use is a group mean to any individual in the group? The SD for any group is much wider than the difference between groups, so the actual real-world value of this information is basically zero.”

    And so we come to what is usually the last fall back position – whatever they test it is meaningless. It is true that anything we know about the group is fairly meaningless for the individual. However there is real-world value in a group result. After all, why should we be taking immigrants from the Caribbean and Africa? That individual may have a high IQ and so fit into British society somewhere above the level of a drug gang. But the chances are his children and grand children will revert to the mean so we will get the average in the long run. So we should have a sensible immigration policy that does not saddle future generations with unsolvable welfare and crime problems. Simple really.

    Also, of course, it is true. So it is important whether it has any useful real-world application or not.

    “That of course isn’t what the racists want to hear, all they want to hear is “Science has proven Blacks are thick”.”

    Which is pretty much what science has done. Not all Blacks. But then if you did not have a strawman you would have nothing.

  29. NiV – “The problem with that is that it uses one characteristic as a proxy for another, only marginally related characteristic.”

    That is not a problem. And it is not so much a proxy as a related cluster. But whatever. Blacks have genes for their skin colour. They also have genes for the muscle growth of interest. Where is the problem?

    “But the number of people with the sprint genes is small and geographically constrained, and happens to be in Africa, so the people with those genes all happen to be black.”

    But more importantly there is a cluster of genes that is racist. We are not actually all the same on the inside. Once we have established that, we can go on to talk about whether we are different in other areas as well. But we are now all agreed – science is racist, because race is actually a real thing and the race differ in some significant ways?

    “But most black people don’t have those genes, and there’s no reason why white people couldn’t acquire them by interbreeding.”

    Sure, if we abolished White people, there would be no more racial differences. Black people would still have those genes. It is just that some of them would have White relatives. See Barak Obama.

    “The scientific problem with racism is that it tests the correlation of every other characteristic with race, irrespective of whether there is any causal link.”

    Oh. My. God. People *test* things! The horror.

    “Science is about finding what the cause of an effect is. Irrelevant correlation studies just muddy the waters.”

    Not since World War Two it isn’t. Science is about deciding what facts the public should be allowed to know. On race especially. Science has been engaged in a three generation long campaign to suppress facts, bully anyone who says otherwise – and assault anyone who even comes close to suggesting something that might say otherwise. E. O. Wilson for instance.

    Even putting that aside, the first step is to find a correlation. Then you go about finding a cause. As Kuhn’s paradigm shift argument goes, first the odd evidence of something being wrong builds up, then someone explores, then a new paradigm is created. So this is science as she is supposed to be done.

    “IQ is really about poverty, education, and culture – and is frequently associated with recent immigrants to a new country. It takes a while to assimilate, be accepted, and to learn how to succeed.”

    How are Black Americans doing then? Been there since the 17th century. Is that long enough to assimilate? Oddly enough a Chinese immigrate fresh off the boat is likely to have no problems assimilating, being accepted and succeeding. They are also likely to test higher on IQ tests. Go figure.

    “Everyone knows the jokes about the Irish being thick – but that’s because at the time the Irish were generally poor and immigrant.”

    And not because they were a non-English speaking colonial population that had every reason not to cooperate with the British authorities? Interesting.

    “It just so happens that in America and Europe black people are more likely to be immigrants and poorer, and that impacts their educational attainment.”

    Is that true in America? There are a lot more people from Africa but I would hazard a guess most of them have been there for a while. Also, of course, the average IQ in Africa is about 60-80 and they sure as hell have been there a while and so have no problems with immigration.

    But I agree, or at least I hope, it is mostly about education. The problem is that we have tried everything and nothing has shifted the needle a single inch. Nothing we do can improve Black educational outcomes. It sure looks genetic.

    “I’d be interested to see some surveys of the IQs of racists, compared to the rest of the population. I don’t know, but I suspect the results would not be congenial to certain parties.”

    Good for you. Anti-racism is mostly about signalling that you are one of the Elite. Thomas Sowell’s Vision of the Annointed. And here you are. Showing us all what a fine upstanding Upper Middle Class urban Liberal you are.

    Although it is odd because if this subject was your favorite idee fixe, Transsexuals, all this doubt and nuance would be out the window. The slightest shadow on am MRI is, for you, absolute proof. But 100 years of IQ tests is meaningless.

  30. Why is any of this a surprise? I don’t believe they can measure the impact of a single day’s visit, but we know that the children of divorce do worse in virtually every outcome:

    https://en.wikipedia.org/wiki/Effects_of_divorce

    In fact divorce seems to kill the children five or so years early.

    So obviously reducing a father’s time with his children is going to have a worse outcome.

  31. Andrew, a tournament is a game with one or few winners, the reward for marginal superiority is disproportionate. It’s why a premier league striker earns so much more than an almost as good player in the championship.

    Now this certainly applies to some more mundane jobs like f500 executives compared to those not quite there, but your iq is far from the sole (or sometimes even relevant) cause of the difference in outcome there.

    Chris, I cannot predict someone’s probability of winning the London marathon from their race. Nor their iq. There will be one winner of the marathon, most likely an east African, but hundreds of millions of east Africans will not win it. So race us also here a useless indicator of performance.

    At the end of the day I don’t care about acknowledged or suppressed group differences, my classical liberal worldview is that there are individuals. There is no such thing as society, as a former pm revered here said. But it is amazing how self-described libertarians are well into identity politics for others, for the “out groups” while defending their own individuality.

    In other words someones race is not a reason to treat that person as anything other than the individual they Are, whatever the group mean of various metrics you care for.

  32. Mine is that they all look like Winston Churchill.

    But isn’t it strange that Winston Churchill was never (to my knowledge) described as baby-faced?

  33. @BiG someones race is not a reason to treat that person as anything other than the individual they are
    Please point out where I ever suggested otherwise.

    There will be one winner of the marathon, most likely an east African
    By George, he’s got it!

    Now this certainly applies to some more mundane jobs like f500 executives
    So all those campaigning for equal representation of races and genders in such positions should STFU.

  34. Bloke in Wales – “But isn’t it strange that Winston Churchill was never (to my knowledge) described as baby-faced?”

    Gary Oldman did so in recent PR nonsense he did for his film on Churchill.

    I doubt he thought it up.

Leave a Reply

Name and email are required. Your email address will not be published.