Blood Will Tell

Most psychological traits are significantly, although not overwhelmingly, heritable.  So if you are given information about the psychology of close relatives of some person, you know more than you did before about that person. On average. If his kids, sibs, parents are all extremely smart ( or extremely dumb, or extremely crazy), you have a better estimate of his true nature than you did in the absence of that information about relatives.  For that matter, group membership will also alter your estimate, if the two groups in question are fairly different.   In some cases, it can make a big difference: if someone tells you that a Pygmy is 6′ 6″  ( and not a pituitary case)  you would assign that a very low probability.

This entry was posted in Uncategorized. Bookmark the permalink.

41 Responses to Blood Will Tell

  1. pyrrhus says:

    Teachers have long been aware that the apples falling from the tree tend to be tightly grouped….

  2. ASR says:

    Regression to the mean is a related issue. Parents often warn their children against marrying the good kid from a bad family because “blood will tell”. My personal observations suggest that generally speaking it does.

  3. Coagulopath says:

    This shouldn’t even be politically controversial. Blank slatism predicts the same thing!

    Family members share similar environments (or at least more similar than those of two random strangers), so even if genes had no input on anything you could predict stuff about me by looking at my brother.

  4. ohwilleke says:

    It is rarely enough information to be very useful, and overconfidence about what to expect without seeing it first hand in an individual can do more harm than humility about what you know.

    It is too easy to think of counterexamples.

    The highly introverted kid with an extremely extraverted sister, mom and dad (an example from my extended family). A bipolar child in a family where no one else seems to have mental health issues (at least to the outside world which I see a lot in estate planning). One sibling likes to debate, another avoids conflict at all cost and listens rather than speaking. There is the child of extremely academically able parents with a severe learning disability. Serial killers and psychopaths rarely have close relatives who are either, even though they may have other issues that came together horribly in that one individual. Leadership rarely runs true. I’ve seen multigenerational ADHD, but I’ve also seen extremely ADHD individuals whose close relatives show no sign of it.

    How traits blend from multiple diverse close relatives can be surprising. You don’t see it a lot in the U.S., but in India, it is not at all uncommon for siblings to have very different complexions. It isn’t uncommon for one mixed race child to pass for one parent’s race and another mixed race child to pass for the other’s. I have a son who is 6’0″ (and weighted exactly what I did when he was 18) when I am 6’1″, even though his mother and maternal aunt are 5’0″, his maternal grandmother is 4’11”, his maternal grandfather isn’t more than 5’6″, his sister is 5’5″, and my parents were only average in height. I have one kid who is thin as a rail and another who was always obese.

    Admittedly there is observer bias. Insiders notice how close relatives differ. Outsiders notice how they are alike.

    Often enough the differences aren’t really nurture, they’re just natural variation that didn’t manifest phenotypically in the other relatives.

    If anything, it is the nurture driven traits (religion, language dialect and accent, tendency to be a hugger, abusive interpersonal skills, fashion preferences, social class cues) run more true in a reliable and predictable way than the genetic ones which can surprise you.

    This isn’t to deny hereditary factors. They are significant and are stronger than most people naively believe them to be. Smarts parents tend to have smart kids. Kids who have strong music talents often have close family members who do as well. In some families, no one is very talkative and in others family gatherings are a bubbling brook. But the correlations are weak enough for many psychological traits that heredity isn’t a reliable compass for most purposes.

    • Rob says:

      Skin color is not like height/IQ, where thousands of alleles each contributes a tiny amount to phenotype. With skin color there are some single genes with large effects. With one bipolar or adhd person in a large psychologically healthy family, you might be looking at a de novo mutation with a large effect. That bipolar girl might be the only person in the world who is bipolar in that particular way. Though maybe other people have different mutations in the same gene.

      Most mutations are detrimental to phenotype, but we live in a evolutionarily new and bizarre environment, so what selection picks up and amplifies might be pretty novel from what most people were like in the past, or are still like now. I’m not saying that everyone weird is a mutant, or we’re all going to be crazy, but some bipolar women have a lot of kids, whereas they’d have been institutionalized a couple generations ago.

      I’ve thought it would be interesting to look for single alleles of large effect on IQ by looking at kids who are a lot smarter than their parents and siblings. Especially if the very smart kid is weird or even crazy. It could be that he has a new mutation that pushes his phenotype far from normal. This may be a population of zero, but if there are families where four kids have IQs around 80, but one kid has a 120 IQ, that could be a single gene of large effect. Sort of like organic vs familial retardation, but the other way. I’m sure they are rarer than people with retardation fom de novo mutations.

      Greg, is that idea crazy?

      • Space Ghost says:

        This may be a population of zero, but if there are families where four kids have IQs around 80, but one kid has a 120 IQ, that could be a single gene of large effect.

        Might be easier to test the milkman’s IQ first.

      • Coagulopath says:

        but if there are families where four kids have IQs around 80, but one kid has a 120 IQ, that could be a single gene of large effect.

        We would have found this gene in a GWAS by now.

        • Rob says:

          Would we, though? If there’s the one kid in the family with 120 IQ in college, he might not want to draw attention to his ‘loser’ family while he attempts to climb socially. Furthermore, the one smart kid is either in his home town helping out his family, where he woukdn’t be participating in GWA studies at the big city university. Or he is at the university, fairly far from his sigblings, and trying to fit in.

          I don’t think this is something that you’d find just GWASing convenience samples. Even twin studies are more likely to have middle class and u people from stable families participating,

          I think this is a phenomenon one would have to try to find. Not one that just falls into your lap,

    • gcochran9 says:

      Sure, you are denying the practical utility of considering hereditary factors. You are advocating ignoring them – which, on average, will give less accurate answers.

      I didn’t say that it tell you everything: I said it tells you something. And it does.

      Generally our social policies are based on something roughly similar to what you advocate, and it causes loads of trouble. Because false.

  5. ohwilleke’s comment about observer bias is important. Insiders in a family will see the individuals as different while those outside will see the similarities. I have five sons. I think of my three adopted children as very different, but the two who are also blood siblings seem more similar when I consider them against the group. The two who are biologically hours are quite similar, even considered twins at times though they are four years apart. The fifth son is biologically a nephew and seemingly different from them – unless contrasted with the two Romanian brothers. Then he seems like my first two.

    Working at a hospital, I see staff and patients from many nations. There are some general commonalities among the Slavs, among the sub-Saharans, and among the dot Indians, though when considered only against each other they are quite different. Coagulapath’s comment that this could just as easily be environmental has some merit, though. One needs large samples to start sorting that out.

    These days we have large samples. The James Lee st al study in Nature (the number of authors is a larger N than many experiments) about ancestry and educational attainment has an N if 1.1 million .

  6. gabriel alberton says:

    They do not need to have relatives who are all and extremely smart, dumb, or crazy, as that doesn’t happen very often. If someone has several relatives (doesn’t even need to be most of them) who are significantly above or below average in a given quantitatively measurable trait, that might be enough for an estimate to be made about them regarding that trait. It’s up to the observer to define ”significantly”, but if the values are distributed normally, more than one standard deviation of difference could already suffice if environmental influence is judged to be low or if it’s similar in all involved populations.

    Having a 1.80 m tall pygmy (that’s around 5’11” in americanese) and who has a similar diet relative to other pygmies could already not be very ordinary (if their average height is actually under 1.50 m), and having several pygmies of a similar height in a pygmy family might strongly suggest something’s up (even if it is something like ”they have a higher incidence of pituitary tumours”). If the observer instead judges that no, 1.80 m is not tall enough, that he’d only find 2.00 m pygmies worthy of note — well, he could be right, but chances are he threw away an opportunity to learn what makes those pygmies taller than average (assuming that’s worth finding out, of course. Not all knowledge is created equal).

  7. shadow on the wall says:

    So if you are given information about the psychology of close relatives of some person

    Of course, in modern society, you are not given this information. In modern society, tightly knit extended families living together at the same place for hundreds of years are as common as horse drawn buggies.

    The last time the Census Bureau calculated this was in 2007, when it found that a typical American will move 11.7 times in their lives. We redid the math using the most recent data (2013 for mobility and 2010 for population estimates) and reached a slightly lower number of 11.3 lifetime moves.

    • gcochran9 says:

      From this I guessed that the reproductive success of people carrying a load of schiz genes is higher now than it once was. Back in the old days, they knew your family had a crazy streak.

      Hundreds of years is not really necessary. And good local knowledge does not require complete immobility: I remember family tales of the 11-mile trek to the county seat.

      • shadow on the wall says:

        In these days, things are not as bad as they used to be.
        If you wanted to make deep background check of family of your prospective marriage partner twenty years ago, you would have to travel to California where mom lives with her second husband, then to Florida where dad lives with his fourth wife, then to find siblings and cousins strewn all around North America and beyond (and they will anyway tell you to fuck off).

        Today, you can just sit home and check their fakebook pages. Social media are bad replacement for old time village gossip, but they are better than nothing. What airplane takes, internet gives back away 🙂

        Of course, nobody is carefully selecting future partner like one carefully picks what refrigerator or washing machine to buy – people just “fall in love” and marry (when love ceases, divorce and repeat the process).

      • cameron232 says:

        Bruce Charlton (if I’m understanding him) has suggested that genes for social adaptation are the most delicate, recent, complex adaptations and are, thus, the first to show mutational issues in our everybody-can-reproduce” society. So more crazy people than elephant men running around.

  8. dearieme says:

    Your hope, dear blogger, that the internet censors won’t spot that this thread is about the Bedins, father and son, is refreshing. But is it realistic?

  9. William says:

    Nice to hear about “blood” again, it has been years. “Bad blood” in a family was a screening tool years ago and as you say, apples not falling far from the tree was also considered a reasonable inference. Looking back, in the two families I heard labeled as having “bad blood”, I would guess some form of schizophrenia was the indicator. For these two , of the twelve off spring over the past two generations, there is one diagnosed schizophrenic, one diagnosed autistic, one in special needs class (mental health) and one suicide. Just an observation.

  10. pyrrhus says:

    In my experience teaching hundreds of kids, the personalities of kids from the same family will often vary quite a lot, even in identical twins….but the intelligence will not vary much…Probably because intelligence is the product of thousands of genes….But teachers are pressured by the bureaucrats to pretend that everyone is the same…

    • Coagulopath says:

      In my experience teaching hundreds of kids, the personalities of kids from the same family will often vary quite a lot, even in identical twins….

      Remember that young people are more susceptible to environmental factors. As we age, most of these effects fade.

      In the 60s there was a stereotype (which you see depicted in The Wonder Years) that the firstborn child of a family would get swept up the social change happening at the time. Woodstock, drugs, radical politics, and so on.

      The second-born child would watch their older sibling’s travails, think “nope”, and cut their hair and enroll in business school. Totally different life paths.

      But those hypothetical kids are now 70. Are they still total totally different people? Or are they now more alike than they are different? I don’t know. But when I talk to a baby boomer I never get a sense that I’m talking to a former hippie. They all seem pretty much the same.

  11. Cpluskx says:

    Off-topic q: Could it be dangerous for the West Eurasians to use the covid vaccine made in/tested on East Asians (maybe Indians too) because of genetic differences?

    • gcochran9 says:

      It is possible for different populations to have somewhat different reactions to a vaccine. Doesn’t seem to happen much, though.

    • adreadline says:

      Isn’t the genetic difference between Europeans and Chinese rather small compared to the difference between Europeans and black Africans, and the difference between Chinese and black Africans even larger than that?

  12. The math is nice and simple if you can restrict yourself to Gaussians.

    For Joe, you can dig up old complaints about him lying about his education, where he was in the 25th percentile at U Delaware and the 10th percentile at law school. Combined with estimates of fraction of men who went to college, I’d peg him at ~75th percentile in the population a priori. Let’s say 110+/-5.

    For the sons, not much public information, BAs in 91 and 92 from UPenn and Georgetown. I dug up old commencement programs and neither of them are listed in any of the honor societies, so somewhere between 0-90th percentile at each institution. So a lot noisier. Both the kids went to harder schools, but with way less information with how they did there. Let’s say 60th to 95th percentile for both of them, or somewhere around 105-125, or 115 +/- 10.

    And hey, we’ve even got a half-sib, which makes it easier to make inferences about the shared parent. Little sister did cultural anthropology at Tulane. So that’s illuminating. Also apparently into cocaine and marijuana – the Post did an expose on her back in 2009. Very little info on her, so I won’t try to nail it down more.

    I was going to say to be fair with the drug/depravity stuff that his son had nasty head injuries when young that might well have caused problems, but well, the half-sib has the same issues to all appearances.

    • gcochran9 says:

      Not only does no-one take advantage of this, any attempt to do would be condemned as naughty, by the usual idiots.

      • arguably wrong says:

        I recall a conversation with Phil Lemoine, where he was wildly overestimating the effects of group priors, in an attempt to be deliberately contrarian, I suppose.

        Take a member of a group with mean 100 and stdev 15, take a measurement of 130 with stdev of 5, and your estimate of the true value is 122.5.

        He was arguing that since the variance in men is higher than in women, the estimates for men and women would be wildly different. But it’s not true: take a group stdev of 16, and your estimate goes up to… 122.9. Go all the way up to a measurement of 160, and the difference in the estimates is almost a whole integer. It’s just not a significant effect here.

        Now, do the same calculations for groups with different means, and you’ll get more of a difference: For a measured value of 130 with measurement error 5, you get:
        Mean 100, stdev 15: 122.5
        Mean 85, stdev 15: 118.8

  13. ERIK says:

    Regression to the mean is acctually regression to the mean of ancestors and if the trait is known for several generations this information should say more than a population average. Nearer ancestors should reasonably have higher weight. Perhaps a reduction by half for each generation as it would give a nice symmetric quality.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s