November 25, 2013

Children of divorce, an unseen historical trend: Is it related to status-striving?

Typically when you look up information on divorce rates over time, you'll see divorces per capita and divorces per marriage, within a given year. Graphing them produces the following:


No matter which rate you choose, there's been a secular increase since the mid-19th C. Apart from that, there's an apparent burst from the early '20s through the mid-'40s, a dramatic recovery through the late '50s, another surge during the '60s and '70s, and another dramatic decline since then.

This cycling allows for a half-full or half-empty debate about the state of things recently and into the future. "Sure, that surge during the '60s and '70s was bad, but it's gotten much better since. Maybe it'll surge again, but who can say?"

No such debate can take place, though, when it comes to a more important statistic -- how likely is a child to grow up without both biological parents? All those stats you read about divorces, marriages, risk of divorce for a given marriage cohort, etc., ignore the main worry that the public has about the topic of divorce -- namely, children of divorce.

The General Social Survey asks a question about who you were living with at age 16. I'll look only at whites to make the point stronger. In the graph below, I've recoded responses into living with both parents (red), living with mother with or without a stepfather (blue), living with father with or without a stepmother (green), and other responses (yellow). I've grouped people into 4-year birth cohorts starting with 1900, which you probably won't be able to read without clicking to make the image larger.


There's a tiny rise in children growing up in intact families, going from those born around 1900 to those born around 1950, but it's only going from 77% to 80%. But it's basically flat at a high level for that entire period. And remember, the question asks about their lives at age 16, so 16 year-olds were still highly likely to be living in intact families into the early '70s (the '56-'59 cohort is barely distinguishable from '40s births).

Then there's a sharp drop with those born in the early '60s, who would've been 16 in the late '70s, and keeps dropping after that, although it may have stalled out at the bottom with Millennials. (Ignore the last cohort, though, since the sample size is tiny.)

So, unlike the picture of divorce rates, which showed a dramatic reversal around 1980 and lasting through today, rates of children of divorce have only gotten worse since those born in the early '60s.

When people argue about whether the divorce problem is getting better or worse, they would seem to be talking past each other. The people who point to the first graph are talking primarily about the fragility vs. stability of the marriage bond, whereas the folks who respond with, I don't care what that graph says, it's getting way worse, probably have in mind the disruption to family cohesiveness.

I put myself in the latter -- marriage is important, but kids growing up in intact families is more. Marriage vs. divorce is about whether two egocentric adults, the husband and wife, can reconcile their differences and stay together. Once kids come into the picture, the egocentric adults are no longer simply husband and wife but father and mother. They have even more other people to take into consideration, and these new ones are defenseless and impressionable to boot. Time to stop thinking only of your own goals.

What's behind these changes? I'm tempted to point to the status-striving / inequality cycle. If some problem started to get bad by the mid-to-late 1970s, and has only gotten worse since, that's a likely culprit. (It's clearly not the cocooning / crime cycle.) We should then see rates of intact families rise during the Great Compression of roughly 1920 to 1970. Or, going from births around 1900 to around 1955. It's there, but very slight. The pattern is a little stronger by adding the red and blue bars together -- that's growing up with at least your mother (both parents, mother alone, or mother and stepfather).

You'd think rising inequality would make mothers and fathers stick together more for the benefit of their children, if they were worried about their offspring doing well in the future. And falling inequality should relax that, with one parent saying, "Meh, it's hard not to earn a decent standard of living these days -- they'll manage."

Therefore, it must have to do with the status-striving that produces inequality to begin with (a la Peter Turchin). One parent starts to feel like they could do better for themselves by splitting off from the other parent, regardless of who gets the kids. This constant monitoring of your prospects for trading up smells a lot more like status-striving. Or, it may not have to do with wanting to leave in order to re-marry or date someone with greater economic resources than your current spouse, but with the drive toward conspicuous consumption and leisure -- both far easier when you aren't held back by the old ball and chain, nor those darn kids who need your time, effort, and money. Whatever the causes, they all seem to stem from status-striving rather than inequality.

Historical research even farther back would be tough because divorce was a much more serious economic disruption before the welfare state. You might look at how common orphans and abandoned children were, though. Those could always reflect parents who just didn't have enough money to raise them -- but then that's always relative to how much the parents feel they need to spend on themselves. "I just wouldn't have enough to buy diapers every month and still shop for groceries at Whole Foods," or whatever the equivalent was in the Gilded Age. And then again, maybe the children were abandoned because one of their parents was too obsessesed with status-striving to rear their own offspring.

Our vague impression of Victorian England and the Gilded Age in America is full of children who don't live with both mommy and daddy --  all those orphaned and abandoned characters from Dickens, and Tom Sawyer and Huck Finn. But some numerical data would be nice to confirm that impression.

I'm becoming more convinced that conservatives who apologize or cheerlead for the Victorian era don't have a clue what kind of society they're defending, except for the cynical rats among the elite who want another Gilded Age where they're on top of a steep pyramid. Below are a few signs that the Gilded Age didn't care hardly as much as the Fifties did about keeping families together. The scenes are of New York in the 1880s, from How the Other Half Lives by Jacob Riis.




GSS variables: family16, cohort, race

47 comments:

  1. Who the fark leaves their wife and kid(s) to seek higher status? What is wrong with people? I hope we don't see a return to the Victorian/Dickensian times if this is how it was.

    ReplyDelete
  2. It's not just the father leaving the kids. The wife may be more interested in "pursuing her career" or "establishing her financial independence" than having to deal with mouths to feed, diapers to change, and lessons to teach.

    In fact, among higher-status women in the Victorian era, if raising a child would have hurt their status (e.g., they were supposed to be a virgin, the child was illegitimate, the father was lower status, or whatever), then they could give it away to a woman to promised to raise it in exchange for money, with the mother's social circle being none the wiser.

    http://en.wikipedia.org/wiki/Baby_farming

    Basically, paying off adoptive parents to take the kid. However, if they were paid a lump sum, it wasn't uncommon for the adoptive mother to just kill the kid -- having collected the lump sum, she had no future costs for raising the kid if it somehow fell ill or didn't get enough to eat. Multiply that by all the mothers who pay you off to "raise" their kids, and bingo, you're in the money now.

    That was a big enough of a problem that there was a moral panic about serial killers who just took in as many babies as they could, poisoned them, and kept the pay-off.

    It was that kind of pervasive corruption and seediness that led the Progressive reformers to formalize and regulate the process of adoption. For one thing, even if adoptive parents do get paid, it's never in a lump sum, where they have no incentive to keep the child alive after getting paid.

    ReplyDelete
  3. The Daily Fail is sensationalised, but...

    http://www.dailymail.co.uk/news/article-2246773/Cougars-21st-century-phenomenon-Research-shows-Victorian-times.html#axzz2JlIVLN8d

    "A study of census data by genealogy website Ancestry.co.uk also showed that there were more single mothers in 19th century Britain than there are in the UK now...These days around five per cent of households are single parent families, however a look at 1841 Census results, reveals back then one in six - or 16 per cent - of homes had just one parent ... But unlike today, the cause of the breakdown of most relationships was down to the high mortality rate - death from childbirth was far more common than divorce."

    Not sure if that is their inference or directly true.
    Victorian orphans and single parent families being due to the awful grindingness of the society (mass child employment, etc.) rather than relationship breakdowns seems plausible.

    Another random website states - "In 1861, it is estimated that 11% of children had lost a father by the age of 10, 11% a mother, and 1% had lost both parents". It seems like people would've noticed before now if it was more due to relationship breakdown?

    Respectfully, it does some like what Neo-Victorians there are around admire mostly the technological trends, certainly not inequality.

    ReplyDelete
  4. Wow that's really sick to me. Thanks for cluing me in Agnostic, I wasn't aware of the "baby farming" racket.

    You're right, there are many more women in the workplace now, for various reasons. But status-seeking is surely one. I think I'm pretty well protected from finding a woman like this though, due to my personality. The woman I'm dating now is the opposite of this, thankfully.

    I'm also happy to know a woman at my part-time gig who is raising three foster children. I can't imagine how tough that must be. So there are some good ones out there, though she is in her mid 40s.

    ReplyDelete
  5. Supposedly Macklemore is straight, but he's highly-annoying. I can't believe I've heard that "Same Love" song in big-box stores.

    ReplyDelete
  6. Divorce levels and bastardy levels are like Siamese twins. Status-seeking? I seek status. Don't you?

    ReplyDelete
  7. For the record, once again, parents and parenting doesn't impact how children turn out. A single vs. two parent family makes no difference to how children turn out, at least today when the need for material support is less a factor in warding off death or serious illness to the child.

    You can choose to believe that or not. I don't care, and more importantly, neither do the facts.

    That said, I think your point about inequality is interesting. In general, it seems individual competition becomes more fierce when there is more inequality. Indeed, this would make sense – the stakes are much higher in a highly unequal society.

    On that, I think it's interesting to see shuttered families when society is highly unequal. Indeed, the consequences of such would be much more dire in pre-modern times.

    That said, I think you can't make many statements in the rise of modern divorce, even with respect to Turchin's theory. The factors that affect divorce and mating (e.g., sexual revolution, welfare state, contraceptives) are unprecedented, and likely lead to a fairly unique set of circumstances today.

    ReplyDelete
  8. "For example, in a 2002 book, For Better or For Worse: Divorce Reconsidered, Hetherington and her co-author, journalist John Kelly, describe a 25-year study in which Hetherington followed children of divorce and children of parents who stayed together. She found that 25 percent of the adults whose parents had divorced experienced serious social, emotional or psychological troubles compared with 10 percent of those whose parents remained together."

    http://www.scientificamerican.com/article.cfm?id=is-divorce-bad-for-children

    Yep, no difference between a 1 in 10 risk and a 1 in 4 risk. Odds ratio = 3, much bigger than 1.

    If we treat adjustment as a bell-curved trait, those figures imply a gap between the averages of 0.36 standard deviations. Or, as though the children of divorce were just over 1 inch "shorter" (less adjusted) than those of intact families, on average.

    ReplyDelete
  9. "The factors that affect divorce and mating (e.g., sexual revolution, welfare state, contraceptives) are unprecedented, and likely lead to a fairly unique set of circumstances today."

    Bantu-speaking Africans show similar patterns for mating, marriage / divorce, child-rearing, allo-parenting, mother pursuing her own career while someone else raises the kids, etc. -- with no welfare state, contraception, sexual revolution, etc.

    Including semi-nomadic groups, not just urbanites living under centralized government.

    ReplyDelete
  10. @agnostic:

    It seems that either you didn't read my link or are ignoring it. For completeness sake:

    "She found that 25 percent of the adults whose parents had divorced experienced serious social, emotional or psychological troubles compared with 10 percent of those whose parents remained together."

    "Yep, no difference between a 1 in 10 risk and a 1 in 4 risk. Odds ratio = 3, much bigger than 1."


    Do I have to tell you correlation does not equal causation?

    Moreover, I hope I don't have to tell you that all human behavioral traits are heritable.

    It's not that children with divorced parents don't have worse outcomes. But those worse outcomes aren't caused by their parents' divorce, but from the traits the children inherited from their parents, traits of people more likely to get divorced.

    Before you ask (again, please see the link), this is not a speculative alternative explanation here. Behavioral genetic evidence makes the causation clear.

    "Bantu-speaking Africans show similar patterns for mating, marriage / divorce, child-rearing, allo-parenting, mother pursuing her own career while someone else raises the kids, etc. -- with no welfare state, contraception, sexual revolution, etc.

    Including semi-nomadic groups, not just urbanites living under centralized government."


    They also have a different evolutionary history, so that's interesting, but of limited relevance.

    ReplyDelete
  11. That link says nothing about how the worse outcomes of children of divorce reflect genetic influences that, when expressed in the parent, led them to break up the marriage.

    You have no experimental results to point to because the proper experiment would take parents at risk of divorce, randomly assign some to get divorced and the others to stay together (under threat), and then compare how the experimental treatment affected the children's outcomes in each group.

    For various reasons, the IRB won't let us run those experiments.

    Harry Harlow did something like that, long before IRBs, where monkeys were taken from their mothers and "raised" by artificial mothers that were cold and metallic, rather than warm and furry like real monkey mothers. Without exception those monkeys grew up psychotic.

    Growing up without both parents is not as extreme of a disruption to familial bonding as the sadism of Harry Harlow, but that just means the effects aren't as pronounced in the human children. They don't wind up completely maladjusted, insensitive to all norms, sociopathic, etc. -- just having more social, emotional, and psychological troubles, like Hetherington said.

    ReplyDelete
  12. I think you've read too much into what behavior genetics tells us about non-genetic influences on adult outcomes.

    Last time you brought this up, you said that a climate of widespread helicopter parenting couldn't lead to any differences compared to a climate of more autonomy-encouraging parenting.

    Remember that those behavior genetics studies compare individuals who were raised in highly similar cultural environments, including what time period or zeitgeist was in the air while they were being brought up. Since the zeitgeist doesn't vary much across the people studied, it can't explain much variation in outcomes.

    That's a restriction of range effect, a statistical artifact, not an accurate description of how strong or weak the zeitgeist is on adult outcomes.

    You're pushing an open door here about how human traits are heritable, and how the garden variety differences in parenting that are picked up in behavior genetics studies don't account for adult differences in personality or intelligence.

    But having one of your parents broken apart from the family, and often moving so that the kids will rarely interact with them, or the broken-off parent moving away on their own -- well, that's not trivial parenting differences like do you read to your baby early or let them start reading in kindergarten. It's profoundly disturbing to children at the time, it takes years just to get over the initial disruption, and evidently they never fully recover in adulthood.

    ReplyDelete
  13. You're also missing the other major point that Judith Rich Harris and Steven Pinker made when discussing behavior genetics. The first major point was that garden variety parenting differences don't influence the kid's adult IQ or personality. And that point takes 200 pages to make, because most people don't believe it, there's a lot of evidence to go through, etc.

    But both of them make a second major point, unfortunately as more of an after-thought, a 5-page epilogue, after trying to convince people of the first major point. I don't think that's because they believe the second point is not important, but not really a social science point, hence only deserving of a few pages in a social science book.

    And that is: what makes a certain style of parenting, or particular acts (spanking, etc.), good or bad is separate from what effects they may or may not produce in the kid when it grows up. It's a descriptive vs. normative difference.

    Like, you may well believe that spanking a child won't influence his adult personality.* Fine -- I'll still spank him for retribution or punitive reasons. He did something very bad, and bad acts must get punished, so he gets spanked.

    Or, not allowing your daughter to wear leggings with a short shirt may not influence how many sex partners she has during her 20s. Fine -- but no daughter of mine is going to create an unwholesome atmosphere in the home and community by parading around leaving nothing to the imagination.

    Social science types are more liberal in their morality (I mean in Haidt's terms), so they are unaware that there are all those *normative* reasons why parents may insist on raising their kids a certain way, or why others in the community may insist that parents raise them a certain way.

    * But behavior genetics doesn't say it won't influence his *juvenile* behavior, which is what the parents are worried about -- getting him to stop smacking his brother here and now, not necessarily molding him into a respectful adult.)

    ReplyDelete
  14. The phrase I'm looking for is "consequentialist vs. deontological" approaches to judging parenting styles.

    ReplyDelete
  15. Agnostic,
    This is such a good post. One thing that Turchin seems to lack in his framework, but which you more than make up for, is that regular people have autonomy and aren't just the recipients of the elite's mood or machinations.
    So, I really like "status-striving" as part and parcel to the inequality gap because it's both true and is a state that everyone is capable of experiencing.

    Jayman,
    I followed your link... I could be getting you all wrong, but are you trying to say that genes, and only genes, matter? Wouldn't this mean that the demographic changes that have occurred after one generation are the result of selection? And when they change back the other way in the next generation, that is selection at work, too?

    In my view, Agnostic is describing a facet of good ol' fashion corruption and it does seem to come in cycles and is universal.
    The Deuteronomic Cycle Or as a teacher taught it to us: sin; fall; repent and get on the straight and narrow; reap the rewards; forget and become proud; REPEAT.

    Closer to Agnostic's post was Phillip Longman in "The Return of Patriarchy" though he focused more on men and with a different take than Agnostic on women's role (though I strongly agree with Agnostic):

    "It [Patriarchy] competes with many other male visions of the good life, and for that reason alone is prone to come in cycles. Yet before it degenerates, it is a cultural regime that serves to keep birthrates high among the affluent, while also maximizing parents' investments in their children."

    "...Caesar Augustus felt compelled to enact steep "bachelor taxes" and otherwise punish those who remained unwed and childless. Patriarchy may have its privileges, but they may pale in comparison to the joys of bachelorhood in a luxurious society -- nights spent enjoyably at banquets with friends discussing sports, war stories, or philosophy, or with alluring mistresses, flute girls, or clever courtesans."

    Return of Patriarchy

    ReplyDelete
  16. Ditto Agnostic's comments on Judith Rich Harris in "The Nurture Assumption". I remember thinking "Man, she's been holding back and is letting the establishment have it!" It was more a reflection of what she was fighting against rather than the fullness of her beliefs. When I read it several years after it came out, it thus already felt a bit dated.

    She wrote a follow-up, which I haven't read, but it was my impression that it was less reactionary against the Blank Slaters.


    ReplyDelete
  17. You've made an about face on this and many other issues.

    I remember you used to say, along with Jason Malloy, that parents don't matter.

    ReplyDelete
  18. Our time isn't like the Victorian era or the Gilded Age in a very fundamental way - we no longer live in an age of rapid economic growth and technological progress.

    Even given all the misery, working class exploitation, and urban dysfunction that characterized the 19th century, it was still saw the greatest improvement in the condition of the masses, whether in terms of income or measures like public health and education, in history up until that point.

    The same cannot be said about our own era. While we may share certain cyclical trends with that period, we do not have high economic growth nor outside of computers, much technological progress.

    ReplyDelete
  19. I'd explain the numbers of kids without two parents as being related to illegitimacy rather than divorce. I think the recent rise in illegitimacy is something distinct and not as explainable with cyclical patterns. There was a very large increase in female participation in the workforce, reducing the need for a male breadwinner. Contraception plus abortion separated sex from marriage. The kids from Victorian literature typically had dead parents, because people died more back then*. I suppose the really status-striving women who get knocked up out of wedlock abort their kids, the "promises I can keep" type who figure they have no hope of attaining middle class status settle for having fatherless children.
    *The predictable response is that the death rate is always 100%, but you know what I mean.

    ReplyDelete
  20. An economic historian discussed labor conditions, child labor & single parent families of the Victorian era at EconLog.

    ReplyDelete
  21. I'm also recalling Greg Clark claiming the illegitimacy rate in England was negligible in the past, but I forget what the bounds for that were and if there was any cyclicality around a low average.

    ReplyDelete
  22. There any twin studies where one twins is in a divorced family and the other isn't?

    One problem with kids of divorce is that, even if genetic correlations go away, it also doesn't really demonstrate that its the absence of a parent, rather than less nurturing from the remaining parent (due to various overwork, distractions, etc.) Once you get into the hotbed of environmental-social treatment being really important (where Jayman isn't), all this stuff becomes "plausible" again. To some degree its potato-potato, but the "real engine" is difficult to quantify.

    "The phrase I'm looking for is "consequentialist vs. deontological" approaches to judging parenting styles."

    I don't think its more about deontology, as such, as deontology is about acts being right or wrong in themselves, so much as you create a particular environment to regulate the kid's behavior in that environment.

    Even though the kid's behavior will immediately change when the environment changes. It's still consequentialist, just not in the long term and developmentally focused, but focused on the kid's immediate behavior as a child.

    Deontological morality comes in more with giving the kid nurturing - it's not because "well nurtured kids are smarter as adults, etc" as much as that treating kids well, in the absence of any *very* compelling reason otherwise, is the right thing to do.

    I'm not sure liberal folks are less consequentialist in this exactly, but I'd expect them to try and look to the long term, since that's "enlightened", despite the fact that unfortunately, long term effects are sparse.

    ReplyDelete
  23. Agnostic, my friend,

    You’re not bringing up anything here I haven’t already heard many times before. (In fact…)

    For one, you’re grasping at straws to plain old making things up in an attempt to “rescue” parenting.

    “Remember that those behavior genetics studies compare individuals who were raised in highly similar cultural environments, including what time period or zeitgeist was in the air while they were being brought up. Since the zeitgeist doesn't vary much across the people studied, it can't explain much variation in outcomes.

    Ummm, it doesn’t take more than common observation to realize that there is a HUGE variation in the way Western parents treat their children. If there wasn’t, why would there even be parenting advice in the first place?

    In any case, even going beyond anecdotal observation, this has been measured directly. There is enormous variation (which has no effect, by the way).

    “That's a restriction of range effect, a statistical artifact, not an accurate description of how strong or weak the zeitgeist is on adult outcomes.”

    Are the styles of Western parents a restricted set with respect to all human variation? Sure. But take that to it’s logical conclusion… I’ll return to this point shortly.

    “You're pushing an open door here about how human traits are heritable, and how the garden variety differences in parenting that are picked up in behavior genetics studies don't account for adult differences in personality or intelligence.”

    No, I’m pretty much reporting what the evidence says. Behavioral genetic findings are, by far, the most solid in all of social science – perhaps even all the human sciences.

    “But having one of your parents broken apart from the family, and often moving so that the kids will rarely interact with them, or the broken-off parent moving away on their own -- well, that's not trivial parenting differences like do you read to your baby early or let them start reading in kindergarten.

    Indeed. But you speak as if these are not included in behavioral genetic studies. Behavioral genetic studies don’t include only children from upper-class two-parent homes (there is a restriction of range from adoption studies, sure), but, the most common form, MZ-DZ twin comparisons, include the full range of family types, of all social classes and all levels of IQ. No shared environment effect is found. If all those things had an impact, including “broken homes”, then it would turn up in the shared environment term, even if it only affected some families and not others. Children growing up in broken homes would be more similar than their genetic relation alone would predict (e.g.: “it takes years just to get over the initial disruption, and evidently they never fully recover in adulthood”). However, this does not happen.

    Please don’t appeal to your own incredulity as if it constitutes some form of evidence (as this guy did). We have data, if you issues with them, please see the source materials.

    Further, to take your point to its logical conclusion, even if, as you admit, good Western parenting doesn’t have an effect, then the majority of parents need not worry about their parenting styles, since it, by your own admission, has no effect. Of course, the evidence takes this a step further, but, there we are.

    “Social science types are more liberal in their morality (I mean in Haidt's terms), so they are unaware that there are all those *normative* reasons why parents may insist on raising their kids a certain way, or why others in the community may insist that parents raise them a certain way.”

    Sure, but does that matter? Not really. As well, it’s largely orthogonal to the topic at hand here.

    ReplyDelete
  24. “* But behavior genetics doesn't say it won't influence his *juvenile* behavior, which is what the parents are worried about -- getting him to stop smacking his brother here and now, not necessarily molding him into a respectful adult.)”

    Well of course. Both Harris and Pinker both note that parents wield considerable (though not unlimited) power to shape kids’ behavior when children are with their parents. No one is claiming that you should abandon your discipline-keeping behavior. However, the topic at hand was whether parenting has a long term impact – particularly, divorce. The answer to that question is, broadly, no. Sure, one may then argue it’s better to favor intact marriages because, even though it won’t really matter to how children turn out, it may influence their happiness today. Of course, even that point is highly questionable (i.e., are children happier with divorced parents or with married and constantly quarreling parents?), but at least it’s an argument on somewhat better footing than appealing to kids’ outcomes.

    As with most things I can’t convince people of regardless of what evidence I have (see my Twitter pic), the parenting thing is a matter of faith for believers – religion. So I don’t expect to change your mind per se, but these are the facts nonetheless.

    ReplyDelete
  25. "You've made an about face on this and many other issues.

    I remember you used to say, along with Jason Malloy, that parents don't matter."

    If I did, that was in the context of individual differences -- personality, intelligence, etc. Not that culture, zeitgeist, etc., don't matter. Or that major disruptions like divorce, orphanages, etc., don't matter.

    That's what I'm trying to get across here -- it's not a silly high school debate topic. Like, "Parents matter: true or false?"

    ReplyDelete
  26. "She wrote a follow-up, which I haven't read, but it was my impression that it was less reactionary against the Blank Slaters."

    No Two Alike. It's about how the non-shared environment causes differences between identical twins. I reviewed it for GNXP a long time ago.

    Here's an example. Two identical twins, Bob and Greg. One day while they're walking to school, Bob trips on the sidewalk and lands face-first in front of all their classmates. Greg doesn't. And it's sheer chance alone.

    All the kids start pointing and laughing at Bob, and it's somewhat harder for him to make friends during the year because he gets tagged with the nickname "klutzo."

    That's not a world of difference, but it's a subtle thing whose effects could last into adulthood. Do that across all possible chance events like this, and identical twins may become noticeably different from one another.

    I chimed in that the same thing happens on a microscopic level -- "developmental instability" -- where a particle zigs instead of zags in the kid's brain, and suddenly they have a slightly different personality.

    ReplyDelete
  27. "Our time isn't like the Victorian era or the Gilded Age in a very fundamental way - we no longer live in an age of rapid economic growth and technological progress."

    Right, we're in for much more of a bummer.

    ReplyDelete
  28. "I'm also recalling Greg Clark claiming the illegitimacy rate in England was negligible in the past, but I forget what the bounds for that were and if there was any cyclicality around a low average."

    I don't know about England, but France saw a surge of infanticide during the late 18th and early 19th C. Paris streets were littered with dead babies.

    That, along with the baby farming of the later Victorian period, point to how tough it is to measure these things in the past. How would you know how many babies were sent out to a baby farmer when the whole motive was to hush up an illegitimate or unwanted birth?

    ReplyDelete
  29. "It's still consequentialist, just not in the long term and developmentally focused, but focused on the kid's immediate behavior as a child."

    I was going for both. It's shorter-term consequentialist, but also deontological -- "bad acts must get punished, so the bad-acting kid must get spanked. That's just what you're supposed to do."

    ReplyDelete
  30. "Ummm, it doesn’t take more than common observation to realize that there is a HUGE variation in the way Western parents treat their children. If there wasn’t, why would there even be parenting advice in the first place?"

    Bring this down to earth. We aren't arguing about whether parenting differs across families. That's too abstract.

    Concretely, we're talking about changes in the zeitgeist -- almost nobody lets their kids play unsupervised around the neighborhood anymore. Certainly not in the upper-middle-class neighborhoods where the adoption studies take place.

    And back in the '80s, almost every kid was allowed to, and actually did, roam around the neighborhood like that.

    Therefore, in this concrete form of "parenting," there is minimal variation during a given time period, and the restriction of range effect applies.

    ReplyDelete
  31. @Agnostic:

    "No Two Alike. It's about how the non-shared environment causes differences between identical twins. I reviewed it for GNXP a long time ago.

    Here's an example. Two identical twins, Bob and Greg. One day while they're walking to school, Bob trips on the sidewalk and lands face-first in front of all their classmates. Greg doesn't. And it's sheer chance alone.

    All the kids start pointing and laughing at Bob, and it's somewhat harder for him to make friends during the year because he gets tagged with the nickname 'klutzo.'

    That's not a world of difference, but it's a subtle thing whose effects could last into adulthood. Do that across all possible chance events like this, and identical twins may become noticeably different from one another.

    I chimed in that the same thing happens on a microscopic level -- 'developmental instability' -- where a particle zigs instead of zags in the kid's brain, and suddenly they have a slightly different personality."


    Kevin Mitchell recently gave a talk when he discussed developmental noise.

    Interestingly, apparently, some twins appear more resistant to environmental disruption of their development. More on that shortly.

    Harris posited an interesting hypothesis in No Two Alike, but, to be honest, there isn't a lot of evidence for it. The effects she chalks up to environmental amplification of pre-existing phenotypic differences (thanks to developmental noise) could easily be purely the product of that noise. I'm not sure that there's huge justification (though there may be some) for not just nixing that step ala Occam's Razor.

    As for Kevin Mitchell's note, it would be interesting to do systematic comparisons of twins that are more similar to each other vs twins who are less so. Are twins who are more similar to each other more symmetric? Does higher similarity for some traits (say IQ) translate into higher similarity for others (like BMI, height, personality)? Are higher IQ twins more similar than lower IQ twins (suggesting genetic load is involved)? These are fascinating questions that may go a long way to answering how we come to be the way we are.

    ReplyDelete
  32. @Dahlia:

    "I followed your link... I could be getting you all wrong"

    Yup...

    "but are you trying to say that genes, and only genes, matter?"

    Put it this way: not everything is in the genes. But, broadly, what's not in the genes is not in the hands of parents.

    Even better, going beyond that, what is not ultimately "genetic", and hence is termed "environmental" doesn't necessarily lead to predictable (or controllable) effects on human development (much of it is developmental noise).

    "Wouldn't this mean that the demographic changes that have occurred after one generation are the result of selection?"

    It would if phenotypic expression were entirely genetically dependent, which they're not, and I'm not arguing such. Just as in the quality of fruit born from the same strain of seeds varies according to the field in which they're planted, human behavior is dependent on the broad environment. Change the environment – most importantly the incentives in that environment – and you see a different set of behaviors even with the same set of genes.

    ReplyDelete
  33. Back to divorce's effects. Again you're being too abstract and general. Just because the generalization holds, er, in general, doesn't mean it holds in every case, to the same extent, and so on.

    Divorce can still affect kids outside of the "shared environment" term, such as gene-environment interactions. For example, a genetically predisposed kid doesn't end up maladjusted because the environment was friendly and did not trigger maladjustment. Whereas a home with conflict would trigger that predisposition.

    I don't really care through what pathways divorce affects how children turn out. That's an "I wonder how that works?" research project for social scientists. The main thing here is -- does it affect them outside of the genetic pathway? That's the concrete question.

    So now to the actual research literature. From "Handbook of Behavior Genetics" (ed. Yong Kyu-Kim, all typos mine since I'm copying by hand):

    "A few studies have examined genetic influences on the relationship between marital conflict and child outcome as well. O'Connor, Caspi, DeFries, and Plomin (2000) examined associations between divorce and child achievement, social adjustment, behavior problems, and substance use in the CAP sample. Genetic and environmental contributions tended to vary based on child outcome. Environmental factors mediated the links between divorce and child adjustment, but results for social adjustment and achievement indicated passive GE [gene-environment] correlation. This work was extended by examining a moderation model for the association among genetic risk, adopted parent separation, and child adjustment (same authors, 2003). Findings indicated that in the absence of marital discord in the adoptive home, genetic risk (estimated from birth parents) was not correlated with child adjustment. However in the presence of marital discord, genetic risk was significantly correlated with adjustment. These results suggest that genetic vulnerability may play a role in the causal pathway between divorce and child psychopathology."

    So right away we see that on average, divorce is harmful to children when they're children. Seems like a "no duh" thing, but there it is. Only counting adult outcomes is a bizarre stance to take, akin to "you can beat your kids up and it won't affect them in adulthood that much."

    More paragraphs from that review coming while I let my hands rest from all that copying...

    ReplyDelete
  34. @angostic:

    "Bring this down to earth. We aren't arguing about whether parenting differs across families. That's too abstract.

    Concretely, we're talking about changes in the zeitgeist -- almost nobody lets their kids play unsupervised around the neighborhood anymore. Certainly not in the upper-middle-class neighborhoods where the adoption studies take place.

    And back in the '80s, almost every kid was allowed to, and actually did, roam around the neighborhood like that.

    Therefore, in this concrete form of 'parenting,' there is minimal variation during a given time period, and the restriction of range effect applies."


    You're falling into the trap that this guy did (like I said, I've heard this all before). You don't actually need to measure parenting per se to determine that parenting has no impact. Indeed, that's broadly far less powerful than looking for shared environmental effects, which serves as a natural quasi-random control.

    Even if one aspect of parental behavior did systematically change, there's still going to be quite a bit of variation in this. And even if there wasn't, there's going to be all sorts of other variation. Behavioral genetic studies show that that variation doesn't matter. There is no secular trend in the shared environment term: it always comes to back to zero, so there is no cohort effect. Differences in parenting didn't matter then and they don't matter now.

    This is quintessential grasping at straws. Think about what you're saying: you're positing that all the variation in parental behavior across Western parents doesn't matter except for a few specific behaviors.(?) Why would those particular parental behaviors be important when the vast majority of parental variation isn't? What makes these different? Hopefully you can see how any explanation you might put forward seriously runs afoul Occam's Razor.

    ReplyDelete
  35. My boldface, original italics.

    "Connections between marital instability and child drug and alcohol use, internalizing problems, and behavior problems have also been examined using a Children of Twins design as part of a large study of Australian twins (D'Onofrio et al., 2005). With this method, the children of adult twins who were discordant for marital difficulty were compared. In other words, a child whose parent was experiencing marital instability was compared to their parent's twin (their aunt or uncle) who was not having trouble in their marriage. This design is useful in identifying the type of GE correlation operating, particularly passive GE correlation theorized to exist between marital conflict and child psychopathology. Results indicated that environmental influences linked to divorce explained the greater rates of psychological difficulty among the children rather than common genetically influenced traits that impact both marriage and mental health. While specific environmental factors were not examined in this study, they were hypothesized to include decline in effective parenting, conflict between parents, loss of contact or inadequate parenting from non-residential parent, and increased economic pressures."

    They explained the logic pretty well, but to elaborate... They looked at kids whose parent was going through marriage troubles. If the kid shows troubled thoughts and behavior, how do we know that isn't an expression of the same genes that, expressed in the parent, are leading to that parent's marital strife?

    We compare the kid to his parent's twin who is not going through marriage troubles. That establishing genetic similarity between the grown-up and the kid. Any extra trouble the kid is having beyond his similarity to the worry-free twin of his parent, could be due to the environmental effects of living with the parent who is going through marriage troubles.

    And that's just what they found, although they didn't dissect "environmental influence" into this or that type of stress for the kid. Again, that's a question of mechanism, not whether a home environmental of marital strife negatively affects the kid -- it does.

    ReplyDelete
  36. Then a concluding paragraph I won't bother to copy about how other non-genetic research has looked into those causes and found effects of divorce on children, how those researchers assume the effect is totally environmental, and how the behavior genetics work shows that there's an interaction between genetic predisposition and environmental stressors.

    As far as policy and personal decision-making goes, we can't alter our genetic predispositions, but we can affect what kind of environment the kid grows up in that would draw out those latent tendencies.

    Conclusion: divorce is bad for kids, and it goes beyond genetic correlations between parent and child, to include a toxic environmental effect.

    ReplyDelete
  37. @Agnostic:

    "Conclusion: divorce is bad for kids, and it goes beyond genetic correlations between parent and child, to include a toxic environmental effect."

    Don't be so sure. I am reading these papers. I already see big methodological flaws, to which I'll address shortly.

    ReplyDelete
  38. Rats! I meant "agency", not autonomy. I paid a high price for my son; have trouble with the most basic vocabulary words. On another thread I said "sports guy" as I could not remember athlete!

    ReplyDelete
  39. The huge decrease in the percentage of children living with dad since 1987 may be due to the passage of the Bradley Amendment in 1986 and VAWA in 1992.

    ReplyDelete
  40. OK Agnostic:

    You have made me work today! But that’s good. It’s important to have one’s claims challenged. I know with something as near and dear to most people’s hearts, parenting, I’ll have many people who will do just that. These challenges are important in help ensuring that we get as close to the truth as possible.

    It’s worth mentioning, as Judith Harris did, that most behavioral geneticists believe in nurture. So to find researchers denying the non-significance of the family environment or desperately trying to find ways to restore the importance of parenting is not difficult. This doesn’t mean the evidence necessarily supports them here.

    And with that, I have analyzed some of the studies you’ve references, and much more. As I expected, when a result sounds too good to be true (finding a reliable effect of family environment), it usually isn’t. In this case, all the studies used to support the case of a deleterious effect of divorce are based on bad methodology. Broadly, they fall under two classes:

    *Adoptive families vs. biological ones

    *The children of twins (CoT) method.

    While these methods sound like clever way of getting around the gene vs. environment causation problem, neither is actually capable of doing this. They are both fundamentally incapable of demonstrating causation. Here’s why:

    The CoT method, as you note, tries to control for genetic influence by using MZ twins. However, like discordant twin observation studies, it can only be used to demonstrate a negative finding, not a positive one. In this case, the reason is simple: it can’t control for the spouses of the twins. If you find that the child of a divorced twin is more likely to have whatever behavioral pathology than the children of the married co-twin, the problem may be due to the genes of the spouse of the divorced twin. Indeed, why did one twin divorce when the other did not? Bad spouses are to be suspected. Indeed, part of this can be driven by assortative mating (by phenotype, at least). The twin more likely to divorce may have sought out a spouse with various disagreeable traits, which that spouse would have passed on to the offspring. Some of the studies claim to have controlled for the traits of the spouse, but in so doing you’re back in the territory of standard observational studies and the problems they have. But there is another problem from these studies also suffer, one which is the fatal flaw in the adoptive/biological studies.

    That problem is the direction of causation. If you find that children from divorced parents (even adopted children) are more likely to have behavioral problems, why do you necessarily assume the causation goes divorce -> problem children? The problem could easily go in the other direction. Little Johnny comes to mind:

    “Johnny came from a broken home.”

    “Yes, Johnny could break any home.”

    Children with behavioral problems could have themselves precipitated their parents’ divorce.

    Beyond this, most of the adoption studies suffered from tiny samples, and often looked at the children when they were young, when shared environment effects are known to exist (which disappear by adulthood).

    As to why this interpretation is the valid one, when you might argue that these researchers’ conclusions might be actually correct, limitations of their research notwithstanding, I again refer you to the lack of a shared environment impact. It’s simply hard to reconcile a broad lack of a lasting effect of parenting – which includes divorce and family dysfunction – with the existence of some limited, particular, and temporary effects. The Occam’s Razor explanation is that those have no effect either.

    ReplyDelete
  41. @Agnostic:

    To your original premise, and that somehow reducing the incidence of divorce (if that were even possible, which I’m not sure it is) would improve children’s outcomes – even in the short term, again, I’m not sure you could make that conclusion. Reducing divorce would not change the people. Likely you would substitute divorced couples with acrimonious ones, and of course not change kids’ adult outcomes, and not even change their childhood happiness substantially.

    ReplyDelete
  42. "What's behind these changes? I'm tempted to point to the status-striving / inequality cycle. If some problem started to get bad by the mid-to-late 1970s, and has only gotten worse since, that's a likely culprit. (It's clearly not the cocooning / crime cycle.)"

    Which is when housewives started entering the workforce thanks to modern home appliances. Pretty soon it took two paychecks to buy a house in a good neighborhood with good public schools. Like standing up in a theater to get a better look, it works for the fist people but ends up forcing everyone to stand up, thus leaving everyone worse off than when they were able to sit..

    I suppose you could describe this as a kind of status striving. Striving to not fall behind.

    Maybe you cover this -- I haven't read to the end. Your main point is strong.

    ReplyDelete
  43. "JayMan said...

    For the record, once again, parents and parenting doesn't impact how children turn out"

    But they sure impact how childhood itself turns out. Those carefree years were the happiest in my life. I would not trade them for anything, not even parenthood itself, every minute of which I have also enjoyed. Come to think of it that would be a second casualty right there, making two huge chunks out of several people's life cycle.

    Anyway you statement assumes we know how to measure emotional happiness or evaluate its importance as an end in itself. I would submit that they are the end for which measurable things like income, educational attainment, etc, are only a means.

    ReplyDelete
  44. Jayman seems to be arguing that since it is all genetics and we don't have free will, nothing really matters. But then science has never been able to establish that there was any point to existence. And never will.

    But then Jayman only feels that way because he was born that way. We who disagree do so for the same reason. And never the twain shall meet.

    I'm not sure I really mean this. I like Jayman. He's a liberal like me.

    ReplyDelete
  45. I realize my above criticisms of Jayman are unfair.

    Maybe he would accept the idea that those who subscribe to the nurture assumption do so only because they have the right genes in the sense that theirs are the children who are more likely to turn out well -- and that these very same genes make them subscribe to the nurture assumption!

    I know so many upper-middle-class parents into nurture, all of whose children have turned out well even though they also turned out very differently.

    It would be interesting to compare the children of these parents with the children of ones that explicitly do not subscribe to the nurture assumption, who deny free will and emphasize the preponderant importance of genes.

    Do the latter turn out "as well" on average in terms of whatever metric choose: rate of reproduction, divorce, socio-economic status, educational attainment, drug abuse, welfare, crime, reported happiness, etc.

    If not then you could say that subscribing to the nurture assumption might turn out to be a proxy for having the right stuff and vice versa. It would be a paradoxical conclusion.

    ReplyDelete
  46. Jayman, an argument that Agnostic almost but didn't quite make is this:

    In a society where divorce is a far more acceptable answer to marital difficulties, the external (non-parental) environment is different from a society where divorce is socially unacceptable. The changes in the divorce rate are obviously too rapid and large to be the result of changing genetics, therefore, something in the culture changed which meant that people who are genetically less sociable became much more likely to divorce; and their kids will find reinforcement in the culture for other sorts of poor behavior. A society where people are pressured into resolving their marital difficulties is probably also one where people genetically predisposed to other bad behavior are pressured into not acting on those impulses (as much). Thus, even though the parents' behavior doesn't directly influence the child's outcomes, if those parents are around enough like-minded parents, they will create the "external environment" which does shape their children's behavior.

    ReplyDelete
  47. The bar chart shows that, for kids born after 1990, something happened to greatly reduce the chance of paternal custody at age 16. I would guess that that something is VAWA.

    ReplyDelete

You MUST enter a nickname with the "Name/URL" option if you're not signed in. We can't follow who is saying what if everyone is "Anonymous."