That is, of people with exactly one sibling who read this blog, about 72% of those are the older of the two children in their family, compared to only 29% who are the younger of the two (where by chance we would expect 50-50).
This was surprising, because at the time lots of studies had shown there weren’t really birth order effects (that is, firstborn siblings had no major personality differences compared to laterborns). I theorized that maybe for some reason it was easier to find by looking in a heavily-selected group of people and asking members about their birth order, compared to getting a random sample and trying to correlate birth order with things. Sure enough, later amateur research revealed strong birth order effects in physics Nobelists and great mathematicians (and potentially Harvard philosophy students). Given that readers of this blog are highly-educated (about 37% have masters or PhDs) and mostly in STEM (41% programmers of some sort), plausibly birth order affects something about intelligence, education, or STEM orientation (somebody should check literature and peace Nobelists!)
Followup research by Less Wrong user "Bucky" determined that the effect fell off with age gaps; the closer in age you are to your sibling, the stronger an effect birth order has:
I’ll be honest - I think I over-reached here. I’m not very good at statistics, and this is a weird statistical problem: the dependent variable is whether the case ended up in the sample at all! I wasn’t able to figure out a good way to use most of the data you gave me. And the stuff I did use, I mostly made work by slicing and dicing so much that the sample size got pretty low, even when I started from 8,000 survey respondents. I’m publishing this in the hopes that it will inspire someone else will more domain knowledge to do a sophisticated re-analysis. But for now, here’s what I’ve got.
Confirming Old Results With The 2020 Dataset
The 2020 dataset also shows a strong birth order effect in people who read this blog.
In 2018, among people with exactly one sibling, respondents were 2.51x more likely to be the older sibling than the younger (72%). In 2020, the number was 2.39x (71%).
Change with age gap is shown below:
Note truncated y-axis
This seems to be broadly similar to the 2018 results. There was an anomaly in 2018 where some categories seemed to drop off surprisingly quickly between 7 and 8 years, which I thought might be meaningful. But Bucky’s analysis showed this was probably a coincidence, and indeed it doesn’t show up in the new data.
Does Sex Matter For Birth Order Effects?
I wondered if there might be a smaller birth order effect for people with opposite-sex siblings. One possible explanation for the birth order effect is children trying to get out of the "shadow" of their older sibling and differentiate themselves in some way. But children are already pretty different from an opposite-sex sibling and might feel less pressure in that situation.
But this doesn’t seem to be true. The percent firstborns in sibships of two on the survey was 70% among people with a same-sex sibling, and 71% among people with an opposite-sex sibling; no real difference.
Do Biological Or Social Factors Produce Birth Order Effects?
All previous results are for biological siblings. But it might be worth asking the question separately for biological vs. social siblings. One could imagine either biological or social causes of the birth order effect. For example, some biologists speculate that pregnancy depletes choline, that it takes a long time for choline stores to recover, and that a second child born within that window will have less choline available to build their nervous system, which could be bad. But also: maybe if you have an older sibling, your parents can’t pay as much attention to you when you’re a kid, and you learn less.
This was very hard to test for. Again, I wasn’t able to use traditional statistical tests because I’m trying to determine whether someone was in the sample at all, rather than whether two variables are related. It was easy to check normal birth order because I could compare people with exactly one older sibling to people with exactly one younger. It was harder to do with things like adoption in the mix, because that could introduce a bias: are parents more likely to adopt out their first child (because that’s when they’re most unprepared for parenting)? Are adoptive parents more likely to adopt when they already have children of their own (because they’re comfortable with child-rearing) or less likely (because they really want kids and can’t have them biologically)? I had no way of getting controls for these questions and so I couldn’t do a lot of the analyses I wanted. But I did two relatively weak analyses instead:
First, I took the entire set of people in weird situations - people who said their number of social siblings was not the same as their number of biological siblings. In this group of 174 people, biological firstborns made up only 61% of respondents with one sibling, notably less than the 71% in the entire sample. That suggests that the unusual social situations are having an effect. You shouldn’t update on the fact that it’s still higher than 50%, because some of the people’s weird social situations don’t affect their status as social firstborns.
Second, I tried to compare people who were firstborn under a biological definition but not a social definition, to people in the opposite situation. There were 40 people in the sample who were biological but not social firstborns, and 60 people who were social but not biological firstborns. Again, this suggests that social firstborn-ness is more important as an explanation than biological firstbornness, although it doesn’t rule out the latter having some effect.
I additionally tried to compare two different types of social firstbornness - one where no older siblings lived in the house when you were growing up, and one where your parents had never parented another child. There weren’t many people discordant on these two measures (29 vs. 20 respectively), but for what it’s worth, the ratio was in favor of the first type.
Since I wasn’t very confident in my analytical abilities here, I asked Bucky, who knows more and who did good work analyzing the last dataset, to look into this (we worked independently and didn’t tell each other our results until we were done). He writes:
It seems to me that the effect is entirely caused by social siblings.
I filtered for only people with 1+ biological but 0 social siblings. There were 24 oldest biological children in this group vs 21 2nd children (or 25 youngest children with a large overlap between 2nd oldest and youngest groups). This significantly differed from the ~0.7 fraction of older children in the general surveys (p<0.05 or p<0.01 depending on whether I use the 2nd oldest or youngest as the comparison) and is close to a 1:1 ratio.
I then filtered for only people with 0 biological but 1+ social siblings. There were 51 oldest social children and 23 2nd children (or 26 youngest children again with large overlap). This differs significantly from a 1:1 ratio (p<0.001 or p<0.01) and matches pretty well with the 70% of the Birth order effect.
I tried doing some filtering by age gap (2-7 years) and the results were compatible with the same result, although the sample sizes got too small to really conclude anything.
For dealing with answers left blank I treated them as 0 unless it looked like the whole section had been missed out. If I ignored any respondents who left something blank I got similar results (smaller sample size but ratios are even further in favour of the social hypothesis).
I checked for categorisation errors by looking at respondents’ descriptions of their families and they mainly matched pretty well with the numbers given so I think the data should be considered reliable. I did chuck a couple of results out which seemed unreliable and there was one row which was a repeat so your numbers might not match up exactly (plus you have the non-public data).
There are a couple of confounders in the analysis such as whether e.g. oldest children are more likely to be adopted or how much you know about your birth family depending on how old one was when the family unit changed etc. I don’t see a realistic way to account for these but I also can’t see any of them being big enough to explain the difference in the results.
Hopefully this matches up with what you found!
I think this suggests birth order effects are social rather than biological.
So What Causes Birth Order Effects?
Based on this analysis, it seems unlikely they are biological. Based on my very weak sub-analysis, and on their tendency to decay with larger age gaps, it seems they have more to do with the social presence of a sibling in the house than with any changes in parenting style (ie your parents learn to parent differently). Two explanations that satisfy both those criteria:
Parents are able to devote their full attention to parenting their first child, but only half of their attention to parenting their second. Firstborns get more quality time with their parents during the first few years of childhood.
Siblings try to differentiate themselves from each other. So for example, since the older sibling will always be smarter than the younger when they’re both young children, maybe the younger sibling is more interested in excelling in areas other than school (like trying to cultivate a specific talent that the firstborn doesn’t have).
You may be able to think of others besides these.
One problem with (1) - wouldn’t you expect smaller effects as age gaps get lower? If it’s about having quality time alone with Mom, someone with a sibling one year younger than themselves only got one year of quality time; someone with a sibling five years younger got five years. But it looks like the birth order effect is stronger for someone with a one-year-age-gap sibling than a five-year one. Either the first year is very important, or I’m missing something.
A problem with (2) - I would expect the effect to be weaker with different-sex siblings in this case - less direct competition - but it isn’t.
If (1) is true, you might expect a strong effect against twins; after all, twins get unusually little of their mother’s attention in the first year. Are twins also under-represented in this dataset?
The average SSC reader is white and 33 years old, so we should expect a twinning rate of 2.4%. But in fact it should be a bit higher than this, because the two strongest risk factors for twinship are maternal age and maternal propensity to use IVF. Maternal age is highly correlated with education (it often means a mother waited to finish a degree before having children) and our readership is nothing if not eager to use new birth-related technologies. So we should expect to have significantly more twins than average. But in fact we have fewer. This isn’t a giant difference, but I think it supports the hypothesis at least a little.
Some previous studies suggested that twins just did worse in general, probably because it was too crowded in the womb and they got fewer resources each. More modern studies fail to find that, at least on later life outcomes. But it’s possible that this is another effect which is too subtle to find in generic studies, but shows up in heavily-selected populations. If so, that would mean this section provides no extra evidence for anything.
If birth order effects are due to parental investment, it would be pretty surprising. The current scientific consensus is that parental investment in the early years of life doesn’t really increase IQ or educational attainment during adulthood. That is, the shared environment has minimal to no effect on later life outcomes (see eg Bryan Caplan’s book Selfish Reasons To Have More Kids). It’s generally agreed that people can put away the Advanced Baby Einstein Educational Toys and just chill.
This research challenges those assumptions. If it’s right, the difference in parental attention between an only child and a child with siblings seems to have noticeable effects later in life.
How do we reconcile this with twin studies? One potentially trollish answer is that all twin studies are necessarily done on twins, so the one component of the shared environment that they can’t vary is whether their subjects have siblings or not! But this doesn’t really get us out of the problem. Some parents give their children much more parental attention and investment than others, regardless of sibling number, and you would expect this to show up in the twin studies as a positive shared environmental effect on those children. It doesn’t.
One possibility is that it doesn’t show up for the same reason birth order effects don’t usually show up: they only matter in heavily selected samples.
Call For Further Research
I wasn’t able to do a very good job on this, partly because of my own limitations, and partly because of the limitations of the dataset.
One potential way to improve on this would be to run very large surveys of the population, such that we could capture the kind of person who does read SSC/ACX and the kind of person who doesn’t, and then do more traditional statistical tests on them.
Another would be to try to find other features in this survey which correlate with birth order, although this would be conditioning on a collider (their presence in the survey at all).
Finally, other people might be able to extract more signal out of this than I was. If you’re interested, you can find the raw data, the original question list, and a description of how to use them here. Note that a few people refused permission for me to republish their data, so you will probably get slightly different numbers than I did.
Subscribe to Astral Codex Ten
By Scott Alexander · Thousands of paid subscribers
P(A|B) = [P(A)*P(B|A)]/P(B), all the rest is commentary.
But they don't always win. They're better at structured reward activities but more maladjusted to social situations seeped in ambiguity - essentially, dealing with peers of various proclivities vs manipulating/pleasung their parents
That doesn't square well with the "path finder" theory. Older siblings have to fight all the fights with the parents, whereas the younger siblings generally profit from the results without having to put in as much of a fight on their own.
I was specifically critical of the "bossing the parents around" statement. Due to the "path finder" aspect of the oldest sibling, I'd argue that younger siblings get to boss around the parents much more often than older siblings.
I agree with your idea though that having to assert yourself on those arguments might provide good practice for later life.
I could have been more clear in my first post though.
For sure! Having watched a 3 and a 5 year old interact for 3 years, it seems like being a younger sibling means spending a LOT more time being bossed around by someone 2 years older than you. Like, multiple hours per day minimum
Oddly, with my kids it is the opposite. The oldest girl got pushed around by the middle girl, now the youngest pushes around the middle. They are all girls so there isn't boy levels of pushing around, but generally when there is "not playing nicely" behavior it is the younger pulling dominance.
Social first-borns are a third parent to their younger siblings. First-borns are replaced as the cutest thing in the household. This is the effect; not more parental attention. They become parent-like and get attention for being helpful instead of cute.
Yes, one of the flaws in the notion of "shared environment effects small -> parenting doesn't matter" is that parents just straight up teach their children differently, not just differentiate in quantity of attention.
Yes, exactly. This is what I wrote above. The older child becomes the leader, hence looks out on the horizon. The younger follows and competes with the older, whilst the older is following the parents.
As as oldest child, my experience was something like I didn’t fully register the "child" category before my sister was born. When she came along it became clear that we were not 3 equals (my illusion before). It was not a demotion into less attention so much as it was a demotion into a different category of being. They had allowed me to believe I was one of them; when the sibling came along that was suddenly obviously not true. Never to return again, except in echoes if the sibling were sleeping over at a friend’s, out of the house.
Also if Harris’ "nurture assumption" book is correct and peers play a larger role in development than parents do, the older sibling would be an in-home influential peer for the younger in a way that wouldn’t be reciprocal (usually) unless the younger sibling were dramatically more competent in some way.
I think I intervened between my sister and my parents, in terms of trying to interpret their behavior for her. I was never intentionally the "boss" figure partly because one parent worked almost all the time (dad) and so the center of gravity of the household was not usually present, so not easily challenged.
I’ve definitely known people who had to babysit/quasi-raise their younger siblings and that was not my situation.
This is the opposite of my experience. My closest sibling is six years younger than me, and yet I distinctly recall, even prior to their birth, that I didn't at all think of my parents as equals, or even the same as me. I remember being astonished when I learned that they were once children.
I'm understanding that birth order effects are strong for a few years, before dropping off around seven years. Some effects seem stronger after a few years, rather than the minimum difference.
Even with that, a full year of difference (pretty much the minimum biologically), can be pretty dramatic. The difference between two and one is huge, and that's when a lot of these effects are being built into the siblings. Sure, by the time we're talking 19 and 20, that's a tiny difference, but by then the relationships are pretty solidly complete.
I remember that Scott said on average, older siblings were only 1 year older than younger siblings (out of the entire pool, obviously, not within the same household) so that is probably not the explanation.
I had this thought too, back when we were looking at the first round of results. I think the fact that the effect is stronger for siblings closer in age, and weaker for siblings farther apart in age, contradicts this hypothesis. If it were just that older people had done more stuff and therefore had had more time to stumble across interesting blogs, then we would expect the effect to get stronger the larger the age gap grows. (Though age might explain the Nobel laureates or whatever.)
Well, it also wasn't arbitrary in the sense that keeping the family's assets intact rather than dividing them is important, and in the sense that having a very clear line of succession minimises infighting, and ultimogeniture would be too likely to put a baby on the throne.
But I see your point that maybe first-borns are meaningfully distinct in other relevant ways too
This is why I want to ask if the thing SSC is selecting for isn’t conscientiousness first and foremost, with programming being full of conscientious people precisely because there really are right and wrong answers
When I think of a field that demands concientious people, I think Civil Engineering, not programming. When you can compile and test your program in seconds it removes much of the need to get it right the first time.
> When you can compile and test your program in seconds it removes much of the need to get it right the first time.
Most programmers can't do that. You can test a few things, but when you're working with dozens of other developpers on a platform that has thousands of users, either you're concientious, or you're going to increase the support workload a lot.
The English in the early 20th Century were quite smug about how their system of primogeniture created a strong, calm, self-confident ruling class prepared by their upbringing to rule. Even the House of Commons tended to be full of near-aristocrats: e.g., Mr. Winston Churchill had the good fortune to be in the now-dominant House of Commons because he was the oldest son of Lord Randolph Churchill who was the second oldest son of the Duke of Marlborough.
- Firstborn siblings spend more of their early years exposed to and learning language from adults versus other children who are only slightly older.
- Similarly, firstborn siblings are more directly exposed to praise and expectations from their parents and other adults, vs the praise and social validation of another slightly older child. Thus they become more oriented to achievements that will impress adults, such as performance in school.
Linguistic research on children has found that the critical period of language learning goes until age 10-12, so any child would have had a lot of exposure to lots of other children and adults in the relevant time period. It’s possible that there is a slight effect on margin, but it seems a bit doubtful.
+1. As the firstborn of 4 and the father of 3, the most obvious distinction to me is that the firstborns are socialized by adults while the younger kids are socialized by siblings. This aspect is consistent with the declining effect when age gaps are large - much older siblings aren't around to socialize younger ones.
In theory I should be able to figure this out - I included a question "How close did you come to not filling out this survey" to see if this could be used to analyze response bias. I don't feel like actually looking at it right now though.
I've read about first-born superiority, so if I were a first-born, I might want to participate to find out more about my superiority. But, in a Harvard class the professor asked the students to raise their hand if they were first born. Most of the room raised their hands. It was a philosophy class. As a second child, I see the oldest one wanting control with their parents in supportive roles. They have good verbal skills and know what to say to get what they want.
Edit: This is wrong! See André Röhm's comment and my response.
Most people are first-born children, if average family size is not too high. Suppose for illustration that 30% of all non-childless families have one child, 40% have two children, and 30% have three children. Then by simple arithmetic, 30%+20%+10%=60% of these children are first-born. And Harvard students come predominantly from well-off families with relatively low birth rates (see https://www.statista.com/statistics/241530/birth-rate-by-family-income-in-the-us/).
I stumbled over it for a bit, but I am pretty confident you made a slight error here.
The real answer in your example is "50% of children are first-borns".
(Do the math for example for 10 families, and calculate the number of total children)
What you calculated was equivalent to answering "If I pick a random family and then pick one of their children at random, what is the chance to select a first-born?". This underestimates the effects of large families.
Even then, I believe your point still stands that first-born's can be more likely than expected. However, if this blog has such a strong selection effect, would we expect Harvard to have none?
You are absolutely right! Doing it properly, the proportion of first-born children is equal to the the reciprocal of the average number of children in non-childless families. In my example, that would be 1/2, but I don't know what the average is over real-life families, or families with a child at Harvard.
Maybe being older you get lots of opportunities to explain/teach things to a younger sibling, this makes you more reflective and self aware hence more likely to go into higher education get a PhD etc. The decline starting at 8 years might be because at 18 you move out and stop teaching your younger sibling things while they are still quite young and can be taught a lot of things, i.e, less than 13.
Absolutely, plus the corollary which is that younger siblings often don't have to learn things for themselves through experience or the hard work of reflection and self-awareness, so they end up getting less practice in those areas. Also, older siblings will often just do hard things for their younger siblings out of impatience, rather than let them figure it out themselves.
There is teaching, but as a second born, I see the first born wanting to control and get the money first, for college. The first born's verbal skills may get their parents to expect more from them and pay more for them. Though they might be the leader, good leaders hire
talented people. A good leader just knows how to control people, time and has a goal. Brains they can hire. They respect time as they usually pay an hourly wage.
My personal hypothesis is that the greater uncertainty involved in raising a child for the first time leads to higher openness to new experiences. I'd guess this can go down a number of different paths in practice, but that cultures situated outside of the mainstream will be disproportionately firstborn as a result.
I've heard plenty of tales of parents being overprotective with their firstborn and then realising that kids aren't actually that fragile and letting the younger ones be more freeform, so I'm not sure the cause you're positing exists in the first place
I speak from experience as a firstborn. I have distinct memories from high school of feeling resentful that it felt like I was the guinea pig. I certainly can't speak for all firstborns, though, and don't really know if my experience was common.
I wonder if that has changed as parenting standards have increased with this generation.
My older brother experienced much of what you did. Unfortunately, most of the experimentation was the cw/assumptions my father made based upon what worked for his own upbringing - go to Catholic school, play football, etc.
My brother did, and even at times felt compelled to profess he liked some aspects of it. But he was miserable. A complete first-born nerd, he would go on to make good friends in college, become a brilliant surgeon, and win an award from his med school, one of the best in the country, for his healthcare policy work.
He was the most miserable adolescent you ever met. His bad experiences at the hands of his peefs haunted him literally for decades.
My parents observed and, once my brother was able to begin to process it all and speak up, accepted his assessment. As a result he moved to an excellent, small, academically oriented school (without a football team) for high school.
I got the benefit of that environment from the third grade, and more importantly of their open-mindedness when I also was not what one would term a typical small-town Louisiana kid.
As a result I was just a more relaxed and amiable kid, comfortable in my skin. Part of it was younger child peer interactions (vs older child pleasing parents). But a big part was I trusted my parents to listen to me and do what was best for me in a way my brother did not experience, and was accepted by my classmates at the smaller, more academic school from a young age.
God bless man. It all works out, but it doesn't mean it's not tough at times
Totally. The first kid into teenage years is the one who triggers all the unresolved teenage trauma in the parents. I had to muddle through it with little guidance. When my sister came along a) they were worn out from fighting me and b) there was a path and c) they realized some of their strategies didn’t work & updated to better ones.
That first born having to fight for what he wants, verbally makes him skilled at this. I have heard about first borns resenting parents, anger fuels their desire to show their parents a thing or two. They see their parents are new at this. Take advantage when possible. I've known first borns who waited to get divorced until after the parent died, so he could show his parents he was a success and right.
Could simply filter for respondents over ~20; There's no reason a 40-year-old reader can't have older siblings, so the effect of younger siblings being too young to be readers would only show up in the youngest readers of the blog.
I remember that Scott said on average, older siblings were only 1 year older than younger siblings (out of the entire pool, obviously, not within the same household) so that is probably not the explanation.
- First borns are part of the "parental teaching structure" of subsequent kids.
- and also competitors, cooperators and imitators. I have a 9 yo and 6 yo and I can tell you they are both forever different, on questions of basic personality/desires because they have a sibling. There are things they do/love/hate because of how their sibling feels about them.
- is there any value to looking at school situations (such as Montessori) that often mix age groups? A class room of all 2nd graders is going to have different "social sibling" effects than one that's 1st, 2nd, 3rd all together.
- and what about the past, when almost all kids were raised more or less in a group - a swarm of cousins of varying ages? Being the first born in a house that only ever has 2 kids is one thing. Being the first born in a generation of cousins probably has different implications.
- did you ask any questions about pets (insert snark about "substitute kids") in the household?
I mean, you're definitely right! At this point with joint dog-human society, you might be able to detect sibling effects on the dog too from having an older/younger "sibling" human. Not sure how to do the conversion to dog years though. ;)
I wonder if what matters is maternal age rather than number of siblings. Younger siblings are born when the mother is older, and that difference in maternal age at birth might explain why younger siblings do less well. It could be that the reduced effect at >7 years is because, for instance, a kid born to a 16-year-old with an unplanned pregnancy won't do as well as their siblings born when the mother is 23.
I suspect that controlling for maternal age wouldn't actually help to resolve this question, given all the usual reasons why truly controlling for such factors proves difficult, particularly given the correlation between maternal age and education.
Edit: I wonder if the children of vegetarians/vegans (who often consume relatively little choline) show less strong birth-order effects. Of course there are a lot of confounders there, but it'd be an interesting experiment.
A young, unmarried mother might have different genes to pass to the child. The more experience a child has at running the family, when the parents are younger and less experienced with more energy to try to be good parents.
But when people have >7 year age gaps between siblings, there's often an unusual family situation explaining it (e.g. teenager has unplanned pregnancy, then 7 years later has more children intentionally). That might cancel out the effects of maternal age.
You're only looking at 1/3 to 1/2 the picture. The interaction between the siblings, especially the role assumptions that go on.
Our children are 18 months apart. Just before our second was born, my wife took our daughter to the toy store so she could have a baby doll of her own, something that looked a lot like a real baby. Mom had the daughter give her baby doll a name (Baby Mike) ... OK ... with the thinking mom would have here baby, and our 18 month old daughter would have her own baby to mimic mom, instead of wanting to hold her baby sister.
Great idea, but wrong. Baby Mike went into the toy box to almost never return. Yes, she wanted to hold her baby sister, and we allowed this with close supervision, the older daughter was only 18 months after all.
But what happened was a total bonding, the younger grew up faster, as the protégé of the older, and the older did revert some especially when it came to diaper training. We had read about, and expected this.
I think there was growth and leadership developed in the older child, far vision thinking if you will ... which is pretty much what this forum is about.
The younger child didn't/ doesn't look to the horizon as much as she follows her older sister still in a lot of ways, they're 29 & 31 now.
And yes, I'm the elder of two, my sister is 16 months younger than I.
I knew that about ACX, but I’m curious how many of the readers are only children, and whether the 10% female readership includes an outsize number of only kids.
In 1976, 11% of American women over 40 had borne only one child, but by 2015 22% were "one-and-done". I have a strong feeling that only children are more or less firstborns, so likely to be over represented here. I’m wondering if that might be especially true for the women, since anecdotally it seems like we’re outliers in various subtle ways.
In my experience youngest siblings can instantly identify each other. They have been strongly socially conditioned; they were always the cutest, they were always smaller than their siblings, and parents had relaxed relative to the first kid. They're used to losing arguments and having things not go their way (being younger and smaller, not necessarily more or less on track), unless they work disproportionately hard for it.
1. My parents *returned* the gameboy I was given as a birthday present, but my youngest sibling was allowed to have multiple consoles.
2. Although I'm three years older than my sister and she's three years older than our youngest sibling, we all learned to read within two years; I was late, she was on time, and the youngest felt left out and taught themself to read at the age of two. (Those early digital books that highlighted and spoke the word at the same work wonders!) Now, of the three of us, 1 and 3 are the most avid readers. (But 2 got a neuroscience degree, she didn't get off easy!)
3. My dad taught us all to program at the same time; we all started with the same level of competence and quickly developed our own styles. Of us, 1 and 3 are employed as software engineers now; 2 lost interest. (This was also part of a larger experiment on us; in the early 2000s there was a theory that girls didn't enter programming because they didn't have a cohort, which my dad found to be a testable hypothesis, so he not only taught us but also most of my sister's school class. It did briefly keep their interest as a communal activity, but it's not clear it had any long term effect.)
I (a first-born) have also surrounded myself with disproportionately first-born friends, I find we're on the same wavelength. But I know I'm pretty unusual, both in general and among first-borns, so I'm cautious about extrapolating my own experience in that regard.
I’m an only child and all my closest friends were either first borns or oldest siblings, or had a much older sibling who was out of the house by the time we met. And even weirder, all our moms had us when they were 38 years old. I subconsciously gravitated toward a whole set of only children born to mothers exactly the same age as mine. It was pretty unusual to have your first kid at 38 in the early 80s, and we didn’t realize we had that in common until years later. I still think that’s kind of wild.
An explanation of the birth-order effect diminishing with age gaps, which still makes the primary driver of the effect parental attention, could be that the first couple of years of a child's life are the ones that require the most parental attention. A second child who is two years younger than their older sibling will get less parental attention than a second child who is six years younger, because in the first situations, the parents are having to deal with a toddler.
I think I was in your "several biological siblings, no social siblings" group, because of the large gap between my older sisters and me (fourteen years). I've often thought I got the best of both worlds growing up. In terms of parental attention, it was like being an only child, but my parents were still experienced parents, and knew what they were doing because of my older sisters.
The driver might be parental mental clarity rather than parental attention. Attention from a worn-out parent is of a different quality. Five minutes with a clearheaded parent can be more beneficial than hours with a distracted & frustrated one.
Yes, but the attention a toddler get would still be diminished by having an infant sibling. We would have to compare firstborns with short birth year gaps and see whether they do worse than firstborns with long birth year gaps.
> One problem with (1) - wouldn’t you expect smaller effects as age gaps get lower? If it’s about having quality time alone with Mom, someone with a sibling one year younger than themselves only got one year of quality time; someone with a sibling five years younger got five years. But it looks like the birth order effect is stronger for someone with a one-year-age-gap sibling than a five-year one. Either the first year is very important, or I’m missing something.
The lived experience of parenthood makes it pretty obvious why small age gaps have larger effects. The mother of a newborn and a 1 year-old has less attention to devote to the newborn, because the 1 year-old is extremely needy. A mom who read dozens of books aloud daily to the firstborn 6 month-old probably reads zero to the second born at 6 months if there is also a toddler in the house. The newborn gets a lot more of Mom’s attention if the sibling is a 5 year-old able to feed herself, use the toilet herself, play without needing Mom for more than ten minutes at a time, etc.
Here's a question: do we actually have much evidence that dietary choline levels affect...anything, as long as the baseline requirements are satisfied? I know the nootropics community likes choline, but they also like pig brain extract.
(They like to note that choline is used to produce acetylcholine, but only a tiny percentage of dietary choline gets used for that purpose, and since some amount of choline can be synthesized by the body, I'm guessing choline deficiency would cause symptomatic liver disease long before it had neurological effects, since the choline produced by the body would probably be used for ACh production before anything else.)
Here's another question: if 12 weeks of choline supplementation is enough to help pregnant women give birth to neurologically healthier children, shouldn't *6 years* of normal choline intake between pregnancies be more than sufficient to replenish the mother's choline reserves for her next pregnancy?
I can kind of understand how a lack of choline would cause birth-order effects in children born (say) a year apart, but not how it would cause effects in children born 6 years apart.
That said, I haven't read the studies, so my objection could be total BS.
Hypothesis: "Firstborn" children are recuited as deputy parents when younger siblings come into the family.
I speculate: Consciously, or unconsciously, the oldest feels more like an adult from that point on, and may feel and behave more responsibly, and also may enjoy feeling more "accepted" by their parents into the world of adult society. Maybe younger children never feel this (although families where i.e. the oldest girl ends up with lots of maternal responsibilities are anecdotally common.) Maybe younger childen have a kind of extended childhood vs an "accellerated" childhood for first-borns.
On the same topic, research published ~5 years ago looked at athletic achievment by siblings in US professional sports. (link below...I think, it seems not to be showing up...) They found a HUGE positive effect for YOUNGER siblings, especially the second child in a family. (Note: this study had vastly more male children and brother-brother pairs, for obvious reasnos, but IIRC the effects were just as strong for the few female athletes in their study.)
Several reasons were posited, the two salient ones were:
-- Practice effects: The younger child is often playing with the older, and with other kids several years ahead, this is very tough competition and an intense learning environment.
-- Parental / Family experience: Parents learn how to navigate schools, leagues, coaching, sports-bureaucracies in general, and the younger kids get fast-tracked into the best opportunities.
For me, the absurd number that stood out was this: In US Baseball, 700 (!) pairs of brothers have played in MLB. The younger brothers, collectively, attempted 10x more stolen bases than the older.
I definitely observe more athletic inclination in my younger daughter than in my older - starting with the age of walking, and propensity to climb, and comfort with climbing. I attribute it to two factors, besides innate proclivity: the older one was heavier as a child and maybe that's why she started walking later; the younger one was also on the ground a lot more and held less, because I had two close together - so perhaps she got more practice and therefore started crawling and walking earlier.
On the flip side, the second is less verbal and started talking and singing later, and is less of a communicator than her sister was at that age. This can be attributable to talents but also to having had less focused conversation time from me, while the first one got it from a full-time caregiver while I worked.
I agree. First borns seem to have very good verbal skills. What about the stage effect. The first born needs to put on a good show for parents and grandparents. First borns had to negotiate with parents to get what they wanted and converse with adults more often. More practice verbally for keeping their parents together, working, so they can send them to college.
Having an older brother with marginal Major League potential would probably be the best upbringing possible for a future baseball star. I knew a Cy Young award winning pitcher who was a third son in a much admired athletic family. His nurture was ideal. His father was a successful lawyer and the best Little League baseball coach around: a terrific guy. He had two highly athletic older brothers. His oldest brother had been an All-Southern California second baseman and played at USC, and then was his youngest brother's high school baseball coach. The older brother protected his younger brother's potential. When the athletic director wanted his star pitcher to start on two day's rest in the Southern California semi-finals, his older brother as his coach blocked that at some cost to his own career. "I'm not going to wreck my brother's career to win a game."
Big time sports are so competitive these days that
Relatedly, Seth Stephens-Davidowitz's new book "Don’t Trust Your Gut: Using Data to Get What You Really Want in Life," has a chapter on sports performance. Identical twins are more common in sports like track and basketball that require some innate skill like sprinting fast or being tall and less common in baseball, where success seems more unfated, being determined among pitchers to a large extent by arm health, while hitting a baseball seems like more of a knack than a phenotypic trait like height or speed, if that dichotomy makes any sense.
Thomas Sowell talks about this in "Discrimination and Disparities". Here's what I highlighted in that chapter:
'A study of national merit scholarships, for example, found that, among finalists from five-child families, the first-born was the finalist more often than the other four siblings combined. First-borns were also a majority of the finalists in two-child, three-child, and four-child families. If there is not equality of outcomes among people born to the same parents and raised under the same roof, why should equality of outcomes be expected - or assumed - when conditions are not nearly so comparable?
Such results are a challenge to believers in either heredity or environment, as those terms are conventionally used.
IQ data from Britain, Germany and the United States showed that the average IQ of first-born children was higher than the average IQ of their later-born siblings. Moreover, the average IQ of second-born children as a group was higher than the average IQ of third-born children.
A similar pattern was found among young men given mental tests for military service in the Netherlands. The first-born averaged higher mental test scores than their siblings, and the other siblings likewise averaged higher scores than those born after them. Similar results were found in mental test results for Norwegians. The sample sizes in these studies ranged into the hundreds of thousands."
"Consider how many things are the same for children born to the same parents and raised under the same roof - race, the family gene pool, economic level, cultural values, educational opportunities, parents' educational and intellectual levels, as well as the family's relatives, neighbors and friends - and yet the difference in birth order along has made a demonstrable difference in outcomes.
Whatever the general advantages or disadvantages the children in a particular family may have, the only obvious advantage that applies only to the first-born, or to an only child, is the undivided attention of the parents during early childhood development. "
Dr. Sapolsky mentions this exact phenomenon in a lecture series on human behavioral biology. The data discounts fetal resource depletion and parental attention. He indicates that the current working hypothesis is that the firstborn is forced into a tutoring or mentorship role.
Hey, I'm one of those edge cases. I'm the oldest of two siblings, but my mother gave a child up for adoption 8 years before I was born. And I'm a classic older child, so one point for some sort of social vs. biological effect.
My guess would be that the difference is variance. Two populations with the same mean will be measured to have the same mean. This is often interpreted as no difference. However, if the standard deviations are different, the tails will look very different. In this situation, that would mean that first born have higher variance. Has anyone looked at the negative end of the spectrum? (As a reader, I’m obviously designating my end of the spectrum as the good side.) what’s the birth order like in prison populations?
>Despite large environmental differences between the areas, we find remarkably consistent results: in families with two or more children, secondborn boys are 20%–40% more likely to be disciplined in school and enter the criminal justice system than are their firstborn male siblings.
Main results (excerpt):
> While we do not find the primary mechanism that results in the higher delinquency outcomes later in life, we do find relatively precise zero results on some of these measures. On others we find decreased direct parental investment in the form of time spent with parents in favor of indirect investment in the form of formal childcare arrangements. Whether and how these differences could affect observed later life gaps is an open research question. This leaves us with parental investments within the home (as opposed to time out of the home and in the labor force) and sibling influences as our leading explanations for the birth order results.
> We find no evidence that second-born children are less healthy, and indeed second-born children appear to be healthier at birth and have lower rates of disability in childhood. We also find no evidence that parents invest less in second-born children’s education. These children attend no-worse schools and are more likely to attend pre-kindergarten. We consider differences in parental attention as a potential contributing factor to the gaps in delinquency across the birth order. Second-born children tend to have less maternal attention than do their older siblings because first-born children experience their mother’s maternity leaves and temporarily reduced labor market participation both following their own births as well as following the birth of the second-born. Therefore, in addition to the fact that first-borns experience undivided attention until the arrival of the second-born, we discovered that the arrival of the second-born child has the potential to extend the early-childhood parental investment in the first-born child
> The firstborn has role models, who are adults. And the second, later-born children have role models who are slightly irrational 2-year-olds, you know, their older siblings. Both the parental investments are different, and the sibling influences probably contribute to these differences we see in labor market and what we find in delinquency. It's just very difficult to separate those two things because they happen at the same time.
In other words, maybe the first-born is a worse role model than the parents.
>One problem with (1) - wouldn’t you expect smaller effects as age gaps get lower? If it’s about having quality time alone with Mom, someone with a sibling one year younger than themselves only got one year of quality time; someone with a sibling five years younger got five years. But it looks like the birth order effect is stronger for someone with a one-year-age-gap sibling than a five-year one. Either the first year is very important, or I’m missing something.
An older sibling one year older is always doing fresh things, and their one year younger sibling is always just redoing the thing that just happened- there is no novelty, less enthusiasm etc. With a greater age gap, the older sibling naturally ages out of intensive parenting at some point during the younger sibling's development and the younger sibling gets more attention, i.e. older sibling goes off to college when younger sibling is in 7th grade, younger sibling is now effectively an only sibling.
Making my first comment because I have actually thought a lot about this lately. Some anecdotal data to introduce my own theory.
My buddies from school are all very analytical, competent, and successful. Three stand out and they are all firstborn. They also stand out because their younger siblings struggled more and were more rebellious. Add to this general trend a half-exception: my brother and I.
I am a twin. Technically, I'm younger, but I have always been more academic and I was the first one to read, so I was the one who started with a natural advantage. My brother has indeed been "more interested in excelling in areas other than school, like trying to cultivate a special talent that the firstborn [or in this case younger twin] doesn't have." I still remember crying when I realized I wasn't as musical as him and he is now a composer, yet alas, not a frequent reader of ACX.
So here's my theory: older siblings start off with a natural advantage in intelligence. So similar to your second theory (in twins that advantage is decided more randomly because they are otherwise so "identical"). However, younger siblings don't immediately start to differentiate themselves. Instead they are PRESSURED to match the performance of their older siblings. Some of them just can't compete and rebel. They are then led away from traditional marks of success. So even if they could have succeeded in school just as well, they are discouraged from making that their priority because they feel they will never be as good as their older sibling. I still remember the tough fights about grades and music school that my brother and mom had, and I was mentioned way too many times.
Also if there is more distance in age, there is less pressure; their lower performance is more expected since they are so much younger and "eventually they'll catch up." This explains why there is a stronger effect for one-year gaps than five-year gaps. The gaps between my friends and their younger siblings are between 1 and 2. In our clique there are also younger siblings, but they are much further apart in age. These exceptions are the way it SHOULD work: younger siblings aren't constantly pressured into feeling naturally inferior and instead they can just learn from the successes and failures of the older sibling.
Finally, you mentioned that twins are underrepresented among your readers and also receive less attention from their parents. Well my mom gave both of us a LOT of attention as kids, so here I am.
What if the actual effect isn't intelligence, but being antisocial? Could it be that siblings with less time with another kid in the house had more introversion or more non-social pastimes, like reading or computers? I think this is a more direct match to the data. This also predicts that only children should be overrepresented; I imagine this is easy to check. I would also be surprised if there wasn't a question about introversion to compare to, though I don't know if a baseline is easy to find.
All speculation based on my own experience as an apparently exceedingly average ACX reader (late 20s, male, white, first born, programmer), but my guess would be that younger children consistently demanding more attention from parents socially conditions older children, especially first borns, to have a figure-it-out/do-it-yourself mentality (higher conscientiousness and industriousness?), which leads to a higher representation in your readership.
It tracks with my own experience that I was more independent and self reliant earlier due to the attention my parents needed to give to my younger siblings, and i think both a high delta in attention % between siblings after a younger sibling is born and the absolute difference in attention % may play a role, in addition to things that other commenters have mentioned like playing 3rd parent.
Interested in others opionions and experiences here, there should be a few other first borns in the comments, does this track for you too?
Every time this comes up I consider commenting, so I guess I'm commenting this time.
A few years ago, my mom (a PhD in developmental psychology) was observing my nephews and said offhand "it's so hard being an oldest child. You have to do everything first."
I asked her about that (being her oldest child) and she pointed out that younger siblings can watch their older siblings, and especially in early childhood, this gives them a lot of information on things like "if I throw a giant fit, what will happen to me?" And "how long can I get away with something before mom and dad notice?" And even "will I grow up? How long will that take? What will happen?"
Oldest siblings (and only children) are pushed to take on new ideas and experiences, that's just how the world works: you start with very limited data and figure it out yourself.
Younger siblings can, if they choose to, observe the results their older siblings got and coast, or at least use that data to limit their options.
So it makes sense for older siblings to be reading blogs about new ideas, technologies, and perspectives. The younger siblings out there will wait to see how it turned out for us before they get into it.
I think we were looking at it backwards. We were looking at this through the lens of, "why is the second born child less intelligent or why is the first born child more intelligent". I think it's worth checking statistically to see if a younger sibling makes the older sibling more intelligent. As in, having a younger sibling provides more learning opportunities from a fly on the wall perspective AT a level that is more easily comprehensible than the level that adults usually speak to their children, which is just outside of the child's current abilities. So, the hypotheses are, 1) having a younger sibling provides more easily understood learning opportunities without direct emotional cost of failure AND 2) a setting that perches the elder sibling in a relationship as an outlier with the potential to compound successes. So, we need to know what baseline IQ of only children are and see if the younger children are actually less intelligent than baseline or the older children are in fact made smarter by having a younger sibling.
I like the quote in the first paragraph, "correlate birth order with things." I use to characterize this kind of psycho research as correlating things with other things, happy to find it here. Folks, please remember, correlations are rarely zero in non randomized groups, so this stuff is much less interesting than you think. Moreover, I think I recognize a pattern, that the author of this blog tries to please us by basically telling us, once in a month, how smart we are. Nah, I'm not.
And, of course, the obligatory call for "we need more research" in the end, like a plumber who wants to solve the world's problems with more pipes.
Did you read the post? I'm not doing correlations, and I'm not just finding that some correlation exists, I'm finding that the first to secondborn ratio is about twice what you would expect, in a sample of thousands of people. Recommend re-reading this and if you're still confused I can try to explain specific parts to you.
Could it be reverse causation, that parents who have a especially smart first child are pleased and are more likely to have a second child whose ability reverts to the mean? Did someone do a comparison of all the first-born and second-born, whether they have siblings or not?
> it looks like the birth order effect is stronger for someone with a one-year-age-gap sibling than a five-year one.
The developing brain is particularly sensitive to stress in the first two or three years. Stress exposure modifies the HPA axis to cause a permanent increase in the sensitivity of the stress response. You can see the effect in this graph from (Essex, 2002):
Transitioning from being an only child to being a sibling is an extremely stressful event, so you'd expect it to trigger that effect. Note that although we think of stress reactivity as 'bad', high stress reactivity is likely linked to later 'overachieving', as well as to a higher risk of mental health disorders in later life.
Another factor that may complicate things is childcare -- having a new baby probably affects whether the older one is placed in daycare. Center based daycare in particular is known to substantially raise cortisol levels. Cortisol rises are actually smaller for younger children (cortisol response hasn't settled down yet + lower bioavailability of cortisol) but the measured long-term effects on younger children are larger. So e.g. one causal study of universal childcare expansion of found
> this policy resulted in a rise of anxiety of children exposed to this new program of between 60 percent and 150 percent, and a decline in motor/social skills of between 8 percent and 20 percent.
First child is more likely to have "high functioning" autism than subsequent children (yes, I know the autistic community doesn't like that term, but unless they can come up with a better one, I'll use it in inverted commas). One theory is that this could be due to elevated testosterone levels in the womb (which then decline for later children due to the fact that the poor woman has other children to look after, which tends to reduce testosterone levels). I would be very surprised if this wasn't associated in some way with the higher readership of ACX. Full disclosure: I'm autistic.
Could it be that younger siblings are more likely to mention the blog to their older brother than the other way around. Because when you're the younger one, you're always looking for things that the older one hasn't done before you. The older one doesn't feel the push. This could be debunked if you know how people came to the blog.
I think, though I'm having trouble thinking about this clearly, that the entire set of older siblings in the world isn't noticeably older than the entire set of younger siblings in the world, except very weakly since some of the older siblings' would-be younger siblings haven't been born yet, and some of the older siblings have died off while their younger siblings are still alive, neither of which seems relevant to the ACX-reading age demographic.
There is a slight bias. Consider a simple model where all families have exactly two children, at a two-year interval, and everybody lives exactly 70 years. Then nobody gets to be an older sibling until they're two years old, so the average age of older siblings is 36; but the average age of younger siblings is 35. However, the average age of older siblings over two years old is the same as the average age of younger siblings over two years old. So you're good, as long as none of your respondents was less than two years old.
Here's how I'm thinking about this. There is some median age at which a person is likely to discover/become interested in ACX. To take one example at random, I started reading ACX when I was about 29, so for the sake of example, let's call that the "cutoff age". The mean ACX reader is 33 (very close to my own age). All younger siblings who are 33 have older siblings who are also above the cutoff and so would be equally likely included in the sample of ACX readers, but many older siblings who are 33 will have younger siblings who are BELOW the cutoff age and so would be excluded from the sample simply by virtue of being too young to "age in".
You'll get SOME effect from this with only two assumptions: there's a de-facto age-in cutoff for ACX readership, and ACX readership is heavily concentrated close to that cutoff, which I think are both reasonable assumptions. I don't know what the R-squared of this effect is but I expect it to be non-zero, and it might be big enough that the question ceases to be interesting.
Here's another example to illustrate the concept: What percentage of retired Americans are the youngest sibling in their family? I have no idea but I'm willing to bet that it's disproportionately high, and that has nothing to do with birth order making you more likely to retire.
I wonder if it is more subtle than just quantity of attention - it's quite exciting doing things for the first time with your oldest, teaching them things etc. but when you get to younger siblings, you feel you have already done whatever it is, and it is slightly boring so you don't bother quite as much - the novelty value has gone.
Even assuming an implausibly hard selection effect where all ACX readers have IQs and openness of at least 3SD above the mean, I can't get such an extreme skew towards first-borns. 64%, not 72%.
But it can certainly explain part of it. The first-born effects of IQ and openness are known, and known to be small (roughly 1/10th of an SD) but they show up more in the tails. If 1/10th is an underestimate (plausible for a few reasons), maybe we can go to 72%, but it's still an implausibly hard selection effect.
But what's the mechanism behind the IQ and openness effects? They don't show up in every culture (Botzet et al. 2021). The study major effects (Barclay & Myrskylä 2016) show up in Sweden, or I would have considered that perhaps in the US with the high cost of tertiary education, first borns are more likely to have their parents be able to pay for it.
> n <- 1e8
> bo <- rep(0:1, each = n/2)
> ops <- rnorm(n, 99, 14.98) + bo * 1.5
> iq <- 99 + 0.3 * (ops-100) + 0.95 * rnorm(n, 0, 14.98) + bo * 1.5
8000 people took the survey. If they're selected out of a worldwide population of at least 250 million English speaking adults, that selection is 4 SD above the mean in whatever traits make people read and fill out surveys on ACX.
If those people are 3 SD above the mean on the traits affected by birth order(IQ, openness etc), that would imply a correlation of 0.75 with the traits that make you an ACX-survey-filler-outer(IQ, openness, conscientiousness?). Sounds reasonable to me.
By the way, when you say 1/10th of an SD, is that an average difference between siblings, or an average between first-born and second-born siblings? From Sandra Black's Swedish study, the difference between the first-born and second-born was the largest(illustration: https://emilkirkegaard.dk/en/wp-content/uploads/black1.jpg), closer to 2/10th of an SD in cognitive ability I believe.
Also, if you're 1/10th SD above average on IQ, and 1/10th SD above average on "personality", then you're more than 1/10th above average on a composite of those traits.
There are several samples in the study though and she gave 10% as the approximate overall effect. But yeah, could be a bit higher. Barclay also has 2.3 IQ points in a very large sample. It still doesn't get us to 72%, but it could explain part. I think Scott has IQ and openness data. I doubt >3SD selection is what he finds, if he has tests with norm data available (easy for SAT, maybe not for openness).
Let's say ACX survey takers are at least 4 SD above the mean on a composite "ACX survey taker"-trait as previously calculated. For older siblings to make up 72% of the people 4 SD above the mean in a would require about a 0.22 SD advantage in the trait.
If that trait is a composite of IQ and openness(or a composite of IQ and many relevant personality traits), then it would only require older siblings to be 0.16 SD(2.4 IQ points) above the mean in each of those two traits, assuming they both matter equally.
Maybe you know enough about birth order personality effects to rule out such an advantage in openness or any ACX-relevant personality composite, but otherwise it seems like decent explanation for the 72% older siblings to me.
> I additionally tried to compare two different types of social firstbornness - one where no older siblings lived in the house when you were growing up, and one where your parents had never parented another child. There weren’t many people discordant on these two measures (29 vs. 20 respectively), but for what it’s worth, the ratio was in favor of the first type.
I can't tell what this means. As far as I understand the terminology, the set of "people discordant on these two measures" unambiguously refers to every subject whose answers to question (1) and question (2) do not match. But that is a single set of people which cannot have population 29 and population 20. Who's in the group of 29? Who's in the group of 20?
> So for example, since the older sibling will always be smarter than the younger when they’re both young children
This claim can't pass a laugh test. How long did it take your little brother to be better than you at playing piano?
Great post, but I’ve got 3 kids and my immediate n=1 explanation didn’t really fit with your 2 suggestions. Seems to me it’s explained by the relationship dynamics of having a sibling but is NOT really to do with differentiation from each other. Young children who are 2nd siblings and relatively close in age to no. 1 seem to get a huge proportion of their social and intellectual development from the older sibling, regardless of gender and differentiation. I think that that close bond and dynamic reduces their drive to seek new sources of information (such as ACT) because the older sibling has such a strong influence in shaping their frameworks for viewing the world - even when they actively/ innately differentiate from the older sibling’s frameworks. In addition, the older sibling becomes accustomed to socializing the younger into their (older sibling’s) developing discourse. They are likely to feel rewarded/ fulfilled for this from the younger sibling’s receptiveness to taking in the info from them. This creates in aggregate a lasting disposition to seek new sources of info. Again, it’s easy to see that this may only play out within narrower samples, such as your readership, or with specific classes of info, but be lost at the general population/ broad context level.
My observation: The first child learns to be a boss by bossing around their parents first.
New parents want to be good parents and attend to their child, so they listen for what the child wants. They have time, attention and no experience at parenting at this point.
That first child can end up running the parents lives. And with that experience, any later children will be less experienced at being the boss and the parents will be more experiences at rearing children. But at the same time, the first child has already learned to dominate and doesn't like giving up that power. The second child might rebel and try to take a different path in opposition to the power of the older sibling, but a third or fourth child has to fit in some how. After all, the decision-makers, oldest child, parents and second sibling will make the rules.
Anyone care to comment on their observations of first child being the boss of the family?
As the technically ‘middle’ child of triplets (ten minute age gap solely due to birth weight) this birth order research always seemed a bit silly to me, but I suppose that’s mostly because of my personal unfamiliarity with the phenomenon of sibling age gaps.
From what I remember (being the 1st, and completing fitting the findings: I read ACT, I am in STEM while my brother choose a more artistic path), we should not look at parent investment or anything parent related (which fit well with parenting generally having little effect), but at direct relation between siblings. We formed a 2-member gang with my brother, with the eldest (me) the natural and not challengeable (at least until we were both past 17) leader. I think it fits my model of genetic>peer>parent in term of relative influence on every psychometric, when peers are siblings and later friends/schoolmates. Your siblings are part of your peers, the most important or even only one before 6yo...
The only thing I find quite strange in the data is that I expected a huge difference between same sex and opposite sex siblings: The brother sister dynamic is very different than between brothers (among sisters I do not know because I have no direct experience nor first hand observation among my friend's families)
good point. I do not have any sister, and among my friends I have observed the brother-sister dynamic only for older siblings, at 12y+ I would say....But now that my friends and I are at the parent age, for pre-school brother/sister, it's indeed more similar. This early they already have very different male/female interests but it's the oldest that is the leader, boy or girl. Still, given the diverging interests, I would have expected a larger difference....except if the key moment is really when the younger child have almost zero non-famlily interraction (before 3yo).
It's often said that all of the original seven Mercury astronauts were first borns or only children and 14 of the next 16 American men into space were the same. But I don't see a link to a study.
As a Catholic Baby Boomer, my being a (rare) only child was a big deal when I was a kid in the 1960s. My friends assumed I got more presents on my birthday than they did (which I probably did). Adults were always telling me I was emotionally deprived by not having any siblings (which I probably was)
Now that only children are common, I never about them anymore.
Maybe for many younger siblings, if they do nerdy things at home, they get teased by their older sibling. And then "not being a nerd" might become part of their identity. But IQ other major life stats stay the same. This hypothesis would weakly predict that male dancers are much more likely to be older siblings.
Say, for example, 20% of older siblings are nerdy compared to 5% of younger siblings. So up to 80% of the time, a younger sibling would get teased for nerdiness. Whereas if a younger sibling tries to tease an older sibling, the older sibling rolls his/her eyes.
It seems kind of strange that the focus is all on intelligence and academic achievement. The effect of birth order on readership of this blog (71/29) seems to be much stronger than for physics laureates (60/40) . While we might like to think that reading ACT is as good a proxy of intelligence as a physics Nobel, it suggests to me that it's a more specific personality trait that first-borns tend to have that correlates with attraction to a weird rationalist blog.
An interesting birth-order sports anecdote is Fernando Valenzuela, the Los Angeles Dodgers baseball pitcher who caused a sensation in the early 1980s. He was the youngest in a family of 12 in Mexico. He had one of the dumpier bodies you ever saw on a famous athlete, but was a baseball genius.
His older brothers were ardent baseball players too, but they had to go to work when still young. But Fernando was the adored baby of the family and was allowed to do nothing but play ball.
Interesting point. Previous attempts here to work out the age effect have assumed a uniform population size. In reality, global population is increasing, leading to a greater likelihood of someone being "not old enough" than someone being "too old". But if it's a subpopulation that's both rapidly growing and just coming of age, this could have a big effect.
*edit* Never mind. There would have been firstborns too old to join the subpopulation.
Is there some "recommended" way of sharing analyses? I'm curious about the the bio/social split, if I find something interesting I'd like to share.
I'm looking into it because I was a bit confused by Bucky's part about the biological vs social effects. If I get it right, the point of their selection (bio > 0, social = 0) is to filter people who had biological siblings at some point, but currently have none? I get that this would "isolate" the purely biological effects, but it also seems to introduce all kinds of other biases (it's not really representative of the general population) and additionally it doesn't even mean that there were no social siblings *at a previous time* in which it could have had an influence.
(btw the link to the CSV file in the old blog seems to be broken, excel file is fine)
Of my three kids, each 21 months apart, the younger two strongly tend to do what their older siblings are doing. They learned early on that to imitate is usually a good choice, so they do it a lot and it serves them well. The oldest imitates adults of course, but not nearly as automatically, because she knows many things we do she cannot or is not allowed to do. What she does do, however, is explain things to her younger siblings and give them instructions. That's what she got from us and it's harder than imitation but I think she has learned to think in that way.
None of them are in school yet, but I expect the oldest will find explanation-heavy subjects like math and physics slightly more intuitive than imitation-heavy subjects like music and physical education, and maybe her younger siblings will tend to go the other way.
Of course the ratsphere is super into explanation and I would argue it grossly neglects imitation. When we try to persuade, we try to craft excellent explanations, rather than build shiny role models to champion beliefs we want others to share. But much, arguably most, of human learning in the general population happens through imitation. Especially on ethical matters.
Could sleep disturbance in infancy be an explanation? When you are the firstborn, your parents try very hard to avoid waking you up, and you enjoy uninterrupted sleep. When you have a toddler sibling, you get waken up all the time. If uninterrupted sleep is important for infant brain development, it would explain why the effect disappear as age gap increase (a 5 year old can keep quiet, a 2 year old can't), and it would explain the twins (twins wake each other up all the time).
Yeah, that is an implication and I should have thought of that. I don't see any discussion about only children in either post, in part because of controls (only children would tend to have more educated parents, so they could be overrepresented here even if the hypothesis is correct) and because the % of only children has increased over time (and is different in different countries).
Could there be some selection effect (at least for ACX readers)? Readership is smart and in an age range such that a smart older bro also reads ACX but the smart younger bro does cooler other stuff(?) Accelerating change of internet, older generations more similar than newer, I dunno.
I guess I'm an aberration. I wasn't around for Slate Star Codex, but I read this blog. I am the third of three siblings in order of birth. I know for certain that one of the others doesn't read this blog because he's dead, and I doubt the other one does because he's not a big blog reader and mostly not interested in even the wide variety of subject matter covered here (if you start covering the NFL and military matters a lot, he might notice).
Eh, I'm the 2nd of 2 and I don't think my brother reads ACX, despite being smarter than me. "Exception" might be a better word than aberration.
Side note, I wish Scott had included families of 1 child in his analysis. Maybe that would have thrown things off or been more complex; it's just that I'm curious. Rumor tells me only children are "like oldest children only more so," and I'd really like to know how likely this is to be true, since I have an only child.
I think that Caplan and others tend to overstate the results of the research he presents. I don't remember the effect sizes of the top of my head, but I think his data indicated that something like 1 SD "better" parenting led to 2-3 point increases in IQ. This is a much more modest effect than most people's intuition, but still significant, especially for tail outcomes.
The social sibling effect could easily raise the IQ of first borns by a couple points, which is going to lead to noticeably skewed ratios for the tails. If, among well educated parents, the average 1st-born IQ is 114 and 2nd-born is 110, then among IQ 140+ individuals who are 1st or 2nd born, 65% will be 1st born.
At least for Nobels, etc., I'm going to blindly guess it's the differentiation thing.
I know both me and my (younger) brother were absolutely determined we weren't going to do STEM related stuff because our dad was an accomplished programmer and adopting his profession in a world of infinite choice felt somehow medieval.
But it turns out genetics is strong. We're both garbage in our chosen careers because young folks stink at choosing careers. But we've both managed to fall into that kind of work later in life.
In a family that was more traditional (as I gather traditionally successful folks' families tend to be), I would have been pressured into my dad's profession (and inherited his business contacts) while my brother would have had to work something else out.
Do I think that's a recipe for general success in a typical family? Force your kid to do what you do? Absolutely not.
But when talking about people who perform at the very top of their fields, probably genetics and nurture both have to be moving in the same direction. And really you can only do that kind of intensive training with one kid, and probably you pick the first one.
It's a tough course to steer. Forcing children into anything isn't going to go well, but seeing that Tommy has talents for X and ushering him along the path towards a career in X is different.
There is so much cultural pressure built up around the eldest son to be the bearer of the family heritage, even if that is relieved a lot in modern Western society, that it's hard to say. Take Leo Varadkar, who is an Irish politician and in government at present. Father a doctor, mother a nurse. Two older sisters - doctors. He's the third child and only son. Trained and qualified as a doctor, then went into politics.
It seems like there is an assumption being smuggled in here, namely that "reading SSC is correlated more with intelligence than some other trait."
What if that other trait is, say, conscientiousness, or being thing-vs-people oriented?
Like, what if older siblings are less likely to be as people oriented because they didn’t have someone a few years older than them as one of the most interesting, salient features of their environment?
I’d love to see if the "first borns are overweighted" thing is true even if you filter out people who put themselves on the autism spectrum. That’s the approach I’d be taking - I’d try and find subsets of the data where the trend dissappears.
This is only slightly related, but... Do first-borns tend to partner with first-borns? This is a trend I've noticed in my family and wondered about before. If there is some set of characteristics associated with being the first-born, it could be that people with a lot of those might be attracted to other people with a lot of those as well. This could explain not only my observation but also why these people clustered together as this blog's community.
With infinite resources, we could also survey a whole university, say, and see if people's chosen major has predictive power over their birth order. I think someone already mentioned this, but I get the feeling first-borns are more likely to be pressured into studying something traditional and "safe" than their younger siblings.
Curiously, there's fairly strong evidence that men with older siblings are much more likely to be gay/bisexual. It used to be thought that only the number of older *brothers* mattered, but IIRC later statistical analyses showed that there was no difference between having older brothers and older sisters. This has been replicated across multiple different cultures, but I'm not sure if we can totally exclude sociocultural explanations.
If this is biological, it raises thorny ethical issues. If we find a way to overcome birth order effects on intelligence, it might also have the effect of changing children's sexual orientation, if the two share the same mechanism.
What about the likelihood of finding only children in your sample? If hypothesis 1 is true, we should observe more only children when controlling for the relevant variables (that is, considering that only children are possibly more likely to come from unmarried couples, short marriages, etc).
Some people commented on the possibility of a "mentoring effect" for first-born children. If only children aren't more likely to appear in the sample, then that hypothesis could be more likely.
I also think about other variables that are not related to intelligence. Like some readers mentioned, the likelihood of the older sibling finding more introspective hobbies. I also consider the likelihood of the younger sibling learning how to be more "charming" and social in order to get more attention and differentiate himself from the always older and more intelligent sibling.
"One problem with (1) - wouldn’t you expect smaller effects as age gaps get lower? If it’s about having quality time alone with Mom, someone with a sibling one year younger than themselves only got one year of quality time; someone with a sibling five years younger got five years. But it looks like the birth order effect is stronger for someone with a one-year-age-gap sibling than a five-year one. Either the first year is very important, or I’m missing something."
Babies and toddlers are higher-maintenance. If the older sibling is already seven or eight when a new baby arrives, the baby obviously has more pressing needs, and the parents will in fact switch tracks to doting on the baby, possibly growing somewhat neglectful of the older sibling. Also, a much older sibling is old enough to, themself, help care for the baby, which will help smooth things over; a two-year-old is pretty useless (and sometimes actively harmful) around a newborn.
Yes, this. Plus if the older sibling is old enough to be at school, then the baby is effectively a firstborn for six hours a day - both in the positive sense that they get more 1:1 parental attention and in the negative sense that they don't have a close peer to play with and imitate.
"a two-year-old is pretty useless (and sometimes actively harmful) around a newborn."
Aha, yeah. I nearly gave my mother a heart attack years back, my younger sister was born when I was two, so when she was about six months or so (EDIT: maybe she was a year old, so I would have been three) my mother put her out in the garden in the pram. She started crying, so like a good Big Sister I took her out of the pram and carried her indoors. She was slipping out of my grasp but I managed to grab her like a parcel, more or less. When my mother saw me proudly carrying her in (with a shaky grasp around her middle), she lunged to take her from me
I was five when my younger brother was born, so I was more helpful and less dangerous then!
My explanation is that first children are required to take the lead on a whole host of situations where the siblings are involved. By taking the lead, out of necessity more often than not, the other children are relieved of that duty. That trains the younger children to be followers and the older children to be leaders. This explains why children a few years older create a stronger tendency, and also why after seven years it drops off - too little difference makes it less obvious who is in charge, while too much difference reduces interactions.
Why does this affect ACX readership? I think this blog interests people who are trying to figure out the world and make decisions about practical and moral effects. For someone who thinks that's important, there's a clear reason to follow the blog. For someone trained to let others do that work, then there's much less reason to follow a blog like this.
I’m an older sibling but I can’t imagine my younger brother, who has a higher IQ, earns more and is generally more successful, taking an interest in this blog. I tend to experiment a little more, have wider interests and deep dive into obscurity whereas he does the normal stuff very well
" Firstborns get more quality time with their parents during the first few years of childhood."
I am puzzled. You seem to have an assumption that reading SSC makes firstborns better in some sense. But maybe it is something not-so-glorious that happens because the firstborns get more undivided attention of their parents: I can imagine the undivided attention makes them prone to navel-gazing and overestimating the extent their personal subjective experience and thoughts in general are super special and important and worth of loudly explaining to everyone around? Or, in other words, increased interest in philosophy (edit, and lets be realistic, strong political opinions)
The effect would be then naturally strongest in people who were the only child in the family. Eldest sibling has to learn to share when the siblings are born.
Does the data show there are more SSC readers who had no social siblings, compared to overall amount of them in the respective age cohorts?
What we need to do is link up the birth order data with the (self-reported) mental illness/psychological problems data and see how that stacks. I think many of us are somewhere on the autism spectrum or thereabouts, so niche interests, obsessiveness, and wanting to talk about our pet subjects at length and in detail all go along with that.
Plonk us down in a group of the like-minded who will pay attention to what we're saying, argue back, and talk about *their* pet subjects at length and in detail, and who won't be "don't you have any *normal* hobbies like going out to bars to have fun, pop music, and movies?", and we have our little acre of paradise right here.
Eldest sibling does have to learn to share but eldest sibling also often gets landed with the caretaker/teacher role, e.g. "help Jimmy with his maths homework; did Susie change her dress?; set the table for tea and call your siblings in" etc. so being bossy and lecturing are habits you pick up.
I'm really surprised that this (IMO very weak) conjecture is getting a pass:
"Given that readers of this blog are highly-educated (about 37% have masters or PhDs) and mostly in STEM (41% programmers of some sort), plausibly birth order affects something about intelligence, education, or STEM orientation"
I didn't take the survey. Did you ask if the siblings have advanced degrees or work in stem? Have you directly addressed the possibility that older siblings might just enjoy reading blogs more? Or as another user points out, be inclined to take surveys more? Or spend time on using the internet more? Or any number if other factors?
I guess I am the only woman and mom here and oldest daughter of a line of oldest daughters, and I had to smile a little because I never bullied my younger siblings: in fact in many times I was their mother and nurturer when I had to babysit for them. Now my sons.. well I see what you are talking about: though i always encourage them to get along: it was the second son who was the bully if any of that went on.. well gotta go.. yes. that is just our little story: circumstances vary.. have doctor's appts. today.
There may be as many as several of us women on here! I don't know about how many mothers, though.
Again, it probably is that girls get steered (pushed into?) the mothering role, also depending on how many younger children; certainly, for our 5+ families, the division of labour with the eldest daughter taking up the slack for the mother is going to be there. Boys may have more social/cultural licence when it comes to 'rough and tumble' and 'fighting your corner' which can slide over the line into bullying.
I think the eldest tend to invoke their delegated authority over the younger ones, in most cases, and if that fails fall back on "I'm telling Mom and Dad", not so much as going for bullying. I can see younger (especially middle) kids bullying because they don't have the same quasi-parental status so to make their younger sibs fall into line, they have to bully.
Anecdotal, but this makes complete sense when it comes to my childhood. Younger of 2, both pretty intelligent.
When at a young age my brother discovered e.g. that you can use strategy in games, he found that it let him win every game against me, so he focused on that discovery, and soon even won some games against my parents.
When at a young age I discovered that I could use strategy at games I found out that it does not matter, I never won a game not entirely on chance no matter what I did (because my brother of course by then was way better at strategy). So I stopped playing Risk, or Chess, or any games like that after a while, because what fun is it if you never win?
Sports was similar.
So my brother learned young that if you dedicate yourself to something, you can excel. Tends to produce a strong work ethic.
I learned young that what I do doesn't really change how things turn out.
You may not be surprised that his academic achievements surpass mine...
Wait, we've got to supplement this one, as we have actual data/knowledge here that goes beyond anecdote.
In sports - at least in basketball where it has been most studied - we know that (i) younger siblings do better and (ii) a factor in a talented kid developing into a pro-level talent is them playing a couple of levels up throughout their youth, with kids a couple of years older than them (something younger siblings can do naturally, obviously).
Essentially, when you're talented it's easy to stagnate, to rely on your physical advantages and fail to develop the finer points of your game. This is especially so with guard skills (three point shooting being a classic David vs Goliath strategy) that you don't need to bother with when you can just overwhelm your opponents inside.
Applied more broadly, older siblings crushing younger, inferior competition shouldn't make them great at those activities - there's obviously a lot of sips between the cup and the lip in your brother's story, in going from beating his kid younger sibling at strategy to beating his parents.
Whether younger siblings develop effective David strategies or just keep getting pummeled and lose interest may depend on the activity. But there's at least some documented advantage for long-term high-level skill development for younger siblings who are able to hang with their older siblings and compete
From my experience as a South Asian/East Indian immigrant (not necessarily high-income) is that among my first cousins on both sides, the oldest of the family is always the most 'traditionally successful' in levels of education, type of field, and income. This is regardless of gender or varying family income or significant life events.
Culturally, there's an idea that the oldest needs to make it for the family and the pressure is intense to do really well in school and everything else. I am not American either, so I can't imagine how much more formalized that gauntlet is with US-style admissions, etc. Perhaps there's something similar for Western families but less explicit and less tear-inducing.
in the developmental psych literature, does parental pressure itself do better than parental investment in the form of Advanced Einstein baby toys or reading books? Its hard to disentangle the two, of course both nature and nurture-wise.
I think some consideration to the ideas outlined in "Deep Nutrition" a book by Dr. Cate Shanahan which talks a great deal about the findings made by Dr. Weston Price should be included. The idea is that nutrition plays an enormous role in physical health and intelligence. Their assertion is that inadequate birth spacing (less than the 4 years they consider ideal) means that the mother is nutritionally deficient which is then passed on to the child creating less than ideal conditions for proper growth and maturation. I am by no means an expert, but I think this merits discussion.
It's fascinating to see this effect in the data - it's *quite* clear in my own life. I'm the oldest, a lifelong nerd. One of my brothers is two years younger and always rejected academic stuff - not necessarily less smart, just much less interested in it. Then my other brother is several years younger than both of us, and pretty nerdy. Point for point reflective of what we see in this data.
I think nerdiness has more to do with it than intelligence or education. I'm about as smart as my sister, and we're equally well educated, but I have a hard time imagining her enjoying this blog. (I'm the younger one, though.)