Lifespan depends on month of birth (2000) | Hacker News Reader

Lifespan depends on month of birth (2000)

pnas.org

116 pointsPetiver55 comments

Leonid Gavrilov and Natalia Gavrilova are a good resource for this sort of demography and epidemiology of aging. [1] In particular they have a lot of material on season of birth effects. [2]

Tangentially related to the seasonal effects is data suggesting that solar cycles also influence longevity, possibly through quite indirect responses to levels of UV exposure in pregnant women. [3] You can see how this might be relevant to season as well, but it is only one of a number of mechanistic theories on the subject.

This sort of thing ties back in to application of reliability theory to aging [4], wherein to make the models fit the observed data individuals have to be born with a preexisting non-zero damage load.

[1]: http://longevity-science.org/

[2]: http://longevity-science.org/Season-of-Birth.pdf

[3]: http://rspb.royalsocietypublishing.org/content/282/1801/2014...

[4]: https://en.wikipedia.org/wiki/Reliability_theory_of_aging_an...

But the data in the post does not seem to be super correlated to seasons. You can see that the decrease in life expectancy occurs way before spring (as soon as February) which is still supposed to be a winter month. So the season link is weak.

I'm thinking it has to do with the continuous day/night cycle. It'll be interesting to see if there are any changes with latitude, in particular, what happens in countries closer to the equator.

But is there any rationale behind that anyway ? Do we see people born during summer being usually in poorer health than people born during November-December, let's say ?

The differential exposure to UVB, that is, sunlight, was the first thing I thought of reading this article. It is especially likely among people born in the industrial-era 19th century, at a time before the nature of the link between health and sunlight was clearly identified.

Around a hundred years ago, the role of UV in preventing rickets was confirmed. By the 1930's, vitamin D supplementation of foods became widespread, and the most obvious manifestations of deficiency (e.g., rickets) essentially disappeared. Studied intensively since the 1930's numerous effects of D (or lack of it) have been reported, and recently interest has exploded.

Increasingly evidence implicates low vitamin D levels in a wide range of conditions, such as autoimmune disorders, cancer, and metabolic disease, that are likely contributors to reduction of life expectancy.

Seasonality of vitamin D level is well known. In northern latitudes (>=45th parallel) the nadir occurs around March-April, and the peak is measured on average in September-October. The timing of these fluctuations appears quite congruent with the article's assertions about life expectancy, so it's reasonable to think there is a connection.

For anyone interested in vitamin D, I highly recommend this excellent review: http://www.tandfonline.com/doi/full/10.4161/derm.24494

The correlation is in terms of fractions of a year so they add up to an average of plus-or-minus a few month over a lifetime. The three curves don't seem to have a very close relationship to each other though there are some similarities.

I assume that if they are getting a meaningful correlation, it is over a large amount of data. It seems logical that at large scale, date born would correlate with a variety of things in various areas - social class and wealth come to mind.

I think it's a matter of taste whether one says "wow, throw enough data together you find unexpected correlations" or "garbage in, garbage out". A life insurance company could eek out a bit of profit changing its price structure on this data - if it didn't have far more to lose if the use of such tactics became public.

I bet someone can use this to claim astrology has validity too.

Also this is only about three countries selected a priori, so not a really good sample, even if the number of individuals is huge...

Would be interesting to look at the 'stresses' that occur within schools. My 11 year old daughter was born on the 30th of August. Ergo she is the youngest in the year at school. She is studying with people that have a year of development on her. She is doing fantastically well but works hard to achieve that result.

My hypothesis is that the age cut off in schools, unduly puts stress on the younger children in a class, that long term, impact on health in later years. In effect which month you are born in affects the stress you will feel during your school years as it will be inherently 'harder' for you.

I'd think the opposite. Since she is used to having to work hard from a young age, she will be more successful in life.

Others who coasted through—whom were used to everything being relatively easier—will not have the experience she has already accrued when faced with a harder challenge.

Not sure that makes sense. She potentially experiences 12 years of increased stress compared to her peers. I'm not implying that later in life, she will have better tools for work, but that her body would have spent most of her formative years under more stress.

This would then reflect in a shorter lifespan.

Note however that this concept she would gain the tools for hard working in itself is flawed. Studies have shown that those born in the Summer months have statistically lower exam scores.

If stress from school caused people to die earlier I think we'd know by now.

How would we know? Why?

This is very hard to measure as you would be looking at teaching practices in the 1920-1940s to determine the effects on life expectancy. I think potentially, nationally, this would be possible as most nations have been quite consistent on their school starting times.

I do think having a breakdown by month, not just quarter could be even more insightful.

I don't have a link, but there was a comparative study done on children in Norway during the debate on lowering school starting age (it was 7 years when I was at school - it's 6 years now, and there's some pressure to lower it to 5).

They compared salaries and other factors based on age at school start and found basically no statistically significant differences later in life between those who had been oldest and youngest when starting school.

They did not look at things like health or stress levels, though.

Since she is used to having to work hard from a young age, she will be more successful in life.

By that logic, howabout we start our kids off one year earlier, so they'll be even more successful? Or howabout 2?

Malcom Gladwell's Outsiders book basically covers this topic. Usually the ones who do best are the ones with an early head start. One example that I remember is that oldest kids in class are usually the biggest, that (generally) makes them better at sport early in life. That leads to extra coaching for sports teams, more practice and the effect snowballs.

I'm from Slovenia, where it's exactly the opposite. Being born in November, I was among the younger people in my class. The cut-off point was January or February (of the next calendar year). So it could be that the more stress you have, the longer you live (better personal/immune system development). Alternatively, younger classmates (i.e. people that go to school with peers that are in average older) achieve accelerated development (because they hang out with "more developed" kids) which could also somehow affect life expectancy.

The cutoff benefits the youngest students academically. They get used to working harder to catch up, and never stop working harder, apparently.

https://news.ycombinator.com/item?id=6421473

>Would be interesting to look at the 'stresses' that occur within schools

For this to be an accurate explanation, given their data, schools in the southern hemisphere would have to start school six months later. Do they?

Your question is interesting nonetheless.

It could also have something to do with their development in school. After all the school year is the same for all. Some kids are up to 11 months younger!

This is an area which I've personally experienced, having spent the first half of my school years in Europe, and being a few days shy of the cut-off point for being held back another year, I was always the youngest in my class (I guess on a large enough population, it would mean that similar kids are the smallest as well, on average).

Later on I switched to a southern-hemisphere school, which was out of phase by 6 months, and I was all of a sudden roughly average age in class. Not that it had a huge impact on me, but it was noticeable, which probably means it has some effect on average.

The effect of school could be isolated by comparing different countries with different cutoff dates. I believe the American and UK systems are based on age when the school year starts, so the youngest in the class were born in August. Sweden groups classes by birth year, so the youngest were born in December. The Japanese school year starts in spring.

This is interesting. They seem to think it is in utero effect. Maybe Vitamin D?

I wondered if they considered if the cause is age when starting school. I know that birth month has pretty powerful effects for educational attainment.

What month should I conceive my child to maximize their educational attainment?

I'm only half joking.

Depends on the rules which vary by location, but you want your kid to be the oldest in the class. Your older kid is going to believe she is very bright in 1st grade, and have a lifetime of confidence to achieve great things. The real reason she was such a good student in 1st grade is because she was just older.

I can't cite any off the top of my head, but more recent research has actually directly contradicted this. Being the oldest is definitely a benefit in athletics due to greater size and strength throughout school. And you're right that older kids have an advantage in early grades in academics as well. In later grades though, that advantage disappears, and in fact, reverses. (The hypothesis I've read is that the reversal is due to greater effort being required in early grades by younger children, resulting in better preparedness for later grades, when the innate advantage of slightly greater age becomes less significant.)

I found two papers that seem to back jcampbell1 up.

http://www.mussioassociates.com/PDF_files/June8.pdf

>Results from this study suggest that a substantial number of students will fall behind their peers in meeting reading and numeracy standard and graduating from grade 12, simply because they are the youngest and most immature in their kindergarten class.

http://www.hks.harvard.edu/pepg/PDF/events/Munich/PEPG-04-24...

>The paper explores the strict school enrolment rules to estimate the effect of age at school entry on school achievement for 15-16 year old students in Norway using achievement tests in reading from OECD-PISA. Since enrolment date is common and compulsory for all students born in a particular calendar year, it is possible to identify the pure effect of enrolment age holding the length of schooling constant. The results indicate that the youngest children (born in December) face a significant disadvantage in reading compared to their older classmates.

Note that the first study track them all the way until age 18, and the second until 15-16.

Hmm, I found this New Yorker article supporting my statement: http://www.newyorker.com/tech/elements/youngest-kid-smartest...

However, it seems to be drawing skewed conclusions from the research it cites.

From the article:

"When a group of economists followed Norwegian children born between 1962 and 1988, until the youngest turned eighteen, in 2006, they found that, at age eighteen, children who started school a year later had I.Q. scores that were significantly lower than their younger counterparts. Their earnings also suffered: through age thirty, men who started school later earned less."

From the abstract of the paper it references [1]:

"We find evidence for a small positive effect of starting school younger on IQ scores measured at age 18. In contrast, we find evidence of much larger positive effects of age at test, and these results are very robust. [...] There appears to be a short-run positive effect on earnings of beginning school at a younger age; however, this effect has essentially disappeared by age 30."

Not quite as sensationalist as the summary in the article! Students starting school younger had slightly higher test scores at age 18. However, the age when taking the test had a greater effect, so students starting older would still presumably have an advantage in tests taken at graduation (given that they would still be older). Students starting younger earned more, but only slightly and only until age 30.

I expect the real conclusion to be drawn is that it's silly to obsess over what age to start your kid in school. Start them when they seem ready, and spend more time focusing on supporting them at home (in their education and otherwise) than worrying about the findings of these studies.

[1]: http://www.nber.org/papers/w13969

Ah, I remember that study and have mentioned it elsewhere in the comments here but couldn't remember where it was from.

In Norwegian newspapers one the authors of the study basically wrote about it with the opposite slant of what the New Yorker did, pointing out what the abstract says: that the long term results are the same either way.

Yes but how much does this explain? Even being convinced of the significance of the effect, I'd like to know the magnitude, as well as how much the magnitude changes conditioned on success. E.g. do the very successful (however defined) have a higher likelihood of being the oldest in their class?

> The real reason she was such a good student in 1st grade is because she was just older.

That may be an intuitive reasoning but not sure it holds true. I have seen numerous times the youngest students performing actually better in class than older ones. Do you have hard data to support this ?

I have no hard data, but I have 10 years of anecdotal data (my wife was a 4th grade teacher for ten years and saw just about 1000 students over her career).

She says the younger kids definitely do worse by all measures in 4th grade, in large part because of their lack of maturity.

> in large part because of their lack of maturity.

Maybe that depends on the topic you are teaching then ?

I know this sub-thread started in half-jest, but your response got me thinking:

1. Month of birth directly determines a child's age compared to their classmates.

2. An older child among younger peers performs better academically.

3. Academic aptitude is correlated with income.

4. Income is correlated with longevity.

Admittedly, hypothesis #2 is iffy. And the good old "correlation is not causation" adage applies for #3 and #4.

> And the good old "correlation is not causation" adage applies for #3 and #4.

It does, but it's not hard to come up with plausible causes for the correlation to support in those cases.

It is probably more accurate to say "Academic aptitutde correlates with job stability which correlates with lower overall stress levels which is part of the definition for better health which correlates with longevity."

You could easily reverse 2.

2. A younger child among older peers needs to work harder to achieve comparable results.

3. Work ethics are correlated with income.

This is particularly true in sports. There is a chapter in Malcom Gladwell's Outliers all about it.

Purely speculative but people whom I know to be the most successful in education and career happen to be born between November and January.

Age doesn't matter though. Some of them jumped a few years ahead of their classmates so they surely weren't the oldest in their classes.

My mother did think that being the oldest helps (or as she would put it: "School and work can wait. Enjoy your childhood while you can"), so my brother and I weren't allowed to go to school until we were 8. We graduated from high school at the age of 19 & 20. Both of us had 4.0 GPA and landed full rides in foreign universities and became the first generation of college educated people in the last 4 generations that we could trace back

Speaking from experience, being the youngest appears to have no upsides: you're expected to perform as well as people older than you which creates tons of pressure, you're probably physically smaller which can impact confidence/sports/romance, socially you're a year behind, and it only gets worse later - in many countries the legal drinking age (18) aligns with the average age of entering university which means your options for social development at a crucial stage of development are restricted further, etc.

By the time you're 30, an extra year saved going to school earlier means nothing, and the price you paid in terms of developing neurotic conditions (e.g. perfectionism) which impact confidence and social skills, is most definitely not worth it. My kids will follow a similar path to the one you outlined, and I will strongly encourage them to take a gap year before considering applying to university.

You're projecting. My experiences as the youngest, and smallest in the class until the last year of high school, left me with the following experience - No impact, no differences based on being younger and smaller.

I would say what you're saying are results of being younger are actually results of your personality traits, and being a year older would have still lead you to having similar experiences at school.

You'll probably enjoy The Biodeterminist's Guide to Parenting[1] by Scott Alexander[2]. It's a good overview of environmental effects on cognition. There's a section on birth month, and the answer is: It's a trade-off. If you're sending your child to public school, August is best. This makes them the oldest in their grade, giving them advantages that compound throughout school. For private school, February/March might be best, as winter babies have slightly higher intelligence and social skills.

1. http://squid314.livejournal.com/346391.html

2. You may have heard of him. Scott is a resident psychologist and writer of Slate Star Codex (http://slatestarcodex.com/).

Well I got the shaft. Early August birthday, nearly the youngest in my grade, public school.

    > What month should I conceive my child to maximize their
    > educational attainment? I'm only half joking.

I'm married to a primary school teacher. We are very definitely aiming to have babies born September to December. We aren't even slightly joking about that.

After that you just can't move somewhere that doesn't have the same timing on school years (e.g. Australia).

You want them to be born in September or October (or, more generally, a month or two after the school year starts). Probably October, so that there's no chance they're born in August. They'll be more developed than their July birthday brethren at every stage of their education - it's a very easy boost. For example, exams at 16 in the UK (GCSEs) that by and large everyone takes, you'd expect 54% of autumn birthday children to get 5 grades A*-C. You'd expect 48% of the summer birthday children to achieve this. At age 11, the gap is a slightly larger 8 points in terms of percentage of children who meet expectations.

This is a nontrivial difference, which makes it a trivial decision.

https://www.gov.uk/government/uploads/system/uploads/attachm...

If school starts in september, around christmas.

I've mentioned elsewhere that someone did a study in Norway, comparing salary later in life to age when starting school, and found no statistically significant correlation.

This was published as part of discussion on reducing the school starting age. Basically there's very little apparent benefit - at the age most kids start school, brain development appears to be the big brake on learning speed.

My team of five software developers were all born in December. I didn't think anything of it until I saw this post on HN.

I was wondering about that. I read somewhere else about the birth month effect for professional athletes. Kids whose birthday falls just after the season of their sport starts are up to a year older, bigger and stronger than their peers born earlier in the year, which boosts their confidence to the point that fewer of them quit.

Very curious. Information about people in the tropics would be interesting since climate doesn't vary as much there. I know in Madras, India it was hot pretty much all of the time, except when it rained.

I didn't fully read it but doesn't it look like there is potentially a big survivorship bias in the data?

They touch the subject with their 3rd hypothesis but in general it's possible that you still have a higher life expectancy being born from April-June in the Northern Hemisphere because you have increased chances of reaching 50.

So why this limit of 50 years? What is the expected lifespan for newborns depending on their month of birth?

Or maybe death on the month?

Indeed. If most deaths occur during mid-summer, then people born in autumn live slightly longer on average than people born in spring.

"The differences in lifespan are independent of the seasonal distribution of deaths" That's important of course.

I would think people conceived in warm weather are more likely to be from less stable relationships...

Reply on news.ycombinator.com