The Unz Review - Mobile
A Collection of Interesting, Important, and Controversial Perspectives Largely Excluded from the American Mainstream Media
 BlogviewJames Thompson Archive
Vita Brevis, Dignitatis Inutilis
🔊 Listen RSS
Email This Page to Someone

 Remember My Information


Bookmark Toggle AllToCAdd to LibraryRemove from Library • BShow CommentNext New CommentNext New ReplyRead More
ReplyAgree/Disagree/Etc. More... This Commenter This Thread Hide Thread Display All Comments
These buttons register your public Agreement, Disagreement, Troll, or LOL with the selected comment. They are ONLY available to recent, frequent commenters who have saved their Name+Email using the 'Remember My Information' checkbox, and may also ONLY be used once per hour.
Ignore Commenter Follow Commenter
Search Text Case Sensitive  Exact Words  Include Comments
List of Bookmarks

Longevity birthday candles

If bright people live longer, is it because of the good things that social status provides?

This question may appear to be very hard to answer, because one thing leads to another, and bright people are clever enough to build themselves agreeable places in which to live. One approach is to “control” for this, on the outrageous assumption that bright people have done nothing to deserve their social status, but have it handed to them on a plate by the random vagaries of fate, or the accumulated greed of their pushy ancestors.

First of all, what influence does intelligence measured at age 11 have on longevity? The good news is that a standard-deviation increase in IQ score is associated with a 24% decrease in mortality risk. So, at IQ 115 lifespan is 24% longer than average*. This is good news, together with the 60% increase in wages above the average level from an OECD study.

On the wages front, the effect of intelligence had already been shown by Charles Murray in his well-known 1998 “Income inequality and IQ” in which he compared the earnings of one child in a family with that of another sibling, showing that the effect of intelligence was powerful at creating later life differences even between siblings brought up within the same family environment.

He extended his analysis in subsequent work, looking at the effects of divorce and poverty, finding them to have contributed relatively little to income measures once intelligence was taken into account.

IQ and Income Inequality in a Sample of Sibling Pairs from Advantaged Family Backgrounds. Charles Murray. AMERICAN ECONOMIC REVIEW VOL. 92, NO. 2, MAY 2002 (pp. 339-343)

Having matched individuals to their nearest siblings, Murray then examined whether within-sibling-pair intelligence differences could predict income differences within pairs. Examining the contribution of intelligence within families circumvented the need to obtain a range of SES measures, as it removed the variance accounted for by the shared family environment (e.g. parental income, occupation and education). Examining outcomes within families therefore allows researchers to account for early-life SES without the need to operationalize and measure it.

Murray found that brighter siblings started out earning less as they carried out further studies, and then surpassed their less bright siblings. Patience and intellect rewarded with delayed but then sustained greater earnings.

Iveson and colleagues have extended the within-family method by linking children in their 1947 national sample to younger siblings in a “Six day” sample, such that they had families tested at the same age with the same Moray House intelligence tests, and longevity measured up to November 2015.

Intelligence and all-cause mortality in the 6-Day Sample of the Scottish Mental Survey 1947 and their siblings: testing the contribution of family background.
Matthew H. Iveson Iva Čukić Geoff Der G. David Batty Ian J. Deary
International Journal of Epidemiology, dyx168,
Published: 21 August 2017

They say:

Background: Higher early-life intelligence is associated with a reduced risk of mortality in adulthood, though this association is apparently hardly attenuated when accounting for early-life socio-economic status (SES). However, the use of proxy measures of SES means that residual confounding may underestimate this attenuation. In the present study, the potential confounding effect of early-life SES was instead accounted for by examining the intelligence–mortality association within families.

Methods: The association between early-life intelligence and mortality in adulthood was assessed in 727 members of the 6-Day Sample of the Scottish Mental Survey 1947 and, for the first time, 1580 of their younger siblings. These individuals were born between 1936 and 1958, and were followed up into later life, with deaths recorded up to 2015. Cox regression was used to estimate the relative risk of mortality associated with higher IQ scores after adjusting for shared family factors.

Results: A standard-deviation advantage in IQ score was associated with a significantly reduced mortality risk [hazard ratio = 0.76, p < 0.001, 95% confidence interval (CI) (0.68–0.84)]. This reduction in hazard was only slightly attenuated by adjusting for sex and shared family factors [hazard ratio = 0.79, p = 0.002, 95% CI (0.68–0.92)].

Conclusions: Although somewhat conservative, adjusting for all variance shared by a family avoids any potential residual confounding of the intelligence–mortality association arising from the use of proxy measures of early-life SES. The present study demonstrates that the longevity associated with higher early-life intelligence cannot be explained by early-life SES or within-family factors.

Here are the results of the contrast between the living and the dead.

Lifespan and IQ

These are scary figures, and worth showing to friends who doubt that IQ has any practical meaning.

Here are the survival rates for 1 sd above average in intelligence, average intelligence, and 1 sd below average intelligence, with the chilling reminder that a good life is probably no more than 30,000 days. Carpe diem.

Longevity survival days

Here is the key paragraph in which they explain the significance of calculating the effects of SES by using the within-family method:

The somewhat attenuating effect of adjusting for family-related SES, although small, is broadly consistent with previous work using discrete proxy measures of early-life SES.19 However, the fact that a substantial IQ-mortality association persists beyond such adjustment is interesting, particularly given that the present study accounts for early-life SES in a different and more comprehensive way. Adjusting for within-family variance is notably more conservative than simply accounting for an explicit measure of SES, as it likely captures shared factors not directly related to SES such as genetic factors, environmental health (air pollution, etc.), childhood diet and exposure to passive smoking. Notably, adjusting for shared family factors led to a slightly larger attenuation of the intelligence–mortality association than adjusting for an explicit proxy measure of SES. However, the advantage of the family-based approach adopted in the present study is that it avoids any measurement error or residual confounding associated with using proxy measures such as parental occupation and income. Where previous studies have used proxy measures, residual confounding may erroneously underestimate the attenuating effect of early-life SES on the intelligence–mortality association.

Here are the additional summary points on this study

• After adjusting for all shared family factors, a standard-deviation increase in IQ score was associated with a 21% decrease in mortality risk.
• Early-life socio-economic circumstances are not sufficient to explain the intelligence–mortality association.

The key finding is that intelligence leads to a longer life, and childhood socioeconomic status does little to alter that effect. This is an important result, so please find a kind way of breaking this news to any sociologist of your acquaintance.

* In fact: So, at IQ 115 your chance of getting to a particular age, say 79, is 24% better than average.

• Category: Science • Tags: IQ, Longevity 
Hide 28 CommentsLeave a Comment
Commenters to FollowEndorsed Only
Trim Comments?
  1. res says:

    Interesting (and encouraging ; ) post. Thanks. Is there any way to separate the contributions of purely physical system integrity from indirect environmental effects like higher income and better decision making? And even “purely physical system integrity” has an indirect accumulative environmental component in its phenotypic realization.

    I don’t think the following holds though:

    The good news is that a standard-deviation increase in IQ score is associated with a 24% decrease in mortality risk. So, at IQ 115 lifespan is 24% longer than average.

    I had trouble finding an analysis to check that, but this looks decent:

    He concludes:

    I read a few weeks ago about a study where vitamin D supplementation reduced all-cause mortality rates by 6%. How many years would that add to life expectancy? I wondered.

    6% of a 75-year life span would mean 4½ extra years, I thought, naïvely.

    I pulled up a mortality table (from the Social Security Admin) and did the calculation in a spreadsheet. The two lines were barely distinguishable. A 6% drop in mortality only increases life expectancy by 7 months.

    Based on the curve he presents it looks like a 24% decrease in mortality rates would result in a 2-3 year increase in lifespan.

    I would welcome a more rigorous analysis if anyone knows of one.

  2. @res

    I agree. When I read this, “…a 24% decrease in mortality risk. So, at IQ 115 lifespan is 24% longer than average…” my immediate reaction was, “This is an incredibly (pick one or more: stupid, naive, innumerate, uninformed) statement. Then I did a back-of-the-envelope check. If the average life expectancy at birth in the USA is about 70 years, is it reasonable to think that people with an IQ of 115 have an average life expectancy of 87 years. Not likely!

    • Replies: @The Alarmist
  3. res says:

    The good news is that a standard-deviation increase in IQ score is associated with a 24% decrease in mortality risk.

    Looking at your link to

    I am curious how linear the mortality risk vs. IQ relationship is. The researchers used R to do their analysis and I have had good results looking at spline fits in survival analysis (e.g. to try to estimate “optimal” ranges for things like BMI) using R. I’m not sure how much of a limitation N=2307 is. That implies only ~3 people at +3SD and at -3SD people probably are in special care so might not even be surveyed (?). The confidence interval for the splines can become quite large at the extremes, but trends are typically discernable.

    If anyone is interested in looking at the nonlinearity I use the coxph function from the R survival package (that package is also used by the researchers in this paper so it should be pretty simple to add nonlinearity to their model) with pspline variables. The cph function from Frank Harrell’s rms package (I use his rcs variables) is also useful because if offers additional analytical and display capabilities. More on the rms package:
    The books recommended in that post provide good examples. I like to use plots of the hazard ratio (with confidence intervals) vs. the variable to look at the results. For large effects logging the y axis tends to work better (especially since at the extremes the tails tend to diverge and have large CIs), but is probably not necessary here.

    One odd feature of their sample is that the SDs of their two component samples were unusually high at 19.05 and 17.60. Does this affect the interpretation of their results (e.g. what SD did they use for their conclusion)? Any comments on the high SD? I was surprised since restriction of range seems more common.

    To put the 24% decrease in mortality risk (hazard ratio = 0.76) in perspective, the HR for the female variable by itself was 0.57. This ties into my earlier comment, and looking at the graph in the link there that estimates an increased lifespan for women of 6-7 years which is pretty close to reality:

  4. Hugh says:

    From the Dead column in your chart it would seem that less intelligent men are the first to go – and by a huge margin.

  5. Seems like a long way to go to imply that IQ is not hereditary.

  6. @res

    Yes, the 24% longer statement is wrong. It is the chance of getting to a particular age. Intend to explain more in a further note.

  7. @James Thompson

    Total mortality risk is 100% … it’s the stretching or shortening of lifespan as expressed by the probability of reaching some age that you are getting at, but that is complicated by the fact that the adjustment to mortality extends across all ages and that might cumulate to the implied 0.24 you cite.

    In the Scottish study, they arrived at a 24% greater likelihood of reaching 79 years of age for 15 more IQ points at age 11. The Scottish study also controlled for socio-environmental factors (Scotland has one of the highest mortality risk factors in the U.K. and Europe for a number of reasons) and determined that socio-environmental factors might explain only 30% or so of any variance. In any case,this IQ effect has been replicated in a number of similar studies globally, so it appears to be real, or at least observable.

    It would be interesting to crunch the numbers and see what this means in terms of lengthening the average life, but that is kind of moot given the finding that the IQ effect appears less correlated to mortality as people get older, and probably considerably less so as people get into their 80s and die from physical limitations rather than from things their intelligence helped them avoid.

  8. @Jus' Sayin'...

    Life expectancy at birth for US is currently 76.3 years for males and 81.2 years for females. The effect would not be to add 24% to those figures. I haven’t done the math, but it might add a couple to a few years to those numbers.

  9. Did they record the parents’ age at the time of the child’s birth? I’m thinking of relatives who were very intelligent, but also had children at later ages than the norm. As western education encourages the very bright to put off childbearing, it could be that the system is effectively selecting for people able to marry, conceive, and bear children at older ages.

    It could mean that the children of people able to bear children at advance ages inherit through selection a set of genes better fitted to a long life. Of course, bright people are not immune to infertility, so it’s a risky gamble.

    I do not think that more intelligent people live a healthier lifestyle than the norm. Well, ok, they generally avoid things like drag racing, going to bars looking for fights, etc. They probably look both ways before crossing the street, and visit the doctor when sick.

    I’m not convinced they floss.

  10. res says:

    OT (but I thought better here than in the Dream House post).

    In an iSteve comment thread I made the following point:

    Test-retest correlations over 6 months for the Wechsler are usually about 0.9:

    This implies a percent variance explained of about 80% implying that almost 20% of the variance for IQ is simply due to error. I don’t consider that argument definitive, but it certainly is food for thought. Especially given that I think the Weschler is one of the better constructed and understood tests out there (e.g. compared to proxies like the SAT).

    And was hoping I could get some feedback here. Surely this train of thought has been explored somewhere? Is there any validity to my statement? Any counterarguments?

    Similar questions apply to the other point I made in that comment:

    Another point I don’t think gets the attention it deserves is that a big difference between within group differences and between group differences is that the between group differences allow for statistical averaging away of both individual measurement noise and individual environmental variation. I think that means there is a non-trivial chance that genetics actually will measure out as having more explanatory power for between group differences than within group differences for groups with relatively similar environments.

    • Replies: @James Thompson
  11. @res

    Quick reply. A test-retest correlation of .93 over 6 months for the WISC is fantastic, and is better than the test-test correlation for height at four months, as found in the Isle of Wight followup study. The reason for the latter differences were that one or two heights were probably written down incorrectly, but also because children don’t always stand up absolutely straight. So, by any reasonable standard, IQ can be measured reliably.
    Now, as to this habit even I have got into, of squaring the correlation coefficient and coming up with a lower “variance accounted for” estimate, I have read (in the last two years or so) some strong criticisms of this rule of thumb. Typically, I can’t find the reference immediately, but will try to look for it.

    • Replies: @res
  12. res says:
    @James Thompson

    Thanks! Do you have a reference for the Isle of Wight test-test height correlation (or even just a number)? I searched and found this:
    But I’m not sure the dates match this comment you made about the Isle of Wight followup study:
    and most of the literature for that study seems to be allergy based and I did not see a discussion of height for it.

    I am very interested in the correlation/”variance accounted for” criticism you mention if you are able to find it.

    Since I don’t have much experience interpreting test-test correlation I found this page helpful:

  13. In my not so modest opinion bright people live longer because they have no desire to become rich, or join the rat race.
    Their brightness gives them the possibility of a quiet job, with work satisfaction.
    Dutch judges complain to have to stop working at 68, even the 70 year limit does not appeal to them.
    On the other hand, I suppose bright people understand what harm they can do to themselves, alcohol, drugs, no excercise, too much food, and especially too much fast food.

  14. Don’t forget that IQ tests don’t measure “intelligence” as such, they measure adaptation to the society in which one lives. Never forget also that IQ 100 is an arbitrarily assigned control figure. Thus, a person who is well adapted to his environment is likely to be more successful on almost every level and also, for that very reason, happier, which also means, to a great extent, healthier than those less well adapted. Healthier people live longer. Thus, IQ tests are certainly not without “practical meaning”. It’s just a question of drawing the right practical meaning from them. The problem arises, not from the tests or their practical meaning, but from the distortion of that meaning for political purposes.

    • Replies: @utu
  15. Joe Hide says:

    Last time I checked my I.Q. was over 1 std above the average. That said, my earlier years were in a very unhealthy poverty stricken environment and a number of health issues thus developed.
    This was decades ago and medical approaches only gave pain relief. Intelligently examining my options, I embraced the new natural healing methods that were just coming out that my average or below average I.Q. peers at that time demeaned, such as whole and organic foods, weight lifting, yoga, superior sleep management, massage, chiropractic, meditation, etc. I haven’t been to a medical doctor in 30 years, am an old guy without old guy diseases, and regularly see the people around me at my age rapidly declining or dying. I think my advantage wasn’t mostly a higher I.Q., but the obsessive desire to know the Truth, instead of just knowing what those who controlled information wanted me to think. Oh yes, my I.Q. noticably went up more after a few years of natural lifestyling. What do You think of this?

  16. Pontius says:

    Less intelligent people drink and smoke a lot more, in my personal experience. They seem easier to influence into partaking in “bad” behavior. At the very least they need more potent diversion from the increasingly desperate lives the non-professional class is doomed to live.

  17. A comment by Santoculto on one of your previous threads, of which I was dismissive, suggested genetically variable sensitivity to environment, which could also lead to different outcomes for a given shared environment. Also, lead tends to bind to phosphates, including the phosphate sheaf of DNA. As such, lead poisoning is a possible cause of correlation of IQ and health. Given its connection to crime, one may expect, and does find, a geographical correlation between murder rates and life expectancy. Under genetically non-determined sensitivity, population mean environment does not enter twin correlations, but if sensitivity is a function of genes, then mean environment will enter the correlations, as it will multiply genetic variability, which would undermine the Falconer/ACE interpretation. Genetically variable sensitivity to lead would be due to presence or absence of bronze age in ethnic ancestry, as copper almost always co-occurs with lead in deposits. Africa skipped the bronze age. China and Mongolia had extremely polluting bronze ages, and their bronzes were quite often copper-lead or copper-lead-tin, rather than exclusive copper-tin.

    • Replies: @Johan Meyer
    , @CanSpeccy
  18. @Johan Meyer

    I forgot to add, regarding the binding of lead to the phosphate sheaf of DNA, a number of points that I wished to add.

    One source of reduced life expectancy is cancer. Lead is a heavy metal, and will thus absorb more background (gamma) radiation, per

    $$\mu/\rho \propto Z^3/E^3$$

    where $\mu/\rho$ is the mass absorption coefficient of the ionizing radiation, where $\mu$ is the absorption coefficient and $\rho$ is the density (mass per unit volume), $E$ is the photon energy of the gamma radiation, and $Z$ is the atomic number of the element (82 for lead).

    Thus lead binding to DNA may increase the lifetime risk of cancer. Lead poisoning, in adulthood, may also lead to heart problems. A relevant question, then, is how IQ relates to life expectancy, in high and low murder rate groups; the murder rate of the year when the population under study was 23 years old should be used. If lead is the cause, then the relationship between IQ and life expectancy should be weaker for lower murder rate groups.

    Also, note that the death rate is about 800 / 100,000 per year, versus murder rates between 0.3 (Japan) to 60s (certain Latin American countries), and thus in most countries on earth, outside two or three in Latin America, correlations between life expectancy and murder cannot be due to direct forcing of the bulk of deaths.

    Also, in Africa, there is little correlation between life expectancy and murder rates, which suggests that lead does not dominate life expectancy, but rather other causes of disease.

    Also, variable sensitivity to lead would be an evolutionary response.

    • Replies: @Johan Meyer
  19. @Johan Meyer

    A further comment. Much of the population that is now at the age where heart problems and cancer dominates death rates, was born and raised during the leaded petrol/gasoline era, in the US. That era ended later in Europe, but east Asia largely followed approximately the US model, albeit removing most of the lead from petrol rather than all of it, in the 70s (Japan; total ban in 80s) and 80s (PRChina, ban in 2000), iirc.

  20. Genetically additive sensitivity to environment, both direct (e.g. lead intake into brain versus IQ) and due to uptake (genetically variable blood lead response to environment) may masquerade as genetically additive causation of IQ, due to (shared) population mean environment, and due to shared deviation from mean environment (common environment, in Falconer/ACE terms) between twins.

  21. CanSpeccy says: • Website
    @Johan Meyer

    one may expect, and does find, a geographical correlation between murder rates and life expectancy

    Yes, being murdered does tend to reduce one’s longevity!

    But you are correct that there are factors such as lead exposure that may give rise to a correlation between IQ and short life that is unrelated to any effect that IQ may have on longevity.

    • Replies: @Johan Meyer
  22. @CanSpeccy

    My point with the death rate is precisely that it is so much larger than the murder rate that murder is usually lost in noise. For a population of 3.2 million, with a murder rate of 5 per 100,000 per year, the annual murder toll would equal the standard deviation of the annual death toll, under Poisson statistics. Unless the death rate is extremely consistent, e.g. at the Poisson limit, and then only on a much larger, e.g. US national scale, I cannot see why murder would dominate the variance of the death rate.

  23. utu says:
    @Michael Kenny

    The problem arises, not from the tests or their practical meaning, but from the distortion of that meaning for political purposes.

    And you think that Thompson is about practical meaning or political purposes?

  24. @James Thompson

    So, how many years of life expectancy does a standard deviation of IQ buy a person? (Not enough, apparently, for the insurance actuaries to include an intelligence estimate.)

    I don’t find it terribly interesting that higher IQ folks live a little longer. After all, insofar as we control our health outcomes, it is largely a matter of judgment, which is a facet of crystallized intelligence, and is directly tapped by the WAIS tests through the Comprehension subtest.

    Now, if it were demonstrated that fluid intelligence predicted life expectancy, that would be more interesting. This talk about “IQ” as a global variable (whatever the [obsolescent] WAIS happens to measure) seems painfully prescientific.

    • Replies: @James Thompson
  25. drumstick says:

    Is it just me or is the title supposed to actually be “Vita Brevis, Dignitas Inutilis”?

    • Replies: @James Thompson
  26. @Stephen R. Diamond

    Years of life associated with 1 sd of intelligence about 5 years, as shown in the post.
    Understandable that it should be thought that the extra years of life were due to intelligent choices. That is part of it, but since simple reaction time accounts for about a third of the effect there must be more to it, hence the hypothesis of “system integrity”.
    The link of lifespan with intelligence is as tested by the Moray House test, which covers a good range of what might be considered fluid and crystallized intelligence.
    There are other intelligence tests apart from the WAIS, and the Woodcock Johnson, each with a body of supportive knowledge.
    As to “pre-scientific”, the field is open to new measures of intelligence.

  27. @drumstick

    Agreed. Would have been better, I think. Aside from the declension, was anyway stuck for something which denoted social class. “Ordo” might have fitted the bill better.

  28. It’s not just IQ. For those of us who reached adulthood during the Vietnam War innate charm was a major factor in life expectancy, the charm necessary to convince an army shrink that you were just too effed up to serve honorably in that great war. I am proud to say that I possessed at least that much charm then and probably still do.

Current Commenter

Leave a Reply - Comments on articles more than two weeks old will be judged much more strictly on quality and tone

 Remember My InformationWhy?
 Email Replies to my Comment
Submitted comments become the property of The Unz Review and may be republished elsewhere at the sole discretion of the latter
Subscribe to This Comment Thread via RSS Subscribe to All James Thompson Comments via RSS