Sunday, 30 November 2014

Intelligence lost at 1.23 IQ points per decade


Michael Woodley of Menie spends much of his time tending his ancestral estate, pacing the linen-fold panelled rooms of the ancient house, warming his hands at the towering stone fireplace and meditating on the collapse of the aristocracy, the paucity of contemporary innovation and the lamentable and persistent downward drift of the national intellect. Now he sends me a barefoot runner with his latest manuscript, which I have read as the autumn mists creep across the Nadder valley, before penning this reply for the poor urchin to carry back to his master.

Young Woodley avers that, not only are we going to hell in a handcart, but we are doing so at a pace which he can predict with some accuracy (1.23 IQ points per decade), composed as it is of two dysgenic effects: the dull have been reproducing with greater fecundity than the bright (.39), and increasing paternal age has increased the rate of deleterious mutations (.84).

In the spirit of the age, and with an ever-present concern for your health and safety, those of you who are of nervous disposition or advancing paternal age should turn away, and listen to light classical music.

How fragile is our intellect? Estimating losses in general intelligence due to both selection and mutation accumulation. Michael A. Woodley of Menie. Personality and Individual Differences 75 (2015) 80–84

General intelligence is an adaptation to solving evolutionarily novel and domain general fitness problems, i.e. problems which occur on an irregular rather than predictable basis throughout the course of evolution and which are complex, requiring the recruitment and coordination of large numbers of specialized mechanisms in solving them (Geary, 2005; MacDonald, 2013).

it has been argued that deleterious mutations exhibiting small effects accumulate within a population – persisting within genomes for long periods of time before substantially inhibiting fitness, thus giving rise to individual differences in general intelligence (g) and other potentially mutation-sensitive traits, such as health and physical attractiveness (Miller, 2000a,b; Penke, Denissen, & Miller, 2007).

The Breeder’s equation (Fisher, 1929) is frequently employed in studies of this kind:

R ¼ S _ h2

In this equation, S constitutes the size of the selection pressure operating on IQ transformed into a phenotypic change (i.e. the degree to which the trait will change over a generation assuming no biological regression to the mean, or perfect heritability). h2 represents the additive heritability of IQ. The product of these two terms gives us the expected responsiveness to selection, or R, which in terms of IQ is scaled as a change in ‘genotypic IQ’, or the degree to which the underlying genetic potential for a certain level of IQ should decline per generation (Lynn, 2011).

Woodley does a meta-analysis of 9 studies in the US and UK (simply for cultural similarity) to estimate this dysgenic effect on heritable g. This comes to an estimated decadal heritable g decline of 0.385 points.

It may be possible to crudely estimate the impact of mutation accumulation on g. Ideal for this purpose is the study of Kong et al. (2012) in which the numbers of de novo mutations in offspring were counted and correlated with the age of their fathers. Kong et al. found a strong linear relationship between the two (r = .97) amongst a sample of 78 Icelandic parent–child trios. They estimated an average increase in the offspring’s number of de novo mutations at 2.01 per year of paternal age. At 35 years of age, i.e. one familial generation (Kong et al., 2012), fathers are producing offspring with an average of 70 de novo mutations. Using this estimate, it is possible to determine the relationship between the increase in de novo mutation and IQ-loss as a function of paternal age.

These calculations are a little more complex, because they require re-examination of previous “corrections” for birth order. Perhaps there is a book to be written about the assumptions which underlie all statistical adjustments in psychology papers, to be called “The Correction of Corrections”.

The meta-analytic aggregate estimate of the loss in heritable g due to selection (estimated in Section 2 at q = .39 points per decade) can be combined with the loss expected from mutation accumulation, which was estimated at .84 points per decade. As the latter estimate was derived using structural equations modelling, error has been controlled, therefore the loss due to mutation accumulation is symmetric in terms of reliability and validity with respect to the meta-analytic loss due to selection.

The sum of the selection and mutation accumulation losses (i.e. the overall dysgenic loss) is therefore 1.23 points of heritable g per decade, or 4.31 points per familial generation. If the 95% confidence interval for the decline estimate due to mutation accumulation (the paternal age effect; i.e. 1.53–.14 points per decade) is generalized to the sum of both estimates, this yields upper and lower bound decline values ranging from 1.92 to .53 points per decade, or 6.72 to 1.86 points per familial generation.

One potential objection to the finding that g is declining by the amount claimed here stems from the Flynn effect, which is associated with an average increase in IQ of three points per decade (Flynn, 2009). The Flynn effect is however least pronounced on the most heritable and also g loaded IQ subtests (Rushton & Jensen, 2010; te Nijenhuis & van der Flier, 2013), which indicates that secular IQ gains occur at the level of less heritable and narrow abilities, rather than on g. Selection effects and the effects of mutation load on IQ are however more pronounced on the most g loaded subtests (Peach et al., 2014; Prokosch, Yeo, & Miller, 2005; Woodley & Meisenberg, 2013). On this basis dysgenic effects and the Flynn effect could co-occur – with dysgenics reducing the level of heritable g, and various environmental improvements raising narrow abilities simultaneously, via their effects on non-g variance. This hypothesis has been termed the co-occurrence model (Woodley & Figueredo, 2013).

So, there is a prompt and visible effect caused by the liberal use of fertilizer (environmental improvements) giving us a 3 IQ point gain, and a less visible and insidious worsening of seed quality (mutation and differential fertility) losing us 1.2 IQ points. All boats rise on the rising tide of affluence, but some are leaky.

Finally, Woodley links this finding to his work showing that reaction times are slowing up, which he sees as confirmation of the same dysgenic trend.

So, in reply to young Woodley, ever conscious that grim calculations of this sort might cause dismay to sensitive souls, I have placed another log on the fire, and sent him the cheerful and uplifting words of Walter Savage Landor:

I strove with none, for none was worth my strife.
Nature I loved and, next to Nature, Art:
I warm'd both hands before the fire of life;
It sinks, and I am ready to depart.


  1. Is he controlling for that big demographic change in the West that we don't talk about?

  2. In the final paragraph Woodley comments:

    Finally, it must be noted that the decline estimate generated here likely underestimates the actual decline in g at the population level in modern Western countries, as there will be an additional contribution stemming from the process of replacement migration (Coleman, 2002) involving immigrant groups sourced from populations exhibiting lower average levels of g and higher rates of total fertility. Nyborg (2012) estimates that in Denmark, replacement migration may be reducing IQ by .29 points per decade (this corresponds to a heritable g decline of .28 points). Assuming parallel trends in the US and UK, this would entail a further increase in the decadal decline to 1.51 points per decade.


    1. Nyborg's estimate is way too pessimistic. His model is assuming that their fertility rates do not change once they reach Denmark, but they do. The second generation has very similar fertility to the natives.

  3. Yes, Woodley covers this in his last paragraph, but I dropped that because it is the importation of another population, so a different though bigger effect.

    Finally, it must be noted that the decline estimate generated
    here likely underestimates the actual decline in g at the population
    level in modern Western countries, as there will be an additional
    contribution stemming from the process of replacement migration
    (Coleman, 2002) involving immigrant groups sourced from populations
    exhibiting lower average levels of g and higher rates of total
    fertility. Nyborg (2012) estimates that in Denmark, replacement
    migration may be reducing IQ by .29 points per decade (this corresponds
    to a heritable g decline of .28 points). Assuming parallel
    trends in the US and UK, this would entail a further increase in
    the decadal decline to 1.51 points per decade.

  4. The effect of increasing paternal age is complex:
    Let's say for example, that there is 30 new mutations in father line for each 20 years and 0 mutations in mother line. If we define the mutation load at time T0, generation G0 zero, then at time T+40 woman has a choice for father of her child - either 40 year old man of generation G0 (having on average 60 new point mutations), or 20 year old man of generation G1 (having on average 30 new point mutations). But generation G1 on average already has higher mutation load compared to G0 (about 30).
    The main benefit of younger fathers to population comes from the fact that the phenotype of younger man is better proxy of the genotype of his sperm (differing only by 30 mutations, compared to 60 for older man). Thus the selection (either sexual or environmental) operates with better accuracy.
    On the other hand the effect of several detrimental mutations may only appear at later age. This may make choosing older fathers evolutionarily advantageous strategy - and indeed in many mammals females prefer older males.

  5. Thanks for this, James. I was just recently wondering about the problem of random mutation, which almost always exerts a deleterious effect on any trait in question. Even to tread water, it's necessary for there to be at least some small balancing selection.

    I do think that Woodley tends to jump the gun and fiddle with statistics to exaggerate effect sizes, and I'd be surprised to know that the rate of loss due to mutation was as high as he claimed. Still, a number is better than no number, and taken as an upper bound, it's quite instructive.

    1. Michael A. Woodley of Menie1 December 2014 at 16:33

      Quote: "I do think that Woodley tends to jump the gun and fiddle with statistics to exaggerate effect sizes"

      This is actually quite a serious accusation, one that can easily be disproved in the context of the present study. Look at the corrections applied to each study in the meta-analysis. One of the corrections involved imposing a uniform conservative generational length across studies of 3.5 decades, which reduces the expected decadal loss when compared with estimates based on the assumption of shorter generational lengths. Another major correction involved reducing the magnitudes of three of the estimates by nearly 50% in order to correct for the method variance between IQ*sibship and IQ*completed fertility correlations. These corrections cumulatively penalized the decline estimates quite substantially in some cases.

      Had it been my intention to bias the estimates upwards, I might have contrived a rationale for adjusting the six studies employing IQ*completed fertility correlations upwards by 50%. I might also have decided on a shorter generational length (such as 25 years). Bias implies consistency.

      That, taken as a whole, my five error corrections were meta-analytically sound is evidenced by the massive reduction in between-study heterogeneity when the 'before' and 'after' estimates are compared.

      The losses due to mutation-accumulation are based on real estimates of the paternal age effect on offspring g, when both parental g and educational level are controlled using structural equation models (from Arslan et al. 2014) and also real estimates of the paternal age effect on offspring mutation load (from Kong et al. 2012).

      Of course the magnitudes of these parameters are open to debate, hence my explicit description of the estimates as 'crude', coupled with my use of the 95% CI from the paternal age effect in bounding the aggregate decline estimate. However, this is something on which future meta-analytic studies can shed more light.

      At the end of the day I cannot help it if my effect sizes are bigger than some people would like. For what it is worth I did not set out to show that the decline estimates converged with those estimated on the basis of the SRT anti-Flynn effect data. That they do however is nonetheless suggestive of the convergent and nomological validity of the >1 point per decade loss claim however.

      The onus is now on the sceptics to show, via reanalysis of existing data or via analysis of new data that the effect sizes have been overestimated. Allegations of 'gun jumping' and 'bias' in my findings are however inappropriate.

  6. Dear Mark, As you say at the end, a number gives us something to argue about, and to refine our measures. It is relatively small when compared to the Flynn Effect, which is still continuing in many countries, but it may be more insidious.

  7. Strangely, that's almost exactly the decline rate I have found by analyzing 5 decades of declining SAT scores in the US, about 1.2 points per decade....

  8. is that allowing for any re-norming or significant changes in pass rates?

    1. Yes, you have to take the reforming into account, but I didn't attempt to quantify more subtle dumbings down of the test, such as a little easier math questions, so my estimate is on the conservative side.

    2. That's "renorming", despite the idiot computer program.

  9. I would add that if there is a Flynn effect in the US, it has not been visible in the results of more than 50 years of standardized testing of school children.

  10. Interesting. Then, where have the intelligence test constructors messed up: normative samples, pass rates, errors in scaled score calculations or what? Rindermann and I assumed there had been some NAEP changes showing Flynn effects. What is your take on that?

    1. I just looked at the 2012 NAEP, which finds progress both absolute and in closing racial "gaps" since 1971 except at the 17 year old level, where it finds stasis. No one actually familiar with school children believes any of that, but it would be impolitic to say so.
      I would point out that the SAT (and I assume the ACT) is a high security, proctored exam.SAT Exams have been cancelled recently in Korea because of security leaks. The NAEP is administered locally. School districts in every large city in the USA have been caught systematically enhancing the scores of
      their students, especially minority students. If scores get too low in the US, the district may be taken over and/or administrators lose their jobs.There are criminal charges pending against administrators and teachers in Atlanta for wholesale alteration of test booklets. In sum, the NAEP data has to be regarded as systematically corrupt.(I strongly suspect that these tests have also been dumbed down, but don't have any data) If these tests were administered by a neutral third party like the SAT College Board, they would show the same, or more probably a worse decline than the SAT itself.
      Many college teachers have commented on the declining quality of incoming students and their preparation, which corroborates these facts.

    2. O/T --I had the impression that the Flynn effect showed up principally on the matrix tests, which are seldom seen in the US.....

  11. Pre-modern, within the Hajnal Line, paternal age effect must have been similar yet we can assume mutational load was negative across generations, correct? Yet much of the culling was environmental (my understanding is that communicable disease must have a large non-shared-environment component), and much selection on hereditary traits may have been on traits with little demonstrated or intuitive relation to load, eg conscientiousness or novel-disease-resistance (probably related via general immune strength but only weakly). So I'm not sold on the Woodley/Charlton model positing huge differences between the pre-modern and modern selection effects limiting load accumulation, because I don't see how the pre-modern could have acted particularly strongly against load to begin with. I'm clearly missing something.

    I'm not fully convinced on the SRT study, either, so the confluence doesn't hold weight yet (esp given the confidence interval on b). Dividing the stdev change in SRT by SRT's g-loading means you're assuming mono-causality, which is yet unproven, no? What about testosterone or physical fitness? I told a neuro-optometry researcher about the SRT study and she was skeptical, something to do with a modern dearth of peripheral vision stimuli; I've yet to follow-up because I'm deficient in follow-up quotient (wink on the acronym). I should note that my prior is actually fully in agreement with a post-Victorian IQ decline; if I'm misunderstanding something I apologize.

    Fascinating topic and I'm glad such sharp minds on working on it, thank you!

  12. More questions:

    -If mutation rate per generation = 2 * Paternal age, and a generation can be treated as the paternal age for mutations given the paternal bias, isn't the mutation rate just 2/yr, regardless of paternal age pattern?

    -Isn't there evidence the mutation rate is considerably lower ?