Tuesday 28 February 2012

Objective and direct evidence of 'dysgenic' decline in genetic 'g' (IQ)


Reaction time (e.g. pressing a button as fast as possible in response to a light or sound) correlates with general intelligence 'g' - faster reactions correlating with increased intelligence.

However, in principle reaction time would be expected to be less affected by socio-cultural factors than standard IQ tests.

Therefore long term trends in reaction time might be an objective and direct measure of true, underlying general intelligence compared with normal IQ tests.


In particular, reaction time trends might resolve the apparent paradox between

1. the long-term decline in inferred underlying general intelligence due to differentially greater fertility among those with differentially lower IQ (the 'dysgenic' argument);

2. and the long term increase in the results of normal IQ tests (the 'Flynn effect' - or more accurately 'Lynn-Flynn' effect, since the observation was first published by Richard Lynn).


The dysgenic argument is that g has declined over the past century and a half and the increase in IQ scores is superficial - in other words genetic intelligence has declined even while phenotypic intelligence has increased; while the Flynn effect argument is that phenotypic increase in measured IQ also reflects underlying an increase in underlying 'genetic' g.


In essence, dysgenics advocates agree with the Flynn-effect advocates that phentoypic IQ has increased over the twentieth century but infer that underlying, genotypic IQ has declined.

However, the dysgenics viewpoint has had no objective, directly measurable evidence that genotypic IQ has declined.

Until now.


Dr Michael Woodley (http://publicationslist.org/M.A.Woodley) has pointed-out to me a paper by Irwin W Silverman of Bowling Green State University from 2010 which resolves this question, and provides convincing evidence to support the dysgenics argument.


Silverman IW. Simple reaction time: it is not what it used to be. American Journal of Psychology. 2010; 123: 39-50.

Abstract: This article calls attention to the large amount of evidence indicating that simple visual reaction time (RT) has increased. To show that RT has increased, the RTs obtained by young adults in 14 studies published from 1941 on were compared with the RTs obtained by young adults in a study conducted by Galton in the late 1800s. With one exception, the newer studies obtained RTs longer than those obtained by Galton. The possibility that these differences in results are due to faulty timing instruments is considered but deemed unlikely. Of several possible causes for longer RTs, two are regarded as tenable: that RT has been increased by the buildup of neurotoxins in the environment and by the increasing numbers of people in less than robust health who have survived into adulthood. The importance of standardizing tests of RT in order to enable more refined analyses of secular trends in RT is emphasized.


In a nutshell, Prof Silverman reviews data from Francis Galton between 1884 and 1893, extracted from a study of visual reaction times in 2,522 men and 302 women. The average reaction times were 183 milliseconds (ms) for men and 187 ms in women.

Silverman notes that in reviews of reaction time studies in 1911 (but not including Galton's work), it is clear that Galton's results were typical of the era - the range being from 151-200 milliseconds - median of 192 milliseconds.

By contrast, Silverman reviews twelve modern (post 1941) studies of visual reaction time (using a comparable methodology to Galton) - and the modern reaction times are very significantly longer - the total number of subjects was 3,836 - the mean reaction time was 250 milliseconds for men (SD 47) and for women was 277 ms (SD 31).

Looked at separately, in only one study, only for men, were Galton's average values contained within the 95 % confidence interval - in other words, in 11 of 12 studies and 19 of 20 comparisons - as well as the overall meta-analysis - the difference in reaction times reaches conventional levels of statistical significance.


Conclusion: Victorians had faster reactions, on average, than moderns.

Implication: Victorians were more intelligent, on average, than moderns.


It therefore seems that average reaction times have become slower over the past 100 years.

Since reaction times correlate with IQ then the measured decline in reaction times is consistent with a significant decline in general intelligence over the past century, as argued by the 'dysgenic' theorists.


Note added 30 April 2012.

I thought of this way of testing the dysgenic hypothesis using historical reaction time data in 2008, as shown by the following e-mail sent to Prof Ian Deary of Edinbugh University. However, at that time I could not find anything relevant in the online scientific literature, and gave-up looking. 


E-mail: Monday 07.07.2008 timed at 09:51

Dear Ian,

I'm sorry to badger you when you are still catching up but I have had an idea for measuring dysgenic change by using reaction times which I am keen to follow-up.

Since reaction time correlates with IQ, and since reaction time is an old physological measurement, it is possible that there are representative data on national population reaction times over the past 100 or so years.

Because high IQ people have a fertility considerably below replacement level, my prediction would be that average reaction time in developed countries should have become longer, and that the standard deviation would have become smaller (due to selective loss of shorter reaction times).

But probably somebody has already done it (perhaps your group?)?

Or maybe reaction time correlates with 'phenotypic' (or measured) IQ (and therefore gets enhanced by the Flynn effect) rather than correlating with underlying 'genotypic' IQ? - I don't know.

If it hasn't already been done - do you know of any databases of reaction times (or somebody whom I might contact about this?

Best wishes, Yours, Bruce



dearieme said...

"... by the buildup of neurotoxins in the environment... ": is there evidence for that claim?

Anonymous said...

It is interesting that concerns about a decline in intelligence have been expressed for quite some time. Take, for example, this comment by Aldous Huxley in 1962 (at about 27:15):

Gyan said...

Curiosity is NOT a traditional virtue and especially the Christian ethics strongly warns against idle curiosity.

So why do scientists study 'g' and possible dysgenic effects thereof.

What is the context?

Valkea said...

So many negative and decaying things in modernity becomes now better understood.

Bruce Charlton said...

@dearieme - not that I know of, not convincing evidence anyway.

@Anon - they date back at least to Galton, which is soon after the demographic revolution became apparent. The logic makes it almost inevitable.

@Gyan - there are many reasons to study intelligence, which is why it is such an old domain of enquiry.

A single IQ test done on a cohort of children is significantly correlated with subsequent life expectancy, health, height, chance of obesity, educational attainment, salary, criminality and so on.

And IQ differs between people and groups and is substantially hereditary.

Of course there are other important factors, such as personality - which is also different in persons and groups and substantially hereditary.

My own interest in intelligence is related to the role in creative genius but also to modernity, atheism, Leftism/ political correctness - I have argued that these are 'clever silly' phenomena, associated with a high IQ elite.


Steve Sailer said...

Does the military find a worsening in reaction times?

Gyan said...

Elite would be of above-normal intelligence, by definition.
Once the elite were religious, now they are secular.

The various correlations of IQ with other things, potentialities, probabilities of this and that could have hardly anything to do with spread of atheism.

Indeed, the study of these correlations is exactly what could be called Idle Curiosity.

Bruce Charlton said...

@SS - I do't know - but Silverman comments that the decline being measured in *this* data set seemed to occur before the mid-20th century. I would guess any military data would be from the second half of the twentieth century.

Bruce Charlton said...

@Gyan - be careful. If you regard an interest in intelligence as 'idle' and are consistent with this line of criticism you will not just throw out IQ research but all of science, all of technology, all of philosophy, art, politics, good company and good cheer - and most of what nearly everybody is doing nearly all of the time (including, of course, blog commenting!).

In an ultimate sense you are correct; and the highest path is indeed that of ascetic monasticism (followed by the via positiva of marriage and family) - but are you really willing consistently to regard as idle and discard all other activities than this?

Bruce Charlton said...

Comment from M.A.Woodley

It must also be noted that reaction times give rise to a Jensen effect, which means that the sorts of IQ subtests that discriminate best between individual differences in reaction times are also the ones that are the most g-loaded (Jensen, 1998). Meisenberg (2010) found in the NLSY that the subtests on which dysgenic fertility differentials were greatest were also the ones exhibiting the highest g-loadings, hence there is an apparent Jensen effect on dysgenesis. Jensen effects also exist on test heritabilities and other genetic or partially genetic variables such as inbreeding depression scores (Rushton, 1999) and fluctuating asymmetry (Prokosch, Yeo & Miller, 2005). This strongly reinforces Charlton's assertion that Silverman's finding provides the first unambiguous evidence for dysgenesis on g, as the presence of a Jensen effect indicates the existence of a genetic nexus amongst these variables.
The most comprehensive study published to date on whether the Flynn effect is a Jensen effect is te Nijenhuis (In press). Here it is clearly demonstrated that across studies in the US and the Netherlands the Flynn effect is an anti-Jensen effect, as it is biggest on the sorts of abilities which have low g-loadings (and hence lower heritabilities). An implication of this is that dysgenesis and the Flynn effect can co-occur as they concern different sources of variance in tests of cognitive ability, and are associated with different causes.
Consistent with this are the results of a recent temporal correlation analysis, in which estimated historical gains and losses in Western genotypic IQ along with the Flynn effect were considered in the context of assorted social and scientific variables (Woodley, in press). It has been found that genotypic IQ estimates based on an increase between 1455 and 1850 and a decrease of between 1 and 2 points per generation between 1850 and 2005 best fitted Huebner's (2005) data on innovation rates, which show a similar pattern (with peak innovation rates occurring in the 1870's followed by decline). The Flynn effect seems to trend in parallel with the growth in GDP (PPP) per capita, as both experienced a sharp upwards trend in the beginning of the 20th century. The strongest predictor of the Flynn effect in regression and path analysis was decreases in a common factor of illiteracy and homicide, which suggests an influence from life history (i.e. lower time preferences and diminished impulsivity).


Bruce Charlton said...

MA WOodley continued...

This indicates that whilst high genotypic intelligence (i.e. g) is necessary for maintaining high innovation rates, the Flynn effect has been a much more significant factor in the generation of wealth. The Flynn effect might therefore have been associated with the acquisition and proliferation of the sorts of specialized, low-heritability cognitive skills which permit people to seek out narrow socio-cultural niches and exploit them in such a way that raises the carrying capacity of the environment and also the aggregate economic efficiency of the population.


Huebner, J. (2005). A possible declining trend for worldwide innovation. Technological Forecasting and Social Change, 72, 980–986.

Jensen, A. R. (1998). The g-Factor: The Science of Mental Ability. Westprot, CT: Praeger.

Meisenberg, G. (2010). The reproduction of intelligence. Intelligence, 38, 220–230.

Prokosch, M., Yeo, R., & Miller, G. (2005). Intelligence tests with higher g-loadings show higher correlations with body symmetry: Evidence for a general fitness factor mediated by developmental stability. Intelligence, 33, 203–213.

Rushton, J. P. (1999). Secular gains in IQ not related to the g factor and inbreeding depression—Unlike Black–White differences: A reply to Flynn. Personality and Individual Differences, 26, 381–389.

te Nijenhuis, J. (In press). The Flynn effect, group differences and g loadings. Personality and Individual Differences. doi:10.1016/j.paid.2011..12.023

Woodley, M. A. (In press). The social and scientific temporal correlates of genotypic intelligence and the Flynn effect. Intelligence. doi:10.1016/j.intell.2011.12.002.

Cantillon Blog said...

Flynn effect related also to left hemisphere utilization bias (Mcgilchrist) and the societal mania for sorting people based on tests, and as a result for education to become a mechanism for coaching people to do better on these, at the expense of all else?

gwern said...

Pretty bizarre. A 1800s-now comparison should show increases, because stamping out iron and iodine deficiencies should be worth anywhere from 0 to 20 points depending on how deficient a population was before supplementation started in the 1920s and later. Are we supposed to believe dysgenics has more than reversed that and all other possible improvements?

Joseph said...

If A is correlated by B and B is correlated with C that does not mean A is correlated with C.

For example, voting for Obama in 2008 was correlated with living in states where the people have higher measured IQs but that doesn't mean Obama voters are brilliant.

Bruce Charlton said...

@Joseph - yes, it is an ecological phenomenon ('the ecological fallacy' is a version of it).

But there is more than a century's worth of evidence connecting reaction times with general intelligence. At the same time, nothing in science, or in biology, especially not in human psychology is logically entailed.

But if reaction times-IQ is rejected as scientifically insecure, then to be consistent in the application of this standard of proof, that would lead to the rejection of a very large amount of other science too - pretty much the whole of psychology.

Gyan said...

I am not arguing for asceticism nor for Puritanism.

My point is specific to the IQ research. Essentially my suspicion is that people in this field do not think with the Church. This is important since it is not physics but a science studying humans. An error in the basic anthropology has potentially disastrous effect on the conclusions.

Bruce Charlton said...

Comment from Scott Locklin:

"Two words nobody wants to talk about: "outbreeding selection." Also, Eugenics was practised by the masses until 1950 or so."

Bruce Charlton said...

@Scott L - If by eugenics you mean selective infanticide of abnormal newborns, I think you are probably correct to a significant extent.

However, the phenomenon described here probably began from about 1800 when child mortality rates began to drop. Until then almost all the children of the poor would die (from infections or starvation) and/ or fail to reproduce - as Gregory Clark describes in Farewell to Alms. Over a few dozen generations the population in many stable agricultural societies became almost replaced by children of the 'middle classes' of crafstmen and the literate.

After this there was a collapse of fertility among the wealthy, educated and intelligent - driven by developments in contraception and greater use of abortion - which continues.

kristor said...

I note that in all the tests, men had faster reaction times than women. Does this mean men have more g than women?

Gyan said...

Why was the Farewell to Arms phenomena
restricted to NW Europe?

Bruce Charlton said...

@Gyan - I don't think it was restricted to NW Europe - I presume it applied to East Asia, for example.

But it may be that only England has a full enough historical record to be able to measure it.

Bruce Charlton said...

@Kristor - yes, men have higher g, on average, in most populations (although, of course, it would be possible to engineer selective scenarios in which thsi was not the case).

Silverman comments that the sample of women (much smaller than that of men) was likely to be exceptionally selective for faster reaction times - the reactions times for Galton's women are way-way shorter than in modern populations - so the average M-F difference was probably larger than Galton measured.

J said...

I dont know if it is relevant but fast paced movies and computer games were supposed to make the new generation much faster on the uptake and reduce reaction time.

I find that early movies are terribly slow and over-acted, while moder movies are exceedingly fast so that it makes it difficult for me to follow. However the environmental neurotoxin explanation seems difficult to prove.

Bruce Charlton said...

@J - Silverman mentions this theory in the paper - but this doesn't really doesn't have anything to do with intelligence variations or with reaction times.

Intelligence is mostly hereditary but can be dragged-down by significant illnesses, diseases etc, and reaction times the same.

There is no known way to enhance real, underlying 'g' (distinct from improving performance on specific types of cognitive tests - which is itself much more difficult to do than most people realize).

The history of claims to have to enhanced general intelligence by particular interventions is a history of environmental irreproducible results, selective reporting, incompetence, dishonesty, wishful thinking, fraud and mis-representation.

See The Raising of Intelligence by HH Spitz.

dearieme said...

"Over a few dozen generations the population in many stable agricultural societies became almost replaced by children of the 'middle classes' of crafstmen and the literate": may I just point out, Bruce, that this phenomenon considerably increases the chances that you and I are mainly of Norman descent?

Bruce Charlton said...

@dearieme - that kind of comment will get you banned...

But seriously folks, I wonder. As I read Clark's argument, it does not seem to apply to the aristocracy (who were Normans) - the men of whom suffered tremendous mortality rates from war, tournaments etc - as to the middle classes; who were Anglo-Saxons.

But probably that level of precision is not possible. The main take home message I got was that in the old days differential mortality was much more important than differential fertility.

Chase Saunders said...

It is theorized that reaction time in older individuals is due to larger amount of information accumulated -- I think a neuroscientist would say the increased noise from accumulated background information would guarantee a slower response time.

It stands to reason this isn't directly due to age, but simply due to the total load of information accumulated by an individual brain. It also stands to reason that modern people accumulate much higher volumes of information at earlier ages. Thus, this could also be a phenotopic, rather than a genetic, phenomenon.

Bruce Charlton said...

@CS - I disagree. I think that g reaches its maximum immediately upon full maturity of the central nervous system - after that it may stay the same or decline, but cannot increase.

Chase Saunders said...

@bgc I was not suggesting that g (general intelligence) was rising due to acquired information load. I was suggesting that response time could be decreasing because of it.

Bruce Charlton said...

@CS - okay - I think I get your point.

But I don't know if it is true - and don't know how we would even conceptualise the brain accumulating information, leave aside measure it. It seems more likely that reaction times decline with age because of reducing general intelligence - or perhaps that general intelligence is actually something like the speed of processing which is crudely measured by reaction times.

Chase Saunders said...

@bgc, I think both neuroscience and information theory offer good models that can both quantify measured information and explain why response time slows for computational reasons.

This video discusses the slow down in response times as a function of noise due to accumulated learnin'


Andrzej T. said...

Have they considered the devastating effect that chronic undersleeping has on reaction times? In Victorian era they slept like 9 h/day, now it's more like 7, perhaps even worse for intelligent people.


Tucker Goodrich said...

"Auditory and visual reaction time in athletes, healthy controls, and patients of type 1 diabetes mellitus: A comparative study" 2006


Balazs Varga said...

While I do agree that the general IQ has dropped, I am not so sure that reaction time is what intelligence is best measured with.

Is intelligence the speed at which we can detect and react to sensory input? Than I'm sure there are a lot of animal life on Earth way more intelligent than us.

Than why are not top athletes also the top scientist? What about care racing. Does this mean Nascar is where geniouses go?

Bruce Charlton said...

@BV - Tere is a vast literature n using reaction time to measure intelligence going back to the late 1800s and forward to Jensen and nowadays Ian Deary's group in Edinburgh.

The great advantage of sRT is that 1. it is an objective physiological measure with a ratio scale measure,


and 2. the data goes back a long way.

rumpole5 said...

I have a Juris Doctor post graduate degree. Yet, when I watch a Shakespeare play that was written to appeal to an unwashed common Londoner, standing in the open air Globe Theater, the play's dialogue dances on my comprehension horizon. One does not need a reaction test to know that some sort of significant cognitive decline has occurred.

Snidely Whiplash said...

I can think of three other mechanisms that would account for the change.
Massive kill-off of the upper classes in WWI and II
Population replacement by Africans and sub-continentals
Selection bias in the original cohort.

Bruce Charlton said...

Snidely - Check out


and word search reaction times on this blog

Bruce Charlton said...

@rumpole - It is indeed striking. During the theatre season, the Elizabethans in London would apparently attend two or three plays per week - they were probably the most experienced and sophisticated theatre audience in human history. Which is why they elicited Shakespeare (and others) - they knew the difference between levels of theatrical excellence, they could appreciate Shakespeare at his best, and knew that he had improved from the early plays (which made the improvement worthwhile).