Saturday, May 3, 2014

What happened in the 1980s to reaction time?

A steady increase in reaction time seems to begin circa 1980 in Sweden, Great Britain, and the United States (h/t to hbd* chick)

Has reaction time been steadily increasing from generation to generation? This was the finding of a paper last year, which argued that mean IQ had fallen in Britain by 13 points since Victorian times (Woodley et al., 2013). The problem here was not the extrapolation from reaction time to IQ, which in any case should not have changed over the past century. The problem was the possibility of sampling bias. The early samples (from the Victorian age) were slanted toward people of elite origin. In one case, they were University of Chicago students; in the other, museum visitors who had paid to take the reaction test. In contrast, the recent samples were much more representative, largely because the educational system had become more universal. Thus, the drop in reaction time over time may be largely, if not wholly, an artefact of better sampling of the general population (hbd* chick, 2013).

This criticism seems less applicable to a study of this drop in more recent times. This study was presented as a conference paper and only the abstract is available. But it does seem interesting:

Here, we show that change in genetic intelligence can be estimated, independently of the Flynn effect, by way of simple reaction time (RT). Data from three studies with different samples from Sweden, UK, and USA converge at an RT increase of 0.7-0.9 ms per year, which corresponds to a decrease in intelligence of between 4 and 5 IQ points per generation, or 1.3-1.7 points per decade in these countries. (Madison, 2014)

That’s a big jump in reaction time. Moreover, most of the increase seems to be squeezed into the last six years of the study, from 1980 to 1985 (see above chart). No selection pressure could have produced such a change over such a short time. So what could possibly be going on here?

Keep in mind that Sweden, the United Kingdom, and the United States are not closed systems. All three countries are open to the world, and the early 1980s corresponded to a time when they became much more open. Consequently, we are not looking at change within a population, but rather the replacement of one population by another, and this change would have been most noticeable in the school classrooms where these tests were conducted.

The other finding of this study is that the Flynn Effect has been masking a decline in intellectual capacity. Yes, we’re getting better at giving standardized answers to standardized questions, but this change doesn’t reflect an actual increase in intelligence. We’re just allocating more and more mental resources to the task of test-taking. The reaction time data suggest that real intellectual capacity has been declining since circa 1980.

At the same conference, Armstrong (2014) likewise argues that the Flynn Effect may be illusory for the most part:

However, "general intelligence", the biological substrata which cause the positive manifold amongst different IQ tests, has not increased, since the sizes of Flynn effects on different tests are inversely related to those tests' g loadings. The same pattern holds amongst items. Thus, for example, vocabulary size (generally the most or among the most g loaded tests) has shown a small Flynn effect, and by some measures even a decline. However, the "Coding" test (from the Wechsler) or the "Draw-a-Man" test both have low g loadings and have shown very large Flynn effects.

To date, Madison’s paper is unavailable. All we have is an abstract that raises more questions than it answers. We will have to wait for the full paper before a thorough assessment can be made.


Armstrong, E. (2014). LCI14 Elijah Armstrong on Rule Dependence, London Conference on Intelligence, Psychological comments, May 1  

Hbd* chick (2013). a response to a response to two critical commentaries on woodley, te nijenhuis & murphy (2013), hbd* chick, May 27  

Madison, G. (2014). Increasing simple reaction times demonstrate decreasing genetic intelligence in Scotland and Sweden, London Conference on Intelligence, Psychological comments, April 25
#LCI14 Conference proceedings  

Woodley, M.A., J. Nijenhuis, and R. Murphy. (2013). Were the Victorians cleverer than us? The decline in general intelligence estimated from a meta-analysis of the slowing of simple reaction time, Intelligence, 41, 843-850.   


Elijah Armstrong said...

I don't believe the Flynn effect is entirely illusory. My abstract opens: "The 'Flynn effect' is the *increase in intelligence* over the past 100 years."

Also, with respect to the sampling error criticism, see:

Sean said...

Visual Reaction Time in Healthy Obese Individuals

Gottlieb said...

The combination motor skill with direct mental response. I mean, if most reaction tests are currently done manually or even through the computer, will most human beings are able to have motor coordination and mental response in the same line.
Poor motor coordination or below the feature can have an effect on the perfect marriage of scoring with a pencil and immediately thought of reaction.

JayMan said...


Woodley's response to the sampling criticism was weak. There's little question that early, paying museum patrons were an elite sample, and trying to match it with a modern purportedly similarly elite sample is just compounding error with error.

Woodley is free to continue to explore that avenue but there will be no resolution because there's no way reexamine the original findings.

Anonymous said...

Aside from the obvious immigration, I wonder if people are getting less sleep--enough to affect reaction times--these days?

redzengenoist said...

I really wish that these abstracts would go into some more detail about where these samples are from. "Sweden, Great Britain, and the United States", very vague. We'll know more about this in due time, but it's really impossible to have any opinion about this result, besides it being impossible to imagine it being the outcome exclusively of natural selection within the population.

Anonymous said...

There's little question that early, paying museum patrons were an elite sample, and trying to match it with a modern purportedly similarly elite sample is just compounding error with error.

No it wouldn't.

Sean said...

Link between intelligence and puberty/speed of maturity

Anonymous said...

A reasonable hypothesis is that the Flynn effect and earlier puberty are both caused by the same environmental factor, which is higher calorie diets which cause both the brain and the body to grow bigger and faster.

In addition to such gross symptoms are more subtle symptoms such as diverting glucose from the brain to fat deposits causing an epidemic of stupid obese people. And far more likely is that such a pathogen would already exist in various human ecologies around the world -- some of which have co-evolved immunity. It would then be taboo to even suspect the vectors or natural history that might allow science to objectively investigate, discover and provide said immunity with populations not so co-evolved.

Peter Fros_ said...


Not entirely illusory, but I would argue that it is mainly illusory. What could be driving a real increase in intelligence? Better nutrition? Reduction in inbreeding depression? Those factors have long been exhausted. Moreover, if mean intelligence has really been increasing in Western societies, we would surely see signs of this increase in popular culture. Yet my impression is that popular culture is becoming less intellectual. TV programs, for instance, have story lines that are less complex and the issues raised require less thought.

Sean, Anon,

Very interesting. The increase in obesity may be the causal factor. Lack of sleep might be another one.


This is a problem with conference papers. Often, the speaker will not divulge too much information because the paper is still being reviewed.


Are you arguing that motor skill has declined since 1980?

JayMan, Anon,

I agree with JayMan here. Since we don't know the magnitude of the sampling bias, how can we correct for it?

Gottlieb said...

Increased consumption of caffeine among other substances, less physical effort because of mechanization. I do not know, maybe yes.
What I meant is that especially for some people (I do not know, for few or many people), specifically this task, the thought may be faster than the ability to score with pencil or mouse at the exact moment.
It's like in the case of stuttering, the thought is faster than the ability to structure sentences eloquently.
I think it's necessary to make a comparison between manual testing of reaction time with tests which the measurement will be taken directly into the brain.

I did not mean that the motor skills declined since the 80s, especially if one relies on studies with small sample and with these possible measurement errors.
Stutterers often think so fast that they can not speak without stuttering. I think the same can happen in some cases to the reaction time.

Bruce Charlton said...

@Peter Frost

Here is further independent evidence (using a different methodology) of a significant decline in simple reaction times in Scottish women - a fuller version is current in press co-authored by Madison and Woodley

Sean said...

Post 1800, there was massive Irish immigration to Scotland. There was also considerable migration out of Scotland.

For a recent fall, I think an excess of calories is the best explanation.

staffanspersonalityblog said...

In 1980 personal computers became common, you had MTV and similar channels. Lot's of noise basically.

Ben10 said...

It's global: More video, more tv, more stress at work, more allergies, higher general inflammation, they eat more crap etc., as a result and perhaps more importantly: people sleep less (or more superficially),.
They can only be more 'tired' even if work has become less physically demanding. Irony of the times.

Anonymous said...

The phrase is Garbage In, Garbage Out. Reliability is a thing and it mathematically imposes constraints on validity.

This issue really is entirely stemming from horrible methodology - lack of randomized blinded studies with proper sample sizes and controlled external conditions. Not a single vaunted piece of work like Woodley's remotely accounts for what medical science and prior work in other fields knows to be true.

You do know (or should know) that individual performance of untrained laypeople on many reaction time tasks varies tremendously with temperature, it can be by ten+ standard deviations. Everything else is practically just noise, only mitigated by the luck that lots of recent researchers will have mindlessly done their work in similar modern climate controlled offices but without actually documenting methods. Still that leaves replication and current usefulness of results almost completely up in the air.

At least for any physical response like pressing a button, I suppose the stuff you'd look for from the MRI people is currently mostly unknown or untested.

In a statistical sense then most psychometric data in the reaction time category, especially from historical work with conditions and controls unknown, is simply worthless.

Sean said...

"'IT makes evolutionary sense that caloric availability would have an impact, not just on brain regions involved in metabolism, such as the hypothalamus, but also on brain regions involved in learning, such as the hippocampus,' says Alexis Stranahan, a professor at Georgia Health Sciences University and Mattson’s co-author on the Nature Reviews Neuroscience essay. 'Your mind needs to be sharp if you are looking for food. At the other end of the spectrum, it also makes sense that an overabundance of food would dull the senses, making it harder to form associations.'"

A fall in reaction time would not be evenly distributed
due to obesity prevalence being commonest in lower social classes. The higher social classes may hardly have been affected at all.

Sean said...

Reading the Mind in Waist/hip Ratios

ben10 said...

Calorie availability may have increase, but the diversity and complexity (in terms of numbers of ingredients in a recipe) of the food has dramatically decrease, unless you are a committed cook who grows his own food.
Kids in the us schools have a particularly poor diet, IMO. First, kids can choose their dishes. I guess it would be undemocrated to do otherwise, since they are even suppose to choose their own gender. Imposing a variety of diverse dishes to school kids would be considered no less than Fascism, perhaps even Nazism.
Add chemical hormonal analogs among other pollutants in the environment and everything I mentioned previously, like
the lack of good, deep relaxing sleep, and everything can be linked to everything.
Such as, "what causes the rise in autism since the 80's". I've heard that 1 boy over 42 in the US may develop autism!. Or what causes the rise in allergies since the 80's? or the increase in 'gender mismatch', etc.

In short, it's gene/culture coevolution, where 'culture' includes all the modifications in our environment and behaviors. I doubt that a single cause or chemical pollutant can be pinpointed. It got to be an ensemble of causes and epigenetic should not be under-estimated here.

Anonymous said...