Tags
There are a lot of problems with how we measure intelligence. Some people measure intelligence using childhood ratio scores (if a 10 year old is as smart as a 12 year old, he gets an IQ of 120, since he’s functioning at 120% of his chronological age). Other people measure IQ with deviation scores, so if a 10 year old is smarter than 90% of 10 year olds, he’s assigned an IQ of 120. Some tests define the population standard deviation at 15, others 16, others 22 or 24.
About the only thing all IQ tests agree on is that the average score is 100. But the average 20 year old is far smarter than the average 3 year old, so assigning them both an IQ of 100 gets confusing. We must clarify that 100 is the average IQ for one’s age. But IQ’s vary enormously from country to country. So an IQ of 100 is said to represent the average in America, or is it Britain? It depends who you ask. And what happens as demographic shifts cause the populations of both countries to change? What long-term consistency is there in defining IQ 100 as the American average. And what about the Flynn Effect? Average IQ has gone up by some 30 points since the Victorian era yet the average IQ is still 100? And what about dysgenics? Average IQ has gone down by 15 points since the Victorian era yet the average is still 100? I get a headache just keeping it all straight.
And is an IQ of 100 really twice as smart as an IQ of 50? Is a 10 point difference between an IQ of 120 and 130 the same as a 10 point difference between IQ 80 and 70?
Clearly, a much simpler scale is needed:
Around the turn of the millennium, a member of the unbelievably brilliant Prometheus society published one of the most interesting and important articles in the history of psychology, and amazingly, virtually no one has ever read it. The article asserted that problem solving speed doubles every 10 IQ points. The Promethean would later go on to revise the figure to every 5 points. Yes, reaction time (information processing speed) has a beautifully Gaussian distribution, but because the human mind operates in parallel, an IQ 105 is not 5% smarter than an IQ of 100, but rather twice as smart! And an IQ of 110 is four times smarter! Although the Promethean never put it in those terms, this inference was based upon the fact that no matter how cognitively homogeneous a classroom, the difference in learning speed is always at least an order of magnitude. Think back to your high school math class. The brightest kid in the class grasped what the teacher was talking about in seconds, while the dullest may take all year.
So I propose a new intelligence scale to reflect these huge differences. We could call the units of the scale BP scores (Brain Power scores) to differentiate them from IQ scores. We must first anchor our scale to some clear definable unambiguous stable level. I say, the intelligence of the average adult ape in its peak years. To the best I can determine, the average ape has a deviation IQ of 40 (sigma 15). So let’s arbitrary assign the average adult ape (IQ 40) a BP score of 1, and then double the BP for every 5 points above 40.
Thus the conversion between IQ scores and BP scores is as follows:
IQ 40 = BP 1
IQ 45 = BP 2
IQ 50 = BP 4
IQ 55 = BP 8
IQ 60 = BP 16
IQ 65 = BP 32
IQ 70 = BP 64
IQ 75 = BP 128
IQ 80 = BP 256
IQ 85 = BP 512
IQ 90 = BP 1,024
IQ 95 = BP 2,048
IQ 100 = BP 4,096
IQ 105 = BP 8,192
IQ 110 = BP 16,384
IQ 115 = BP 32,768
IQ 120 = BP 65,536
IQ 125 = BP 131,072
IQ 130 = BP 262,144
IQ 135 = BP 524,288
IQ 140 = BP 1 million
IQ 145 = BP 2 million
IQ 150 = BP 4 million
So the average American (IQ 100) would have a BP of about 4000, meaning they’re 4000 times smarter than an ape. The average Ivy league graduate (IQ 130) would have a BP of about 262,000, meaning they’re 262,000 times smarter than an ape. And the average academic Nobel Prize winner (IQ 150) would have a BP of 4 million, meaning they’re 4 million times smarter than an ape. And if intelligence gaps really are as huge as this scale implies, then no wonder we have so much economic inequality!
Brain Size 
Would you provide a reference to this remarkable article from the Prometheus society? I’d like to check it out, because it sounds incredibly wrong, at first glance!
It seems to me that the difference between an IQ of 105 and 100 is impossible either to measure or to notice at the individual level – and pretty difficult to discriminate at the group level.
Maybe the key point is “because the human mind operates in parallel, an IQ 105 is not 5% smarter than an IQ of 100, but rather twice as smart! ” – I don’t think this is necessarily true. When it comes to g – I suspect that the brain works in a ‘series’ kind of way – as fast as the slowest step, with any inefficienmcies being ‘rate limiting’.
In sum, I think the range of real, ratio scale human intelligence is a bit more than double – as represented by the minimum simple reacyion time of c 150 mseconds and high values of some 3-400 ms.
http://iqpersonalitygenius.blogspot.co.uk/2013/02/the-ordinal-scale-of-iq-could-be.html
Would you provide a reference to this remarkable article from the Prometheus society?
It’s been many many years since I read it so remembering the exact reference is difficult. I seem to recall it was in issue 92 (January 1998) of the Prometheus society journal “Gift of Fire”, page 11 to 14? But my memory could be wrong. The article imagined how fast a 400 deviation IQ creature would solve a rubik’s cube (without training), arguing that problem solving speed doubles every 10 IQ points (the author would later revise it to every 5 IQ points). I’m not even sure if the article is still available.
It seems to me that the difference between an IQ of 105 and 100 is impossible either to measure or to notice at the individual level – and pretty difficult to discriminate at the group level.
Well psychometrics is not an exact science, but if for example, only 50% of the population can solve a puzzle in less than 2 minutes, but only 33% can solve it in less than 1 minute, then by definition, IQ 105 is twice as fast as IQ 100, because the top half and top third are assigned IQ’s of 100 and 105 respectively.
Maybe the key point is “because the human mind operates in parallel, an IQ 105 is not 5% smarter than an IQ of 100, but rather twice as smart! ” – I don’t think this is necessarily true. When it comes to g – I suspect that the brain works in a ‘series’ kind of way – as fast as the slowest step, with any inefficienmcies being ‘rate limiting’.
In sum, I think the range of real, ratio scale human intelligence is a bit more than double – as represented by the minimum simple reacyion time of c 150 mseconds and high values of some 3-400 ms.
I think it might depend on the level of analysis. If you measure intelligence at the neurological level, using crude proxies like brain size and reaction time, you’ll find that the biggest brains and fastest reaction speed are seldom more than double the average. Even at the level of one dimensional psychometrics tests like rote memory, the longest memory span for digits are almost never more than twice the average.
However at the level of complex behavior (i.e. solving a rubik’s cube, learning calculus, making money, creating technology), it seems the top people are often at least many thousand times more efficient than average.
So the fastest brain might be be only twice the speed of the average brain, but this difference might get multiplied by the parallel processing of the human mind, causing high IQ people to be orders of magnitude more intelligent than average people. However I’m not exactly sure how this would work mathematically, so I can understand your skepticism.
The problem I perceive is that ‘g’ is supposed to be ‘general’ intelligence and to underpin all cognitive abilities – and measuring a single domain can be terribly misleading; especially when there are large differences in that domain and when there are massive parctice effects (as with most types of problem solving – such as the Rubik Cube).
Yes, in order to test this hypothesis, one would have to try to rule out practice effects by giving a representative sample of people relatively novel problems to solve (in more than one domain) and then observing whether the fastest quarter of the sample (by definition IQ 110 on each particular test) solves each problem twice as fast as the top third (IQ 105) who solves it twice as fast as the top half (IQ 100) etc.
I’m not sure whether problem solving speed doubles every 5, 10, or even 15 IQ points, but I suspect that small percentage differences in reaction time become huge percentage differences in speed of actual complex novel problem solving.
But it’s a difficult model to test across a wide range of ability because the brightest solve easy problems so quickly that it’s hard to differentiate the speed of their mind from the speed of their hands (on say a block design test). By contrast, the dullest solve difficult problems so slowly that it would be impossible to get them to stay motivated until they complete it. Comparing the speed of the very bright & the very dull on the same task would be challenging, though one could focus on smaller IQ differences & extrapolate from there.
I like the concept. But such large number might not be very easy to use and accepted. Also the measurement of intelligence at this moment is still not very precise as height or weight. But this is still quite revolutionary idea for better understanding of intellectual power.
Yes, it’s a radical fascinating idea that needs thorough investigation.
Pingback: Are Canadians twice as smart as Americans? | Brain Size
Pingback: Intelligence
Pingback: Measuring Intelligence of Students
An IQ of 120 most certainly does NOT mean that someone is operating at 120% of their capacity. IQ is a closed order ranking system with a Gaussian distribution. That’s it. ALL it means is that the people who have an IQ of 102 are smarter than the people who have an IQ of 101, who are smarter than than the people with an IQ of 100.
I had hoped you were referring to Prometheus Society member Grady Towers’ note on Rasch measures of intelligence, which are provably the only absolute measures. (See the book Measurement Essentials by Wright and Stone, 1999, full text pdf)
Rasch scores such as used on recent versions of Riverside Publishing’s Woodcock-Johnson and Stanford-Binet are true ratio measures, with equal intervals and true zeros. The range of such tests is much less than you posit, e.g. on the SB5 the highest scorer in the norming population being at 592 while an average 10-year old is at a defined 500 (the only arbitrary choice in converting raw scores to scaled scores, setting 10y.o. = 500 sets the size of the “change sensitive scale” (CSS) unit. A 16 y.o. is at about 510, A 2.25 y.o. is at about 435 CSS. The range of intelligence on an absolute scale is much lower what you posit.
There is a logarithmic transformation involved in constructing Rasch measures, which you might regard as showing the underlying quantity is exponential, but that quantity is actually no good for ordinary arithmetic operations. The overall CSS development curve from early childhood to adulthood is also approximately a log function, improving rapidly at first, then falling off dramatically with a “corner” at around age 8 to 12.
On the other hand, in more intuitive terms there is a great range of intelligence in the population. The adult CSS s.d. is about 8 points, but can be up to 12 at age 5 and as little as 5 at age 8. A +3 s.d. 5 y.o is about as smart as a +0.5 s.d adult, and a +3s.d. adult is as far above an average adult as that average adult is above an average 5.5 y.o..
LOL.
“The average Ivy league graduate (IQ 130) would have a BP of about 262,000, meaning they’re 262,000 times smarter than an ape. And the average academic Nobel Prize winner (IQ 150) would have a BP of 4 million, meaning they’re 4 million times smarter than an ape. ”
Interesting..I live in poverty and yet my BP beats an average Ivy League grad.
I seriously doubt the average NP winner is 150…more like 140
“Nevertheless, scattered information offers some suggestive hints. Consider the IQs of three past American Nobel laureates in physics. In 1965, the year he won his Nobel, Richard Feynman gave a talk at his former high school in New York and told the students that when he took an IQ test at school, around 1930, he scored 125. A ‘merely respectable’ figure (notes his biographer), somewhat below the level of about 130 generally considered to indicate ‘giftedness’. William Shockley was tested twice at school in the 1920s and failed to reach the IQ of 135 required to join Lewis Terman’s pioneering survey of gifted children, started in 1921 at Stanford University. Luis Alvarez, too, was rejected by Terman–which meant that Terman’s programme lost a further opportunity to ‘discover’ a future Nobel laureate. In fact, none of Terman’s gifted children went on to win a Pulitzer prize or a Nobel prize.
The much older Einstein was of course never tested as a youngster. But when he visited the US in 1921, he was informally subjected to a question set by the inventor Thomas Edison (see picture), an advocate of intelligence testing for his prospective employees. Some journalists asked Einstein: ‘What is the speed of sound?’ He confessed he did not know, and replied patiently that there was no need to carry this information in his head, as he could look it up in a book. Next day, the inevitable headline was: ‘EINSTEIN SEES BOSTON; FAILS EDISON TEST’.”
Apparently I think like Einstein..
“…and replied patiently that there was no need to carry this information in his head, as he could look it up in a book.”
I really, really dislike Edison too btw.
Finally, consider that there is a ceiling to processing results. What point is finding the answer in 5 seconds if implementation takes years to actually accomplish? I would suspect figuring the problem out in a few months would work just as well. This is why science proficiency isn’t as correlated with intelligence as we would assume.
Ooops..here’s my source for the quotes:
https://www.psychologytoday.com/blog/sudden-genius/201101/is-high-intelligence-necessary-be-genius
These articles on intelligence and IQ are based on archaic bullshit. The reality is cognitive science is moving towards models of intuition and heuristics.
You mean to say you would probably get an average rating…wouldn’t you?
I have discovered in my experiments that intelligence grows like a plant grows(Neurons, Dendrites, and Axons, and Myelin all grow). This of course takes time, the same way a plant takes time to grow. But if your a good farmer so to speak, it is possible to increase IQ 5 to maybe 35 points, if the right kind of conditions so to speak, are pressent. From the Greatest Scientist Ever
This is interesting. But IQ only correlates to the speed of doing (relatively) simple tasks. As you make those simple tasks more (relatively) complex, the correlation increases until you reach a certain point and then there’s suddenly no longer a very good correlation. For instance, a complex reaction time task correlates better to a simple reaction time task, but when you ask people to solve a word problem, the time they take to solve it doesn’t really predict IQ as well as their performance on the complex reaction time task. For highly-speeded IQ tests, there’s usually a confounding speed factor that reduces the g-loading of the test.
No. Ultimately, your ability to understand the world is limited by IQ. Higher IQ means you can develop a more accurate understanding at full maturity in addition to doing it much faster.